Cuprins Tehnomus XVI

TEFAN CEL MARE UNIVERSITY of SUCEAVAFACULTY of MECHANICAL ENGINEERING, MECHATRONICS AND

MANAGEMENTTECHNOLOGY AND MANAGEMENT DEPARTMENT

TEHNOMUSNEW TECHNOLOGIES AND PRODUCTS IN MACHINE MANUFACTURING

TECHNOLOGIES

2011

Copyright © Technology and Management DepartmentFACULTY of MECHANICAL ENGINEERING, MECHATRONICS AND MANAGEMENT

TEFAN CEL MARE UNIVERSITY of SUCEAVA

Editura Universit ii tefan cel Mare din SuceavaStr. Universit ii nr.13, 720229, Suceava, Romania

Telefon: 0230-216147 int. 273

Editura este acreditat deConsiliul Na ional al Cercet rii tiin ifice din Înv mântul superior – CNCSIS

COD 115

Authors instructions

The full-length paper for publication should not be less than four pages and not more than eight, in

English. Please follow the conference format. A maximum of two papers are accepted, by either a main

author or a unique author. Prospective authors are kindly invited to submit full text papers including results,

tables, figures and references. Full text papers will be accepted by email [email protected]

Organizing Committee can accept only *.doc or *docx format.

All the papers will be evaluated by two reviewers. Selected papers will be published in the

TEHNOMUS New Technologies and Products in Machine Manufacturing Technology Journal.

Web address http://www.tehnomusjournal.fim.usv.ro

Contact:Faculty of Mechanical Engineering, Mechatronics and ManagementDepartment of Technology and ManagementStefan cel Mare University of SuceavaStr. Universit tii nr. 13, Corp B,720229 SUCEAVA, ROMANIAPhone: (+40) 230 216147 int. 260E-mail: [email protected] web site: www.tehnomusjournal.fim.usv.roFaculty web site: www.fim.usv.roUniversity web site: www.usv.ro

mailto:tehnomus:@fim.usv.ro.The

http://www.tehnomusjournal.fim.usv.ro/

mailto:tehnomus:@fim.usv.ro

http://www.tehnomusjournal.fim.usv.ro/

http://www.fim.usv.ro/

http://www.usv.ro/

CONTENT:

1. COMPARISON OF A MACHINE OF MEASUREMENT WITHOUTCONTACT AND A CMM(1) : OPTIMIZATION OF THE PROCESS OFMETROLOGY.WOLFF Valery, TRAN DINH Tin

9

2. EFFICIENT DEVELOPMENT OF HIGH PERFORMANCE POLYMERPARTS USING ADVANCED SIMULATION METHODSAchim Frick, Timo Dolde, Dorota Sich

15

3. SUPERIOR VALORISATION OF THE METALLIC SCRAPS FOROBTAINING BIMETALLIC STEEL-BRONZE PARTSAni oara Ciocan, Florentina Poteca u

21

4. STUDY FOR THE VARIATION OF HARDNESS AND RESILIENCECHARACTERISTICS AFTER HEAT TREATMENT FOR STAINLESSSTEELSManuela Cristina Perju, Drago Cristian Achi ei, Carmen Nejneru, Anca Elena

rgeanu, Mihai Axinte, Ion Hopulele

27

5. AN INTELLIGENT SYSTEM SEARCH FOR INDOOR COLORRECOGNITION: A COMPARISONMário António Ramalho

31

6. MATERIAL SIMULATION OF POLYMERS CONTAINING SPHERICALSTRUCTURESAchim Frick, Melissa Matzen, Timo Dolde

37

7. MONTE CARLO PIVOTAL CONFIDENCE BOUNDS FOR WEIBULLANALYSIS, WITH IMPLEMENTATIONS IN RJurgen Symynck, Filip De Bal

43

8. MODERN GREEK LANGUAGE FREQUENCY COUNTS FOR TEXTENTRY DEVICESIlias Sarafis, Anastasios Markoulidis

51

9. ERRORS OF MECHANICAL TREATMENT PROVOKED BY THERMALDEFORMATIONS OF TECHNOLOGICAL SYSTEMV.I.Savulyak, S.A.Zabolotniy

59

10. IRON REMOVAL FROM WASTEWATER USING CHELATING RESINPUROLITE S930Petru Bulai, Elena-Raluca Cioanca

63

11. ANALYSIS OF CONDITIONS FOR OBTAINING A KEYHOLE REGIMEFOR LASER WELDINGRemus Boboescu

69

12. ASPECTS REGARDING THE QUALITY OF CAST IRON PRODUCED INS.C.FONTUR S.A. SUCEAVAGramaticu Mihai; Cosmaciuc Vasile, Costiug Gina, Hri cu Ionu Marius

77

13. OPTIMIZATION OF THE PLASTIC INJECTION PROCESS THROUGHTHE MODIFICATION OF THE PROCESS FUNCTIONALPARAMETERSTeodor - Daniel Mîndru, Ciprian Dumitru Ciofu, Dumitru Nedelcu

87

14. A PROBABILISTIC DYNAMIC MODEL ASSOCIATED TO A JACKSONNETWORK, WITH SERIES QUEUES; THE SOLUTION OF THEASYMPTOTICALLY STABLE STEADY STATEPopa Marin, Dr gan Mih , Popa Mariana

93

15. TECHNICAL SOLUTIONS TO DESIGN AN EQUIPMENT FOR ICEPARTS RAPID PROTOTYPINGNicolae Ionescu, Aurelian Vi an, Alexandru Savin, Mihai Trif nescu

101

16. COMPOSITE SHAPE MEMORY ALLOYS USED IN ENERGYDISSIPATION APPLICATIONSI. Cimpoe u, S. Stanciu, A. Enache, D. Dan , V. Manole, P. Paraschiv

107

17. IRESEARCH ON THE CORROSION OF CONSTRUCTION MATERIALSFISSURE MACHINERY HYDROCARBONS IN CONTAMINATEDENVIRONMENTS WITH STAFF FROM CORROSIVE BY CRUDE OILDISTILATIONMoro anu Marius , Ovidiu Georgescu

113

18. SOURCES OF POLLUTION IN SIDERURGY AND TECHNIQUES FORREDUCING NOXIOUS EMISSIONSElisabeta Vasilescu , Ana Doniga , Marian Neacsu

117

19. THEORETICAL CONSIDERATIONS ABOUT THE PRECISIONINCREASE OF CONICAL DRAWING PIECESLucian V. Severin, Traian Lucian Severin

123

20. METHOD FOR THE CONTROL OF THE GEOMETRIC ANDCONSTRUCTIVE ELEMENTS OF THE SPIRAL-FACE AND ENDMILLING CUTTERStr jescu Eugen, Pavlov Olimpia, Dogariu Constantin

127

21. ZINC COATINGS ON STEEL SUBSTRATE ATTAINED BY DIFFERENTELEMENTS ADDEDRadu Tamara, Vlad Maria

133

22. ACCELERATION TEST MACHINEMarc Juwet, Koert Bruggeman, Filip De Bal

139

23. DEVELOPMENT OF WASTE MANAGEMENT SYSTEMS IN ANINTEGRATED SHIPYARDDaniela Buruiana

145

24. METHOD FOR DETERMINATION OF THE DISTANCE BETWEENELECTRODES TO ELECTROHYDEAULIC UNDERWATER THROUGHSCINTILLAS DISCHARGEDumitru Iacob

151

25. SUPERIOR WASTE RECOVERY IN THE METALLURGICALINDUSTRYElisabeta Vasilescu , Ana Doniga, Alexandru Chiriac

155

26. IMPACT OF STREET’S LANDSCAPE ON PEOPLE ORIENTATIONAND ROAD SAFETY IN THE CITYA. Polyakov, V. Shvets, O. Veremiy, M.Grabenko

161

27. IMPLEMENTING A SPC INTEGRATED SYSTEM TO IMPROVEMANUFACTURING PROCESSES IN AUTOMOTIVE INDUSTRYL., Lobon , C. V., Kifor, C., Oprean, O., Suciu, Lucian Blaga

165

28. A NEW APPROACH OF THE MAIN LANDING GEAR EQUATIONSDaniel BOSNICEANU

173

29. RESEARCH ON HOLLOW CATHODE EFFECT AND EDGE EFFECTAVOIDANCE IN PLASMA NITRIDING TREATMENTAxinte Mihai, Nejneru Carmen, Perju Manuela Cristina, Cimpoe u Nicanor,Hopulele Ion

181

30. AN ALGORITHM FOR SOLVING QUALITY PROBLEMS, USINGQUALITY TOOLSMarius B u, Mircea Ciobanu

185

31. ABOUT NATIONAL STRATEGY OF ROAD SAFETY IN PRESENT ANDFUTURECiubotariu Danut, Neculaiasa Vasile

191

32. RESEARCHES CONCERNING THE CORRELATION STRUCTURE OFPROPERTIES OF STEELS FOR METAL CONSTRUCTIONSSTRONGLY REQUESTEDStela Constantinescu, Maria Vlad

195

33. DETERMINATION OF THE EFFORTS THAT OCCUR IN THE CASE OFPERFORMING INNER THREADS WITH FORMING ROLLERSIon Cristea , Crina Radu

203

34. CONSIDERATIONS REGARDING THE USE OF CHUCK COLLETS INMECHANICAL SYSTEMSTraian Lucian Severin, Vasile Rata

207

35. THE CORROSION OF TUBULAR FURNACES IN PLANTS.ATMOSPHERIC DISTILLATION OF CRUDE OIL AND IN VACUUMFOR FUEL OILMoro anu Marius, Ovidiu Georgescu

211

36. APPLICATIONS OF FULL FACTORIAL DESIGN EXPERIMENTS FORLASER WELDINGRemus Boboescu

215

37. U.S. QUENCHING AND DIMENSIONAL STABILITY IN TIME OF100Cr6 STEELNicolai Bancescu, Constantin Dulucheanu, Traian Lucian Severin

223

38. INFLUENCE OF POLYMER CONCENTRATION ON THEPERMEATION PROPERTIES OF NANOFILTRATION MEMBRANESBalta Stefan, Bodor Marius, Benea Lidia

227

39. STATIC DEFORMATION OF A WORKPIECE FIXED IN UNIVERSALCHUCK AND LIFE CENTRELauren iu Sl tineanu, Lucian T caru, Margareta Cotea , Teodor Tilic , MihaiBoca, Irina Grigora (Be liu),

233

40. RESEARCH REGARDING THE USE OF QUALITY TECHNIQUES ANDINSTRUMENTS IN VIEW OF MAINTAINING AND IMPROVINGQUALITY MANAGEMENT SYSTEMSLiliana Georgeta Popescu, Mihai Victor Zerbes, Marilena Blaj, Radu VasilePascu, Roland Blaj

239

41. PROPERTIES OF HARD ALLOYS SINTERED FROM METALLICCARBIDESConstantinescu Stela

245

42. RESEARCH ON THE CORROSION BEHAVIOR OF ZINC COATING BYMEASUREMENT OF POLARIZATION RESISTANCERadu Tamara, Istrate Gina Genoveva

251

43. MATHEMATIC MODEL FOR OPTIMIZATION OF ZINC-NICKELALLOY CO-DEPOSITION PROCESSVioleta VASILACHE, Marius BEN A

257

44. MODULAR DESIGN FOR A FAMILY OF MECHANICALANTHROPOMORPHIC POLY-MOBILE GRIPPERS WITH 4 FINGERSFOR ROBOTSIonel Staretu

261

45. THE INFLUENCE OF ENVIRONMENT TEMPERATURE VARIATIONON THE STRENGTH CHARACTERISTICS OF COMPOSITEMATERIALS TYPE “ALUCOBOND"Constantin Dulucheanu, Traian Lucian Severin, Nicolai Bancescu

267

46. MATHEMATICAL MODELS FOR SMOOTHING PROCESS BYMECHANICAL SHOCKS OF CYLINDRICAL SURFACESRaluca Tanasa, Traian Gramescu

271

47. EXPERIMENTAL RESEARCH ABOUT INFLUENCE OF INCLINATIONANGLE DIRECTION UPON THE SURFACE ROUGHNESS IN BALLEND MILLING OF OLC45 (C45) MATERIALPa ca Ioan, Lobon iu Mircea

277

48. PARTICLES DIMENSIONAL ANALYSIS AND MICROSCOPICCHARACTERIZATION OF HYDROXYAPATITE POWDERAurora Anca Poinescu, Rodica Mariana Ion

283

49. OVERVIEW OF LATEST MINERAL CARBONATION TECHNIQUESFOR CARBON DIOXIDE SEQUESTRATIONBodor Marius, Vlad Maria, Balt tefan

289

50. DEVELOPMENT OF MEANS AND A PLATFORM FOR RESEARCHLiliana Georgeta Popescu, Mihai Victor Zerbes, Radu Vasile Pascu, LucianLobon , Livia Dana Beju

297

51. IDENTIFICATION AND QUANTIFICATIONS OF CHEMICALELEMENTS AND MICROSCOPIC CHARACTERISATION OF THECOPPER BASE ALLOY AT DIFFERENT TEMPERATURE OF MOULDSCristina Camber Iordache, Maria Vlad, Simion Balint, Vasile Basliu, MihaelaGheorghe

303

52. MANGEMENT OF TECHNICAL DATA IN THE STUDY OF THEBRAKING SYSTEM OF CARSCiubotariu Danut, Neculaiasa Vasile

309

53. COMPOSITE SHAPE MEMORY ALLOYS FOR METALLICSTRUCTURESA. Enache, I. Cimpoe u, S. Stanciu, I. Hopulele, R. M. Florea, N. Cimpoe u

315

54. STUDY OF THERMOMECHANICAL FATIGUE FOR SHAPE MEMORYALLOYS TYPE CuZnAlAchitei Dragos Cristian, Hopulele Ioan, Perju Manuela Cristina, Mihai Axinte

321

55. STUDY REGARDING THE AUDIT OF MANAGEMENT PRINCIPLESCostel Mironeasa, Silvia Mironeasa, Georgiana Gabriela Codin

325

56. INTEGRATION OF SIX SIGMA AND LEAN PRODUCTION SYSTEMFOR SERVICE INDUSTRYMilitaru Emil

329

57. OPTIMAL PERTURBATION FOR MIXED CONVECTION HEATTRANSFER IN RECTANGULAR CHANNELRachid SEHAQUI

335

58. STUDY ABOUT VIRTUAL AND ACTUAL MANUFACTURINGPROCESS WITH THE ROBOTTraian Lucian SEVERIN, Romeo IONESCU

341

59. CONSIDERATION OF REFLECTION COEFFICIENT AS A FUNCTIONOF FREQUENCY FOR WEDGES OF DIFFERENT MATERIALSLiliana Petre-Cainiceanu, Constantin D.Stanescu, Tudor Burlan-Rotar

349

TEHNOMUS - New Technologies and Products in Machine Manufacturing Technologies

9

COMPARISON OF A MACHINE OF MEASUREMENTWITHOUT CONTACT AND A CMM (1) : OPTIMIZATION OF THE

PROCESS OF METROLOGY.

WOLFF Valery 1, TRAN DINH Tin 2

1 IUT Lyon 1, University of Lyon, [email protected] University of Lyon 1, [email protected]

Abstract: In the field of the control of manufactured products, the process of measurement usuallyproceeds on a type of machine (for example CMM) and the totality of measurement is then carried out.We propose here to study the capacities of 2 types of measuring machines in order to guide the operatortowards an optimized choice of the process of measurement. The taking into account of the capability ofthe machine associated with the feasibility (facility of use) of measurement will be also supplemented bythe taking into account the time necessary to realize measurement. A better knowledge of the limits ofeach machine will allow an optimization of the process of control.

Keywords: metrology, control, CMM, no-contact measuring system, industrial vision

1 IntroductionThe three-dimensional metrology is one of the

main problem for many industries for achievingdimensional and geometric product control. Thecontrol of technical parts for the plastic industry,automotive or electronics is more and more usingthe optical measuring systems without contact. Theother way is to use a CMM (coordinate measuringmachine) for a more classical control.

We have chosen to work with these two typesof machine to develop our study. We try to definea methodology to optimize the use of thesedifferent types of measuring machines.

2 ProblematicThere are three main steps during the

realization of manufactured products: the stages ofdesign, manufacture and control.

The designer defines the functionaldimensioning of the parts by using the GPSstandards.

The traditional method, for the control of theparts in metrology, consists in defining ameasurement plan in adequacy with the measuringinstruments available (We have to separate, forexample, the measurement of the roughness of a

surface, and the 3D control of geometricalspecifications).

We can note that an operator usually prefers touse only one machine when it is possible. Forexample, in the case of the use of a CMM, theentirety of measurement will be carried out on themachine (dimension, form, and positionspecifications).

The idea we want to develop here, relates to thepossibility of a distributed control: using differentmeasuring instruments for different specification.

Is it preferable, in term of cost, to carry out thetotality of a control on the same machine? Or is itbetter to carry out measurements on variousmachines?

This question is the base of the processplanning in manufacturing activity. But, it is a nonstudied point for the control of the parts.

3 Study

The context of the study is the industrial visionsystems [2] and we are using the optical methodsfor dimensions measuring. In our study, the systemis equipped with an optical device (camera) givingto the operator an image of the measured part. Theoptical machine is a TESA V300 (Figure 1). It is amanual machine with triple lighting: diascopic,episcopic and LED. Each axis (X/Y/Z) has aresolution of 0.05 m.

mailto:valery.wolff:@univ-lyon1.fr

mailto:tdtinbd:@yahoo.com


10

Figure 1 : Tesa Visio 300

3.1 Measurement method

For this study, we use a reference standard ring.So, we can compare the results of the measurementto the theoretical value. This is a standard datumring with diameter accuracy less than 10-3 mm.

The TESA Visio provides 3 possibilities:manual selection of each point, automaticacquisition for each point, global acquisition of aprofile (automatic sector detection).

The first step of the study was to measurediameter dispersion by using different ways tocapture a geometrical element on a standard ring.The obtained results [8] show that the minimumdiameter dispersion is given by the manual method(Figure 2).

Figure 2 : accuracy of measurement methods

In this article, we have chosen the manualmode. It is the best way to get accurate results forour experiments.

By measuring the reference standard ring, wecan obtain several values: the diameter, the XYposition of the center and the circularity default(form). The figure 3 shows the definition of thecircularity.

Figure 3 : circularity default

The real line (we can call it: “real circle”) mustbe included into an area: the Tolerance Zone. Inthe case of the circularity, the tolerance zone isdefined by two circles with the same center, andwith a radius difference lower than the value of thespecification.

If we are using the Tesa Visio machine in aclassical way, we can obtain each circle wemeasure with only three data: diameter, form andposition.

But, for the form information, we have only aglobal result without any possibility to have adetailed measure. For example, we choose toacquire 20 points for a circle. We could expect tohave 20 pairs of XY coordinates. In the Tesa Visio300, the machine calculates a theoretical “best fit”circle associated to the 20 points and gives only thefinal result: the default of form (circularity).

Because we want to have a detailed study of theresults of each measurement, we have chosen tomeasure all points independently. After that, weuse an Excel Sheet to calculate the sameinformation we could have obtained with the TesaVisio. The method of calculation is presented intothe next paragraph.

3.2 The calculation method

In the mechanical engineering field, themetrology often needs to calculate mathematicalelements associated with real elements. In the GPS(geometrical product specification) concept, theterm of skin model represents the real surfaces(figure 4).

The most common method used to associate atheoretical element to a skin model is the methodof least squares [1]. This method allowsestimating the numerical values of the form.

15 15,005 15,01 15,015 15,02 15,025 15,03 15,035 15,04

1

2

3 Automatic

Global profile

Manual acquisition

Theoreticalvalue

Tolerance

tol


11

$Figure 4 : skin model and mathematical

association

It exists with several variations: it’s simplerversion is called ordinary least squares (OLS), amore sophisticated version is called weighted leastsquares (WLS), which often performs better thanOLS because it can modulate the importance ofeach point in the final solution. Recent variationsof the least square method are alternating leastsquares (ALS) and partial least squares (PLS). [1]

The oldest (and still most frequent) use of OLSis linear regression, which corresponds to theproblem of finding a line (or curve) that best fits aset of points. In the standard formulation, a set ofN pairs of coordinates {xi, yi} is used to find a pairof parameters , and (figure 5)

iii xye

Figure 5 : least square method for a line

The least square method defines the parameters and as the values which minimize the sum of

the distances ei² (hence the name least squares)between the measurements and the model. Whatleads to minimize the function:

2)(i

ie (1)

This is achieved using standard techniques fromcalculus, namely the property that a quadratic (i.e.,with a square) formula reaches its minimum valuewhen its derivatives is equal to zero. This gives thefollowing set of equations (called the normalequations):

0 and = 0 (2)

Solving these 2 equations gives the least squareestimates of , and .

The OLS method can be extended to more thanone independent variable (using matrix algebra)and to non-linear functions.

In this article, we are calculating the default ofform of the datum ring. The result of theexperiment (coordinates of points) to determine thecircle requires to use the polar coordinates.

Figure 6 : least square method for a circle

The figure 6 shows the parameters we use todetermine a theoretical circle associated to a realset of points.

Zi: distance is measured from the nominalcircle

ei: error after calculationR: radius variation of the experimental

circle compare with the nominal circleu, v: distance between the two centers

(small displacement)

Based on the figure 6, ei is calculated asfollows:

Rvuze iiii sincos (3)

After calculation, we replace u, v and R intothe equation (3). The values (emax - emin), determinethe default of form of the circle.

3.3 The design of experimentIn the aim to compare the two machines we

have, the experimentation was carried out in a

Nominal element Associated model

Skin model

i


12

X Y u 0,008,504 -0,038 v 0,008,128 2,512 dr 0,002

… … 0,010… … eimin 0,005-

8,066 -2,678 eimax 0,005X Y u 0,001

8,504 0,137 v 0,0018,126 2,503 dr 0,002

… … 0,010… … eimin 0,005-

8,068 -2,685 eimax 0,005

similar way on the Tesa Visio 300 and the tri-mesure CMM.

We used the manual mode for TESA Visio 300with the setup parameters as follows: down light,zoom at 50%, and manual acquisition of the points.The CMM allows to measure circles and to exportall points after the measurement operation.

We measured the datum ring several times. Foreach machine, we get the points from fivemeasures. Each measure, is composed of twentypoints with Xi, Yi coordinates.

The figure 7, shows the position of the centre ofa measured circle obtained by the calculation of thetwo parameters u and v. If we want to compare(with a graph) our results on the circularity default,we need to have all of the points xi,yi expressedinto the same datum coordinate system.

Figure 7 : nominal (dot) and optimized (red line)circles

On the CMM machine, it’s easy to obtain allcoordinates with a defined origin. The softwareallows defining an origin before to save all thecoordinates (xi, yi). In this case, each circle is usedas the origin of the system. The calculation of the

least square method gives u = v = 0 (table 1). Theonly one parameter is R.

For TESA V300 machine, we measured 8points on the circle to determine the center of thecircle. This center is then defined as the origin ofthe system. This is necessary before to obtain 20points by second measurement operation. In thiscase we can have small displacement: u and v havesmall values (Table 1). This is one of theassumptions of the least square method.

After performing the experiments, we calculatewith an excel sheet the parameters of each circlecomposed of 20 points.

3.4 The statistical approach

We made five measurements for each type ofmachine. For more accurate results during theanalysis, we took the average value with twomethods (Figure 8).

The first approach: Each set of 20 pointsmeasurement determine a circle by using the leastsquares method. Each measurement allowsobtaining a circularity default. We will thencalculate the average value of the circularityparameter.

The second approach: After performing theexperiment, we calculate the average (xi, yi)coordinate for each point of each measurement.The least square method is used only after thisstep. We determine the parameters of a circlecalculated with the average set of points.

Figure 8 : statistical method

The aim of these two approaches was todetermine the best methodology (in term of time ofmeasurement and calculation).

4 Results and analysisThe quality of the measured circle (circularity

default) has the same average value for the twomachines we used (0,007 mm on the CMM and0,010 mm on the Visio300).

Table 1 : CMM and Visio300 datum system

CMM

Visio 300


13

The repeatability of the measure between eachset of 5 measurements allows us to say that theresult can be studied (figure 9).

Figure 9 : graphical results

The most important result we have noticedrelates to the graphics on figure 10 (case of theVisio300).

The graphical observation shows that the areawhere we can obtain the outer and the inner pointof each measure is at a different place. For theCMM machine, the results are good at thepositions: 0, 90, 180, 270° and for the TESAVisio300 machine, the best results are at points 45,135, 225, and 315°.

This can be explained by the way we acquirethe points.

On the CMM machine, the move of the sensoris controlled by a numerical system with two Xand Y axis. The direction of the move gives morereliability if we use only one movement at thesame time: X only or Y only.

Figure 10 : best result area for the Visio300

Figure 11 : visio300 acquisition

The combination of the two axis for asimultaneous move involve a different dispersiondue to the detection of the measured surface by thesensor.

In the case of the Visio300 machine, it is moredifficult for the operator to be precise when he tryto measure a point with the tangent position thanwith a clear intersection of the cursor and the circle(Figure 11).

5 ConclusionThis article is a first step for a more global

study. It shows that the choice of the machine wewant to use is maybe linked to the capability of themeasurement machine (average value of thecircularity in our case), but also linked to the timewe want to spend for the measurement operation.

The choice of the points (number and position),and the way to acquire each point is a subject thatwill be experimented in a next study.

We will also extend the study by taking intoaccount the velocity of the machine (in particularfor the CMM machine).

References[1] Herve Abdi, “Least Squares”, The University of

Texas at Dallas, Richardson, TX 75083–0688,USA.

[2] P. Bourdet, C. Lartigue, A. Contri. “Les capteurs3D état de l’art, problématique liée à la précisiondes systèmes de numérisation 3D." 3ème congrèsNumérisation 3D/Human Modeling, Paris, 27-28mai 1998.

[3] J.-L. Charron. Mesure sans contact – Méthodesoptiques (partie 1). Technique de l’Ingénieur,2004.

[4] J.-L. Charron. Mesure sans contact – Méthodesoptiques (partie 2). Technique de l’Ingénieur,2004.

small dispersion

large dispersion

CMM

Visio300 0, 90, 180 or 270° 45, 135, 225 or 315°


14

[5] R. Christoph, H. Neumann. Multisensorcoordinate Metrology: Measurement of form,size, and location in production and qualitycontrol. Verlag modern industrie, 2004.

[6] D. Gava. Vision conoscopique 3D : calibration etreconstruction. Thèse de doctorat, UniversitéRené Descartes – Paris V, 1998.

[7] C. Mehdi-Souzani. Numérisation 3D intelligented’objets de formes inconnues basée sur descritères de qualité. Thèse de doctorat, EcoleNormale Supérieure de Cachan, 2006.

[8] Valery Wolff, Arnaud Lefebvre, Dimitri Pachel,Jasper Thijs, “Capability of measuring machine:Case of an optical measuring machine withoutcontact”, Proceedings of IDMME – VirtualConcept 2010 Bordeaux, France, October 20 –22, 2010.

[9] H. Zhao. Multisensor integration and discretegeometry processing for coordinate metrology.Thèse de doctorat, Ecole Normale Supérieure deCachan, 2010.


15

EFFICIENT DEVELOPMENT OF HIGH PERFORMANCEPOLYMER PARTS USING ADVANCED SIMULATION METHODS

Achim Frick, Timo Dolde, Dorota Sich

HTW Aalen, University of Applied Sciences, Institute Polymer Science and Processing,Beethovenstr.1 73430 Aalen, Germany, PHONE: +49/7361/576-2169,

e-mail: [email protected]

Abstract: The optimization of part development process should lead to an improved time to marketperformance, raise cost and resource efficiency and at the same time it has to ensure product quality andreliability. The quality and performance of polymer parts, in contrast to metal parts, depend not only onselected material but additionally they are determined by processing conditions during manufacturing e.g.injection moulding process. Thus, the optimization of development process of polymer parts is of highimportance.The nowadays available simulation techniques are necessary tools in the virtually based developmentprocess and by the creation of a reliable simulation results the development process can be considerablyaccelerate and the competitive quality parts can be in short time realized.This article shows the advantages of application and combination of the current available simulationtechniques in the development of single stage gear made of plastic. The evaluation and verification of thesimulation results will be presented and discussed.

Keywords: CAE, FEM, Structure Optimization, Filling Simulations, Rapid Prototyping, VirtualDevelopment

Figure 1: Flow chart of the virtual development process of polymer products

mailto:Achim.Frick:@htw-aalen.de


16

1. Introduction

For the development of highly loaded parts, theselection of a suitable material is of the firstimportance while it determines the final productproperties and the whole development process. Thedevelopment process depends on the chosenmaterial class, namely for polymer, in comparisonto metal, it needs special polymer adapted methods(e.g. polymer design, filling simulations). Thus, thecomplexity of development process of plastic partsis much higher than steel parts. Nevertheless,polymer materials, due to their lower density incomparison with metals, show a special potentialin the field of lightweight construction. They offeralso flexibility in the processing and design.Additionally, polymers offer the chance tocombine different properties (multi-functionality,multi-component parts, “tailor made” polymerse.g. fibre reinforcement), which can lead to someadvantages in comparison to steel.For a complete use of the potential of polymericmaterials, the virtual material development andsimulation have to be involved in the polymer partdevelopment. The flowchart [Figure 1] shows thenecessary steps in the development process ofpolymer parts and the necessary interfaces tosimulation and development of the material.Consecutively the steps in the flow chart will beshown in details on the example of a gear wheelmade of engineering thermoplastic -polyoxymethylene (POM). Additionally, the aim isto replace the steps “prototyping” and “prototypeevaluation”, which include the cost expensive andtime consuming manufacturing and testing ofhardware prototypes, by virtual assessmentmethods.

2. Development of a highly loaded polymerpart

According to the flow chart [Figure 1] the firststep in the development process is the creation ofthe product idea. Principally in the creation ofproduct idea following aspects have to be takeninto consideration: function, economy and ecology.

2.1 Product idea

A single-stage spur wheel gear made ofpolymeric material is intended for application inelectrically driven vehicles. A single stage gearneeds a high gear transmission ratio between the

pinion and the gear wheel, and therefore the gearwheel is big-sized. The application of polymericmaterials enables the lightweight constructionswhat results in the reduction of both, the rotatingmasses and the mass of the car. The lightweightleads to a higher acceleration, energy saving andthus less CO2 emission. Another important aspectfor using a polymer material is that for the largescale production injection moulding is competitivemethod due to the short cycle time and the readymould part “without post handling”. Hence, asignificant cost reduction can be achieved incomparison to a gear wheel made of steel.

The material of choice for the gear wheelconstruction is polyoxymethylene (POM). A gearwheel made of POM can provide low friction andlow noise emission as well as the dampingbehaviour. In order to reduce the mass of vehicle,the polymer gear wheel shall substitute the safetyclutch between the electrical engine and the gear.Due to sufficient impact toughness, POM is thesuitable material for this application.

2.2 ConceptTo concept the product it is necessary to

analyze the problem and clearly define therequirements. Thus, all boundary conditions of theelectrical vehicle and the polymer gear arespecified:

- Rear wheel drive- Max. velocity = 120 km/h- Worst load case: Blocking of the electrical engine- Consequence of the blocking: Blocking of the rear wheels which leads to deceleration of the car- Dynamic axle load during the deceleration: 300 kg on the rear axis- Friction between the tires and the street: µ = 0.8- Diameter of the wheel: 0.62 m- Rotational speed of the electrical engine: 3000 rpm- Material pre-selection: Polyoxymethylene (POM)- Manufacturing process: Injection moulding

Based on the basic information, the transmissionand the worst case loading with respect to torqueas well as the required gear wheel dimensions canbe calculated as follows.


17

2.2.1 Calculation of the transmission [1]

Circumference of the tire in [m]:2 = 0.6 × = 1.884

Velocity in [ ]:

= 120 ÷ 3.6 = 33.33

Rotations per second:33.33 ÷ 1.884 = 17.69 /

Rotations per minute:17.69 × 60 = 1061.46

Max. rpm engine: 3000rpm

Selected transmission: i = 3:1

2.2.2 Calculation of the maximum torqueMax. friction force:

= × = 0.8 × 300 × 9.81 = 2354

Max. torque:= × = 2354 × 0.3 = 706

Max. torque, gear wheel:= 706

Max. torque, pinion:= 706 ÷ 3 = 235.33

2.2.3 Conception of the gear

The basic concept of the gear is shown in Figure 2.

Data of the gear:

Transmission: i = 3Pressure angle: = 20°Module: m = 12Pinion: 10 teethGear wheel: 30 teethPitch circle pinion: = 12 × 30 = 360Pitch circle gear wheel: = 12 × 10 = 120Centre distance: C = 245 mm

Figure 2: Basic concept of the single stage gear

2.2.4 Calculation of tooth root stressParameters:

E-Modulus of POM: E = 2700 MPa [2]Temperature: 40°CLoad cycles: 105

Safety factor: s = 1.4

Dynamic strength of the root [3]:= = 46.4

Circumference force [2]:= × = 3922.22

Tooth root stress [2]:=

×× × × = 46.4

width of the tooth: w = 40 mm

2.2.5 Calculation of surface pressure [2]

Max. surface pressure: = 100Temperature: 40°C, in oilLoad cycles: 105

Safety factor: s = 1.4= = 71.4

Surface pressure:

= × × × × ×

= 71.4

= 64.26 71.4

2.3 Basic DesignOn the basis of the calculated, required

dimensions of the gear wheel, the construction ofthe basic design can be done. With the 3D-CAD-program ProEngineer, 3-D models of the gearwheel [Figure 3] and the pinion [Figure 4] are cons-tructed. In addition, the design of the gear shafthub is realized. By assembling the single parts, thefitting of the parts can be proven.The models will be used in the dynamic impactcalculation with LS-Dyna. For this purpose onlythe gear teeth are important and that is why thefirst draft of the wheel and the pinion contains noribs.

Advantages of 3-D-CAD systems:

With virtual assemblies, the functionality of thesystem can be ensured.


18

Iterative course of action:

Depending on the results of the dynamic strengthcalculation, the basic design can be improved untilthe necessary strength is achieved.

Figure 3: Basic design of the gear wheel

Figure 4: Basic design of the pinion

2.4 Dynamic strength calculationThe FEM-calculation of the gear wheel

impacting the pinion is executed with the explicitFEM-solver LS-Dyna. For the simulation of anabruptly stopping engine, the pinion is stoppedafter a rotation of 180°. Like in the reality, the gearwheel is loaded with the momentum from thesliding wheel of 607 Nm. In consequence of theabrupt stopping, the teeth of the gear are impactingeach other. But also the hub-rim-connections arehighly loaded [Figure 5].

Advantages of dynamic simulations:

Load peaks which results from dynamic loads canbe calculated. In addition, the functionality isproven by examination of the sliding behaviour ofthe teeth.


If necessary, the results of the dynamic FEM-calculation must be taken into consideration for theimprovement of the basic design. On the exampleof the gear, the teeth are strong enough, only therim-hub-connection of the pinion requiresimprovement [Figure 5].

Figure 5: Dynamic simulation of the gear wheelimpacting the pinion (Ls-Dyna)

2.5 Structure OptimizationThe aim of a structure optimization is the

improvement of the properties. These propertiescould be e.g. the eigenfrequencies, the stiffness ofthe part and its weight.

On the example of the gear, the aim is to reachthe lowest possible weight, under the condition,that the gear wheel will still resist the worst caseload. The optimization is realized with theoptimization program "Optistruct" from the AltairCompany. It is an implicit solver and theoptimization works with a static FEM-calculation.The course of action is that the FE-meshed gearwheel is loaded with the circumference force onthe teeth. During the calculation, the programidentifies elements with low stresses anddeformations. These elements will be deleted andthe calculation runs again. After some steps ofiteration, the program offers a proposition of animproved shape of the part. In addition, there aresome tools for defining boundary conditions. Forplastic parts, the opportunity to propose shapewithout undercut which guarantees good releaseproperties is really important tool. Anotherimportant tool is the pre-setting of a constant wallthickness. This new shape can be converted backinto a CAD-part, in this case a ProEngineer file. Incase, that the result, e.g. the weight reduction, isnot satisfying, there is the possibility to switch


19

back to the material simulation and try to get e.g amaterial with a higher stiffness.

Advantages of structure optimization:

The structure optimization enables great progressesin the shape of a part, because it identifies theunloaded sections and carries out much iteration.This is much more efficient than a manualimprovement of the shape.


On the one hand, the results from the optimizationcan require improvement of the materials, but onthe other hand, the results of the filling study canrequire changes of the design, so that theoptimization has to be processed again.

2.6 Final Design

Based on the optimized CAD-part, the finaldesign is reached by adding the necessary details tothe construction. These details are e.g. radiuses,chamfers and relief grooves. In the example, thisstep is realized with ProEngineer. For the next stepin the development process, the simulation of theinjection moulding process, some versions of thegear wheel with different sprues should bedesigned. For the gear wheel, the best location forinjection is from the hub, maybe with single spruesor with a film gate.

Figure 6: Final design of the gear wheel

2.7 Filling study

The injection moulding simulation is used forthe assessment of the draft design concerning theability to reach good material and morphologicalproperties. The program which is used is"Moldflow", a product of the Autodesk Company.The purpose of Moldflow is a complete virtualtesting of the production process (injection

moulding). It is possible to edit all necessaryparameters for the simulation.

- Selection of the material with all relevant properties:

Matrix materialFilling material (fibres, spheres)Viscosity of the materialThermal properties of the material

- Adjustment of the tool temperature- Setting of injection and holding pressure- Setting of switchover point- Setting of holding time

The results of the Moldflow simulation areinformation about the following parameters:

- Cooling rate of the part- Building of the morphology (amorph, semi-crystalline etc.)- Process time- Fibre orientations- Appearance of voids [Figure 7]- Weld lines [Figure 8]- Mould shrinkage- Deformation of the part triggered by different cooling rates and the resulting differences in morphology

Figure 7: Voids in the gear wheel

Figure 8: Weld lines in the gear wheel


20

Interpretation of the filling study:

The voids signify a weakening of the material.With the knowledge about their appearance in thepart, it can be assessed, if they are critical or not.

The weld lines [Figure 8] mean considerableweakening of the part, particularly in combinationwith fibre reinforced materials, because theorientation of the fibres at the weld lines aredisadvantageous. Furthermore, the weld lines inthe gear wheel are localized in the critical areaaround the highly loaded hub.

Advantages of filling study:

The filling study enables the optimization of theshape, the mould, the material and the processparameters in a way, which cannot be achievedeasily with normal methods. It means an enormouscost reduction compared with the manufacturing oftesting-moulds. Moreover, it allows assessing ofthe morphological properties of the manufacturedpart. Normally, the analysis of the morphologicalproperties would require the expensive analysis ofhardware prototypes with a computer tomography.


According to the flow chart [Figure 1], the resultsof the filling study can require an iterative backstep to the material simulation and the final designof the part.

4. OutlookThe subject of prospective researches will be

the analysis of the influence of the voids on themechanical strength of polymer parts. Therefore,the real parts with voids will be scanned in thecomputer tomography. Using the methods ofreverse engineering, a CAD-model of the part withthe voids will be created. Based on this virtual part,all the necessary FEM-calculations will beexecuted. For the purpose of evaluation andverification, the simulation result will be comparedwith the results of mechanical tests on the physicalparts.

5. References[1] Grote, K.-H., Feldhuse, J., Dubbel, Springer

Verlag, 2007.[2] Ehrenstein, G.W., Mit Kunststoffen

konstruieren, Hanser-Verlag, Munich, 2007.[3] Hermann R., Matek, W., et al.,

Maschinenelemente, Verlag Vieweg undTeubner, 2007.


21

SUPERIOR VALORISATION OF THE METALLIC SCRAPSFOR OBTAINING BIMETALLIC STEEL-BRONZE PARTS

Ani oara Ciocan1, Florentina Poteca u2

1Dun rea de Jos University of Galati, [email protected] rea de Jos University of Galati, [email protected]

Abstract: The possibility of bimetals producing by direct melting of the bronze machining chips intosteel support is considered. The aim is the manufacturing of the bimetal that is formed by joint of abronze layer on steel support. By this way the recycling of the metallic wastes as final products withvalue added is possible. In this paper the flow sheet for the bimetal manufacturing by this method wasshowed. Also the parameters that influenced the quality of bimetal were analysed. The quality of thebimetal was analysed in term of contact zone quality. The interface zone of two alloys was investigatedby methods of metallographic and microscopic analyses.

Keywords: bimetals, recycling, bronze chips, steel

1. Introduction

In many industrial applications, the workingsurfaces of the parts are simultaneously exposed toa very different kind of stresses. In these types ofworking regime, the bimetallic materials can beused. These materials are obtained using thedepositing process of one material, as layers withdifferent thicknesses, on top of another materialconsidered as a base. These metallic materials, alsonamed bimetallic can satisfy some requests of theworking conditions impossible to be obtainedusing one single metallic material. The coveredsteels with non-ferrous alloys layers are anexample and the bimetallic pieces are obtained bythe different methods production. The bimetals areobtained by metallic layers deposition on thesurface of the support parts. The joining of themetallic layer on the solid support is determined bydiffusion and thermal processes [1]. The metaldeposition most is makes on one or both surfacesof the solid support. The thickness of coveringlayers is 8 – 20% of the support alloy. The firstdates about the bimetals are early mentioned. Inthe published papers is show that 1858 is the yearwhen the first patent (USA Patent) for the bimetalobtaining is mentioned. For the Germany theresearches for the production of the cladded steelsheets begins in the seventh decade of the XIXcentury and a patent was realized. The first specialindustrial application for bimetals is dated from

1930. This is from USA and is referred atutilization of some steel sheets with nickel claddedfor construction of the tank wagon for the chemicalproducts transports. The production of the greatquantity of the carbon steel sheets cladded withstainless steel starts around 1938 [2].

One method for obtaining of the bronze layeradded on the steel support for the bimetalmanufacturing is by welding process. For this isnecessary the remelting of the bronze wastes andmoulding in the form of bars. The gas-shielded arcwelding process with wolfram electrode can beused for welding bronze on the parent steel [3, 4].During the bimetal obtaining by this method a lotof problems are associated: the lost of metal isgreater; the costs of labour and energy or materialsare increased as well as the problems ofenvironment protection are higher.

The present paper presents the researches forbimetal manufacturing from waste products. Tocreate a bronze layer on the steel surface, bronzemachining chips were used. In this case the chipswaste of the bronze with aluminium complexalloyed with iron, nickel, manganese were used.

2. Experimentals and materials

In the experimental work, steel samples assupports were utilized, Figure 1.

mailto:aciocan:@ugal.ro

mailto:mihaela_potecasu:@yahoo.com


22

Figure 1: Experimental steel sample used that steelsupport into experimental

In accordance with the complex work of thisbimetal that be used as bearing parts (ability tocarry heavy loads under friction conditions withoutexcessive wear and resistance to corrosion) weused selected bronze chips in accordance withcertain chemical composition, Table 1 [5, 6].

Table 1: Chemical composition of the bronzeElements,

[wt%]Al Ni Fe Mn Cu

Rangeexpected

14-15

4-4.5

4-4.5

0.8-1.2

bal.

Chips 14.7 4.4 4.1 1.1 bal.

The fluxes with multiple actions were used:borax as neutral cover fluxes, glassy fluid coverflux, graphite as reducing flux containingcarbonaceous materials and, copper-phosphorusmetal mixed with bronze chips that self-fluxing oncopper-base metals [7].

A resistance furnace was used for heating thesamples, Figure 2.

Figure 2: Furnace used to melting the bronze chips intosteel support

The quality of bimetal is discussed in term ofcontact surface for both alloys (steel and bronzewith nickel, iron and manganese). The structuralaspects associated with heating process werepresented. An Olympus microscope

Metallographic was used. Optical microscopyincluded standard methods for preparing samples.

2. Results and discussions

Although the method proposed for bimetalmanufacturing is cheap and simple this involvesany problems. The main are associated with thequality and the properties of the bronze and thebronze chips. Other are associated with processesthat accompanying the copper alloys melting andjoining with steel support.

The first condition for a quality process isreferred on the properties of the recycled material.The success of any melting system depends on thephysical characteristics and nature of thefeedstock, therefore, the integration of durable,efficient pre-treatment is crucial in achieving fullprocessing efficiency and hence, a high metalrecovery. This is particularly important whenrecycling the machining chips, which have a highsurface area per unit volume. Even small residuesof water/oil soluble fluid will have a significantimpact on metal recovery. Such contaminantsinterfere with the bronze and may lessen the jointstrength or cause failure. For this reason the firststep of the flow sheet is the pre-treatment.

For good bronze melting into steel support arenecessary small and uniform chips. The machiningchips have diverse forms. The types of chips arecategorised or subdivided into followingcategories, Figure 3.

To obtain the uniform chip size, these arecrushed. The cleaning can be considered to be atwo-stage process. In the first process of the stagethe centrifuge separation to minimized water andcoolants is applied. Also the magnetic separation toeliminate ferrous parts is sometimes necessary.Second process of this cleaning stage is based onthermal processes. The object is to remove theorganic compounds from the surface of the chipsby converting them into a gaseous state. Thisprocess requires a low temperature. Any water thatis present within the chips will be vaporized. Then,at a higher temperature is removed the carbon-based deposit that remains on the surface of thechips.

The second condition for a quality process isreferred on the physical and chemical processesthat are developed in accordance with thermalconditions that are occurred at the melting of thecopper alloys chips and at the heating steelsupport.


23

a.

b.Figure 3: Chip forming classification: a - not

desirable; b – good [8]

Copper is considered a half noble metal but witha high solubility for oxygen in the liquid state.Generally an oxygen and hydrogen pick-up canlead to very negative effects on mechanical andphysical properties of copper and copper alloys.These gases have a high solubility in liquid copperalloys that decreases sharply during solidification.This can lead to bubble formation, i.e. porosity inthe solid material. Oxygen can also form cuprousoxide (Cu2O) above its solubility level thatimmediately reacts with the moisture of the airforming water vapors. Dissolved hydrogen andoxygen (or Cu2O) will react with water underextreme pressure in the lattice and will form cracksand lead to embrittlement [9].

Borax as neutral cover fluxes are used to reducemetal loss by providing a flux cover. This melts atcopper alloy melting temperatures to provide afluid slag cover (borax melts at approximately7400C). The glassy fluid cover flux is used also.This flux agglomerates and absorbs nonmetallicimpurities from the input material (oxides,machining lubricants, and so on). A reducing fluxcontaining carbonaceous materials such as graphiteis used. Its principal advantage lies in reducingoxygen absorption of the copper and reducing meltloss. Carbonaceous flux material should always beused with copper alloys to avoid gaseous reactionswith sulfur or hydrogen from contained moisture.Use the copper-phosphorus metal mixed withbronze chips that are considered self-fluxing oncopper-base metals is utilized. Some of these

fluxes are mixed with bronze chips and somecovered the input materials after their preliminaryheating. Also the steel samples were heated beforethe filling with bronze chips and fluxes. Theamount of flux used was established in accordancewith their effect on the porosity and mechanicalproperties of bronze alloy. This is more that 5% ofmetal charge.

Figure 3: Preparation of the samples beforeintroduction in the heating furnace

The heating temperature ensures only themelting of the bronze chips and also for the bronzebounding on the surface of the steel. The heatingtemperature is very important for obtaining a goodadhesion of the bronze layer on steel surface. Thiswas established in accordance with the thermalprocess that is developed in the samples. Thethermal regime should ensure the melting of thebronze, the superheating of this melt fordeveloping the diffusion zone at interface withsteel support. Certainly, this was correlated withthe Cu-Al binary diagram and with the influence ofother elements (iron, nickel and manganese) thatare present in the aluminium bronze composition.After heating the samples were maintained atoptimum temperature and then were slowly cooledtogether with the furnace.

The experiments were carried at different worktemperatures between 1200 and 13000C. Theexperimental shows that heating of the samples at12000C cannot ensure a complete melting of thebronze chips, Figure 4.

The choice of 13000C as optimum temperatureis confirmed by the experimental samples, Figure5.

Figure 4: Bronze chips incomplete melted that wereheated at 12000C


24

Figure 5: Good adherence of bronze on the steelsurface for the bimetal sample obtained at 13000C,slowly cooled with furnace (surface after exposure tometallographic attack)

At surface of the support steel, the aluminiumbronze with nickel, manganese and iron wasadhered to obtain the bimetal. The appearance ofthe joint zone shows that the process wasconducted with optimum parameters. The surfacelayer had no defects such as oxide films, orporosity, Figure 6 and 7.

The bonding takes place during heating andmelting processes. In this way, with participationof diffusion processes good bimetals can bemanufactured.

Figure 6: Aspect of bimetal at interface (100X withoutexposure to metallographic attack)

Figure 7: Microstructure of bronze and steel supportafter of exposure to metallographic attack (100X)

4. ConclusionsUsually, the recycling of the chips presents

many difficulties [10, 11]. Firstly, the metal chipscan be classified in the category of hazardouswaste because have residues of water/oil solublefluid with significant impact on metal recovery.The generation of pollutants are considerable. Inthe method proposed some of these negativeproblems are suppressed. By other hand the metallosses at remelting are higher. The direct meltingof the bronze chips into steel supports for thebimetal manufacturing is possible. It is a cheap andsimple solution to transform the machinery wastesas final products that can satisfy the requirementsin industrial applications. The important factorsthat influenced the manufacturing process are thedimension of the chips and the cleaning. Also, theheating and the melting parameters in accordancewith the physical and chemical properties ofmaterials most correlate. Especially, all of thesehave particularity in respect to the materialsutilised, the manufacturing processes and thebimetal applications.

References

[1] Popovits, D., Subu, T., Bimetale, EdituraFacla, Timi oara, 1982.

[2] Radu, T., S. Constatinescu, S., Balint, L. –Materiale metalice rezistente la coroziune,Editura F.M.R., Bucure ti, 2004.

[3] Ciocan, A., Bratu, F., Chemical and structuralchanges for bimetallic materials obtained bythe welding process, The Annals of Dunareade Jos University of Galati, no2, 2005, p.15-22.

[4] A. Ciocan, A., Potecasu, F., Drugescu, E.,Constantinescu, S., Characterization of theDiffusion Zone Developed in a Bimetallic


25

Steel-Bronze at Interface, Materials ScienceForum; ISSN 0255 – 5476, Trans TechPublications LTD Switzerland, UK, USA III,ISSN 0255 – 5476, 2010, Vols. 636-637, pp556-563.

[5] Campbell, H., Aluminum bronzes alloys.Corrosion resistance guide, Publication 80,Copper Development Association, U.K., pp.1-27, July 1981, www.cda.org.uk.

[6] Aluminium Bronze alloys for industry, CDAPublication No.83, 1986.

[7] Ciocan, A., Scraps Recycling. Technology forProcessing and Valorization of Copper Scraps,Editura Funda ia tiin ific MetalurgiaRomân , Bucure ti, 2004; ISBN 973-8151-29-5.

[8] Standard test method of evaluating machiningperformance of ferrous metals using anautomatic screw/bar machine, ASTM E618,American Society for Testing and Materials,Philadelphia, USA.

[9] Oprea, Fl., Teoria proceselor metalurgice,EDP, Bucure ti, 1966.

[10] Gronostajski, J., Chmura, W., Gronostajski, Z.,Phases created during diffusion bonding ofaluminium and aluminium bronze chips,Journal of Achievements in Materials andManufacturing Engineering, Volume 19ISSUE 1 November 2006, p.32-37

[11] Copper Development Association, Cost-Effective Manufacturing Machining Brass,Copper and its Alloys, Publication TN 44,October 1992 – Revised for inclusion on CD-Rom, May 1994

http://www.cda.org.uk./


26


27

STUDY FOR THE VARIATION OF HARDNESS AND RESILIENCECHARACTERISTICS AFTER HEAT TREATMENT FOR STAINLESS

STEELS

Manuela Cristina Perju1, Drago Cristian Achi ei1, Carmen Nejneru1, Anca Elenargeanu1, Mihai Axinte1, Ion Hopulele1

1Technical University “Gheorghe Asachi” of Ia i-România, E-mail address: [email protected]

Abstract: The paper presents tests of the stainless steel. This material was subjected to heattreatments, afterwards conducted impact bend tests and hardness measurements. Using the data wasdrawn graphic variation of hardness and resilience depending on the condition of structural material(hardened, annealed, solution treated and aged).

Keywords: hardness, resilience, heat treatment

1. IntroductionMetallic materials structure researches are

obvious by the fact that structure is very importantdetermining the properties, the behavior ofmaterials in service, and their use areas. In modernmetallurgical technology, based on thoroughknowledge of the structure, it can be obtained newmetallic materials, with mechanical and chemical-physical features precise, requested the use of thesematerials [5].

Material behavior is determined by its reactionon stress. Material property is defined as measuredbehavior on given test. By the exterior stressesthere are three properties categories:

- Mechanical properties, which reflects thematerial behavior on the mechanical forces.- Physical properties which measures thematerial behavior, under the temperature field,electrical or magnetic field effect.- chemical properties which characterizes the

material behavior in an more or less aggressiveenvironment [6].

The components behavior regarding mechanicalstresses produced by the exterior forces depends onthe certain specific material properties, calledmechanical properties.

Usually, mechanical properties for a metallicmaterial is determined by specific tests, that meanssample testing (some parts with well definedconfigurations and dimensions), in adequateconditions for the needed properties, [8]. Withmechanical test aid, qualitative data on the materialbehavior is obtained. Also with mechanical test aid

we can obtain the physical quantities, calledmechanical characteristics, which can be used asquantitative parameters for mechanical properties.

2. Experimental results

For the experiment stainless steel was used.Standard samples were heat treated to modify theirproperties. Afterwards the samples were tested byimpact bending and hardness.

2.1 Austenitic stainless steel analysisAustenitic stainless steel is used in different

domains (fig.1), [10].

Figure 1: The use of austenitic stainless steel.

Energyproduction

Transports

Environmentprotection

Chemicalindustry

Medicin andpharmacy

Machinebuilding

FoodIndustry

Use stainlesssteel

mailto:cryss_ela:@yahoo.com


28

Austenitic stainless steel, X5CrNi18-10(AISI 304), STAS EN ISO 3651-2 was subjectto chemical composition determination, presentedin table 1.

Table 1: Steel chemical composition, %C Si Mn Cr Mo Cu Ni

0,025 0,49 1,6 18,38 0.43 0.49 8.12

Foundry Master Spectrometer qualitativeanalysis was used.

2.1.1 Heat treatment applicationIn order to modify the resilience and hardness

properties, a solution treatment was applied [3, 4],followed by aging heat treatment, using the Uttisheat treatment furnace.

Solution treatment temperature was 1000°C,heating the samples with the furnace, maintaining30 minutes, and the cooling made in warm water at60°C. Heating rate was 10°C/minute.

Figure 2: Solution treatment and aging treatmentdiagram at 400°C.

Figure 3: Solution treatment and aging treatmentdiagram at 600°C.

After the solution treatment, followed the agingheat treatment, at two temperatures:

- aging heat treatment, at 400°C, maintaining 2hour, followed by air cooling. Aging heattreatment diagram in figure 2.

- aging heat treatment, at 600°C, maintaining 2hour, followed by air cooling. Aging heattreatment diagram in figure 3.

2.1.2 Hardness testingThe tests for hardness were conducted on

Wilson Wolpert hardness test instrument. TheBrinell method was used. The pressing force was187,5 daN, with steel ball 2,5 mm diameter.

The samples hardness was tested before theheat treatments; data in table 2. For a betterprecision on hardness values an arithmeticalaverage for three tests was calculated.

Table 2: Untreated and heat treated stainless steelhardness values

Stainless steelUntreated

Hardness, HB

174,1 177,0 175,3Average: 175,46

Stainless steelSolution treatment

Hardness, HB

163,3 164,9 161,5Average: 163,23

Stainless steelAging heat

treatment at 400 C

Hardness, HB

208,0 221,6 216,7Average: 215,43

Stainless steelAging heat

treatment at 600 C

Hardness, HB

188,2 193.4 191,6Average: 191,06

175,46 163,23

215,43191,06

0

50

100

150

200

250

HB

untreated solutiontreated

aged 400 aged 600

Comparative analysis of hardness

Figure 4: Comparative analysis for stainless steelhardness untreated and heat treated.


29

Comparative hardness analysis is presented infigure 4.

2.1.3 Impact bending testTo achieve impact bending tests were used

standardized samples, parallelepiped shapedimensions: 100x100x550 mm: samples with “V”shape indentation.

Testing instrument is Charpy-type pendulumhammer. The test is conducted as follows:

- positioning the sample on the fixed bearingdevice, so as to be hit right behind the indentation;

-hammer rises releasing its action space;-positioning the pointer at its maximum value;-fall firing;-after sample breakdown the hammer is stopped

with a computer command;-analyzing the samples.After the impact bending, the untreated

stainless steel and the heat treated one were subjectto macroscopic analyses; this is for highlightingthe fracture character. The fracture character couldbe brittle or ductile [1, 7, 9]:

-the brittle section looks crystalline grain andbright.

-the ductile fracture section (tenacious), lookingfibrous, mat.

Austenitic steel, beside the corrosion resistance,is characterized by high resilience, tenacity andstrength.

The untreated stainless steel, X5CrNi18-10(AISI 304), was called group 1 (the grouprepresents the average from three samples tests, atimpact bending test) (figure 5), and the average forthe two tests gives a high value for resilienceKV=357,81 J.

Figure 5: Macrostructure for untreated stainless steeltested at impact bending.

Austenitic stainless steels are generallycharacterized by low yield and good cold plasticdeformation properties. They present a strongerstress hardening on cold plastic deformation due tomartensite formation in metastable austenite [7].

The fracture structure is determined by theexistence of hollows or indentations. Theycharacterize the material ductile fracture and theyhave a concave structure resulted from vacanciesfracture, which are initiated and grows during theplastic deformation process. From a structuralpoint of view is presented as a solid solution withprecipitates, chemical compounds like complexcarbides, and intermetallic compounds.

According to the corresponding ternarydiagrams Cr carbides are of the form: Cr3C2, Cr7C3,Cr23C6, Mo compounds like: Fe2MoC, Fe2Mo,Mo2C, and Mn compounds like: Mn3C, Mn5C,Mn7Cr3.

In figure 6 is presented the group 2 fracture.Group 2 represents the solution treated materialfollowed by artificial aging, at 400 C. Thismaterial was subject to impact bending test,obtaining a lower value for the resilienceKV=193,84 J.

Solution treatment generally seeks plasticityvalues needed for processing by cold forming andaims at removing the secondary phase crystalstructure (complex carbides of complex elements).Resilience decreasing is possible by increasinghardness from aging treatment.

Figure 6: The heat treated stainless steelmacrostructure – solution treatment 400 C agingsubjected to resilience test.

The stainless steel samples, solution heattreated and aged at 600 C resulted with a very hightoughness, (group 3), having a maximum resilience


30

at 375 J. Samples were partially fractured (figure7) when hitting with the force of 300 daN.

The reason for this is given by artificially agingat 600 C, when the temperature exceeds thesolubility variation line, so that the precipitatesremain dissolved in material, even on cooling inair. The obtained structure will be a supersaturatedsolid solution, soft and plastic, with a very highresilience.

Figure 7: The heat treated stainless steelmacrostructure – solution treatment 600 C agingsubjected to resilience test.

357,81

193,84

375

0

100

200

300

400

[J]

lot 1 lot 2 lot 3

Comparative analysis of resilience

Figure 8: Comparative analysis of resilience.

Comparative analysis of resilience is presentedin figure 8.

3. Conclusions1) The heat treatment importance is given

through their application.A structural transformation occurs in materials,

which alters the properties and gives themopportunities for rational use. Metallic materialssubjected to a correct heat treatment have a relativelow cost and higher durability.

2) After the hardness testing it was found thatthe hardness value for austenitic stainless steel is175,46 HB, and after the solution treatment thehardness value decreased (163,23 HB), correspondingto this treatment. Because the supersaturated solutiondecomposition occurs during maintaining, at thebeginning the hardness grows gradually, as theprecipitation of crystals related to the matrix. Sincethe consistency fracture between the precipitatesand the matrix, and the extent of their growth bycoalescence, the hardness decreases.

3) The stainless steel hardness values, artificiallyaged, are: at 400°C, 215,43 HB, and 600°C, 191,6HB, thus obtaining a major cases difference.

4) The impact bending test, on austeniticstainless steel gives a resilience value KV=193,84J for the heat treated group, followed by agedtreatment at 400°C. The conclusion was thatsamples from this group have a much better valuethan other groups from this class of steel.

4. References[1] u Gheorghe, Propriet ile materialelor

metalice, Note de curs, 2007-2008.[2] Geru N., Teoria structural a propriet ilor

metalelor, Editura Didactic i Pedagogic ,Bucuresti, 1980.

[3] Hopulele Ion, Alexandru Ion, G lu Dan-Gelu, Tratamente termice i termochimice, volI, 1983.

[4] Hopulele Ion, Alexandru Ion, G lu Dan-Gelu, Tratamente termice i termochimice, volII, 1984.

[5] Mitilea Ion, Bud u Victor, Studiul metalelor,Editura Falca, Timi oara, 1987.

[6] Verme an G., Tratamente termice. Îndrum tor,Editura Dacia, Cluj-Napoca, 1987.

[7] Schumann, H., Metalurgie fizic , EdituraTehnic , Bucure ti, 1962.

[8] Shackelford F. J., Introduction to materialsscience for engineers, Macmillan PublishingCompany, New York, 1991.

[9] Zaharia Luchian, Teoria deform rii plastice,Editura „Gh.Asachi”, Ia i, 2001.

[10] http://zeus.east.utcluj.ro/~hvermesan/doctorat/cap1-1.html

http://zeus.east.utcluj.ro/%7Ehvermesan/doctorat/


31

AN INTELLIGENT SYSTEM SEARCH FOR INDOOR COLORRECOGNITION: A COMPARISON

Mário António Ramalho 1

1TU Lisbon, IST, IDMEC,Av Rovisco Pais, 1049-001 Lisboa, Portugal

[email protected].

Abstract: This work addresses a map-based method for the localization of mobile robots (wheeled orlegged) in indoor environments, using monocular vision. This method is comprised of four stages: imageacquisition, image feature extraction, image and model feature matching and camera pose computation. Aspecial attention is given to the matching phase. In particular we research the advantage of including aretinex pre-processing stage in neural networkthe color classification schemes. Results presented suggeststhat this is correct.

Keywords: Image Processing Retinex Neural Networks

1. IntroductionAutonomous guided vehicles are increasingly

common these days and will be more in a nearfuture. Effectively their use has been extended todomestic applications, such as cleaning robots orfor assistance of disabled persons, but also inpublic buildings, such as museums, serving asvisitor guides. These applications all have incommon the fact that they are indoorenvironments. And many more applications can beenvisaged. In these environments, the robot mustbe autonomous, being able to fulfill a navigationaltask, which involves reaching a goal with nohuman intervention, while at the same timeavoiding obstacles and people..

There are mainly two types of AGV’s, free-pathor free-ranging (1). Fixed path guidance refers to aphysical guide path (e.g., wire, tape, paint) on thefloor that is used for guidance, and free-rangingguidance has no physical guide path, thus it iseasier to change the vehicle's path (in software),but absolute position estimates (optical or laserguided) are needed to correct dead-reckoning error(2)

A human do not use neither a path way on thefloor for navigation (eventually for orientation),neither requires a laser to sense the area in front.

Rather than that it uses vision, from which he canextract the relevant information for the short pathin front, and memory for getting orientation andreference points.

We believe that a robot can work with thisinformation only.

To be autonomous, the robot must be able tocompute and update position and orientationrelative to a fixed global frame.

This work addresses a map-based method forthe localization of mobile robots (wheeled orlegged) in indoor environments, using monocularvision. This method is comprised of four stages:image acquisition, image feature extraction, imageand model feature matching and camera posecomputation. A special attention is given to thematching phase.

Results are presented for a real indoorenvironment, suggesting being adequate towardsthe envisaged applications

mailto:mar:@dem.ist.utl.pt


32

2. Previous workThe problem of indoor localization has been

addressed by a variety of techniques. Among themmap based approaches techniques are often used.Typically there are three main types of maprepresentations: topological maps, grid maps andgeometric models. The environment model caneither be pre-stored or built online simultaneously.

When performing map-based localization, anusual approach is to match the local representationof the environment, with the model map.

The matching phase is a critical step, since itinvolves determining the correct correspondencebetween image features and model features. Toaddress the problem of map-based robotlocalization using visual sensors, Kosaka et al (3)proposed a method where robot orientation andposition were updated from initial conditions,according with navigational commands. The newrobot pose is used to project model features on theimage plane and compare them with the actualimage being viewed. Uncertainty on the new robotposition and orientation is propagated to predictuncertainty zones in the image, where a search forcorresponding features is required. This methodrequires a high sampling rate and not every imageacquisition may provide enough information toaccurately update the robot pose. Talluri et al (2),introduced the notion of Edge Visibility Regions.These are used to restrict the matching process to asmall subset of features in the global model.However it only focuses on horizontal features,representing rooftop edges.

In this work a map-based localization methodusing monocular vision is researched. The focuswill be on the critical matching phase, betweenimage and model features. The approachresearched is based on color correspondence. Theenvironment model is constructed offline and pre-stored. The map is of the geometric type andcontains the main parts of the environment, whichare not supposed to move, such as wall and dooredges.

3. Indoor Map Building

Indoor environments considered in this workare man-made and semi-structured. These aretypically delimited by flat surfaces, such as walls,ground and ceiling. The intersections betweenthese flat surfaces tend to be straight lines, theedges.

Figure 1 Interior scene with light gradient

The built model contains the main parts of theenvironment that are not supposed to move, suchas walls and door edges, and is of the geometrictype, that is, it contains the lengths of each edgeand its position relative to a global fixed frame.

Figure 2 Interior scene model in VRML

4. Color Models

The basic system is color is the RGB (red-Green and Blue), used by the majority of thecameras. A more suitable system for computerprocessing is the HSV (Hue –saturation-brightness). The later correspond to an axe rotationfrom the RGB, making the lightness value to bedefined by the extreme points (black and white).

In computer vision segmentation refers to theprocess of partitioning a digital image into multiplesegments (sets of pixels). The goal of segmentationis to simplify and/or change the representation ofan image into something that is more meaningfuland easier to analyze (5). Image segmentation istypically used to locate objects and boundaries(lines, curves, etc) in images.


33

More precisely image segmentation is theprocess of assigning a label to every pixel in animage, such that pixels with the same label sharecertain visual characteristics.

In the particular case of this work, imagesegmentation refers to the process of identifyingcomponents in the image. And one logicalapproach is to search for color areas, which ininterior environments can be many times linkedwith a particular type of objects (e.g. doors arecoherent in color).

This approach seems logical; theimplementation is a little more difficult, as theillumination gradients make the colors vary, and inparticular the triplet of values. This is not evidentall the times due to the ‘quality’ of the humanvisual system. Color constancy mechanismscompensate this characteristic. Unfortunately theseare not reproducible in computer vision. So welook for a method to correct the brightness thatallows that similar color appears similar in theimage.

4.1. Retinex model of visionIn the early 70’s Land and McCann (5), (6)

introduced the retinex model for the computationof lightness. Since then, variants have beenpresented mainly to improve the computationalefficiency of the model. Among them, Funt et al(7) (8) have presented a MatLab implementationwhich is used through this paper.

Retinex calculations aim to calculate thesensory response of lightness. According to theauthors, “For testing the retinex model it is crucialthat the data be calibrated in the sense that theimage digit values must be a logarithmic functionof scene radiance and they must be representedwith sufficient precision”. This is a problem, as weuse, and intend to use standard cameras.Nevertheless we are investigating the ability tohomo

Retinex algorithms present this characteristic,and thus Funt et al algorithm was tested, presentinga poor performance. However during the testsperformed it seems to perform well in some imageswhere the color didn’t have much variation.

A small variation from the strategy proposedgives much better results. Effectively, retinexalgorithm is being using for reassigning of thebrightness component of the image. We split theimage in the HSV sub-matrices, apply the methodto the v-image (as monochromatic), andreconstruct the image with the H, S and, correctedV.

4.2. Artificial Neural NetworksArtificial Neural Networks (ANN) are

structures that mimic the human brain and whohave the capability to learn. They have been usedfor edge detection (10)and edge detectionarbitration, among other applications. (10) (11)

Among the several approaches that could beused in implementing this process, an artificialneural network seems to be particularly suited tothis job.

Effectively, artificial neural networks canhandle incomplete or corrupted sets of data thusthey can be applied to the recognition of images,can be used to infer the position of missing edgesor misplaced edges based on the knowledgeapplied by the different edge detection techniquesdescribed or color identification. To this purpose amulti-layer back-propagation neural networks arebeing used. These types of networks are capable ofreproducing an input/output relation, learned froma repetitive exposition to a set of examples. Theyalso have an inherently parallel structure allowingfor a parallel implementation.

4.3. ResultsExamples of the execution of the algorithm

are presents in the following pictures. Figure 3until 8 are from interior scenes, corridors atTechnical University of Lisbon, under normallight. Images were taken with a digital camera,Olympus 750UZ, with 4 Mpixels, and resized.They are originally jpeg images from the camera.No further processing was done.

The processed images present more objects,darker areas are visible, without ‘burn’ the morebrilliant areas.

Figure 3 Interior Scene - Original


34

Figure 4 Interior Scene – Processed

Figure 5 Interior Scene - Original

Figure 6 Interior Scene - Processed

ANN are trained with a sample of the colors.The colors are manually handpicked. The networkis then applied to the original image and theRetinex processed image

Visually, colors seem to be morehomogeneous, without shifting from the originalhue. Details in dark areas are visible. Tones seemto be more coherent.

Examples are presented in Figure 7 through11. The Retinex processed images will presentmore colors detected and better detail. By theopposite original color images are less consistentin the results.

Figure 7-Corridor after retinex (Figure4) processedby the ANN

Figure 8- Corridor (Figure4) processed by the ANN

Figure 9 Office corridor after retinex processed byANN

Figure 10 Office corridor processed by ANN


35

5. Conclusions and Work under Development

This work presented studies towards acollaborative frame-work for the localization andcooperation of mobile robots in indoorenvironments. We studied one method for coloridentification for a segmentation of interior scenes.The integration of the system into a robust coloridentification algorithm, e.g. Neural Networks, wasstudied. It can be seen from the examples that thepreprocessing with the retines algorithm willincrease the ann function. Results are verypromising.

6. AcknowledgmentsThis work was supported by Fundação para a

Ciência e a Tecnologia, through IDMEC underLAETA

7. Bibliography[1]. WikiMehda . http: //www.wikimheda.org

/wiki/AGV. [Online] [Cited: 09 21, 2010.]http://www.wikimheda.org/wiki/AGV.

[2]. Industrial Trucks. About the MHETaxonomy. [Online] [Cited: 09 21, 2010.]http://www.ise.ncsu.edu/kay/mhetax/TransEq/IndusTr/index.htm#Light%20load%20AGV.

[3]. Kosaka, A. and Kak, A.C. Fast Vision-Guided Mobile Robot Navigation Using Model-Based Reasoning and Prediction ofUncertainties. Computer Vision, Graphics andImage Processing - Image Understanding.1992, Vol. 56, 3, pp. 271-329.

[4]. Talluri, R. and Aggarwal, J.K. Mobilerobot self-location using model-image featurecorrespondence. IEEE Trans. Rob. Autom.2005, Vol. 12, 1.

[5]. Shapiro, L.G. and Stockman, G.C.Computer Vision. New Jersey : Prentice-Hall,2001.

[6]. McCann, Edwin Land and John. Methodand System for Image reproduction BAsed onSignificant Visual Boundaries of OriginalSubject. 3 533 360 1971. US Patent.

[7]. Edwin Land, John McCann. Lightnessand Retinex Theory. Journal of the OpticalSociety ogf America. 1971, Vol. 61, 1.

[8]. Briean Funt, Florian Ciurea, JohnMcCAnn. Retinex in MatLab.http://www.cs.sfu.ca/~colour/publications/IST-2000/retinex_mccann99.m. [Online] [Cited: Sep15, 2010.]

[9]. Brian Funt, Florian Ciurea, and JohnMcCann "," Proceedings of the IS&T/SIDEighth Color Imaging Conference: ColorScience, Systems and Applications, 2000, pp112-121. Retinex in Matlab. Proceedings ofthe IS&T/SID Eighth Color ImagingConference: Color Science, Systems andApplications. 2000, pp. 112-121.

[10].Integrating Parallel edge detection Schemesusing Neural Network Arbitration. Ramalho,M and Curtis, KM. s.l. : ICSPAT'93, USA,1993.

[11].Edge detection using neural networkarbitration;. Ramalho, Mário and Curtis, K.s.l. : IEE Conf. Pub. 1995, 514 (1995),, 1995.

[12].Ramalho, M and Curtis, KM. NeuralNetwork Arbitration of Edge Maps. [bookauth.] A, Jane, M and Marini, D. (Eds),Gloria. Transputer Applications andSystems'94;. s.l. : IOS Press, Netherlands,1994.

[13].Borenstein, J., Everett, H.R. and Feng, L.Navigating Mobile Robots: Systems andTechniques. s.l. : Wellesley, Mass.:AK Peters,1996.

[14].Borenstein, J., Everett, H.R. and Feng, L.Navigating Mobile Robots: Systems andTechniques. Wellesley, Mass : AK Peters,1996.

[15].Aider, O.A., Hoppenot, P. and Colle, E. Amodel-based method for indoor mobile robotlocalization using monocular vision andstraight-line correspondences. Robotics andAutonomous Systems. 2005, Vol. 52, 2, pp.229-246.

[16].Kim, J.S. and Kweon, I.S. EstimatingIntrinsic Parameters of Cameras using TwoArbitrary Rectangles. Proc. 18th Int. Conf.Pattern Recognition. 2006, Vol. 1, pp. 707-710.

[17].Simsarian, K.T., Olson, T.J. andNandhakumar, N. View invariant regionsand mobile robot self-localization. IEEETrans. Rob. Autom. 1996, Vol. 12, 5, pp. 810-815.

[18].Valente, Filipe. Movement and ImagePrediction. Lisbon : MSc Thesis, IST,TULisbon, 2009.

http://www.wikimheda.org/wiki/AGV.

http://www.ise.ncsu.edu/kay/mhetax/TransEq/In

http://www.cs.sfu.ca/%7Ecolour/publications/IST-


36


37

MATERIAL SIMULATION OF POLYMERS CONTAININGSPHERICAL STRUCTURES

Achim Frick1, Melissa Matzen2, Timo Dolde3

1,3Aalen University, Inst. Polymer Science and Processing iPSP, Aalen (Germany),2University of Bremen, Bremen (Germany);

Email: [email protected]

Abstract: One of the keys for successful polymer product development is proper material selection. Itis recommended to apply tailor sized polymers in order to realize best product performance with respectto the requirements. Material simulation is a possibility to predict material data from modified polymers,e.g. particle filled polymer, for strength calculation purpose, which is eminent part of the product designprocess. A method to simulate stress-strain behaviour of polymers containing spherical structures isshown for TPU material.

Keywords: Polymer, Morphology, Material Simulation, Computer Aided Product Development

1. General Introduction

A successful polymer product requires a propermaterial selection. Therefore it is necessary tointegrate both a material based strength calculationand structural optimization as the two mainelements of a computer aided product developingprocess, Figure 1. Material simulation is a possibletool to provide material data from any polymermodification for strength calculation.

Figure 1: Product development process from the idea tothe final production

Material properties of thermoplastic polymersdepend on the one hand on their thermal history,due to the processing conditions during moulding,

as well as on a possible tailor sizing by adding e.g.fillers or reinforcement to the polymer matrix.

To study the possible consequences of suchvariations for the strength behaviour of thepolymer a material simulation is recommended. Itprovides a fast and helpful method for advancedproduct development.

2 Investigated Material

Segmented thermoplastic polyurethanes TPUbelong to the class of TPE and they show soft andrubberlike behaviour at RT. Nevertheless theyperform comparably strong, are oil resistant andbehave superior abrasion resistant, whichrecommend them for e.g. sealing applications. Butthe properties of TPU are highly influenced bytheir morphology developed during the meltprocessing and subsequent solidification process.The stress-strain behaviour of TPU is observed tobe strongly dependent on the present crystallinedomain structure, which shows to be sphericalshaped. Also for semi crystalline polymers thespherical crystalline fraction determines the enduse properties of the material. Concerning that amaterial simulation can provide to better materialunderstanding basically and can also suggestoptimal morphological structure for the polymericmaterial in order to perform best inapplication.Thermoplastics can further containfillers such as glass balls or minerals particles toenhance the strength and stiffness of the basic

mailto:Achim.Frick:@htw-aalen.de


38

material. Material simulation gives a possibility topredict the material behaviour of modifiedpolymers for strength calculation purpose. It canprovide information for improved materialselection and qualification for a demanded productunder development.

Material simulation requires materialparameters for input in order to model a polymericmaterial as a part of product development process.There are two possibilities to obtain theparameters. Either a theoretical based materialsystem can be generated or the material data arereceived by analyzing a real polymer sample.

3 Modelling

3.1 GENERATED STRUCTURETo study theoretically on a possible influence

of spherical inclusions in a polymeric materialmatrix a determined distribution of the sphereswith a chosen radius, Figure 2, is input into thesoftware, e.g. Digimat® from Xstream-Engineering, and a representative volume element(RVE) is generated. The software then providesthe stress-strain curve of modeled material. It is away to simulate material behavior depending on itsspherical inclusion fraction and inclusion size anddistribution, where the inclusions may representspherulitical crystalline structure in semicrystalline polymers or any filler with an aspectratio near 1.

Figure 2: Virtual material simulation based ongenerated structure

A virtual section view of this RVE can be takento visualize and study the morphology and it alsocan be compared with a micrograph taken from areal cross section of a material sample microstructure, in order to validate how realistic themodeled morphology is.

3.2 DERIVED STRUCTURETo approach on the evidence of a certain

morphological structure on the mechanicalbehavior of a polymeric material sample it ishelpful to model this real existing materialmorphology by means of material simulation in areverse procedure.

As a starting point a light microscopy (LM) oreven transmission electron beam microscopy(TEM) micrograph is observed and the distributionof the detected spherical inclusions, e.g. spherulits,is identified, see Figure 3.

Figure 3: Real material modelling based on the derivedstructure

Using these information received a RVE isgenerate and is basis for further materialsimulation procedure, in order to establish thestress-strain behaviour of the modelled materialstructure.

The challenge hereby is on one hand to analyseproperly the inspected morphological structure inthe 2D micrograph and on the other hand toreconstruct the size and dispersion of the observedinclusions in the regarded volume.

3.2.1 Image analysis of micrographsA computer assisted image analysis of

morphological micrographs seems problematicregarding a proper detection and reproducibleresults.

As a basic requirement for computer assisteddetermining the diameter of a spherical structureembedded in a grey environment it needs to have asufficient contrast between the both structures nexteach other. The grey values of each of them mustdiffer strongly in their brightness. Otherwise anautomatic detection is not possible, whether due toinsufficient contrast of the structure intended to getmeasured or due to random variation of the


39

structure in geometry and deviation from thepredefined geometry.

If an automatic evaluation of the micrograph ofa material morphology is not feasible and theinspection is done visually then there is nonecessity for any contrast optimization andcorrection of inhomogeneous grey valuesdistribution in the image due to the varying sectionthicknesses of the previous inspected microtomesection under transmission light microscope. Toavoid inconsistencies in the measuring results byvisual evaluation of morphologies it is recommendinvolving always the same person for inspectiontask.

3.2.2 Relation between measured circle radiusand the radius of the sphere

When observing on a cross section of a materialfilled with spherical inclusions randomlydispersed, and even all the inclusions have samediameter, the circle shaped cross sectional areas ofthose inclusions show different diameters.

The mathematical calculation of the diameterdistribution of spherical inclusions in a matrixmaterial from the estimated radius distribution ofthe measured circle shaped areas is mathematicallyan ill-defined problem [3], which needs to besolved. It is quite ambitious and requires expressedmathematical skills. It is like “tomato saladproblem” and to estimate the diameter of a tomatoby measuring the diameter of its slices. Due to thedifficulties to solve this mentioned problem someapproximation methods have been suggested.Wicksell [4] was the first who described in 1925 ina treatise titled “corpuscle problem” the relationbetween a measured radius of an arbitrary sectionof a sphere and its real radius. Bach [5] adopted in1958 the problem and gave refined solutionregarding the size distribution of inspectedspherical sections in translucent slices of finitethickness.

The approach of Bach [5] uses the secondVolterra integral to describe the relationshipbetween the size distribution of spherical radiusand circle radius, see equation 1,

R = 2 R RR R

R R+ R

(1)

It means:G R Sphere radius distributiong R Circle radius distribution

F Cross section areaR Sphere radiusR Circle radius

Sample thickness

The calculation method according to Fullman[6] is simple compared to the suggested method ofBach. Fullman assumed that the sphere radius isidentical to the average radius of Gaussiandistributed circle radii, Figure 4, which are detect.

Figure 4: left: the sphere size distribution (distinct);right: detected cross section radius distribution

According to that assumption, Fullmanconcluded the sphere radius to be equalthe circle radius .

(2)

Mathematically, the Bach method is relativelyelaborative, but it provides excellent results. Thesolution of Fullman, which bases on the simplestapproach to determine the relationship between theradius of a two dimensional circular section andthe three dimensional sphere radius, provides atminor calculation effort also useable results.

3.2.3 Relation between measured cross sectionalarea and the volume

To simulate a filled material it is essential toknow on its composition (filler volume fraction) aswell as on the inclusion size and distribution. Sothe content volume of filler has to be estimatedfrom a micrograph of the morphological structureof a sample, Figure 5. It is necessary to estimatethe 3D volume of an inclusion from the diameterof its two-dimensional (2D) cross sections. Figure5 depicts the problem.

According to [7] the sum of circular areas is equivalent to the total volume of the

spheres . Thus the following equation 3expressing the relationship between the areafraction and volume fraction is valid.

Sphere radius [ m] Circle radius [ m]


40

=(3)

Figure 5: Left: micrograph of the cross sectional areaof samples with inclusions; right: reconstructed volume(F1: circular area, V1: sphere volume)

By means of image analysis the cross section ofa material containing spherical inclusions both, thesphere radius distribution of the inclusions, as wellas the volume fraction of the spheres, can beestimated.

4. TPU Material Modelling

Material modelling requires first amathematical equation which is adapting to thespecific material deformation behaviour and thusfits to its stress-strain behaviour best. For instanceMooney-Rivlin equation can be used for modellingelastomeric materials such as TPU. Themathematical concept behind this model is a hyperelastic approach.

In case of segmented thermoplasticpolyurethanes TPU the macromolecular chainsexist from sequentially linked hard and softsegments. The hard segments can build rigiddomains in the soft segment matrix, depending ona possible phase separation between both, which isruled by the melt processing conditions duringmoulding. If the material will be processed at highmelt temperature the hard segments does notsegregate and shape distinct domains, the TPUtherefore shows almost transparent. In such casethere is no expressed crystalline super structure inthe transmission light micrograph visible. Thematerial is considered to be amorphous in itsmorphological structure.

Figure 6 shows a measured stress-strain curveof an amorphous TPU sample (nominally 94 ShoreA durometer hardness) processed at a melttemperature of 250 C and its best fit curveconsidering Mooney-Rivlin equation. The materialmodel describes the measured data quite well,especially at higher strain more than 40%.

Based on the fit curve, considered as thematerial model representing the amorphous matrixmaterial, it is now possible to study theconsequences of a certain crystalline volumefraction in the amorphous matrix on themechanical properties of the considered TPU bysimulation. The material simulation of partiallycrystalline TPU has to approach the stress-strainbehaviour of a RVE containing an amount of hardand spherical shaped inclusions, which representsthe crystalline structure.

Figure 6: Modelling stress-strain behaviour ofamorphous TPU by Mooney-Rivlin approach (fit curvecalculated with software ABAQUS 6.7)

The further simulation of partially crystallinematerial requires the mechanical properties, e.g.the Young’s modulus of the pure hard segments ofTPU. Unfortunately the mechanical properties ofthe pure hard segments or even of the harddomains cannot be determined easily because it isalmost impossible to approach those separately andsimply. Therefore it needs to find approximatematerial data for crystalline structure.

It is well known that a homo-polymericpolyoxymethylene (POM-H) represents a stronglycrystalline polymer and has crystalline content upto 90% [8, 9, 10]. Thus in a first order approachthe TPU hard segments are considered to have asimilar stiffness than POM-H. The Young´smodulus is taken to be 3000 MPa for the hardsegments.

4.1. SIMULATION OF CRYSTALLINECONTENT

In order to simulate the influence of hardsegment domains in amorphous TPU materialrepresentative volume elements (RVEs) of TPUwere generated containing spherical shaped

0 20 40 60 80 100 120 140 1600

2

4

6

8

10

12

14

Stre

ss [M

Pa]

Strain [%]

Material model Mooney-Rivlin tested material

1

1F1 V1 1

11


41

crystalline fraction content. The RVE generationprocedure was as described before and RVEcreated using Digimat software. The diameter wasconsidered 10 m for simulating crystallinestructure when the volume fraction content variesfrom 0 to 30 vol%.

Figure 7 shows the influence of growingcontent of crystalline structure on the stress-strainbehaviour of TPU, the material becomessignificantly stiffer and its strength rises. Itbecomes obvious that the TPU must perform inservice application quite different depending on itspresent crystalline domain fraction. The stress-strain behaviour of this material depends hardly onthe melt processing conditions duringmanufacturing process.

Figure 7: Simulated stress-strain curves for TPU withdifferent vol.% crystalline domain content (d: spherediameter of inclusion)

Figure 8: Simulated stress-strain curves for TPU,amorphous state and with 20 vol.% crystalline domaincontent (mono and bi-sized inclusions, d: spherediameter of inclusion 1, D: sphere diameter ofinclusion 2)

Assuming the crystalline domain fraction intotal to be 20 vol% and it exists from 2 differentdiameter sized inclusions (10 m and 30 m) it iscalculate by simulation, that the bi-modalcrystallinity gains better material strengthbehaviour than a mono-modal one, Figure 7

5. Summary

Material simulations allow to model polymericmaterial with incorporated inclusions, such ascrystalline domains or fillers, in order to study theresulting mechanical behaviour. The data receivedcan be used for within the product developmentprocess and also for material science purposes.

The simulation possibility was approached onsegmented TPU material.

References[1] A. Frick, M. Mikoszek: Predispositions to failure

of Polyurethane Sealings: Relation betweenMorphology and Visco-elastic Properties, p. 413-414, full paper: S1205_A0273 in Gomes, J.F.S.;Meguid, S.A. (editors): Integrity, Reliability andFailure: challenges and Opportunities, EditionINEGI, 2009 (ISBN 978-972-8826-22-2)

[2] A. Frick, M. Mikoszek, M. Kaiser: Qualität vonTPU-Dichtungen: Über den Einfluss derMorphologie auf die Gebrauchseigenschaften inDichtungstechnik Jahrbuch 2010, S. 356-372

[3] A. Louis: Inverse und schlecht gestellte Probleme;Stuttgart; Teubner Verlag; 1989.

[4] S. Wicksell: The corpuscle problem, Amathematical study of a biometric problem;Biometrika; 1925, A-17; S. 84–99

[5] G. Bach: Über die Größenverteilung vonKugelschnitten in durchsichtigen Schnittenendlicher Dicke; Diss; Universität Gießen; 1959;S. 2ff.

[6] R. Fullman: Measurement of approximatelycylindrical Particles in Opaque Samples;Transactions AIME;1953; A-197; S. 1267ff.

[7] W.Rostoker, J.Dvorak: Interpreatation ofmetallographic structures; Academic Press Inc.;1965; S. 198.

[8] G. Becker, L. Bottenbruch, D. Braun: TechnischeThermoplaste: Polycarbonate, Polyacetate,Polyester, Celluloseester; München; HanserVerlag; 1992; S. 333.

[9] [Online] Wapedia: 5.07.2010; http://wapedia.mobi/de/Kristallisation_(Polymer).[10] G. Ehrenstein, G. Riedel, P. Trawiel: Praxis der

thermischen Analyse von Kunststoffen; München;Hanser Verlag; 2003; S. 74.

0 20 40 60 80 100 120 1400

5

10

15

20

25

30

Stre

ss[M

Pa]

Strain [%]

d=10; 0 Vol-% d=10; 5 Vol-% d=10; 10Vol-% d=10; 15Vol-% d=10; 20Vol-% d=10; 30Vol-% d=10; 40Vol-%

0 20 40 60 80 100 120 1400

5

10

15

20

25

30

Stre

ss[M

Pa]

Strain [%]

d=10; 0 Vol-% d=10; 20Vol-% d=10; 15Vol-%; D=30; 5 Vol-%

http://wapedia.mobi/de/Kristallisation_(Polymer).


42


43

MONTE CARLO PIVOTAL CONFIDENCE BOUNDS FOR WEIBULLANALYSIS, WITH IMPLEMENTATIONS IN R

Jurgen Symynck1, Filip De Bal2

1KAHOSint-Lieven,[email protected], [email protected]

Abstract: To gain expert insight in the inner workings and pitfalls ofcommercial lifetime analysissoftware, the authors created an open source alternative with asubset of analysis tools and made it freelyavailable as a package for the R statistical software, called the Weibull Toolkit for R. This articlefocuseson creating pivotal confidence bounds using Monte Carlo simulation for B-lives from a Weibull model.These bounds were suggested by Lawless (ref. [1]) and are recommended for small sample sizes of(nearly) complete databy Abernethy and Fulton(ref. [2], [3]). Fully functional and annotated R code ispresented, derived from the toolkit’s codebase. For the latest version of this document, check ref. [4].For more in-depth treatment of the Weibull analysis with R, check ref. [5].

Keywords: Weibull, R, pivotals, Monte Carlo, confidence bounds.

1 Introduction

1.1 The FATIMATProjectFATIMAT (FATigue In MATerials) (ref. [6])is

a completed PWO project (ref. [7]) supported byKAHOSint-Lieven (ref. [8]) that investigatedcheaper alternatives to servo hydraulic dynamictesting machines, like servo pneumatic (ref. [9])and electromechanical spindle actuator machines.

Products can be tested in various ways: forexample an engine mount for fixing car engines toa chassiscan be tested by cyclically compressing itbetween 10 [mm] and 20 [mm] on a Zwick/RoellEZ020 high speed electromechanical spindleactuator ata load frequency of 3 [Hz]. After sometime, the specimen will show signs of wear-out (inthis case oil leakage) that are detected by themachine. The test is then halted and the failuretime is recorded. This test is repeated (underidentical loads and circumstances!) for a number ofspecimens.

One of the subprojects of FATIMAT was toanalyze these failure data using the Weibulllifetime distribution for drawing conclusions on thegeneral reliability of the specimens.Commercialpackages such as Reliasoft’sWeibull++ (ref. [10])and superSMITHWeibull (ref. [3]) were evaluated,but for gaining expert insight the authors createdtheir own software: the WeibullToolkit (ref. [11])for the statistical software R (ref. [12]).

1.2 The Two-Parameter WeibullDistribution

The two-parameter Weibull model is widelyused in the field of reliability engineering, becauseit allows useful analysis with extremely smallsample sizes (two failures or less). The Weibullmodel covers many other distributions (eitherexactly or approximately) like the(log-)normal andexponential.Also, an informative graphical plotcan be created that helps to convey the analysisresults to non-statisticians like engineers and theirmanagers. In many cases, a two-parameterWeibull model is sufficient for accuratelydescribing failure data. Its cumulative distributionfunction (c.d.f.) is given by:

( ) = 1 e (1)

The ‘shape’ parameter indicates the type offailure: < 1 is a sign of infant mortality while > 1is a sign of wear out failures. = 1 suggests aconstant failure rate, not related to lifetime. The‘scale’ parameter , also called the ‘characteristiclife’represents the age at which 63.2 [%] of thespecimen have failed.

1.3 The R Statistical Programming Language

(from the Rhomepage, ref. [12]:) “R is alanguage and environment for statistical computingand graphics. It is a GNU project which is similar

mailto:jurgen.symynck:@kahosl.be

mailto:filip.debal:@kahosl.be


44

to the S language and environment which wasdeveloped at Bell Laboratories (formerly AT&T,now Lucent Technologies) […] R provides a widevariety of statistical [...] and graphical techniques,and is highly extensible.”

R can be downloaded with no cost from itshomepage and can be installed on most computers.It is essentially a console-like application wherethe user enters commands at the prompt. Multiplecommands can be scripted and stored in a plaintext file, making complete applications possible.Some graphical interfaces for R, and somededicated R code editors like Tinn-R (ref. [13]) areavailable, making R easier to use.

1.4 The WeibullToolkit in R

The WeibullToolkit is a package for the Rstatistical programming language. It was initiallyconceived for doing analysis on the very simplereliability problem of complete, uncensored data,but its feature set and capabilities are continuouslyupdated.In this paper, its code is demonstratedusing simplified toolkit functions: just copy andpaste the code in an open R console.More detailson the Weibull toolkit can be found under ref. [5]and [14].

The WeibullToolkit is hosted online atSourceforge (ref. [11]) and can be downloaded asan installable package. The source code can bebrowsed online (ref. [15]) or downloaded using the‘Git fast version control’ system (ref. [16]).

2 Entering Data in the R Console

Let us assume that eight engine mounts weretested under the previously mentioned conditions(cyclic compressive sine wave load alterationsbetween 10 [mm] and 20 [mm], 3 [Hz], sameenvironmental temperature and humidity, …).Therecorded observations are 149971, 70808, 133518,145658, 175701, 50960, 126606 and 82329 cycles.All specimen were tested to failure (in this case:failure by oil leakage) creating a so-calledcomplete dataset. The data is imported in R byopening the R console and entering the followingcode at the prompt:

### Entering failure data ###d <- data.frame( time=c(149971, 70808,133518, 145658, 175701, 50960, 126606, 82329), event=1)

This creates a table-like variable named d of the‘dataframe’class, with two columns: d$time andd$event.The latter represents the event at thetime of observation: here, all the specimens failed,corresponding with event ‘1’ or ‘died’. Just type dfollowed by <ENTER> at the prompt to displaythe contents of d(note that help is available fordata.frame and all other functions by typing?data.frame followed by <ENTER> at theprompt).

2.1 Censored Data

In many cases, mechanical-dynamical testingmeans only a few specimens are tested (3-8) untilfailure, after which the failure time is recorded.This creates a ‘complete’ or ‘uncensored’ dataset.Sometimes an upper test duration limit is enforced,after which unfailed specimen are taken from themachine and labelled as ‘survived’. This kind ofcensored data is called ‘right censored’ data or‘suspended’ data.

The WeibullToolkit can handle this type ofcensored data, but for simplicity, this (important)subject is not treated in this article and thepresented code does not support it; check out ref.[5] for a more detailed treatment.

3 CreatingaWeibullPlotThe goal of the Weibull plot is to provide a

useful 2D representation of the observations. Onthe vertical axis, the ‘unreliability’ of thespecimens is found while on the horizontal axisone finds the observation time. The doublelogarithmic scale of the Weibull plot's vertical axistogether with the horizontal axis’ logarithmic scalemakes the Weibullc.d.f. (eq. [1])appear as astraight line (fig. [1]).

3.1 Median Rank RegressionTo create the straight line representing the

sample’s and , a vertical plotting position isassigned to the ordered observation times(‘ranking’).These ranksare equivalent to the‘unreliability’ of the specimens’ population. Then,by means of simple linear least-square regressionon transformations of the observations and medianranks, the Weibull parameter estimates arecalculated from the data points.Several methodsexist for the rank assignment (Hazen’s, Bernard’s,mean ranks) but here the inverse of incompletebeta function is used, which is considered the bestmethod.


45

For small and moderate sample sizes (2-100)with few or no censoring it is always best practiceto determinethe estimates usingmedian rankregression (MRR) in favour of other techniqueslike maximum likelihood estimation (MLE) (ref.[2]).Continue by running the following code toassign median ranks to the previously enteredfailure data (note that for this uncensored data thed$event column is not used or needed,simplifying the ranking process):

### Ranking failure data ###mrank.ob <- function(j,f){ r <- qbeta(0.5,j, f-j+1);r}mrank.data <- function(d){ n <- nrow(d) d$rank <- rank(d$time,"first") d <- d[order(d$rank), ] d$mrank <- mrank.ob(d$rank, n);d}print(d <- mrank.data(d))

The ranking is done by theqbeta() function,the inverse of the incomplete beta function and partof the standard libraries of R. The dataframed nowholds the median ranks for each observation in thed$mrank column, as shown below.

time event rank mrank6 50960 1 1 0.082995962 70808 1 2 0.201131198 82329 1 3 0.320518977 126606 1 4 0.440155203 133518 1 5 0.559844804 145658 1 6 0.679481031 149971 1 7 0.798868815 175701 1 8 0.91700404>

After ranking the observations, a line can befitted through the data points, calculating theWeibull parameter estimates.Continue by runningthe following code to fit a line through the datapoints:

### Median Rank Regression ###F0inv<- function(p){log(log(1/(1 - p)))}fwb <-lm(log(d$time)~F0inv(d$mrank),d)beta <- 1/coef(fwb)[2]eta <- exp(coef(fwb)[1])print(paste("beta=",signif(beta)))

print(paste("eta =",signif(eta)))

The lm() function fits a straight line by fittinglog(d$time) on a transformation of the medianranks.The F0inv()transformation, together withthe logarithmic transformation of the observationsallow treating the soughtWeibullc.d.f. as a straightline, making a line fit possible (the transformationscan be derived from eq. [1] and will result in

( ) log log ), which is of thesame structure as the standard line equation

).Note that here it is the observations that are

fitted on the ranks (X-on-Y regression), and notvice-versa (Y-on-X, as is standard in mostimplementations of least square fitting).It is goodpractice to fit the variable with the most variability(the observations) on those with less variability:the values of the median ranks are exactly definedand do not depend on the actual values of theobservations, only on their positions in the list ofordered lifetimes, and the total number ofobservations!

Call:...[1] "beta= 2.58128"[1] "eta = 132512">

3.2 The WeibullPlot

The WeibullToolkit automates all of the abovesteps. For convenience however,the stand-alonecode for a simplified version of the Weibull plot ispresented here. Because R lacks native support forautomatically transforming the vertical axis to thedouble logarithmic scale, this must be donemanually with the F0inv() function (logarithmictransformations are supported by the log=”x”parameter).It will be used frequently with mostgraphical functions in this article.Continue byrunning the following code to produce aWeibullplot for the given example:

### Simplified Weibull plot ###options(scipen=10) # no scientific # notationplot(y=F0inv(d$mrank),x=d$time, log="x",axes=F,lwd=2,cex=1.2, main="Engine mount cyclic test", xlim=c(5000,500000), ylim=F0inv(c(0.01,0.99)), xlab="time",ylab="Unrel. [%]")


46

curve(add=TRUE,lwd=2, beta*log(x)-beta*log(eta))ygrid <- c(1:9,seq(10,90,10),91:99)axis(2,F0inv(ygrid/100), ygrid,lwd=2);axis(1,lwd=2)abline(v=c(5000,seq(1e4,1e5,1e4), seq(1e5,5e5,1e5)), h=F0inv(ygrid/100))abline(lty=2, h=F0inv(ETA <- 1-exp(-1)))

Figure 1:SimplifiedWeibull plot, showing the medianrank regression (MRR) fitted line, representing theWeibull parameters and .

After plotting the median ranks versus theobservations, the curve()function draws astraight line withbeta = andeta = . Note thatan identical line can be plotted by substitutingcurve(...) with the following code, where inthe argument of F0inv() one recognises theWeibullc.d.f. from eq. [1].

### Alternative curve() method ###curve(add=TRUE,lwd=2,F0inv(1-exp(-(x/eta)^beta)))

Grid lines are plotted at horizontal and verticalplot position (vertical positions are listed inygrid). The 63.2 [%] (dashed) unreliability lineis also plotted: the point of its intersection with thefitted line provides the characteristic life of thedistribution, .

4 B-life

When the Weibull plot is created, predictions ofthe failure behaviour of the specimenspopulationcan be made.

The B-life is the age at which a certainpercentage of the investigated populationisexpected to fail (based on the analyzed sample!).For example, the B10-life for the population ofengine mounts can be read from the plot andisapprox. 55500 cycles (the number following thecapital ‘B’ indicates the unreliability percentage).The B1-life is approx. 22300 cycles. Just draw ahorizontal line at the 10 [%]unreliabilityand findthe intersection with the fitted (straight) line.Then,read the age from the horizontal axis.Run the lowercode for marking the B10 life:

### Marking the B10 B-life ###abline(lwd=2,lty=2,v=55500,h=F0inv(0.1))

In R, the B-lives are very easily calculated bymeans of the qweibull(p,beta,eta)function which is part of the R standard libraries.Itcalculates the pthquantile from the Weibull modeldescribed by and . Executingqweibull(c(0.1,0.01),beta,eta)calculates the B10 and B1 life, respectively:

[1] 55415.93 22299.16>

5 Confidence in Predicted B-lives

5.1 Definition of Confidence Interval

It is evident that B-livesbased on Weibullparameter estimates from large sample sizes are tobe taken more seriously than those based on two orthree observations; small samples contain verylimited lifetime information.To get an indication ofthe confidence that one should have on anestimated B-life, a confidence interval can becalculated. A90 [%]confidence interval for a B-life has the following meaning:

“When the B-life would be estimated over andover again from samples similar to the originalone, then the real, unknown B-life of thespecimens population will, with a 90 [%]frequency, be situated inside the 90 [%] confidenceinterval.”

A 90 [%] confidence interval is limited by twobounds, who can also be described as – at thelower side – thelower confidence bound of a


47

95 [%] confidence interval with no upper limit(100 [%]), and – at the upper side – the upperconfidence bound of a 95 [%] confidence intervalwith no lower limit (0 [%]).

This means that the real, unknown B-lifeexceeds the lifetime at the lower confidenceboundwith a 95 [%] frequency. Also: the real,unknown B-life will,with a 5 [%] frequency, besmaller than the lifetime at the lower confidencebound. The latter shows that there is still a smallchance that the actual B-life turns out to be worsethan anticipated by the confidence bounds!

5.2 Usage

According to the method demonstrated in thisarticle the real, unknown B10 life for the providedexample lies between approx.24500 and 87000cycles, with 90 [%] confidence level. The B1-life(estimated at 22300 cycles) liesin a big 4700 –50200 cycles interval, with 90 [%] confidencelevel.

Generally, confidence levels of 90 [%] are usedand silently implied. In the automotive industry, aconfidence level of 50 [%] is often used, meaningthat the confidence concept isnot used at all but B-lives are read straight from the fitted line. In thiscase they rely on large sample sizes for reliablepredictions. Much higher confidence levels areused (90 [%] to 99.9 [%] or more) in the medicaland aircraft industry.

The presented confidence limits widendramatically when the sample size decreases.Although still valid, it is clear that the interval willbe so wide that it could be of little practical use:increasing the sample size is the only remedyagainst wide confidence intervals.

5.3 Calculating Confidence BoundsTo obtain a lower bound for a 90 [%]

confidence bound for a B10 life, one must find the5th percentile of the distribution of the B-lives atthe given (10 [%]) uncertainty. This distribution isnot easily described and it is, except for completeor ‘Type II’ censored data (where all the unfailedspecimens are censored at the time of the latestfailure) difficult or impossible to calculateanalytically (ref. [1]). Some methods approximatethis distribution by a well known distribution orapply transformations to the B-lives.

Confidence bounds come in a great variety:beta-binomial bounds, Fisher’s matrixbounds,likelihood ratio bounds, Monte Carlo pivotalbounds and bootstrap bounds are the most popular.

The first are very easily calculated (even by hand)but cannot be extrapolated to lower unreliability’s;precisely where the reliability engineer needsthem. Abernethy (ref. [2]) concludes that MonteCarlo pivotal bounds are best practice whenmedian rank regression is used, given that theneeded computer power is available. For largersample sizes (> 400), likelihood ratio bounds forMLE or MLE-RBA based Weibull estimates arefaster to calculate. This article describes how tocalculate the Monte Carlo pivotal bounds.

5.4 Straightforward Monte Carlo BoundsIt would make sense to calculate the confidence

interval of a B10 life using the following MonteCarlo based method:

Calculate the parameter estimates andfrom the original sample , , …, withsample size n, as explained earlier.Create a lot (2000 <= R<= 5000)of syntheticsamples (samen) by randomly generatingsynthetic observations based on and ,resulting in = ( , … , ) , ( , … , ) ,… , ( , … , ) . In R this is accomplished byrepeating therweibull(n,beta,eta)functionRtimes.Find the synthetic Weibull parametersestimates and for all the syntheticsamples, , … , and =

, … , .Calculate the B10 life for each syntheticWeibull plot, resulting in B10 .. = B10 ,B10 , … , B10 by repeatingqweibull(0.1,beta,eta)for all and

.Calculate the 5th and 95thpercentile (for a90 [%] confidence interval) from the empiricaldistribution of 10 using thequantile() function. These valuesrepresent the lower and upper confidencebounds.

The above straightforward method turns out toprovide too optimistic (narrow) intervals,especially for small sample sizes (n< 20). Thereason is that the distribution of the synthetic B10lives (from which one calculates the 5th and 95th

percentiles) depends too much on the real,unknown values of and . The abovecalculations are based on estimates and , whothemselves become less accurate with smallersample sizes. The deviations of these estimates


48

from theirreal, unknown values should be reflectedin the range of the confidence intervals, but theyare not. When bigger sample sizes are used, theestimates are closer to their true values, makingtheseconfidence bounds more realistic, however.

5.5 Monte Carlo Pivotal Confidence BoundsA better approach is to derive the 5th and 95th

percentiles from a ‘pivotal’ quantity: this is aquantity that does not depend on the underlyingtrue and unknown distribution parameters and .Lawless (ref. [1]) supplies a pivotal quantity fordetermining B-life distributions:

= (2)

Where = log ( ), = log and= 1/ . isthe time where the unreliability is= ) from eq. [1]; the Weibullc.d.f. for

the real, unknown parameters and . Otherwisesaid, ( ) which is the inverse of theWeibull c.d.f..Practically, is the B(100*p)-lifewhere a confidence bound is to be calculated for.

By definition, ZP does not depend on unknownparameters, meaning that its distribution formstaysthe same regardless of the actual values of and .

The pivotal can therefore be calculated byarbitrary setting = = 1 (or u = 0 and b = 1) forfurther calculations.

Recall: ( ) (3)

Let = log , and with = = 1:

= log ( ; 1, 1) = log ( ) (4)

Eq. [4] can now be calculated for it represents theinverse of the standard Weibullc.d.f. (eq. [1]) andresults in:

= log log (5)

To determine the empirical distribution of ,Monte Carlo methods are used again:

Create a lot (2000<= R <= 5000) of syntheticsamples (same size as the original sample) byrandomly generating synthetic observationsfrom the standard Weibull model ( = = 1).Find the synthetic standard Weibull parameterestimates = log ( )and = 1/

for all the synthetic samples,giving )and ).

Calculate all pivotals using a combinationof eq. [2] and [5]:

( ) = ( )

( )(6)

The empirical distribution of these pivotals cannow be calculated. To calculate the bounds for a90 [%] confidence interval for the B10 life, ontakes the q=5th and q=95thpercentile of Z(p=0.10)and calculates the corresponding B-life with arearranged eq. [2]:

) = log ( ) ) (7)

) = e ) (8)

Continue with the next code block to load thepivotals() function:

### Pivotals function def. ###MC<- function(n){ std <- data.frame(time=rweibull(n,1,1),event=1) std <- mrank.data(std) fwb <- lm(log(std$time)~F0inv(std$mrank),std)c(u0_hat=coef(fwb)[1],b0_hat=coef(fwb)[2])}pivotals<- function(r,R,unrel){ piv <- as.data.frame(t(replicate(R,MC(r)))) wp <- F0inv(unrel)Zp <- function(wp){((piv$u0_hat-wp)/piv$b0_hat)}piv <- cbind(piv,sapply(wp,Zp)) names(piv) <-c("u0_hat","b0_hat",signif(unrel))piv}

The MC() function calculates and returns onepair of and by fitting a Weibull line trougha synthetic sample from the standard Weibulldistribution. The pivotals() function repeatsMC() for R times (usually R=2000),applies eq.[6] with sapply() on all and and bindsthe pivotalsto the pivdataframein an extracolumn.


49

Note that wp<- F0inv(unrel)gives thesame results as wp<-log(qweibull(unrel,1,1)), showing therelation with the standard Weibull model and thecalculation of B-lives.Proceed by calculating thepivotals for determining the B10 life’sconfidenceinterval, based on the exemplary samplewith sizen=8:

###B10 pivotals calculation ###piv <- pivotals(8,2000,0.1)head(piv,3);tail(piv,3)

In the third column of the pivdataframe, onecan find the 2000 pivotal quantities (only a few aredisplayed with the head() and tail()functions).Continue by calculating the pivotalpercentiles and the actual confidence intervals:

### B10 pivotal conf. bounds ###tp_low <- exp(log(eta)-quantile(piv[,'0.1'],0.95)/beta)tp_upp <- exp(log(eta)-quantile(piv[,'0.1'],0.05)/beta)print(paste("B10 90[%] CI = (",signif(tp_low),",",signif(tp_upp),")"))points(pch=3,lwd=2,cex=2, x=c(tp_low,tp_upp), y=rep(F0inv(0.1),2))

Eq. [8] is executed twice for the 5th and 95th

percentile. Finally, the pivotal confidence intervalfor the B10 life is calculated and displayed, andadded to the plot:

[1] "B10 90[%] CI =( 23931.3 , 86688.7 )">

Note that recalculated bounds will never beexactly identical because of the Monte Carlovariability. For lower B-lives like B1, manyreplications like R=5000 or more may benecessary. If repeatable results are needed, onecan run set.seed(1) before all code to lock therandom number seed value.

Repeating the above steps for a number ofunreliability levels and connecting the points sothat a curve emerges is the next logical step; theseare the pivotal confidence bounds as plotted in theWeibullToolkit:

### Plot pivotal conf. bounds ###pivotal_CB<- function(piv,CL){unrel <- as.numeric(names(piv[,c(-1,-2)]))rdf <- data.frame(unrel=unrel,row.names=unrel) Tp <- function(Zp,conf){ exp(log(eta)-quantile(Zp,conf)/beta)}rdf <- cbind(rdf,lower =sapply(piv[,c(-1,-2)],Tp,1-(1-CL)/2),upper =sapply(piv[,c(-1,-2)],Tp,(1-CL)/2))rdf}piv<- pivotals(8,2000,ygrid/100)CB <- pivotal_CB(piv,0.90)lines(lwd=2,CB$lower,F0inv(CB$unrel))lines(lwd=2,CB$upper,F0inv(CB$unrel))

Figure 2:Weibull plot with Monte Carlo pivotalconfidence bounds for a 90 [%] confidence level.

The B10 confidence intervals can now be readfrom the graph in the same way as the regular B10lives.

6 Conclusion

The article demonstrates the inner workings ofthe Weibull Toolkit for R, an open sourcereliability and lifetime data analysis package.After demonstrating R code for calculating theWeibull parameters for complete lifetime data


50

using the median rank regression method,MonteCarlo pivotal confidence bounds are discussed andcalculated. Where appropriate, alternativecalculation methods are briefly mentioned anddiscussed. The calculation of the pivotal quantitiesis clarified in detail and applied in functional Rcode, culminating in a simplified version of theWeibull plot as generated by the Weibull toolkit.

7 References[1] Jerald F. Lawless, Statistical Models and

Methods for Lifetime Data, 2nd edition,Wiley-Interscience, Hoboken N.J., 2003.

[2] Robert B. Abernethy, The New WeibullHandbook, 5th Edition, 2008.

[3] http://www.barringer1.com/wins.htm[4] http://mechanics.kahosl.be/fatimat/index.php

/downloads-and-information/40/198[5] JurgenSymynck, Filip De Bal, The Weibull

model in fatigue and reliability analysis: anintroduction & implementation in R’, KAHOSint- Lieven, 2011http://mechanics.kahosl.be/fatimat/index.php/downloads-and-information/40/77

[6] http://mechanics.kahosl.be/fatimat[7] http://www.kahosl.be/site/index.php?p=/nl/p

age/153/[8] http://www.kahosl.be[9] http://www.youtube.com/watch?v=g2IczVy

uQkQ[10] http://www.reliasoft.com[11] http://sourceforge.net/projects/weibulltoolkit/[12] http://www.r-project.org/[13] http://www.sciviews.org/Tinn-R/[14] JurgenSymynck, Filip De Bal, Weibull

analysis using R, in a nutshell,NewTechnologies and Products in MachineManufacturing Technology, Stefan cel MareUniversity. of Suceava, 2010http://mechanics.kahosl.be/fatimat/index.php/downloads-and-information/40/171

[15] http://weibulltoolkit.git.sourceforge.net/git/gitweb-index.cgi

[16] http://git-scm.com/http://code.google.com/p/msysgit/

8 Further readingChi-Chao Lui, A Comparison Between TheWeibull And Lognormal Models Used ToAnalyse Reliability Data, dissertation fromUniversity of Nottingham, 1997.William Q. Meeker and Luis A. Escobar,Statistical Methods for Reliability Data,Wiley-Interscience, New York, 1998.http://www.weibull.com/Reliasoft’s ‘Reliability subjects’ websitehttp://cran.r-project.org/web/packages/survival/The homepage of the ‘survival’ R package forgeneral survival and reliability analysis.http://cran.r-project.org/web/packages/boot/The homepage of the ‘boot’ R package, forbootstrapping functions.

http://www.barringer1.com/wins.htm

http://mechanics.kahosl.be/fatimat/index.php


http://mechanics.kahosl.be/fatimat

http://www.kahosl.be/site/index.php?p=/nl/p

http://www.kahosl.be/

http://www.youtube.com/watch?v=g2IczVy

http://www.reliasoft.com/

http://sourceforge.net/projects/weibulltoolkit/

http://www.r-project.org/

http://www.sciviews.org/Tinn-R/


http://weibulltoolkit.git.sourceforge.net/git/g

http://git-scm.com/

http://code.google.com/p/msysgit/

http://www.weibull.com/

http://cran.r-/

http://cran.r-project.org/web/packages/boot/


51

MODERN GREEK LANGUAGE FREQUENCY COUNTS FORTEXT ENTRY DEVICES

Ilias Sarafis1, Anastasios Markoulidis2

1 Technological Educational Institute of Kavala Agios Loukas, GR65404, Kavala, [email protected]

2 Technological Educational Institute of Kavala Agios Loukas, GR65404, Kavala, GREECE.

Abstract: In this paper we present and analyze tabulated case-sensitive single letter, digram and trigramfrequency counts from a 450 Kword modern Greek corpus. Special characters, such as space, comma,paragraph change and full stop, used for text input, are also included. Data analysis revealed moderatecorrelation between English and Greek frequency counts, justifying a different approach when studyingGreek keyboards. A keystroke-level model using the digram frequency data and applying Fitts’ law wasemployed to predict entry speed rates for Qwerty, Fitaly, and Telephone virtual keyboards, both inEnglish and Greek layouts. The results of this paper could be used when designing new software orhardware keyboard layouts and in order to improve their efficiency and text entry speed.

Keywords: digram frequencies; linguistic data; modern Greek; text input devices; virtual keyboard

1. Introduction

Moving towards mobile computing andportable devices has brought to the fore, the needfor smaller, more versatile, and space efficient textentry methods. The use of T9 (http://www.t9.com,Tegic Communications 1998, James and Reischel2001) on mobile phones for SMS editing is acharacteristic example of a smart and effective wayto enter text on a keyboard initially designed fornumeric input. Only 9 alphanumeric keys (12including space and special characters), and adisambiguation algorithm and lexicon smallenough to fit into mobile phones memory, providemillions of users the ability to communicatequickly using text messages.

Apart from hardware mini-keyboards (Green etal. 2004, Clarkson et al. 2005) there are also manyvirtual (soft-type) keyboards, which have beenstudied and applied on mobile devices (Masui1998, Goldstein et al. 1999, Zhai et al. 2005).Research for better keyboard layouts anddisambiguation techniques is an on-going process,emphasizing both in better modelling algorithmsand in designing smaller and more user-friendlylayouts with higher text entry rates.

Much effort was made to develop theoreticalmodels, in order to evaluate the performance of

such virtual keyboards, in terms of maximumwords-per-minute. These models mainly use thewell-known Fitts' law (Fitts 1954) and keystroke-level timing together with linguistic data (digramfrequencies) (Mackenzie et al. 1999, Soukoreff2002).

Concerning research on Greek languagekeyboards, extended literature review has revealedtwo matters: first, no published research existsabout optimized and efficient Greek languagekeyboard layouts, not even actual performancemeasurements of existing Greek keyboards.Second, there are no linguistic data published forGreek digram and trigram frequencies, but only forletter frequencies, despite the fact that there is alarge (47 Mwords) corpus available.

Thus, our research effort was focused, on onehand to compute and publish Greek digram andtrigram frequencies and on the other hand to applya keystroke-level model using Fitts’ law to anumber of Greek keyboards. One of the modelledkeyboards is a novel design, based on FITALYEnglish layout (http://fitaly.com/, TextwareSolutions 1996) and applying the philosophy ofminimizing key distances between the mostfrequent digrams.

mailto:isarafis:@teikav.edu.gr

http://www.t9.com/

http://fitaly.com/


52

2. Modern greek linguistic dataA large (47 Mwords) corpus is available for

Greek language, by the Institute of Language andSpeech Processing (ILSP - websitehttp://hnc.ilsp.gr). This corpus is known as"Hellenic National Corpus"™ (HNC) and it is notpublicly available as a whole, rather through a webinterface for searching words and lemmas.

Mikros et al. (2005) have published quantitativecharacteristics based on HNC corpus, mainly forlinguistic purposes. Thus, their data do not includefrequencies for special characters, such as space,comma, paragraph change and full stop, used fortext input. Moreover, they do not provide data fordigram and trigram frequencies.

In our study, we decided to use corpora fromtwo nationwide Greek newspapers, "Ta Nea" and"Macedonia". The corpora are available throughthe Center for Greek Language (http://www.greek-language.gr/). We have selected the "articles"section and created a combined corpus having453.407 words, 2.574.047 characters withoutspaces, 3.018.705 characters with spaces and 8.807paragraphs.

Although our corpus is much smaller thanHNC, it is certainly larger than the limited 20Kwords corpus (Mayzner and Tresselt 1965), usedby keystroke-level models for English language(Mackenzie et al. 1999), thus the results can beconsidered reliable enough. Further on, wecomputed Pearson product-moment correlationcoefficients in order to find out if there is asignificant statistical difference between HNC andour corpus. We compared single letter frequenciesand ranking from the two corpora and we found tobe highly correlated (p = 0.9996 & p=0.9991).Therefore, one could assume that the digramfrequencies and other linguistic data we presenthere would be reliable enough, at least forkeystroke-level modelling.

2.1. Word-length and single letter frequencies

Word length distribution for the studied corpusis presented in Figure 1. The average word lengthof 5.551 characters is very close to that reportedfor HNC (5.33 total and 5.56 for "miscellaneous"media type). We have also tested three, randomlyselected subsets of the 450k corpus, one with 20Kwords, one with 70 Kwords, and one with 180Kwords. The word length for these subsets showedno significant differences (5,529, 5,552 and 5,557characters respectively). The word length

distribution shown in Figure 1, is also consistentwith the one reported for HNC.

Single letter frequencies for lower-case, upper-case, accented and non-accented letters, are shownin Table 1. Notice that in Greek language there aretwo types of accent: (a) the "tonos" ( ' ) that can goover all the seven vowels and (b) the "dialitika" ( ¨) that go over " " or " " in certain cases. In order totype an accented letter on a keyboard, one shouldfirst press the accent key, prior to the letter key.There are also cases where "tonos" and "dialitika"are combined, e.g. and . In these cases also, oneshould first press the proper accent key, prior tothe letter key.

Upper case letters relative frequency is 2.85%,the seven vowels count for 54.83% of the corpusand accented vowels are the 14.83% of the totalletters. Correlation of the single letter frequencieslist between the 450k corpus and the subsets of20k, 70k and 180k corpora, was high (0.99950.9997 and 0.9999 respectively).

By comparing upper and lower-case counts wefound a Pearson correlation of 0.7697 which ismoderate but higher than the reported for Englishlanguage (0.6337) (Jones and Mewhort 2004).Correlation between upper-and lower-case vowelsis 0.7953 (accented 0.4969 & non-accented0.8249), between upper- and lower-caseconsonants 0.7350 and between accented and non-accented vowels 0.8564.

In order to compare the Greek with theEnglish frequency counts, we used the standardQWERTY character mapping, as in Table 2. Wehave used English linguistic data published byJones and Mewhort (2004). We can observe thatthere are 13 totally identical letters and 7 letters thatare similar (Delta, Fi, Gamma, Lamda, Pi, Ro,Sigma).

Using this mapping, correlation between the twolists of single letter frequency counts gives ap=0.8157. This result reveals the need for adifferent approach when designing and evaluatingGreek keyboard layouts. Any proposal foroptimized Greek layouts could not be based onwork done for English ones.

After reordering the mapping to achieve a highercorrelation (p=0.9703), we concluded to the resultsshown in Table 3, referred as "Greek reordered"layout. It is obvious that such a mapping would betotally confusing for the user of hardwarekeyboards, where two letters (English and Greek)are printed on the same key

http://hnc.ilsp.gr)./

http://www.greek-/


53

2.2 Special Characters

Special characters are of significant importancefor keyboard design and efficiency. As shownin Table 4, special characters, count for a totalof 17.98% of the corpus. Space only, counts for

14.74% of the total characters, in comparison to18.65% reported for English language(Mackenzie et al. 1999). Other significantspecial characters are the full-stop (0.75%),comma (0.91%) and paragraph change (0.29%).


54

2.3 Digram and Trigram Frequencies

Digram and trigram frequencies are useful in manycases. From anagram solutions (Dominowski andDuncan 1964) and word recognition to keyboardmodeling and even for developing disambiguationalgorithms (Maragoudakis et al. 2002). AnAppendix that contains case sensitive bigramfrequencies, including space, full stop and comma,together with a list of the 500 most commontrigrams (case independent), is available by theauthors, upon request.We have tested the correlation between thefollowing lists: (a) single letter counts, (b) letter aspredecessor counts and (c) letter as successorcounts. We found high consistency between themthat was over 0.9997 in the worst case. There ismoderate agreement between single letter countsand the frequency counts of a letter at thebeginning or the end of the word (0.6992 & 0.6838respectively).High inconsistencies are observed between the listsof (a) the first letter and (b) the last letter of a word(p=0.2377), even after ignoring the final , whichcan only be found at the end of the word. Thiscould affect the positioning of the "space" key,since space is the dominant character, both before(97.9%) and after a word (89.8%). Other charactersfollowing the end of a word are paragraph change(i.e. the "Enter" key), full stop and comma.We have tested the correlation between casesensitive bigram counts and compared them withthe values reported for English language (Jonesand Mewhort 2004). Correlation betweenupper/upper and upper/lower digrams exhibits poorconsistency at 0,4316 (English 0,372). Correlationbetween upper/upper and lower/lower is also lowat 0,3526 (English 0,665) and there is moderateagreement between upper/lower and lower/lowerdigrams at 0,5945 (English 0,587). The totalnumber of instances for each bigram type is:upper/upper: 9.919 occurrences upper/lower:52.152 occurrences lower/lower: 1.962.339occurrences lower/upper: 28 occurrences.

3. EVALUATION OF GREEK SOFTKEYBOARDS

We have used a keystroke-level model proposedby Mackenzie et al. (1999), to predict theperformance of three different Greek keyboardlayouts and compare them to their Englishequivalents. The model is valid for one-hand typingon soft (virtual) keyboards.

At first, a spatial layout of each keyboard isdrawn and the x-y coordinates of the keys arederived. Hence, the distance between every twokeys in the bigram table can be calculated:

where i is the predecessor and j is the successorin a bigram.By applying Fitts' law (Fitts 1954), which modelsthe time a human needs to move from one key to theother, we could calculate the movement time foreach possible bigram

where RT is the reaction time a human needs tolocate a specific key on the keyboard, BW is thehuman motor bandwidth as defined by Fitts, Aij isthe distance between the centers of the two keys, asin (1) and Wj is the width or the size of the targetkey.

Based upon the above prediction, we can thensummate the movement time of all bigramcombinations, after weighting them with theprobability Pij of occurrence of a digram with firstletter i and second letter j (Mackenzie et al. 1999):

Equation (3) gives the mean movement time inseconds, between two characters. To convert thisto words-per-minute (WPM) we use:

where CPW is the mean characters per word fora specific language. In order to have comparativeresults, we used CPW=5 as proposed by Mackenzieet al. (1999), both for English and Greek language.

In our calculations we have used the dimensionsof a common hardware keyboard with key width W= 18mm but with zero clearance between keys. Wealso focused on maximum entry speed prediction,thus we used RT=0 (expert user) and the valueBW=4,9 proposed by Mackenzie et al. (1999). Forthe key repeat time (bigrams with same letters), weused the 0,153 seconds value, also proposed byMackenzie et al. (1999). English bigram data were


55

downloaded fromhttp://www.dynamicnetservices.com/~will/academic/bit95.tables.html.

All the above model equations and bigram tableswere entered into a spreadsheet and produced thefinal wpm results for each of the followingkeyboard layouts.

3.1 QWERTY layoutsWe used typical QWERTY layouts with

English-to-Greek mapping as in Table 2. ForEnglish language we used case independentbigram tables (a 27X27 matrix including thespace character). For the Greek layout, we applied

in addition, an analytical model with case-sensitive bigrams with accented letters, includingfull-stop and comma, as well as the spacecharacter. The analytical bigram table is given inthe Appendix and leads to a 66X66 matrix withbigram probabilities.

In the analytical model, we assumed that if thesecond bigram letter is in upper case (e.g. at thebegin of a sentence), a shift key should be pressedbefore the letter key, thus increasing the traveltime. In the same manner, the accent key shouldbe pressed before any accented letter that issuccessor in the bigram.

The Greek QWERTY layout for the analyticalmodel is shown in Figure 2. In order to calculatethe distance between the space key and the letterkeys, we divided the space key length to threeequal parts and used the distance from the partthat was closer to each key.

We have also tested the EN-GR mapping inTable 3, which results from a reordering of lettermapping that gives a maximum correlationbetween English and Greek single letterfrequency counts.

The predicted time for QWERTY layouts, forexpert users (RT=0), was as follows:

English layout simplified: 31,23 wpmGreek simplified model: 29,91 wpmGreek reordered-simplified: 30,75 wpmGreek analytical model: 27,73 wpm

As we can see, reordering the Greek keysresults in a typing rate close to the EnglishQWERTY. The lower speed rate for the analyticalmodel is mainly due to accented and upper caseletters that require an extra key-press before theletter key.

3.2 FITALY layouts

FITALY layout shown in Figure 3, is acommercially available software keyboard layout(http://fitaly.com/, Textware Solutions 1996), thatwas designed for minimizing travel distancebetween the most frequent letters. According toTextware , a 56,7% of total keystrokes occurs inthe central area of the keyboard (letters T-A-N-E-O-R and the two space bars)

We have computed speed rates for thefollowing FITALY-like Greek keyboards: (a) oneusing the typical QWERTY mapping as in Table 2and (b) one with reordered keys as in Table 3(Figure 4). We have tested layouts using thesimplified bigram tables (case independent withoutcomma, full stop and accent keys) and a Greekreordered layout using the analytical bigram tablesof the Appendix.After modifying the spreadsheet to calculateoptimum distances from the two space keys, weobtained the following results for expert users:English layout simplified: 40,57 wpm

http://www.dynamicnetservices.com/%7Ewill/academi

http://fitaly.com/


56

Greek simplified model: 37,22 wpmGreek reordered-simplified: 3 9,67 wpmGreek reordered-analytical: 34,81 wpmAs expected, FITALY layout is much more

efficient than QWERTY, both in English and inGreek languages. Entry speed rate improvement is29,9% for English layout, 24,43% for Greeksimplified model, 29% for Greek reordered-simplified model and 25,53% for the analyticalGreek model.

3.3 Telephone pad layouts

The telephone keyboard is used extensively forexchanging short text messages via mobile phones.As there are three or more letters assigned to eachkey, text entry is done (a) either in a multi-tapmode, or (b) using a disambiguation algorithm.In multi-tap mode the user has to press a key oneor more times until the desired letter appears. I.e.one should press three times the number 2 key, inorder to enter the letter "C". In worddisambiguation mode, each letter key is pressedonce and the algorithm selects and suggests thepossible word combinations from a built-inlexicon. In case of more than one ambiguouswords, the user can select the desired word from alist of suggestions.

As an example, to enter the word "HELLO" thekey sequence is "4-3-5-5-6" in worddisambiguation mode. In multi-tap mode thiswould be "4-4-3-3-5-5-5-1 second pause-5-5-5-6-6-6". Notice that in multi-tap mode a delay of onesecond is needed if the subsequent keys are thesame.

It is obvious that multi-tap mode is highlyinefficient in terms of text entry speed.Furthermore, modern mobile phones and deviceshave enough memory and power to implement adisambiguation algorithm. Thus we focused ontelephone keypads incorporating such analgorithm. In addition, we assumed an "ideal"algorithm and we have not taken into account thetime a user needs to select the desired word from alist of suggestions.

With the assumptions above, modelling thetelephone keyboard (Figure 5) for maximum entryspeed (the case of an expert user), produced thefollowing results:English layout simplified: 40,86 wpm Greeksimplified model: 41,30 wpmAccent is automatically entered when typing inlower case with disambiguation. Thus, ananalytical model for telephone keypads was not

computed. Further on, since the telephone keypadfollows a simple logic of alphabetically orderedkeys, a reordered Greek layout for this type ofkeyboard would be meaningless and thus was nottested.

4. CONCLUSION

One main difference between English andGreek language is the presence of accented letters.The high frequency (~15%) of these charactersshould be taken into account when designing,modeling and evaluating keyboard layouts, sinceaccent is an extra key that should be pressed priorto the letter key. An exception is telephone pads,where there is no provision for special accent key

Our predictions using an analytical model thattakes accent and letter case into account, showed a7,3% (QWERTY) to 12,2% (FITALY-like)decrease in entry speed, comparing to a simplifiedmodel.

Another key finding is the moderate agreementbetween English and Greek single letter counts,when using the typical QWERTY mapping. Afterreordering the keys so that the two frequency listscorrelate, we predicted a slight 2,8% increase intyping speed for QWERTY-like keyboard, but anoticeable 6,6% increase in the FITALY-likelayout.

Though reordering the keys on hardwareQWERTY keyboards might be confusing for expertusers, novel software keyboards could benefit fromit.

Finally, a very promising keyboard layout is thetelephone pad, being very efficient, both in terms ofspace and in terms of entry speed rates, whendisambiguation algorithms are used. A vast majorityof the population is very familiar with it andespecially young people, which in many cases makeuse of a mobile phone before they try any othermobile device.

REFERENCES

[1]. Clarkson, E., Clawson, J., Lyons, K., andStarner, T., 2005. An Empirical Study ofTyping Rates on mini-QWERTY

[2]. Keyboards. In: ACM CHI 2005 Conferenceon Human Factors in Computing Systems,April 2-7 2005, Portland Oregon

[3]. USA. ACM Press, 1288 - 1291 Dominowski,R.L., and Duncan, C.P., 1964. Anagram


57

solving as a function of bigram frequency.Verbal Learning & Verbal

[4]. Behavior, 3, 321-325. Fitts, P.M., 1954. TheInformation Capacity of the Human MotorSystem in Controlling the Amplitude ofMovement.

[5]. Experimental Psychology, 47, 381-391.Goldstein, M., Book, R., Alsio, G. and Tessa,S., 1999. Non-keyboard QWERTY TouchTyping: A Portable Input Intereface

[6]. For The Mobile User. In: ACM CHI 1999Conference on Human Factors in ComputingSystems, 15-20 May 1999

[7]. Pittsburgh PA USA. ACM Press, 32-39.Green, N., Kruger, J., Faldu, C. and Amant,R.S., 2004. A Reduced QWERTY Keyboardfor Mobile Text Entry. In: ACM

[8]. CHI 2004 Conference on Human Factors inComputing, 24-29 April 2004 Vienna Austria.ACM Press, 1429 - 1432. James, C.L., andReischel, K.M., 2001. Text Input for MobileDevices: Comparing Model Prediction toActual Performance.

[9]. In: SIGCHI 2001 Conference on Humanfactors in Computing, 31 March - 4 April2001 Seattle WA USA. ACM Press,

[10]. 365-371. Jones, M.N., and Mewhort,D.J.K., 2004. Case-sensitive letter and bigramfrequency counts from large-scale Englishcorpora.

[11]. Behavior Research Methods, Instruments,& Computers, 36 (3), 388-396. Mackenzie,I.S., Zhang, S. X., and Soukoreff, R. W.,1999. Text entry using soft keyboards.Behavior & Information

[12]. Technology, 18 (4), 235-244.Maragoudakis, M., et al., 2002. ImprovingSMS Usability Using Bayesian Networks.In: I.P. Vlahavas and CD.

[13]. Spyropoulos, eds. Lecture notes inartificial intelligence (LNAI, Vol. 2308.Springer-Verlag, 179-190. Masui, T., 1998.An Efficient Text Input Method for Pen-basedComputers. In: ACM CHI 1998 Conferenceon Human

[14]. Factors in Computing, 18-23 April 1998Los Angeles California USA. ACM Press,328-335. Mayzner, M.S., and Tresselt, M.E.,1965. Tables of single-letter and digramfrequency counts for various word-length and

[15]. letter-position combinations. PsychonomicMonograph Supplements, 1 (2), 13-32.Mikros, G., Hatzigeorgiu, N., and Carayannis,G., 2005. Basic Quantitative Characteristics ofthe Modern Greek Language

[16]. Using the Hellenic National Corpus.Quantitative Linguistics, 12 (2), 167-184.Soukoreff, R.W., 2002. Text entry for mobilesystems: Models, measures, and analyses fortext entry research. Thesis (MSc).

[17]. York University. Tegic Communications,1998. Reduced keyboard disambiguatingcomputer. US Patent no. 5,818,437. TextwareSolutions, 1996. Method for designing anergonomic one-finger keyboard andapparatus thereof. US Patent no.

[18]. 5,487,616. Zhai, S., Kristensson, P.O., andSmith, B.A., 2005. In search of effective textinput interfaces for off the desktop computing.

[19]. Interacting with Computers, 17, 229-250


58


59

ERRORS OF MECHANICAL TREATMENT PROVOKED BY THERMALDEFORMATIONS OF TECHNOLOGICAL SYSTEM

V.I.Savulyak1, S.A.Zabolotniy2

Dr.Sci.Tech., [email protected],Cand.Tech.Sci., [email protected]

State Technical University Vinnitsa, Ukraine

Abstract: The thermal deformations of a detail and boring bar during the cutting out precise slotswere investigated by the means of modeling of compassing the method of final elements. The mainfactors, which define the magnitude of thermal errors and their influence on distortion of the profile of adetail were also established.

Keywords: keywords list

Introduction

One of the most widespread processes ofcasting machine detail slots remains the process ofcutting of boring bar. The preciseness of slots isdetermined by the complex operation of a numberof factors, which canse the emergence ofsystematic and casual errors of treatment.

The important factor of forming of macro-geometry of slots are the errors which arise due tothe action of heat the springs up in the placecutting [1-4]. The most actual is the taking intoaccount the thermal deformations of the cuttinginstrument as well as the surface of a detail duringthe final operations.

It is very difficult to calculate most analyticallythe defined errors, for the details having complexconfiguration or being heterogeneous as for theirphysical and mechanical peculiarities for examplehaving functional coatings. For these cases it isvery advisable to take into consideration thecalculation of errors using the models based on themethod of final elements and the realizations ofthem by the means using a computer.

Organization of a Task of Investigation

To the thermal deformations of thetechnological system, which influence the arisingof treatment errors, they regard the deformation ofmachine mechanisms, a cutting edge of the

instrument (a cutter, boring bar, etc) as well as asurface coat of the being treated allotment of adetail in the place of cutting under the influence ofthermal currents.

In this paper the influence the current of heat,that appears during the cutting a slot in the detailshaving different configuration.

The profile distortion is extremely undesirablewhile forming of cutting of the slot surfaces, whichare the bases for the roller-bearings or those thatoperate conjugating with pistons, plungers etc[5-8].

The modeling was carried out with using themethod of final elements. The algorithm ofcalculations made on a computer are described inthe paper [9]. The preliminary results of modelinghave shown that the thermal deformation of thedetail surface being treated mechanically directlydepends on general configuration of a detail.

In this paper the results of modeling of cuttingof precise slots in the details having differentthickness of a side are also brought into account.

Taking into consideration, that the distortion ofa slot profile is the result of the combined action ofsuch factors as: thermal deformation of a detailcaused by a source of a thermal stream, that wasarised by the process of cutting having intensity ofqd and moving along the slot axis with the sped offeeding (fig.1, a) and the thermal deformation of ainstrument under the influence of the thermalstream having the intensity of qi (fig.1, b).

mailto:vsavulyak:@mail.ru

mailto:zab82:@mail.ru


60

For the simplification of the model suchassumptions were made: thermal streams into thedetail qd and the instrument qi are constant in time:the whole period of treating is divided into thesteps periods during the next in turn step thecertain new allotment of the detail is heated by thethermal stream caused by the process of cuttingand if is being known that during each next stepthermal influences of previous thermal insertingwere taking into account.

Figure 1: Thermal streams in the slot cuttingprocess: a) – a scheme of putting a thermal stream intoa detail operation; b) – a schema of putting a thermalstream into a cutter into operation.

The quantities of thermal streams, thatinfluence the instrument and the detail, arecalculated with the help of the well knowndependences [1]. For example, in this paper theresults of investigations used for cutting based on0,5 mm/ feeding and 100 mm/min cutting sped areclearly shown.

Having aimed on determing the range ofinfluence of the thermal heat capacity of a detail,that occur in the process of slot cutting, modelingfor two types of details were carried out: 1 – detailshaving relatively small thickness of sides, thatmakes possible of increase the temperature out theexternal surface during the cutting process; 2 –details having relatively big thickness of sides, thatpractically doesn’t allow to change the temperatureon the external surface and that is quite typical forframe details.

If we call the detail of the first type a bushwhich is modeled by a cylinder having suchparameters: the slot diameter d=60mm; theexternal diameter D=100mm; the length L=40mm;the material – steal 45. Let’s call the details ofother type a frame which is modeled by a cylinderhaving such parameters: the slot diameterd=60mm; the external diameter D=300mm; thelength L=40mm; the material – steal 45.

The results and their discussion

The positional of the thermal source qd on theeach step of thermal loading are corresponding tothe position of the instrument during the slottreating. The character of the slot profile distortioncaused by the thermal deformations of the detailwill be determined by the movement of somepoints of material in the zone of cutting. Figure 2and 3 show the radial shift of slot surface points inthe process of treating under the influence of thethermal stream.

The investigation of the details having slots hasshown that the macrogeometry of the treated slotsdepends on the their configuration and the thermalheat capacity of their sides.

For the models of a bushing type, i.e. when thethickness of the side is relatively small, theextention of the slot sizes due to the heating causedby cutting process is obviously seen and that leadsto the diminution of metal layer, which is beingcut. On account of this fact, the slot diameter of thecut detail appears to be smaller after cooling than itwas calculated during the desposing the instrumentto a certain size.

Absolutely another result is observed while slotcutting of frame details, and it is very characteristicor them that the thickness their slot sides is rathersignificant correlated to their diameter. As thethermal heat capacity and the hardness of theirslots sides are rather big, external layers of detailsare almost never become deformed, and that’s whysignificant strain pressure is developed in the zoneof cutting, and the deformation of the slot surfaceis directed to its axis.

As a result at this, the bigger than necessarylayer of metal is cut off and that leads to theincreasing of a slot after it’s treating and thencooling.

Residual distraction of the slot profile is theresult of the total sum of the thermal shifting ofslot surface points and of the instrument top. Thedistortion character of longitudinal slot profileachieved in the result of superposition of shifting isshown in figures 2 and 3 applied to the details suchas the bushing and frame relatively.

The main component of the thermaldeformation of a instrument is its lengthening inthe direction perpendicular, to the axis of thetreated slot which appears under the influence ofthe thermal stream, and has an effect on theoperating part of the cutting edge.


61

The distortion of the slot profile in the processof treating owning to the lengthening of ainstrument is shown in figure 2 and 3 and on curve1. It is clearly seen that the maximum distortion ofthe slot radins caused by the tool instrumentdeformation makes up 0.022 mm.

Figure 2: Thermal deformations of the bushing inthe process of cutting: 1 – cutter top shifting, 2 – slotsurface points shifting in the zone of cutting, 3 – slotprofile distortion after treating and cooling.

Figure 3: Thermal frame deformations in theprocess of cutting: 1 – cutter top shifting, 2 – slotsurface points shifting in the zone of cutting, 3 – slotprofile distortion after treating and cooling.

The maximum slot distortion for the details ofthe bushing tupe is twice bigger than the radialshifting and makes up 0,138 mm, and for the framedetails it makes up 0,069 mm.

Conclusion

1. It is expedient to calculate the quantity of theerror caused by the thermal deformation of theinstrument-detail system using the models basedon the method of the final elements.

2. The character and parameters of slot profiledistortions appeared in the process of cuttingdepend on the construction and sizes of the detail

as well as the sizes and physical and mechanicalqualities of the cutting instrument.

3. To reduce the errors of mechanical treatingappeared in the process of cutting, it is expedient touse the system of adaptive guidance together withadjusting the instruments accordingly to the resultsof model calculations.

References

[1] Reznikov A.N., Reznikov A.N., Thermalprocesses in technological systems,Machinebuilding, Moscow, 1990.

[2] Sylyn S.S. Limilarity method of metal cutting ,Machinery, Moscow, 1979.

[3] Bobrov V.F., Theory basis of mental cutting,Machinery, Moscow, 1995.

[4] Yascherytsin P.I., Feldstein .,Kornyevych MA, Theory of cutting, NovoeKnowledge, Minsk, 2005.

[5] Petrakov Y.V. Automatical management ofprocesses of treating materials by cutting,UkrDNIAT, Kyiv, 2004.

[6] Kosilova A.G., Meshcheryakov RK Referencebook of technologist- and Machine-buildingengineer. Vol.1 Machinery, Moscow, 1985.

[7] Kurchatkin V.V., Telnov N.F., Ivanov K.A.Reliability and maintenance of automobiles,Kolos, Moscow, 2000.

[8] TChernoivanov V.I. Organization andTechnology Restoration of automobiles :HOSNYTY, Moscow, 2003.

[9] Savuliak V.I., Zabolotny S.A, Shenfeld V.Y.Thermal fields and deformations duringrestoration details of transport techniquesBulletin DREAMS them. Vladimir Dal., Luck,2009.


62


63

IRON REMOVAL FROM WASTEWATER USING CHELATINGRESIN PUROLITE S930

Petru Bulai1, Elena-Raluca Cioanca1

tefan cel Mare” University, Faculty of Mechanical Engineering, Mecatronics and Management, Department ofTechnology & Management, no. 13, Universit ii, 720229, Suceava, Romania

[email protected]@yahooo.com

Abstract: The work presents the sorption characteristics of Iron (II) on iminodiacetic resin PuroliteS930 in various operating conditions such as initial pH, copper concentration, contact time, temperature,ionic form of the resin and resin dose. The percent of Iron (II) removal has a maximum at pH 5.0, andincreases with the increasing of resin dose, of the contact time and decreases with increasing initialconcentration of solution.

Keywords: iron, sorption, chelating resin, wastewater

1. IntroductionIn water (drinking water, water supplies and

wastewater) the Iron causes problems, such asgiving reddish color and odor [3].. In general inwater, iron exists in soluble form as Iron (II) and isoxidized to ferric iron according to eq. 1 [3].

2Fe2+ + (1/2) O2 + H+ 2Fe3+ + H2O (1)

The state of iron in water depends on thehydrogen concentration and the redox potential.With the increasing of the hydrogen concentration,dissolved iron (Fe2+or Fe3+) hydrolyzes to formprecipitates. The ferrous ion hydrolyzes to producethe array of mononuclear species FeOH+ toFe(OH)4

2 between pH 7 and 14 [10].There are several methods for removal of iron

and other heavy metals, used in water purificationprocesses (drinking water treatment andwastewater treatment. All techniques have theiradvantages and limitations in application [4].

The most important techniques are chemicalprecipitation, coagulation–flocculation, flotation[6], [4], membrane filtration: ultrafiltration,nanofiltration [1] and reverse osmosis;electrochemical treatment techniques: electrodialysis, membrane electrolysis andelectrochemical precipitation, electroextraction [9];

sorption treatment techniques: ion exchange [7].,adsorption [12], biosorption [11].

As other heavy metals the presence of iron inthe environment is a major concern due to theirnon-biodegradability, bioaccumulation tendency,persistence in nature and toxicity to many lifeforms. Therefore, treatment of wastewaterscontaining heavy metal ions before discharge is animportant component of water pollution controland becomes more important with the increasing ofindustrial activities

Iron (II) ions have a high solubility in theaquatic environment and can be absorbed by plantsand living organisms. The maximum acceptableconcentration of Iron (II) in drinking waterrecommended by World Health Organization(WHO) is 0.2 mg/L. The limit of Iron (II) intowastewater 5 mg/L according to Romanianregulation (NTPA001/2005).

Compared with other usual methods, ionexchange provides some advantages and is one ofthe most widely used techniques for treatment ofwastewaters in many chemical process industries(Macoveanu et al., 2002). The efficient removal ofmetals ions requires an exchange material whichcan function in severe environments, exhibit a highaffinity for the metal of interest and can be readily

mailto:bulaipetru:@yahoo.com

mailto:cioancaraluca:@yahooo.com


64

regenerated. Due to their physical and chemicalproperties chelating resin are generally efficient inthe removal of heavy metals ions from wasteeffluents.

2. Experimental

2.1. Materials

In the experiments was used the S930chelating resin obtained from Purolite InternationalLimited (Hounslow, UK). The main physical andchemical properties of the resin are presented inTable 1.

The stock solution of iron (II) (500 mg/L) wasprepared from analytical-reagent grade coppersulphate (FeSO4.7H2O) in distilled water. RequiredpH adjustments were performed with sulphuricacid.

2.2. Sorption experiments

To avoid the iron precipitation in alkalinesolution the hydrogen resin form was convertedinto sodium resin form. For the conversion of theresin was used 10% HCl solution. The process wasfollowed by washing with distilled water until thepH of the effluent dropped to neutrality. Thehydrogen resin form has been dried at 60 ºC usingan oven.

Sorption of iron (II) ions on Purolite S930in hydrogen (S930-H) form was carried out inbatch experiments using 50.0 mL of iron (II)solutions with different initial concentrations(10–300 mg/L) that where added toErlenmeyer flask already containing 0.05 g ofdry resin. The initial pH of the solution wasadjusted by using diluted solutions of H2SO4or 10 mL of acetate buffer solutions and

measured with WTW pH/cond 340i. Theflasks where mechanically shacked at severalfixed temperatures and at the rate of 120cycles min 1 using Orbital Shaking IncubatorGFL 3031. After equilibrium (24 hours), theresin and solution were separated by filtrationand the iron content of the solution and, also,the final pH of solution (pHe) were measured.In the experiments concerning the effect of theresin dose, a range of resin samples from 0.01to 0.2 g was used. For contact timeexperiments, the procedures were the same asthose presented in equilibrium experiments butthe samples were analyzed after a specifiedperiod of contact time.

The concentration of Fe(II) in solutions wasmeasured using a spectrophotometric method with1,10-phenantroline and hydroxilaminochlorohidrat

=510nm) using Hach DR/2000spectrophotometer [8]. Because in the presence ofO2 from air a part of Fe(II) becomes Fe(III),allexperiments measured the concentration of totaliron as Fe(II) [7].

The sorption of Iron (II) by Purolite S930-Hresin was quantitatively evaluated by percent ofIron (II) removal R (%) and by amount of Iron (II)retained on resin, q (mg/g):

0

0

100eC CR

C(2)

0( )eC C Vq

m(3)

where C0 and Ce are the initial and theequilibrium concentration of Iron (II) in thesolution (mg/L), V is the volume of solution(L) and m is the mass of the resin (g).

Table 1. Characteristic proprieties of the chelating resin used*

Polymer matrix structure Macroporous styrene divinylbenzeneFunctional groups Iminodiacetic acidIonic Form (as shipped) Na+

pH range (operating): H+ form, Na+ form 2 - 6; 6 - 11Maximum operating temperature 70ºCParticle size range + 1.0mm <10%, -0.3mm <1%Total exchange capacity 1,9 meq/mL* Manufacturer supplied.


65

3. Results and discussion3.1. Effect of solution pHSorption of Iron ions on chelating resins

Purolite S930 involve an ion exchange process inconjunction with the formation of a coordinatebond. The ability to bind metal ions through aminegroups lead to the grater selectivity for heavymetals from wastewaters. The formation andstability of the chelates with many metal ions arevery dependent on the solution pH. Because thefunctional iminodiacetate groups of thePurolite S930 are weak acids, they are veryselective for hydrogen ions (their dissociationis dependent on pH) [2].

In order to investigate the influence of initialsolution pH on the Iron (II) removal by S930 resin,experiments were carried out using a resin dose(S930-H) of 1g/L and the initial iron concentrationof 300 mg/L at 293.15 K. The initial solution pHwas adjusted using H2SO4 diluted solutions. Thecorrelation between the initial pH of solution andIron (II) equilibrium concentration and the percentof Iron (II) removal, respectively, is illustrated inFigure 1.

Figure1: Effect of solution acidity on Iron removalonto S930 resin in H+ form:

C0 = 300 mg/L, T = 293.15 K, time = 24 h, resindose = 1g/L

The results reveal that in strong acidic media(pH=2) most carboxylic functional groups areprotonated but with increase of pH up to 4-5 theystart to dissociate and can react with Iron ions. Theequilibrium Iron concentration decreases from theinitial value of 300 mg/L to 296.82 mg/L and235.27 mg/L for initial pH 2 and pH 5,respectively.

Figure 2 illustrate the variation of sorptioncapacity of S930-H resin for Iron (II) ions as afunction of the initial pH of solutions.

Figure 2: Effect of solution acidity on loadingcapacity of Purolite S930 resin in H+ form: C0 = 300 mg/L, T = 293.15 K, time = 24 h, resin

dose = 1g/L

The loading capacity decrease significantlyfrom 64 mg Iron (II) per gram resin for pH = 5to 3.14 mg per gram resin for pH = 2. Figure 3present a correlation between initial pH of thesolution and equilibrium pH.

Figure 3: Comparison between initial hydrogen ionsconcentration ( pHi) and equilibrium hydrogen ions

concentration ( pHi):C0 = 300 mg/L, T = 293.15 K, time = 24 h, resin dose

= 1g/L, S930 resin in H+ form

Figure 3 shows that equilibrium pH are lowerthan initial pH and the difference from the initialpH and equilibrium pH decrease with decreasingof initial pH. The decrease of equilibrium pH canbe explained by hydrogen ions releases in solutionfrom ion exchange process (Eq. 4.).

222 930 ( 930) 2S H Fe S Fe H (4)

3.2. Effect of metal ion concentrationMetal ions sorption onto resin is strongly

influenced by the initial metal ion concentration.The sorption data for Iron (II) ions on S930 resin inhydrogen form, namely removal percent andequilibrium concentration, as a function of initial


66

concentration of iron are showed in Figure 4 andfigure 5, respectively. Experiments were carriedout in solutions of pH=5 adjusted with sulphuricacid, at a resin dose of 1g/L and temperature293.15 K.

Figure 4: Effect of metal ion concentration on theIron (II) removal percent onto S930-H:

pH=5, time = 24 h, T = 293.15 K, resin dose = 1g/L

Figure 5: Effect of metal ion concentration on thePurolite S930-H loading capacity for Iron (II): pH=5,

time = 24 h, T = 293.15 K,resin dose = 1g/L

Figure 4 reveals that with the increasing of Iron(II) concentration the percent of removal decreasebut even for an initial concentration up to 25 mgIron (II)/L, the percent removal was of 80%.Figure 5 shows that the Purolite S930-H loadingcapacity decreases with increasing of initial ionconcentration.

3.3. Effect of contact timeThe effect of the phases contact time on Iron

(II) sorption onto hydrogen ionic form of the resinPurolite S930 is illustrated in Figure 6. Theexperiments were conducted using solutions of 100mg Iron /L at temperature 293,15K; the initial pH= 5 was adjusted using H2SO4 diluted solutions.

Fig.6. Effect of contact time on the Iron (II) removalon S930-H resin:

C0 = 100 mg /L, pH = 5, T = 273.15 K, resin dose =1g/L

The results reveal that within first 10 hours40% sorption was occurred. The equilibrium wasconsidered attained (reached) after 24 hours; afurther increase in contact time has a negligibleeffect on the percent removal. The faster initialsorption rate may be explained by the greaternumber of resin sites available for the sorption ofmetal ions.

4. Conclusions

The present study shows that Purolite S930 isan effective sorbent for the removal of Iron (II)ions from aqueous solutions. The percent of Iron(II) removal has a maximum at pH 5.0 andincreases with the increasing of resin dose, of thecontact time and decreases with increasing ofinitial concentration of solution.

5 References

[1] Bernat X., Fortuny A., Stüber F., Bengoa C.,Fabregat A., Font J., Recovery of iron (III)from aqueous streams by ultrafiltration,Desalination, 221, 413–418, 2008.

[2] Bulai P., Balan C.D., Bilba D., Macoveanu M,Study of the copper (II) removal from aqueoussolutions by chelating resin Purolite S930,Environmental Engineering and ManagementJournal, 8 (2), 213-218, 2008.

[3] Cho B.Y., Iron removal using an aeratedgranular filter, Process Biochemistry, 40 (10),3314-3320, 2005.

[4] Kurniawan T. A., Chana G.Y.S., Loa W.,Babel S., Physico–chemical treatmenttechniques for wastewater laden with heavymetals, Chemical Engineering Journal, 118,83–98. 2006.


67

[5] Petrov S., Nenov V., Removal and recovery ofcopper from wastewater by a complexation-ultrafiltration process, Desalination, 162, 201-209 2004.

[6] Polat H., Erdogan D., Heavy metalremoval from waste waters by ionflotation, Journal of Hazardous Materials.,148, 267–273, (2007).

[7] Popa C , Bulai P., Macoveanu M., The study ofiron (II) removal from 34% calcium chloridesolutions by chelating resin Purolite S930,Environmental Engineering and ManagementJournal, 9 (5), 651-658, (2010).

[8] Popa G., Moldoveanu S., QuantitativeChemical analysis using organics reagents, (inRomanian), Technical issue, Bucure ti,Romania, 1969.

[9] Smara A., Delimi R., Poinsignon C.,Sandeaux J., Electroextraction of heavy

metals from diluted solutions by a processcombining ion-exchange resins andmembranes, Separation and PurificationTechnology., 44, 271–277, 2005.

[10] Tekerlekopoulou A.G., Vasiliadou I.A.,Vayenas D.V., Physico-chemical andbiological iron removal from potable water,Biochemical Engineering Journal, 31 [1], 74-83, 2006.

[11] Tuzen M., Uluozlu O. D. , Usta C. , Soylak M.,Biosorption of copper(II), lead(II), iron(III)and cobalt(II) on Bacillus sphaericus-loadedDiaion SP-850 resin, Analytica Chimica Acta,581, 241–246, (2007).

[12] Veli S., Alyuz B., Adsorption of copper andzinc from aqueous solutions by using naturalclay, Journal of Hazardous Materials, 149,226–233, 2007.


68


69

ANALYSIS OF CONDITIONS FOR OBTAINING A KEYHOLEREGIME FOR LASER WELDING

Remus Boboescu1

1University Politehnica Timi oara, [email protected]

Abstract:. For laser welding using high intensity of laser beam produce the keyhole phenomenon inwelding pool. For keyhole welding regime there is a particular of the weld cross-section shape aspectgiven by the heat affected zone area. It analyzes the laser welds made with Nd: YAG laser on low alloysteel plates using irradiation in continuously regime. The power, welding speed and defocusing arevaried. Experiments were conducted after the full factorial experimental design 22. This showed the effectof laser power for setting keyhole welding regime.

Keywords: laser keyhole welding, full factorial design, laser beam defocusing, heat affected zone.

1. Introduction

Laser welding has two distinct regimes ofconduction welding and keyhole welding regimewhich differ in the weld penetration and weld crosssection shape.

Conduction welding regime. This regime iscalled conduction limited welding. The solid-liquidinterface heat transfer mechanism is exclusively byconduction. Heat is conducted directly from theheat source that occurs due laser radiation from theworkpiece surface within the material. Conductionwelding regime is characterized by irradiationsufficient to produce piece surface melting andvaporization but not enough to producevaporization in the depth of material. Forconduction welding regime values of F ratiobetween the weld width and weld depth are higherthan 1, 1F . As an indicative value for the welddepth in conduction regime is 1.5 mm [1]. Defectsthat are obtained in conduction welding regime arecracks on the weld surface.

Keyhole welding regime. This welding regimeinvolves the keyhole phenomenon in weld poolwhich involves the values of laser beam intensitiesthat exceed threshold needed to produce a vaporfront propagation in the material. Heat transfermechanism at solid-liquid interface involves bothconduction and convection due to the meltmovement in the weld pool.In weld cross sectionthere is a welding defects gas bubbles, the presenceof this pore led to the designation of keyholeregime (“key hole”). F ratio between weld widthand weld depth has a smaller value than 1, 1F .

Defects in welds are obtained for the keyholewelding regime are gas porosity due to dissolutioninto the melt a gas bubbles and vacuum large areas.The presence of the Keyhole and vapor within thematerial made pushing melting front (solid-liquidinterface) inside the material. This providesincreased weld penetration [2], [3], [4]. For weldsmade in experiments the two welding regimes areshown in Figure 1 by the weld cross section.

a) b)

Figure 1: Image of weld cross-section of the weld a)weld in the conduction regime b)weld in keyholeregime

The paper proposes the identification and use ofquantities that can characterize welds underkeyhole welding regime. It looks like those incharacterizing keyhole regime occurs sizesassociated with both weld cross section and weldsurface. Experimental factorial design 22 usedpresents for higher levels of power keyholewelding regime and for the lower power levelpresents the conduction welding regime. It willdiscuss variation of the response functions relativeto varied parameters power and welding speed.

mailto:remus_boboescu:@yahoo.com


70

2. ExperimentsThe experiment consisted in made lines of

fusion (welds), 110 mm long, on Dillimax 500steel plates with thickness of 10 mm (carbon steel,carbon content 16.0 %). Was used a Nd: YAGTrumph Haas 3006D laser source with 3kWmaximum power on a continuous wave regimeCW. Laser beam was transmitted through a opticalfiber with core diameter of 0.6 mm

The focus system made a focal spot with 0.6mm diameter. Lens focal length was 200 mm. Asprotective gas argon was used with a flow rate of20 l / min. Were used sheets of material withdimensions of 10130100 mm for which weremade between 5 and 8 welds, with a distance ofover 10 mm between welds. Parameters varied inthe experiments are presented in Figure 2.

Figure 2: Scheme of keyhole welding pool with variedparameters in welding process

In experiments was varied the laser power,welding speed and distance between focal planeand piece surface (defocusing or defocusing depth)figure 2. Welds were cut in the stable part of theweld near the place where welding process wasstopped. Weld section was processedmetallographic. Weld width, near the piecesurface, and weld depth were examined using amicroscope with precision of 0.01 mm. Meltedarea was measured directly by its footprint..Defocusing values are considered negative if thelaser beam focus inside the piece.

In the experiments were varied power andwelding speed. To statistically analyze the effectsof parameters was necessary to introduce adimensionless parameter values.Transformations between the two systems ofparameters values (actual values and codedvalues) are based by following relationships:

2PA [-] (1)

vB 22.233.2 [-] (2)

The experimental plan is presented in Table 1 withactual values that coded for power and cuttingspeed.

Table 1: Varied parameters values in experiment

Analysis procedure consisted of presenting theresults of the mathematical model, ANOVA tableshowing the correlation coefficients associatedwith the mathematical model, Pareto chart showingthe hierarchy of effects and response surface is agraphic representation of mathematical model. Forthe mathematical model were presented two formsfor real values laser power and welding speed andfor coded system values. The first allows rapidapplication of formulas and the second allowsdirect analysis of the values of regressioncoefficients. Based on these values were achievedPareto charts.Figure 3 shows the analysis weldscheme and analyzed sizes that characterizing theweld.

Figure 3: Scheme of weld with measured sizes

The paper analyzed variations the followingsizes: weld width, shape ratio of weld cross sectionand the heat affected zone area on the weld crosssection.

Weld width L [mm] was obtained as a result ofthree measurements on the weld surface at thebeginning, middle and end of the weldingprocess.The weld width characterized in general


71

the weld and it is independent of the area in thiscross section in weld was performed.

F ratio (w/h) is the ratio beetwen weld widthand weld depth on the weld cross section.This ratiois associated with the welding regimecharacterization.Values of F ratio below unityshows keyhole welding regime.

A heat affected zone area S[mm2] was measuredon the weld cross section. It is given by theisothermal line of transformation metal structurewhich is well below the melting temperature. Areaheat affected zone containing molten zone area.Heat affected area can be measured with greateraccuracy than the area of molten zone.

3.Weld width

The mathematical model for weld width atdefocusing 0 is given by relations (3) and (4).Statistical analysis by ANOVA method is given inTable 2. Figure 4 shows the Pareto chart for weldwidth at defocusing 0 . It is noted that the weldwidth increases with power and decreases withwelding speed.

The decrease effect with welding speed isgreater than the increase effect of power. Theinteraction between power and welding speedincreases width of the weld width so the overalleffect of increasing with power exceeds thedecreasing effect with welding speed. The Paretodiagram presented three effects have similar valuesand have no statistical significance. It looks likethat along the weld there is a change for the firsteffect between power and welding speed.

BABAL 275.0375.029.0086.2 (3)

PvvPL

61050.00535.293075.064859.0 (4)

Table 2: ANOVA table for weld width L at 0

Figure 4: Pareto Chart for weld width at 0

Figure 5: Response surface for weld width at 0

Figure 5 shows the surface response for weldwidth at defocusing 0 . It is noted that theweld width has high values on the experimentalfield. Lower values for the weld width are obtainedat low power and high welding speeds. There is asharp increase in weld width values on thediagonal of experimental field from low power andhigh welding speed on the situation in whichpower is high and low welding speed. It isrecommended for high value of weld width. Thesevalues can be associated with deep welds.

The mathematical model for weld width atdefocusing mm2 is given by equations (5)and (6). The correlation coefficients for themathematical model associated with themathematical and statistical method of ANOVAanalysis of variances are given by equations (7)and (8).


72

Figure 6 shows the Pareto diagram for the weldwidth at defocusing. It is noted that the weld widthincreases with power and decreases with weldingspeed. Effect of power have statistical significance.The interaction between power and welding speeddecreases weld width. It is noted that thecumulative effect of decreasing of the weld widthwith welding speed is lower than the increase withpower. This shows the strong dependence of theweld width with power.

BABAL

325.0545.0875.00966.2 (5)

PvvPL

72150.023310.063225.110195.0 (6)

95.02R (7)

87.0)..(2 fdforadjR (8)

Figure 6: Pareto Chart for weld width at mm2

Figure 7 shows the response surface for theweld width at defocusing mm2 . It is notedthat on the experimental field weld width increaseswith power and decreases with welding speed.Maximum values for weld width are obtained forthe maximum power and minimum welding speed.The type of increase seen in Figure 5 exists in thiscase for focusing within piece. The difference thatarises is that this type of growth is valid throughoutthe experimental field.

Focusing the laser beam inside the piece bringsthe following changes relative to the weld width.By focusing a laser beam inside the piece producea control of weld width through its dependence onpower and association of this dependence tostatistical significance. Focus inside piece changessign of interaction between power and speed.

To focus on the piece surface effect of theinteraction is associated with speed effect, forfocus within the piece the interaction effect isassociated with power effect. This shows theincreased contribution of dynamic phenomena thatoccurring in the weld pool. It shows that the laserbeam focus inside the piece provides stability onweld width and for overall the weld surfacecharacteristics of welds.

Figure 7: Response surface for weld width atmm2

4. Ratio F

The mathematical model for the ratio F atdefocusing 0 is given by equations (9) and(10). Study of variation by the ANOVA method ispresented in Table 3.

ABBAF 425.0375.0025.1535.1 (3)

PvvPF

9435.07195.201525.243925.6 (4)

Table 3: ANOVA table for Ratio F at 0

Figure 8 shows the Pareto chart for the ratio Fat defocusing 0 . Decreased of ratio F showsgetting keyhole welding regime and increase itsintensity. The figure notes that the experimentalfield F Ratio strongly decreases with power. Effectof interaction between power and welding speed


73

decreases the ratio F. Welding speed increasesRatio F. The overall effect of power is much higherthan the welding speed effect. The three effectshave not shown statistically significance.

Figure 8: Pareto Chart for ratio F at 0

Response surface shown in Figure 9 indicatesthat the F ratio decreases with the power andwelding speed on the experimental field. Lowestvalues for the F ratio are obtained at high powerand low welding speeds. At high power weldingspeed has no influence. In this situation getskeyhole welding regime. Conduction weldingregime is carried out at low power and high speedwelding.

It notes that there is a sudden transition fromconduction welding regime to the regime keyholewelding.

Figure 9: Response surface for Ratio F at 0

The mathematical model for the ratio F atdefocusing mm2 is given by equations (11)and (12). Study the variations by ANOVA methodis presented in Table 4.

ABBAF 2275.02425.00925.171.1 (11)

PvvPF

505050.0548450.1622575.1520175.5 (12)

Table4:ANOVA table for Ratio F (w/h)at mm2

Figure 10 shows the Pareto chart for the ratio Fat defocusing mm2 . It is noticed that theexperimental field F ratio decreases with powerand also decreases with welding speed. The onlyeffect that increases the ratio F is given by theinteraction between power and welding speed. Theeffects have not shown statistically significance. Itis shown that both increased with power andwelding speed leads to increased intensity ofkeyhole welding regime. Focusing the radiationinside the piece provides propagation andabsorption of radiation in Keyhole. High weldingspeed leads to the production of two spots in whichthe laser radiation absorption takes place. So thismay lead to loss of the radiation out of keyhole byreflection, situation shown by the interactionbetween power and welding speed.

Figure 10: Pareto Chart for Ratio F at mm2


74

Figure11:Response surface for Ratio F at mm2

Response surface in Figure 11 shows thevariation ratio F at defocusing mm2 . It isnoticed that on experimental field the F ratiodecreases with power. Welding speed do notproduce significant changes. It looks as if the laserbeam focus within the piece setting of keyholewelding regime depends solely on power.

Focusing the laser beam inside the piece,compared with focusing on the piece surface,increased the role of power in the setting of thekeyhole welding regime. It notes the changingrole of the interaction effect between powereffect and welding speed effect.

5. Heat affected zone area on the weldcross section

The mathematical model for heat affected zonearea on weld cross section at defocusing is givenby equations (13) and (14). ANOVA statisticalanalysis method is presented in Table 5.

BABAS 125.3875.3875.625.9 (13)PvvPS 9375.62725.540625.003375.1 (14)

Table 5: ANOVA table for HAZ area S at 0

Figure 12: Pareto Chart for HAZ area at 0

Figure 12 shows the Pareto chart for the area ofheat affected zone to focus on the piecesurface. 0 . It is noted that the area of heataffected zone increases with power and decreaseswith welding speed. Power effect is statisticallysignificant.The interaction between power andwelding speed decreases the area of heat affectedzone. The overall decrease effect with weldingspeed is stronger than the increase effect withpower. It is noted that all three effects areimportant contributions for determining the heataffected zone area.

Figure 13: Response surface for HAZ area at 0

Figure 13 shows the response surfacefor area ofthe heat affected zone at defocusing 0 . It isnoted that on the experimental area of heataffected zone increases with power and decreaseswith welding speed on the field experiment. Thehighest values for the area is obtained at highpower and low welding speed.


75

The mathematical model for heat affected zonearea on the weld cross section at defocusing

mm2 is given by equations (15) and (16).ANOVA statistical analysis method is presented inTable 6.

BABAS 625.2875.3125.84166.9 (15)

vPvPS

82750.505250.324125.1403715.10 (16)

Table 6: ANOVA table for HAZ area S at mm2

Figure 14: Pareto Chart for HAZ area at mm2

Figure 14 presents the Pareto chart for the area ofheat affected zone at defocusing mm2 .It isnoted that the area of heat affected zone increaseswith power and decreases with welding speed. Theinteraction between power and welding speeddecreases the area of heat affected zone. Power isthe only effect that has statistical significance.

Response surface in Figure 15 shows thevariation in heat affected zone area on weld crosssection at defocusing mm2 .It is noted that inthe experimental field heat affected zone areaincreases with power and decreases with weldingspeed. Variations with welding speed are strongestat high power than at low power.

Response surface deformation on experiment alfield is due by the interaction between power andwelding speed effect.

Figure15:Response surface for HAZ area at mm2

For heat affected zone area laser beam focuswithin the piece led to an increase in the role ofpower. This increase is relatively small. At focuswithin the piece the effect of the interactionbetween power and welding speed decreases.

6. Discussion

Laser welding is widely applied in particular formetals and steel. The laser beam is a concentratedheat source.

In laser welding Keyhole phenomenon has twoimportant consequences. The first relates toincreasing the energy coupling between laserradiation and material because the laser radiationcan propagate in the Keyhole and its absorptionoccurs inside the material. This creates thepossibility of formation of heat sources inside thematerial. It is considered that the presence ofKeyhole will increase the energy coupling betweenlaser radiation and material.The second issue concerns the very strong growthof Keyhole in the depth of material. The presenceof a strong heat source that produces suddenchanges in time and space heating favors strongevaporation. Vaporization is followed by a rapidincrease in pressure followed by high invaporization temperature. Such gas-liquid interfacebecomes overheated. Rapid disappearance of theirradiated heat source will produce rapidvaporization of superheated liquid by passing thevapor.

In this paper we studied the variation of thethree sizes that characterize welds in keyhole


76

welding regime for two situations for the depth offocus. Keyhole welding regime was obtained forhigh values of power. The Ratio F shows subunitvalues for keyhole welding regime. These F ratiovalues are associated with high values of weldwidth and with high values heat-affected zone areaon the weld cross section.

It looks like that all the three sizes analyzed inthe paper, weld width, Ratio F, heat affected zonearea can be used to separate the welds in theconduction regime by the welds in keyhole regime.Although the F ratio criterion defines theseparation between the two welding regimes, themathematical models for the weld width and areaof heat affected zone are useful in predictingwelding regime obtained. Focusing on the surfaceof the piece introduces instabilities in the welds.These are associated with the lack of statisticalsignificance for the mathematical modelsassociated with one dimensional size for welds.Weld instabilities produced by focus on piecesurface are more pronounced for weld width, sizethat characterizes the weld surface. Heat affectedzone area shows small with defocusing.Mathematical models have shown high levels ofcorrelation coefficients.

7. ConclusionsThe paper presents two important issues for the

study of welds made with laser beam the presenceof keyhole regime and the effect of defocus forlaser beam. For analysis of the welding processwas used a full factorial experimental design andwere used mathematical modeling and statisticalanalysis of variations. The following were showed:- The weld width and heat affected zone area onthe weld cross section can be used to characterizethe welding regime as well as the F ratio.- For high power levels were obtained welds inkeyhole welding regime. - Focusing the laser beam to the workpiece surfaceproduces instability in the welding process.-Defocusing change did not affect the weldingregime obtained as evidenced by effects on theheat affected zone area.

The aim of experimental study was theassessing contribution of power and speed ofwelding to weld characteristics. From the analysiscarried out were the following: power has maineffect on all response function and in conclusionfor laser melt capacity; power increase lasermelting capacity of steel and welding speeddecrease this capacity; mathematical models for

the weld width and weld depth provided a goodstatistical correlation and statistical significance.The experimental modeling achieved consideredonly the main part of the laser processing. Therehave been lines fusion lines and not welded joints.Were varied only main factors of influence, powerand speed, while focal point position in relationwith the piece surface and assistant gas parameterswas maintaining constant. The plates used weresufficient thick to not be penetrated duringwelding. It is important that experimental researchshould be applicable in slightly modifiedexperimental conditions. To fill the gap betweentow pieces using at laser welding using addedmaterial in the form of wire. Mathematical modelsperformed make predictions for the laser weldingresults for Dillimax 500 steel or other steels withclose thermal characteristics. Experimentalresearch methods used are generally applicable inmaterial processing.

References[1] Alexander F. H. Kaplan, Masami Mizutani,

Seiji Katayama, Akira Matsunawa, KeyholeLaser Spot Welding, International ConferenceApplications of Lasers and Electro-Optics2002

[2] Yukimichi Sasaki , Tadashi Misu , ShunroYoshioka , and Toshiyuki Miyazaki,Monitoring of YAG Laser Spot Welding-Detection of Porosity Defect by AcousticSignal International Conference Applicationsof Lasers and Electro-Optics 2002.

[3] Inoue, Miamoto, Ono, Adachi, Matsumoto, Inprocess monitoring of penetration depth in 20class CO2 laser welding of thick sections,International Conference Applications ofLasers and Electro-Optics 1999.

[4] Jae Y. Lee, Sung H. K., Dave F. Farson,Choong D. Yoo, Mechanism of keyholeformation and stability in stationary laserwelding, J. Phys. D: Appl. Phys. 35 (2002)1570–1576.


77

ASPECTS REGARDING THE QUALITY OF CAST IRONPRODUCED IN S.C.FONTUR S.A. SUCEAVA

Gramaticu Mihai1; Cosmaciuc Vasile2; Costiug Gina3; Hri cu Ionu Marius1

1 tefan cel Mare” University of Suceava, Faculty of Mechanical Engineering, Mechatronics andManagement, University Street, No. 13, 720225, Suceava, Romania,EU. Suceava, [email protected]

2S.C. EXPROTERM S.A. Suceava, [email protected] 3S.C, FONTUR S.A. Suceava,[email protected]

Abstract: The authors analyze the quality of produced cast iron and molded forms fromSC FONTUR SA Suceava. For this purpose samples were cast from different batches of samples, fromwhich were made tensile specimens, samples for chemical and metallographic analysis and hardnesscontrol. Based on analysis conducted authors show corrections that have to be made tothe chemical composition and of cast iron structure to obtain high mechanical strength. Also they giverecommendations on improving technology development.

Keywords: cast iron, chemical composition, microstructure, mechanical properties, moulding

1. General aspects

Increased production of iron castings is dueprimarily to the fact that cast iron is the cheapestcasting metallic material with great technologicalproperties and tensile strength, that for nodularductile cast iron is simillar to those of steel. Themoulding cast iron technologies are more simpleand give better reproductibility of results.

Gray cast iron structure is made of: basicmetallic mass (mold) and graphite.

Basic metallic mass is simillar to hipo oreutectic steel (Cmb 0,77%). It can be:

- ferritic, carbon of metallic core mass is 0.1% C, polyhedral aspect, soft, HB = 100 ...

120;- ferito-perlite, Cs0,77 > Cmb > Csol made

of white polyhedral ferrite grains (aroundgraphite) and lamellar pearlite, HB = 140...180 ;

- perlite, Cmb = Cs0,77 , with lamellar aspect,hard and resistant, HB = 250...340.

Where graphite is grossly spread , its influenceon cast iron strength is prevalent, the metallicmass being insignificant, therefore primary coarsegraphite cast iron have reduced mechanicalstrength [1].

From a structural point of view, graphite ischaracterized by: shape, size, quantity anddistribution.

The graphite amount (occupied area) shows thefilling degree of graphite formations. It dependson the total amount of carbon in cast ironcomposition and especially the amount of spreadgraphite in crystallization process. We need toconsider a minimal value of graphite quantity

restricted by moulding conditions. Therefore,total carbon content of gray cast iron is takenbetween 2.4 ... 3.8 [2],[3], [4].

With density four times lower than perlite, theactual volume occupied by graphite is muchhigher than the Cgr value (free spread graphitecarbon form).

For exemple, for a cast iron with CT=3%, thathas basic perlitic mass (Cmb=0,77%), Cgr=2,23% itcoresponds a volume of 9% graphite.

Shape appears to be the most importantgraphite classification criteria, characterized bydensity degree of graphite spreads, respectively thecutting degree of basic metallic mass and thetensions focusing effect. The size of graphitespreads also influences the mechanical proprietiesespecially lamellar graphite.

The prevalent role of graphite concerning thereduced level of cast iron mechanicalcharacteristics derive from its negative effects,namely:

- cutting effect, respectively decreasing basicmass usefull section in takeover mechanicalprocess of solicitations;

- isollating effect of basic metallic massportions or areas, favorizing a distribution and aunequal takeover – discontinuous of solicitations,especially in graphite interdendritic distribution incompact or discontinuous network form;

- focussing effect due to values of 10 to 100times bigger regarding the medium value ofmechanical tensions, especially in lamellar spreadpeaks area, causing premature rupture duringtablets.

mailto:gram:@fim.usv.ro

mailto:exproterm:@yahoo.comS.C

mailto:contact:@fontur.ro


78

The mechanical characteristic values indicatethe total unfavorable influence of lamellar graphitewith sharped peaks – with maximal cutting effect –on rupture restistence – (Rm=100...350MPa), andalso on specific elongation (A<0,5%), those castirons being considered fragile[4].

The experimental researches and currentpractics showed that the most resitent cast iron isthe one with perlitic metallic mass, that containsthe most fine and uniform spread graphite.

The most important metallurgical processesleading to increasement of mechanical resistenceand tenacity of cast iron are following, by primarystructure modication, the purposes given below:

a) graphite modification, by:- obtaining the most compact form;- the most uniform distribution into metallic

mould;- dimensions decreasement of graphite

spreads;- decreasement of graphite quantity.b) metallic basis modification, by:- crystalline refinement (grains dimension

decreasement and consequently grains numberincreasement on volum unit);

- percentage decreasement of unmetallicphases and segregations due to grains limit.

In conclusion, for a better characterization ofcast iron mechanical behaviour is necessary aninterdependence analysis between structure andproperties.

In this paper authors have analyzed the qualityof cast irons produced by SC FONTUR SASuceava Company, following the purpose offinding the technical solutions that would lead tothe increasement of its strength.

2. Experimental researches

S.C. FONTUR Suceava S.A. Companyregularly delivered banks of cast samples,

according to SR EN 561 from which have beenmade:

- samples for chemical and microstructuralanalysis, and hardness control;

- samples for tensile test.

2.1 Chemical analysis

Chemical analysis has been made by S.C.EXPROTERM S.A. Suceava on a Quantovac 3460machine.

The results of chemical analysis are given intab.1.

In obtaining gray cast irons with lamellargraphite, with superior mechanical strength, basicelements content (C, Si, Ph, Mn, S) are chosen in away that assures a minimal quantity of graphite incristalization, and on the other hand, when theeutectic transformation is finished, we have toachieve a complete perlitic structure.

In cast iron characterization from chemicalcomposition’s point of view due to Fe-C binarydiagram are beeing used two notions: equivalentcarbon and eutectic degree of saturation. Those aremeasures that allow position determination of acast iron given in relation to Fe-C eutectic.

The most common formula in equivalentcarbon determination is [5]:

Cech = Ct + 0,3 (Si + P)where Ct is total carbon from cast iron

(chemical determinated). Cech of cast iron fixes itsposition due to stabile eutectic (4,26%) ormetastabile (4,30%) of Fe-C alloys.

Eutectic saturation degree SC is another way toput a given cast iron in relation to Fe-c eutectic insilicon and phosphorus presence, the eutecticsaturation degree being equal to 1. Saturationdegree can be determined with one of thefollowing formulas:

Tab.1.- Chemical composition of analized banksNo.crt.

Date2010

Chemical composition, %C Si Mn Ph S Cr Mo Ni A1 Co Pb Sn Ti V W

1 20.04 3.79 1.75 .436 .113 .153 .119 .048 .092 .005 .236 .015 .020 .026 .017 .007

2 25.06 4.01 1.95 .447 .085 .152 .155 .048 .104 .005 .341 .013 .027 .024 .027 .007

3 31.07 3.92 1.86 .414 .102 .148 .151 .049 .101 .005 .197 .011 .021 .023 .015 .007

4 25.08 4.00 1.88 .413 .085 .126 .161 .050 .101 .005 .319 .011 .028 .022 .025 .007

5 06.09 3.86 1.88 .407 .075 .125 .146 .041 .098 .004 .203 .009 .016 .022 .015 .007

6 07.10 3.95 2.05 .428 .059 .089 .132 .041 .106 .023 .141 .010 .017 .025 .015 .007


79

EC

EtC CC

CCS or

)(3,026,4 PSiC

CCS t

C

tC

where CC =4,26-0,3 (Si + P) is eutectic carboncontent.

The equivalent carbon, eutectic carbon and theeutectoid one are calculated with relations givenbelow:

The equivalent carbon:Cechiv. = C + 0,3( Si + P ) – 0,03 Mn + 0,4 S +

0,07Ni + 0,05Cr + 0,074Cu + 0,25AlThe eutectic carbon:Ceut. = 4,26 - 0,3( Si + P ) + 0,03 Mn - 0,4 S -

0,07Ni - 0,05CrC'eut. = 4,30 - 0,3( Si + P ) + 0,03 Mn - 0,4 S -

0,07Ni - 0,05CrThe eutectoid carbon:Cs = 0,68 – 0,15Si – 0,05(Ni + Cr + Mn – 1,7S)Cs' = 0,80 – 0,15Si – 0,05(Ni + Cr + Mn – 1,7S)Saturation degree in carbon :Sc = C / [ 4,26 - 0,3( Si + P )]

Cast iron structure moulded in samples withdifferent diameters is determined into the diagramby knowing the limits of structural domaines dueto K characteristic and by sample diameter, shownin tab.3.

This constant is given by formula:

SiCSiK

351

43

Examining the determinated values for carboncontent results that those are bigger thanrecomended to cast irons Fc200 i Fc250:

- sulfur and phosphorus contents are abovemaximal limits normally recomended by standards,

--

most of it due to usement of waste from cast ironradiators as raw material, in which content isallowed more phosphorus to increase liquid castiron fluidity;

- equivalent cast iron carbon is above 4,26%,which leads to primary and coarse graphiteappearance, and consequently to decresement ofmechanical strength and cast iron tenacity.

- calculated K constant is between limits0,85...2,05, which coresponds to perlitic cast iron,fact that results of metalographic analysis ofanalyzed cast iron structure.

2.2 Metalographic analysis

From each bank have been mademetalographic samples with 20x25mm that hadbeen polished, glossed and chemical attacked withnital.

2.2.1 Bank 20.04.2010

a– no chimical attack

Tab.2.- Maximal saturation value and of K characteristic in carbonNo.crt. Cechiv.. Ceut. C´eut. CS C´S K

1 4,428 3,641 3,681 0,398 0,516 0,812

2 4,707 4,035 4,075 0,366 0,486 0,940

3 4,583 3,920 3,960 0,379 0,499 0,883

4 4,665 3,334 3,374 0,375 0,495 0,902

5 4,513 3,206 3,243 0,376 0,496 0,886

6 4,628 3,271 3,311 0,351 0,471 0,985


80

b– attack with nital 4%.Fig.1.- Sample microstructure of bank 20.04 2010Cast iron structure: P + F + G; where P – perlite; F –ferrite; G – graphite.

Microstructure shown in fig.1.a presentslamellar graphite as forms of isolated separationsin metallic mold.

2.2.2 Bank 25.06.2010

a – no chemical attack;

b – attack with nital 4%.Fig.2.- Sample microstructure of bank 25.06 2010

Cast iron structure: P + F + EP + G; where P– perlite; F – ferrite; EP- phosphorous eutectic;

G – graphite.Microstructure shown in fig.2.a presents

lamellar graphite as forms of isolated separationsassociated with punctiform graphite, in lightcoloured metallic mold.

2.2.3 Bank 31.07.2010


Tab.3.- K for different structural domaines

Sample diameter,mm

Kamotley

Kperlitic

Kbperlito-ferritic

30 0,65-0,85 0,85-2,05 2,05-3,10

20 0,75-1,10 1,10-2,25 2,25-3,40

10 1,05-1,50 1,50-2,35 2,35-3,50


81

b – attack with nital 4%.Fig.3.- Sample microstructure of bank 31.07 2010

Cast iron structure: P + F + EP + GWhere P – perlite; F – ferrite; EP- phosphorous

eutectic; G – graphite.Microstructure shown in fig.3.a presents

lamellar graphite as forms of isolated separationsassociated with punctiform graphite, in lightcoloured metallic mold. Graphite arched andsemiarched. Microstructure shown in fig.3.bcontains: perlite (dark colour), 94,26%; ferite(light colour), under 2%; phosphorous eutectic –shiny white coloured grains, very well shaped;lamellar and punctiform graphite.

2.2.4 Bank 25.08.2010

a – no chemical attack

Cast iron structure: P + EP + GMicrostructure shown in fig.4.a presents

prevalant lamellar graphite as forms of isolatedseparations associated with punctiform graphite, inlight coloured metallic mold.

2.2.5 Bank 06.09.2010


b – attack with nital 4%.Fig.5.- Sample microstructure of bank 06.09.2010

Cast iron structura: P + EP + GWhere P – perlite; EP- phosphorous eutectic;

G – graphite.Microstructure shown in fig.5.a presents

prevalant lamellar graphite as forms of isolatedseparations associated with punctiform graphite, inlight coloured metallic mold.

2.2.6 Bank 07.10.2010



82

b – attack with nital 4%.Fig.6.- Sample microstructure of bank 07.10.2010

2.3 Mecanical tests – tensile and hardness

Of each given sample bank had been made:- tensile samples with calibrated part diameter

of 20mm, calibrated part length of 100mm.- samples for hardness determination, same

used in metallography.Tensile test was made on universal machine of

FIMMM mechanical testing laboratory, Universitytefan cel Mare, din Suceava.

Number of samples : at least 3 of each bank .Gross molded samples diameter: 30 mm.

Tensile samples diameter, after chip removingprocess: 20 mm, as standard SR EN 561.

Hardness samples – same prepared inmetalography.

Mechanical test results are given in tab.5.Tensile tests were done as standards SR EN

561. on processed samples with Ø 20 mm.Hardness has been measured universal harness

testing machine CV-700 with ball penetrator of 2,5mm, downforce 1839 N, F/D2=30, where D is balldiameter.

Tensile and harness test results are given intab.5.

Based on numerous determinations it could beestablished several correlations between saturationdegree, tensile strength and cast iron hardness:

29,3613912230 CCr SS

Cr S5,8210230 or

Cr S806,10030

HB30=538-355 SC or HB30 = 465-270 SC

where 30r i HB30 are respectively tensile teststrength and Brinell hardness on standard rods withdiameter of 30 mm.

Cast iron structure: P + F + GMicrostructure shown in fig.6.a presents prevalant lamellar graphite as forms of isolated separations

associated with punctiform graphite and annealing graphite, in light coloured metallic mold.In quantity determination of graphite, perlite, ferrite and phosphorous eutectic, were used

microstructures captured with OPTIKA digital camera of Science and Material Engineering Laboratory, ofwhich overlapped measurement grids transposed by camera program. The quantity of each constituent hasbeen determined by points method:

Qconstituent = 100t

c

NN

(%)

where: Nc – number of points (nodes or intersections) that fall into determined inner constituent; Nt –total network number of nodes or intersections.

Quantities of each constituent, for cast iron samples of each analized bank are given in tab. 4.

Tab.4.- Experimental determined values on constituents quantities of analyzed cast iron structureNo.crt.

Date/2010 Structural constituents quantities, %QF QP QG QE

1 20.04 4,72 90,59 4,69 -

2 25.06 2,34 92,70 3,03 1,93

3 31.07 - 94,26 3,92 1,82

4 25.08 - 93,76 4,12 2,12

5 06.09 4,40 91,24 4,36 -

6 07.10 4,72 90,79 3,40 1,09


83

On this aspect, Namur [6] proposed as quality

apreciation cast iron criteria the mechanicalmassiveness invariant:

]./[ 23

2

kgmmHB

i r

That is characteristic to a chosen cast iron,because it does not depend on sample dimensiononto which is being measured of (on condition that

r i HB have to be measured on the same sample).

The following paragraphs presents tipicalvalues for i (in m2/kg, that eliminates 10-6 ):

- cast irons for stoves with 1-1,5% Pi=20-40;

- semiphosphorous cast irons (0,5-0,9% P)i= 35-50 ;

- unmodified hipoeutectic hematite cast ironi=50-75;

- hematitical cast irons modified with 0,1-0,2% SiCa i=65-90;

- perlitic mechanic cast iron i=80-110;- cast irons with 0,25% Ti i 0,25% V

modified i= 90-120;- soft steel i=l100-1300.At the same saturation degree, tensile test

strength can vary into a wide domain and it

allowed to Patterson [7] to propose a newcriteria of quality cast iron apreciation namelyrelative strength RR or ripe degree (Reifegrad):

C

r

SRR

5,82102 or

C

r

SRR

806,100 ,

where counter ( r) represents determined teststrength of chosen cast iron, and denominatorrepresents theoretical strength. Quality indexshows the increasement or decreasement of givencast iron strength due to calculated medium value.

Variation limits of relative strength are between0,6-1,3 and in vacuum overheated cast irons, thismay decrease up to 0,4. In the same way we candefine another index, called relative hardness,RH:

CSHBRH

355538 or

CSHBRH

270465 ,

that shows how much given cast iron hardnessvaries due to medium value, theoreticallycalculated as on formula .

Combining the ecuations we can determinerelative hardness due to strength:

Tab.5.- Tensile and hardness test results

No.crt.

Tensile Brinell hardness, daN/mm2

Cast iron brand asstandard

SR EN 561

Samplediameter,mm

Breakingstrength,(N)

Rm,(MPa)

Strength’smedium value,(MPa)

Mesuredvalues

Samples’mediumvalue

Burdenyield’smediumvalue

11 20 68452 218

216,66199 ;192 ;211 200,6

201,9 FGL2002 20 70336 224 206 ;210 ;212 209,33 20 65312 208 194 ;198 ;196 196,0

21 20 76105 242.37

231,42248; 244, 252 248

248 FGL2502 20 68314 217,56 246; 244, 254 2483 20 73650 234.35 240; 256, 252 249

31 20 60393 192.33

203,27204 ;202 ;211 205

203 FGL2002 20 50882 159.49 206 ;210 ;212 2093 20 81015 258.00 194 ;198 ;196 196

41 20 76105 242.37

242.37287 ; 308,297 297

254 FGL2502 20 79542 253.31 248 ;234 ;241 2413 20 84877 270,31 214 ;234 ;224 224

51 20 60882 193,89

206,80300;356 ;328 328

256,6 FGL2002 20 70704 225.17 246 ;249 ;247 247,53 20 63224 201,35 187 ;202 ;194 194,5

61 20 79542 253.31

237,10287 ;308 ;297 297

254 FGL2502 20 75400 240,12 248 ;234 ;241 2413 20 68450 217,89 214 ;234;224 224


84

r

HBRH3,4100 or

r

HBRH4,3125 .

Value of RH can vary between 0,8-1,3 anddecreases by modification. Cast iron overheatedwith lower RH, leads to its increasement.

Because lower hardness of a given strength isan index of cast iron superior quality (greaterplasticity and tenacity, better processability),general cast iron quality index (IC) is therelation between relative values of cast ironhardness and strength, for example:

2.4 Conclusions From chemical composition analysis results

following conclusions:- carbon content is higher than recomended

for cast irons Fc200 i Fc250;- sulfur and phosphorus contents are above

maximal limits recomended by standardprocedures, most of it due to usement as raw

material in processing waste from cast ironradiators, being admitted in their contents morephosphorus to increase liquid cast iron fluidity;

- cast iron equivalent carbon is above4,26%, that leads to primary and coarse graphiteappearance and consequently to a decreasement ofmechanical cast iron strength and tenacity.

- calculated K constant is between limitso,85...2,05, that corresponds to a perlitic cast iron,fact resulted from metalographic analysis ofanalized cast iron structure.

- analyzed cast iron structure is mostlyperlitic, ferritic content being reduced. Thisstructure offers good mechanical strength toprocessed cast irons.

- Graphite has lamellar shape and itsdistribution as form of isolated separations; somesamples also have punctiform and annealinggraphite;

- Banks 2, 3, 4 i 5 – have phosporouseutectic in their structure, constituent that weakencast irons, reducing their mechanical strength toshocks.

By determined strength, cast irons ofprocessed banks fit into standardized brands :

.5,82102

355538

C

Cr

SS

HBRHRRIC

Cast iron quality is improved as relative strength value increases and relative hardness decreases.By knowing different technological parameters (chemical composition, eutectic cell number, relative

thickness of wall piece) we have to determine major mechanical characteristics based on different empiricformulas established by processing numerous tests (tab.6).

Tab.6.- Mechanical characteristics calculation of cast iron with lamellar graphite (Fgl) based onsimplified relations[8]:

The characteristic Measurement unities

Calculation relation Observations

Flowing limit, Rp0,2 MPa Rp0,2 = (0,75-0,85)Rm Rm – breakage strength, MpaTensile breakage

strength on rods ofØ30, Rm30

MPa Rm30 = 122 – 139 Sc + 36,9 Sc-

1Sc – saturation degree in

carbon

)(3,026,4 PSiCS c

Idem, Rm30 MPa Rm30 = 102 – 82,5 Sc

Idem, Rm30 MPa Rm30 = 100,6 – 80 Sc

Hardness HB30 MPa HB30 = 538 – 355 Sc IdemIdem HB30 MPa HB30 = 465 – 270 Sc

Tensile strength ona sample or on a piecewith unknown diameter(Ø=X, RmX)

MPa RmX = Rm30 (HBX/ HB30) 3/2 Rm30, HB30 – known forsamples Ø30; also, RmX knownfor sample Ø=X.

Tensile strength, Rm MPa Rm = 7(13,5 – 2CE – 2,3 log g) CE – equivalent carbon, %;G – wall piece thickness, mm


85

- FGL 200 (Fc200) – banks 1 ; 3 i 5 ;- FGL 250 (Fc250) – banks 4 i 6.

Recommendations:- processing temperature increasement in

cupola ;- use of air enriched with oxygen in

processing;- supplementary desulphurisation and

dephosphorusation;- changing processing aggregate, purchasing

an induction furnace for processing that will assuresuperior qualities of processed cast irons and theposibility of modifying those.

2.5. References:

1.- Piwowarski, E., Fonte de înalt calitate,Editura tehnic , Bucure ti, 1967.

2.- Sofroni , L., tef nescu, M.D., Fontemodificate cu propriet i superioare, EdituraTehnic , Bucure ti, 1971.

3.- Sofroni , L., Ripo an, I., Chira, I., Fontealbe rezistente la uzur , Editura Tehnic ,Bucure ti, 1987.

4.- Sofroni , L., tef nescu, M.D., Fontespeciale, Editura Tehnic , Bucure ti, 1974.

5.- Sengupta, P.K., Jordan, D.E., DieScheweibarkeit von austenitischen Gusseisen mitKugelgraphit. In> Giesserei, nr. 26, 1971,p.801…805.

6.- Volianik, N., Contribution a l etude de lagraphitisationdes fontes malleables. În: Revue deMetallurgie, Iunie 1967, p 567…584.

7.- Iatan, N, .a.- Considera ii privind influen aprincipalelor elemente de aliere asupra durit iifontelor destinate turn rii cilindrilor cu crustextradur . În ;: Metalurgia, nr.9, 1971,p.556…559.

8.- Henke, F., Verschleissbestandige weisseGusseisen. În: Gusseiesen Praxis, nr.1, 1973,p.1…21, nr.2,1973, p.32…40, nr.3, 1973, p.1…23,nr.4, 1973, p.69…74.

9.-Farge, J.C.T., Chollet,P.,Yernaux, J. Effectof composition, cooling rate and heat treatment onproperties of new-resistant White iron. În :Foundrz Trade Journal, nr. 2836,1971, p. 319-327


86


87

OPTIMIZATION OF THE PLASTIC INJECTION PROCESSTHROUGH THE MODIFICATION OF THE PROCESS FUNCTIONAL

PARAMETERS

Teodor - Daniel Mîndru, Ciprian Dumitru Ciofu &, Dumitru Nedelcu

”Gheorghe Asachi” Technical University of Iasi, Faculty Machine Manufacturing and IndustrialManagement, e-mail: [email protected]

Abstract: The paper presents the optimization of the plastic injection process through the reduction ofthe cooling and removing times. For the deployment of the experiment the PA 6 polyamide will be used.The SolidWorks and Moldflow software packages will be used for CAD and process simulation, thematerial flow into the mold, as well. The experiment aims to obtain the optimal variant of PA 6 polyamideinjection through the modification of the mold as well as the melt temperature in order for the moldfilling to be complete and the cooling time to be as short as possible, therefore generating an increase ofproductivity.

Keywords: cooling channel, flow, melt, mold, temperature

1. Introduction

In order to plan a plastic injection process wastaken into consideration that the most importanttechnological parameter is the temperature of themold. We may consider that the process developsunder normal condition only the temperature of themold is stationary and may be controlled.

Depending of the plastic materials used, theinjection molds can be heated [1].

The plastic injection process is a cyclic one,formed by several operations such as:

- The gauging [2];- The heating and the melting of the material

in the machine cylinder;- The mold clamping;- The mold charge with under pressure

material;- The solidification and the cooling of the

material;- The opening of the mold;- The removal of the injected piece from the

mold.This paper will study the injection operation as

well as the charging and the cooling operations ofthe material from a mold through the optimizationof the inlet temperatures of the injection process.

Cooling of the injected piece from themaximum value of the temperature to the room

temperature value is an operation determined bythe thermal conductibility of the plastic material.

After the mold is opened, the cooling operationcontinues. In plastic injection cycle, the coolingtimes are the longest, getting up to 68% of the totalperiod of the process [2].

The injection technology aims to obtaining theshortest cooling time possible that assures theprescribed quality of the piece.

The figure 1 presents the temperature variationinside a plastic component during the coolingoperation.

Figure 1. Temperature variation inside a plastic pieceduring cooling operation; Tc – environmentaltemperature, TD – mold release temperature; Tr –cooling temperature.[2]

mailto:mindru.daniel:@gmail.com


88

As a result of the analysis of figure 1, we mayexpress the calculus formula of the total time of theinjection process as:

drut tttt (1)Take into consideration the relation (2):

dru ttt (2)Ktd (3)

and replacing relations (2) and (3) into relation(1), this one may be written as:

Ktt rt (4)

tt - Cycle time

ut - Injection time

rt - Cooling time

dt - Time CleanersK - Constant

2. Simulation process, results and discussions

The cooling system of the mold must bedesigned according to the overall dimensions ofthe injected parts or, as the current analyzedsituation, when the injection is made in cavitydistribution channels, thus obtaining sixcomponents, the cooling channels must cover theentire surface of the model.

The piece used for process optimizing ispresented in figures 2a and 2b. The figures arerealized using the SolidWorks software package.

a)

b)Figure 2: The bushing type piece drawing, a) 2DModel, b) 3D Model

The six cavities along the distribution channelsand their placement in the mold are represented inFigure 3.

It was considered the cold type of sprue. Theinitial parameters are as follow:

- Conical ledge: 23 mm;- Inlet diameter: 5 mm;- Outlet diameter: 9 mm;- Semicircular runners type: 10 mm;- Conicalgate: inlet diameter 5 mm. outlet

diameter 3mm;- Dimensions of mold active parts:

165x105x50mm;- Used material for mold: P20.

Figure 3: The 3D general view fo injection mold

The material used during the simulation is PA6polyamide. In figure 4 it may be seen the variationof the viscosity depending on the workingtemperature. In figure 5 is presented the variationof the specific volume by temperature.

Figure 4: The viscosity variation by temperature

Figure 5: The specific volume by temperature

In order to obtain accurate components, specialattention must be applied to the placement of thecooling channels in relation to the component, to


89

the injection sites, the filling direction of the moldetc. For a better cooling there were realizedseveral types of runners: with rectangular shapedsection, square shaped section and triangularshaped section. As a result of the experiments, itwas noticed that the most efficient runners areexperimentally the square shaped ones andtheoretically the rectangular ones, see figure 6, aand b. Constructively speaking, the easier to makeare the circular ones, because they are just drilled,see figure 7 [2].

Figure 6: Runners. a) Rectangular shaped runner; b)Square shaped runner.

Figure 7: Drilled runners

The most often used cooling circuit is made outof straight runners. The connections between therunners are made with the help of supplementaryrunners, applied to the mold plates. The role of therunners in the injection process is to maintain thetemperature of the mold plates at the establishedvalue, without considerable variations.

The injection process’s inlet parameters are:mold temperature 70-110oC, melted materialtemperature 230-300oC and maximum injectionpressure 180 MPa.

2.1 Process simulationThe analysis with Autodesk Moldflow of the

injection process offers the possibility ofvisualizing the process evolution without beingnecessary to build the mold and to solve theeventual problems that can occur in the designingprocess of new parts [3].

Due to the fact that the simulation is made foran injection cycle, the experimental plan will nottake into consideration the runners but only the

temperature variation of the mold and of themelted material.

The mechanical properties of the material usedfor the simulation part are presented in table 1. Intable 2 are presented the process parametersrecommended by the manufacturers [4].

Table 1 The mechanical proprieties of the used plasticmaterialElastic modulus 2910 MPaPoisson ratio 0.386Shear modulus 1050

Table 2 The process parametersMold temperature range – minimum 70oCMold temperature range – maximum 110oCMelt temperature range – minimum 230oCMelt temperature range – maximum 300oCAbsolute maximum melt temperature 340oCEjection temperature 133oCMaximum shear stress 0.31 MPaMaximum shear rate 100000 1/s

Table 3 highlights the parameters of thesimulation and the number of experience. As itmay be noticed, during the first nine experimentsthe temperature of the melted material is constantand we varied the temperature of the mold. Duringthe following 8 experiments the temperature of themold is constant and we varied the temperature ofthe melted material 10o like step adh 5o for moldheating temperature.

Table 3 Parameters and experiments number

Exp.Nr.

Moldtemp.[oC]

Melttemp.[oC]

Exp.Nr.

Moldtemp.[oC]

Melttemp.[oC]

1 70 265 10 90 2302 75 265 11 90 2403 80 265 12 90 2504 85 265 13 90 2605 90 265 14 90 2706 95 265 15 90 2807 100 265 16 90 2908 105 265 17 90 3009 110 265

The paper presents the highlights experimentsfrom table 3

2.2 Results of the simulation

Following the simulation, we may analyze theresult presented in tables 4, 5, 6, 7 and 8 as well asin the figures attached to each experiment.


90

Table 4 The simulation resultsExperiment 1 – Figure 8,9,10

Actual filling time 0.32 (s)Actual injection pressure 28.440 (MPa)Clamp force area 24.8234 (cm^2Max. clamp force during filling 5.216 (ton)Velocity/pressure switch-over at% volume

99.02 (%)

Velocity/pressure switch-over attime

0.31 (s)

Estimated cycle time 7.38 (s)Total part weight 11.028 (g)Shot volume 18.4877 (cm^3)Cavity volume 11.8609 (cm^3)Runner system volume 6.6268 (cm^3)

Figure 8: The filling time distribution

Figure 9: The temperature at flow front

Figure 10: The pressure at end of fill



99.11 (%)


0.31 (s)







99.11 (%)


0.31 (s)





91

Figure 16: The pressure at end of the



99.00 (%)


0.64 (s)




Figure 19: The pressure at end of the fill



99.09 (%)

Velocity/pressure switch-over at 0.16 (s)

timeEstimated cycle time 7.56 (s)Total part weight 10.717 (g)Shot volume 18.4877 (cm^3)Cavity volume 11.8609 (cm^3)Runner system volume 6.6268 (cm^3)




2.3 Discussions

As a result of the experiments we can drawgraphics for the variation during the injectionprocess. In the figure 23 is presented a graphic forthe first 9 experiments, during which thetemperature of the melt temperature is constant.The figure 24 presents the graphic display of thevariation during the last 8 experiments, for whichthe temperature of the mold is constant.

Figure 23: The variation of the process time by themold temperature.


92

Figure 24: The variation of the process time by the melttemperature.

Analyzing these graphics we may observe thatthe optimal variant for designing the injectionprocess is by setting the technological parametersto: melt material temperature, 265oC moldtemperature, 90oC. The optimal values were usedwhile performing five experiment. The values forall the inlet parameters of the process may befound in table 5. The simulation may be visualizedin figures 11, 12 and 13.

If go back to Eq. 1, and set the experimentaldata into this formula, the cooling time for theinjection process using PA 6 polyamide may beexpressed as:

dr tt32.063.7 (5)Applying Eq. 3, we obtain:

Ktr 31.7 (6)Following there are several information

resulted from the simulation of number 5experiment during the multi-cavity injectionprocess.

Figure 25 graphically presents the variation ofthe temperature during the injection process. Thedifference between the maximum value of thetemperature and the minimum is 3,735 oC.

Figura 25: The temperature variation

Figure 26 presents the variation of the coolingtime. The difference between the maximum andthe minimum obtained values is 0,5337 s.

Figura 26: The cooling time distribution

During the simulation of the 10th and the 11th

experiment appeared a production fault, namedincomplete filling, presented in figure 27. As aresult of these two experiments, we had wastage.

Figure 27: The incomplete filling

3 ConclusionsAfter the experimental research the main

conclusions are as follow:- The time of deployment for an injection

process is increasing by maintaining constant thetemperature of the melting material and theincrease of the mold temperature;

- When the mold temperature is constant andthe melting material temperature is increasing, thedeployment time is decreasing;

- The injection time, as minimal value could beobtain using the overlap of figure 23 and 24 andthuts result the process parameters;

- The obtained results confirm that the injectionpressure decreases by the increase of the injectiontime if the melted material temperature is constant.The value of the pressure in this case decreasesfrom 28.440 MPa down to 26.145 MPa;

- If the mold temperature is constant, theinjection pressure decreases by the melted materialtemperature from 111.486 MPa down to 23.683MPa.

AcknowledgementThis paper was realised with the support of POSDRUCUANTUMDOC “DOCTORAL STUDIES FOREUROPEAN PERFORMANCES IN RESEARCHAND INOVATION” ID79407 project funded by theEuropean Social Found and Romanian Government.

4 References[1] Fetec u, C., Injectarea materialelor plastice,

Bucharest, 2005, pp 127 - 156[2] eres, I., 1996, Injectarea materialelor

termoplastice, Imprimeria de Vest Publisher, Oradea(1996), pp 254 - 300

[3] Fetecau C., et all. Overmolding injection moldingsimulation of tensile test specimen, InternationalJournal of Modern Manufacturing Technologies,Vol. II, No. 2/2010, Politehnium Publishing House,ISSN 2067-3604.

[4] *** Plastics Data Charts - www.plastemart.comAccesed: 23.12.2010

http://www.plastemart.com/


93

A PROBABILISTIC DYNAMIC MODEL ASSOCIATED TO AJACKSON NETWORK, WITH SERIES QUEUES;

THE SOLUTION OF THE ASYMPTOTICALLY STABLE STEADYSTATE

POPA MARIN, DR GAN MIH , POPA MARIANAUniversity of Bucharest, Technology Department, [email protected]

University of Bucharest, Technology Department, [email protected]

Abstract: This article studies in detail a subclass of Jackson networks class, namely the subclass ofcomputer networks with queues in series and comes to prove a theorem which is called the Jackson'stheorem, which provides a formula that gives an analytic expression of the probability distributions forthe asymptotically stable equilibrium state of this subclass.

The solution of this model in asymptotically stable equilibrium state will provide, at every moment,the probability that in the network nodes to be a certain number of transitions, in the waiting tail of thenode or in processing by the server of the respectively network node.

Keywords: computer network, JACKSON network, network with series queues, server, sorting process.

1. Preliminary

The behavior of computer networks ischaracterized by the presence of some congestionpoints of transitions, called network nodes [1, 2].

In every network node forms a waiting queuewhere the transitions arrived in the node wait to beselected according to the discipline associated tothe waiting queue, in order to be processed.

So we consider a network node as an entity madeof a server or processing unit and a waiting queue,depicted as in Fig. 1[1].

Collection of such network nodes and interactionbetween them forms a computer network withwaiting queues.

Definition 1.1 It’s called Jackson network acomputer network with waiting queues whereevery server is preceded by a waiting string and

where every transition, after has been processed bythe server, is sent in the queue of another server.

The name of such a network class comes fromthe name of the one whome, using stochasticprocesses, like multi dimensional processes ofbirth and death, discovered the solution ofmathematic model associated to the net in asimplified shape of product.

Jackson network class contains subclasses ofnetwork with series queues, with parallel queue,acyclic, with feedback and with queues in localbalance.

We’ll build a mathematic model associated to aJackson type computer network, obtaining adynamic probabilistic model.

The solution of this model will give, in everymoment, the probability that in network’s nodes to

arrivals Waiting queue server

departures

Figure. 1 Network node

mailto:marpopa2002:@yahoo.com

mailto:dragan_mihaita:@yahoo.com


94

be a number of transitions in the waiting queue ofthe node or in processing by the server of net node.

2. Theoretical resultsThe main result of the article requires the

following result due to Burke's demonstration,whose demonstration is in [2].

Theorem 2.1. If in a net node the process oftransition arrivals is Poissonian with parameter andthe process of transition processing is Poissonianwith parameter, then the process of transitiondepartures of net node is Poisson withparameter.

Theorem 2.2 We consider a computer networkmade of N net nodes in series, to which the influxtransitions is Poissonian with rate and theprocessing flow in net node i is Poissonian with rate

i, i= N,1 . If utilization factors 1,0i

i for

every i= N,1 , then the solution of model network inseries proper to the asymptotically stable steady state

is: p(n1,n2,….nN) =N

i

nii

i

1.1

ProofWe note iQ the number of transitions in node i

at time Ni ,1, [1]Sought to determine the probability that at timeto have ni transitions in node i, Ni ,1 , that is the

probability:NNN nQnQnQPnnnP ,...,,,,...,, 221121

In the asymptotically stable state for anyNini ,1,0 there is ,,...,,lim 21 NnnnP and

is finite, value noted Nnnnp ,...,, 21 .The probability distribution Nnnnp ,...,, 21 ,

proper to the asymptotically stable steady state,shows a description of the computer net averagebehavior on long-term.

We noteNnnnS ,...,, 21

the network state proper to

the case in which, for Ni ,1 , in node i there are ni

transitions.Thus Nnnnp ,...,, 21 =

NnnnSP ,...,, 21 that is the

probability distribution of the asymptotically stablesteady state represents the probability that thenetwork be in state

NnnnS ,...,, 21. Will obtain these

probabilities considering the mathematical modelassociated to the network as a stochastic process typebirth and death multi-dimensional.

Next, we demonstrate the assertion fromenunciates through induction after N(N 2).

Case N=2.Consider a JACKSON network with two

series queues, as in Fig.2 in which the transactionsarrived at the end of queue 1,after a Poisson flow ofrate , waits to be processed by server of node 1.After a transaction is processed by this server, willgo at the end of the queue of node 2, where waits tobe processed by this node’s server, while server 1selects a new transaction from it’s queue, accordingto the order chosen for the queue. As per BURKEtheorem, the arrivals to queue of node 2 are alsoPoisson of parameter + 1, where 1 is the rate oserver’s 1 processing. Thereby the arrivals to queueof node 2 are a Poisson flow of rate + 1. Becausethe rate of transitions processing by server 2 is 2according to BURKE theorem, transitions that leavethis network are a Poisson flow of rate : + 1 + 2.

Figure. 2.

Follow to determine the probability that at timeto have n1 transitions in node 1 and n2 transitions innode 2, that is:

P(n1,n2, ) = P(Q1( ) = n1 Q2( ) = n2).Therefore, we will calculate successively the next

probabilities:P(0,0, + ) = probability of not having

transitions in any node at time . P(0,n2, + ) =probability of not having only

in node 2, n2 transitions at time(n2 1).

P(n1,0, + ) = probability of having only innode 1, n1 transitions at time (n1 1).

P(n1,n2, + ) = probability of having n1transitions in node 1 and n2 transitions in node 2 at

time 1( 1n i ).12n

queue2

queue1

Nodeof net 1

node ofnet 2

Arrivalspoissonn

departures

server2

server1


95

Using the hypothesis of a stochastic process typebirth death, we obtain the following:

P(0,0, + ) = P(0,0, ) ·(1- ) + P(0,1, ) 2 · (1- ) + 0( )

P(0,0, + ) = P(0,0, ) – P(0,0, +2P(0,1, + 0( )

),0,0(),0,0( PP

.)(0

),1,0(),0,0( 2 PP

Passing to limit after 0 obtain:

),1,0(),0,0(),0,0( 2 PPPdd

P( 0,n2, ) ==P(0,n2, )[1- )[1- 2 ]+P(1,n2–1, )· 1 [1-- )[1- 2 ]+P(0,n2+1, )· 2 [1- )+0( ) .

),,0(),,0( 22 nPnP =

= -( 2 + )P(0,n2, )+ 1P(1,n2 – 1, ) + 2P(0,n2+

+ 1, )+ )(0 .


dd P(0,n2, )=-( + 2)P(0,n2, )+ 1P(1,n2-1, )+

+ 2P(0,n2 +1, )Analogue :

P(n1,0, )=P(n1,0, )[1– )[1– 1 ]++P(n1–1,0, )· [1- 1 ]+P(n1,1, )· 2 [1- )[1-

1 ] + 0( )),0,(),0,( 11 nPnP =

= -( 1+ )P(n1,0, )+ P(n1-1,0, )+ 2P(n1,1, )+ )(0

Thus:

dd

P(n1,0, )=-( + 1)P(n1,0, )+ P(n1-1,0, )+

+ 2P(n1,1, )Finally:

P(n1,n2, )=P(n1,n2, )[1– )[1– 1 ][1– 2 ]++P(n1 – 1,n2, )· ·[1– 1 ][1– 2 ] ++P(n1,n2 +1, )·[1 – )[1 – 1 ] 2 ++P(n1+1,n2 –1, )·[1- ) · 1 ·(1 – 2 ) + 0( )

),,(),,( 2121 nnPnnP

=-( 1+ 2+ )·P(n1,n2, )+ P(n1–1,n2 )+

+ 2·P(n1,n2+1, )+ 1P(n1+1,n2 –1, )+ )(0


dd P(n1,n2 , ) = -( + 1 + 2)·P(n1,n2, ) +

+ P(n1–1,n2, ) + 1P(n1 +1,n2 –2, ) + 2·P(n1,n2 +1, )Conclusively, for the model of computer network

with two series nodes, is obtained the followingsystem of differential equation:

),1,(),1,1(),,1(

),,()(),,(

),1,(),0,1(),0,()(

),0(

),1,0(),1,1(),,0()(

),,0(

),1,0(),0,0(),0,0(

21221121

212121

12111

1

222122

2

2

nnPnnPnnP

nnPnnPdd

nPnPnP

nPdd

nPnPnP

nPdd

PPPdd

0,1),,( 20,

121

nnPnn

(complimentarily

condition)In general case, this system is extremely difficult

to resolve. In the stabile asymptotically balance isknown that for any n1,n2 0 there is, and id finite :

),(),,(lim 2121 nnpnnPThe probability distribution p(n1,n2) of stabile

asymptotically balance delivers a description of theaverage behavior of the computer network long-term. In this condition, the system above becomes:

1

01111

01010010110

01000

20

1

212211

212121

12111

222122

2

21

)n,n(p

)n,n(p)n,n(p)n,n(p)n,n(p)(

),n(p),n(p),n(p)()n,(p)n,(p)n,(p)(

),(p),(p

n,n

We consider:p(n1,n2) = 0, n1 < 0 or n2 < 0

1p(0,n2) = 02p(n1,0) = 0.

relations that are natural if we think of theirinterpretation.

Thus 1p(0,n2) is the probability of processing atransition in node 1, but in this node we have notransition, and so 1p(0,n2) = 0.

Analogue 2p(n1,0) = 0.In these hypothesis, the system of the first 4

equations proper to the state of stabile


96

asymptotically balance is reduced to a singleequation[1, 2], which is:

p(n1 –1,n2) + 1p(n1 +1,n2 –1) + 2p(n1,n2 +1)== ( + 1 + 2)p(n1,n2)

From this relation, deduce that in case of stabileasymptotically balance, the flow of each state isconserved, that is for any state, the input flowcoincides with output flow of this state.

To this relation adds the obviousrelation 1),(

0,21

21 nnnnp .

If note21 ,nnS the network state appropriate to the

case in which in node 1 there are n1 transitions and innode 2 there are n2 transitions, in the relation abovesee that the states to which and from which have thetransitions with state

21 ,nnS are:

1,,11,1 212121,, nnnnnn SSS as inputs of state

21 ,nnS

1,1,1,1 212121,, nnnnnn SSS as outputs for

21 ,nnS .The interaction of these states is given by diagram

in Fig. 3.

Fig 3Because in any system in stabile balance, the flow

rate of input in a state is the same with the flow rateof output from that state, and that is for any state ofthe system, results that in our case, by writing thisrelation of flow’s preservation for state

21 ,nnS ,obtain:

Flow rate of input for :

21 ,nnS = p(n1–1,n2)+ 1p(n1+1,n2-)+ 2p(n1,n2+1).Flow rate of output from :

21 ,nnS = ( + 1 + 2) p(n1,n2).

By condition of flow’s preservation in21 ,nnS obtain

relation:

p(n1-1,n2) + 1 p(n1 + 1,n2-1) + 2 p(n1,n2 + 1)= =( + 1 + 2) p(n1, n2)

which is actually the relation equivalent to theprobability system in case of stabile asymptoticallybalance.

Notice that if we put into a rectangular networkn1On2, the neighbor states for

21 ,nnS ,then every flow

rate is well targeted, which is: from left to right,

2 downwards and 1 direction SE-NV.Since in the considered computer network with

series queues, the two net nodes are independent,Jackson [1] comes with the idea of looking forsolutions to the system corresponding to stabileasymptotic balance as a product, that isp(n1,n2)=p1(n1)·p2(n2), where pi(ni) is the probabilityof having in node i, ni transitions in state of stabileasymptotic balance and so:

022

011

21

11nn

npsinp (*)

With this choice, the equation above becomes:1p1(n1 + 1) p2(n2 – 1) + p1(n1 – 1) p2(n2) +

2p1(n1) p2(n2 +1)=( 1 + 2 ) p1(n1)p2(n2) (1)For n1 = n2 = 0 and using the hypothesis above,

obtain the equation:1p1(1)p2(-1) + p1(-1)p2(0) + 2p1(0)p2(1) =

= ( + 1 + 2) p1(0)p2(0)2p1(0)p2(1) = p1(0)p2(0)

p1(0)[ 2p2(1) – p2(0)] = 0.We have the following situation possible:

1.If p1(0) = 0 in relation (1) above n1 = 0 obtain:1p1(1)p2(n2 - 1)+ p1(-1)p2(n2)+ 2p1(0)p2(n2 +1) =

( + 1 + 2) p1(0)p2(n2)Using the hypothesis above and p1(0)=0 obtain:

1p1(1) · p2(n2-1) = 0If p1(1) = 0 is demonstrated by induction that

p1(n1) = 0 for any n1 0.If p2(n2 – 1) = 0 results that p2(n2) = 0 for any n2

0.Both situations are false, because they contradict

relations (*) above.

2.If 2p2 (1)– p2(0)=0 p2(1)=2

p2(0) (2)

With this value go in relation (1) after n2 = 0.Have: 1p1(n1 + 1)p2(-1) + p1(n1 – 1)p2(0) +

+ 2p1(n1)p2(1) = ( + 1 + 2) p1(n1)p2(0)Results: p1(n1 - 1)p2(0) + 2p1(n1)p2(1) ==( + 1) p1(n1)p2(0)

Replacing p2(1) with it’s value from relation (2)obtain:

Sn11,n2+1

Sn1,n2+1

Sn1,n S

n1+1,n

Sn1-,n2

Sn1+1,n2-1

Sn1,n2-1


97

p2(0)[ p1(n1–1)– + 1)p1(n1)]+ 2p1(n1)2

p2(0)=0

p2(0)[ p1(n1 – 1) – p1(n1)] = 0.If p2(0) = 0 is shown that above are obtained null

values for p1 and p2, which is false.

Thus, results that for p1(n1) =1

p1(n1 - 1) for any

n1 1.

Writing this relation as :)1(

)(

1

1

kpkp =

1, k 1

and doing the product:1

1 1

1

)1()(n

k kpkp

=1

1

n

obtain1

11

11

)0()(

n

pnp

and so )0()( 11

11

1

pnpn

(**)

From relation:01

111n

np

01

01

1

1

,1

1

111

11)0(10n

n

n

n

pp ,

if1

< 1, so that the series is convergence.

Note 1 =

1

and call 1 utilization factor in net

node 1.So for 1 (0,1) have p1(n) = (1 – 1) 1

n, n 0,(3)

Use this result in relation (1) after n1 = 0.Obtain:

1p1(1)p2(n2–1)+ p1(-1)p2(n2)+ 2p1(0)p2(n2 + 1)==( + 1 + 2) p1(0)p2(n2).

From hypothesis result:1p1(1)p2(n2–1)+ 2p1(0)p2(n2+1)–( + 2)p1(0)p2(n2) = 0

Replacing p1 with value found in relation (**)obtain:

11

p1(0)p2(n2–1)+ 2p1(0)p2(n2+1)–

-( + 2) p1(0)p2(n2) = 0p1(0)[ p2(n2 –1)+ 2p2(n2 +1)–( + 2)] =0We’ve seen above that p1(0) 0 and so:

[p2(n2–1) – p2(n2)] = 2[p2(n2) – p2(n2 + 1)], n2 1.Writing this relation as:

[p2(k –1) – p2(k)] = 2[p2(k) – p2(k + 1)], k 1and making the sum of k from 1 to n2 obtain:

2 2

1 122222 1)()1(

n

k

n

kkpkpkpkp ,

relation equivalent with:[p2(0) – p2(n2)] = 2[p2(1) – p2(n2 + 1)]

But from (2) p2(1) =2

p2(0). Replacing this

above and reducing the term p2(0) obtain:

p2(n2 + 1) =2

p2(n2)

In doing the above we obtain successively:

00

10

1

22

2222

22

2

1

0 2

2

22

2

22

22

pnp)(p)n(p

)k(p)k(p

k,)k(p

kp

nn

nn

k

from

02 02

2

101 22

22n n

n

pnp

,11)0(2

2

2

02

2

n

np

if2

< 1 so that the series to be convergence.

Note2

2 and call 2 factor of utilization

in node 2.Thus for 2 (0,1) obtain:

nnp 222 )1()( , .0n (4)Form relations (3), (4) and from :p(n1, n2) = p1(n1) · p2(n2) obtain the solution for

the state of stabile asymptotic balance of the modelas being:

0,),1)(1(),( 2121212121 nnnnp nn ,

where 1,0,1,02

21

1 .

Case N>2.Assuming now that the relation form the

statements true for 1,2 Nk .For k= N - 1 take the first N – 1 series nodes

and form the net 1 for which we know thesolution:

p(n1,n2,….,nN – 1) = ,11

1

N

i

nii

i according the

hypothesis of induction.


98

Thus is obtain connecting in series 1 withnet node N, as in Fig. 4.

NoteNnnS , state corresponding to :

n = (n1,…,nN – 1) and nN andp(n,nN)= p(

NnnS , ) = )( ,,..., 11 NN nnnSp ==p((n1,…,nN – 1),nN) = p(n1,…,nN–1,nN).

As passing from 1 in net node N is madethrough net node N –1, it can be built the nextdiagram of flow between states (similar to case oftwo series net nodes)(see Fig.5) in which

)1,,...,(1 121 NN nnnn and)1,,...,(1 121 NN nnnn

Write relation of flow preservation in state

NnnS , and for resolving the obtained system welook for solutions like:

p(n,nN) = p(n)pN(nN).The preservation relation is:

N –1p(n + 1)pN(nN – 1) + p(n – 1)pN(nN) + N

p(n)pN(nN +1) = ( + N –1 + N) · p(n)pN(nN) (5)

Sn-1,nN+1

Sn-1,nN

Sn,nN+1

Sn+1,nN

Sn+1,nN-1

Sn,nN

Sn,nN-1

N-1N-1

N-1 N-1

NN

N N

Fig. 5

Arrivalspoisson depart

nod N

nod N-1nod1

1

Figure. 4


99

Using that p(n) = inii

N

i)1(

1

1 (results

from induction hypothesis) it will be obtained arelation of recurrence of order I for pN andtherefore: Nn

NNNN np )1()( , where

)1,0(N

N .

Thus : p(n1,…,nN –1 ,nN) = p(n,nN)=p(n)pN(nN)=

= iNi nii

N

i

nNN

nii

N

i)1()1()1(

1

1

1.

Indeed, replacing p(n) in relation (5) obtain:

2

111

2

1111

)()1()1(

)1()1()1(

N

iNNNNi

ni

N

iNNNNi

niN

npnp

npnp

i

i

2

1111

2

111

)()()1()(

)1()()1(

N

iNNNNi

niNN

N

iNNNNi

niN

npnp

npnp

i

i

We divide this relation to2

1)1(

N

ii

ni

i and

)()1()(

)1()1(

)()1(

)1()1(

111

11

11

1

11

11

1

1

1

1

NNNnNNN

NNNnNN

NNNnN

NNNnNN

np

np

np

np

N

N

N

N

Divide through )1( 11

11

NnN

N and obtain:

)1()()1(

1

211

NNNN

NNNNNN

npnpnp

)()( 11 NNNNN npReduce the term )( NN np with

)()(1

111 NNN

NNNNN npnp , divide

with1

1N

N and obtain:

)()()1()1(11

NNN

NNNNNNN

npnpnp

)]1()([)]()1([

NNNNN

NNNN

npnpnpnp

a relation similar to the one form which weobtained p2. Forwards, proceed as for p2 and finallyobtain pN as being Nn

NNNN np )1()( where

)1,0(N

N .

In order to obtain this result, it was also needed

the relation )0()1( NN

N pp which is get if in

relation (5) put n = 0 and nN = 0.Obtain:

)0()0()()1()0()0()1()1()1(

1

1

NNN

NNNNN

pppppppp

)0()1(

0)]0()1()[0(

NN

N

NNN

pp

ppp

because p(0) 0.

Observations[1].

1) Jackson has shown that forma the product ofthe final relationship of a mathematical model isavailable also in a more general case, in which areallowed transitions between states form:

Nj nnnS ,...,,...,1 Nj nnnS ,...,1,...,1, a transition enters

from the environment directly to the queue ofprocessor j.

Ni nnnS ,...,,...,1 Ni nnnS ,...,1,...,1, a transition leaves

the system (computer net) through net node i.Nji nnnnS ,...,,...,...,1 Nji nnnnS ,...,1,...1,...,1

,a transition passes from net node i straight tonode j.

2) Also, Jackson has shown that the finalsolution is obtained as a product for a more generalclass of networks in which is allowed processingthe transitions several times in a certain net node,that is allowing cycling the transitions betweenentering and leaving a certain net node.

3. Conclusion

This article presents a concrete modality tomodel systems for the particular case of computernetworks with series queues, which is themathematical method that leads to a equationssystem, sometimes impossible to resolve withoutimposing supplementary hypothesis which are, insome cases, extremely restrictive, hypothesis thatease up the model from the real modeled system.

In this case, working in the asymptoticallystable equilibrium state, the differential equationssystem that’s obtained, impossible to resolve ingeneral, is reduced to a resolvable equationssystem, which will provide, through the probabilitydistribution p(n1, n2, ..., nN), a description of the


100

medium behaviour on long term of the computernetworkwith series queues.

In the same way the other subclasses of Jacksonnetwork can be modeled, and also network typeBCMP, BUZEN, etc and the theoretical results canbe compared by mathematical models forperformance indicators, such as: using a node, usetwo or more node in the same time, residence timeof transactions in network, medium length ofwaiting tails, a node’s efficiency and many other,with the results given by modeling through othermethods, such as modeling trough Coloured Petrinets [3,4].

References

[1]. R. F. Garzia, M. R. Garzia, “Networkmodelling, simulation and Analysis”, Ed.Marcel Rekker, New York and Basel, 1990.

[2]. M. Popa, “Modelarea matematica a retelelorde calculatoare”, Ed. Univ. of Bucharest, 2004

[3]. M. Popa, Mariana Popa., M. Dragan,“Modelling patient flow in a medical office bystochastic timed Coloured Petri Nets”, the 3rdInternational Conference onTelecomunications, Electronics andInformatics, May 20-23, Ed. UTM, Chisinau2010.

[4] M. Popa, M. Dragan, Mariana Popa,”Study theMarkings Trap in the Coverage Graph of aColored Petri Nets(CPN)”, 16THInternational Conference on Soft Computing,June 23-25, Brno, Czech Republic, 2010.


101

TECHNICAL SOLUTIONS TO DESIGN AN EQUIPMENT FORICE PARTS RAPID PROTOTYPING

Nicolae IONESCU1, Aurelian VI AN2, Alexandru SAVIN3, Mihai TRIF NESCU4

1 POLITEHNICA University of Bucharest, [email protected] POLITEHNICA University of Bucharest, [email protected].

3 POLITEHNICA University of Bucharest, [email protected] POLITEHNICA University of Bucharest, [email protected].

Abstract: The paper presents some detailed technical solutions to design an equipment for ice partsRapid Prototyping. The authors designed two alternative solutions of the equipment applying amethodology developed within the TCM Department of the POLITEHNICA University of Bucharest.Both alternative solutions are based on the operation principle of selective freezing with liquid nitrogen ofwater sprayed on the working surface according to section image conveyed by computer by slicing aCAD model. The article focuses on the methods of solving several issues of conception, design andachievement of the high complexity equipment.

Keywords: Rapid Prototyping, ice, equipment, design

1. Introduction

Rapid Prototyping procedures has gained widerapplicability recently, as a modern manufacturingtechnique, due to short time machining of a largetype of parts by using this method. As it is wellknown, there are currently applied with goodresults several rapid prototyping techniques, suchas [7]: Stereolithography - SLA, Selective LaserSintering – SLS, Solid Ground Curing – SGC, 3DPrinting, Fused Deposition Modelling – FDM,Laminated Object Manufacturing – LOM and Ink-Jet Printing.

Each of these methods had both advantages anddrawbacks, yet what all have in common is the factthat the part is achieved by material deposition notremoval, as it is the case with classicaltechnologies.

Not withstanding all the latest improvements ofthe recent years, all these rapid prototypingmethods have a major drawback in common that isthe very high cost of parts achieved, due mainly tohigh cost of equipment and consumables.

One way to cut down the rapid prototypingcosts is to use less expensive raw materials.Therefore, rapid achievement of ice prototyping isbased on using very cheap material, namely water.At this moment, world wide achievements in thisdomain are but at an early stage of experimentalresearch [2, 3, 4, 5, 6].

This article presents the principal aspects of theextended research work carried out by the authors,ending up with the achievement at final stage of anequipment for ice parts rapid prototyping.

Considering the current research carried out tothe present moment, the authors are entitled tostate that the achievement of this equipment wouldbring about special effects, both in scientific terms,and practical applications. Among the mostimportant domains of application for the ice partsrapid prototyping that are currently applied withspecial outcome, it should be mentioned here ofthe achievement of the fusible casting moulds [1,8] and medical bio-models.

2. The principle of ice prototypes productionMost of the rapid prototyping techniques are

still very expensive; in addition, some of themgenerate dust and smoke emissions, thusendangering the operators’ health. One of them isthe method „Rapid Freeze Prototyping – RFP”,which is a remarkable technique [5, 6], featuringhigh technical characteristics,

This method helps to obtain the intendedtechnical characteristics, as well as higheroperation speed, without generating noxiousagents, while the working material – water, ismuch less expensive than any material used for theother rapid prototyping techniques [6].

mailto:ionescu_upb:@yahoo.com

mailto:aurelianvisan2003:@yahoo.com

mailto:savinalexandru:@yahoo.com

mailto:trify_tl:@yahoo.com


102

Based on the RFP technology, a three-D modelcan be obtained of ice by using the CAD model,through selective deposition of water and waterfreezing layer by layer, in conformity with theprinciple technical diagram shown below in figure1.

In this case, according to figure 1, water issprayed on a table and freezed by simultaneousspraying of liquid nitrogen. The part formationtakes place in a tank with constant low temperaturemaintained through thermostat control system.After deposition of the first layer, the nozzle islifted, or the table is lowered with the thickness ofthe deposition layer. At the end of the indicatedperiod when total water freezing takes place, theprocess is resumed with the deposition of a newwater layer on top of the first.

Drop – on demand nozzle

Feeding pipe

Water droplet

Ice part

Figure 1: Principle of Rapid freeze prototyping [2]

When the entire part is ready, the part thusobtained can be used immediately, or it may bekept in a refrigerator for a distant use.

3. Conceptual design of the ice partsrapid prototyping equipment

The conceptual design of the RFP equipmenthas been achieved based on a methodology thathad been developed during several years by theTCM Department of POLITEHNICA University.This methodology has been applied with specialresults, both in didactic events, and in graduationdiploma studies of the master degree, as well as inprojects that are being in progress.

This methodology started to be applied whenthe need has been identified and defined to rapidlyobtain cheap prototypes. Based on a questionnaire,the client’s demands regarding the rapid

prototyping equipment have been examined andgrouped according to ten primary requirements andsequenced in hierarchy.

The next step was the analysis of competingproducts, that is, in principal, the rapid prototypingequipment corresponding to the most relevanttechniques, as mentioned in Introduction to thisarticle. The authors followed this procedure toensure competitive design of this type ofequipment, in comparison with the most advancedexisting international technologies; the equipmenttechnical specifications have been determinedaccordingly.

The next step was to prepare the functionaldesign, by determining the general function of theequipment that is ice prototyping in a wide rangeof standard dimensions. The general function wasthen submitted to a general examination procedure,resulting in the first place the principal functions,and then the complementary and supplementaryfunctions.

Considering the principal functions, a list wasset up of the critical functions, which determine thecommercial success of the product whichcorrespond to the size and demands with relativelyhigh importance: ensure the movement on the threeaxes, conveying of information from computer toequipment, water spraying on strictly controllablesectors, simultaneous spraying of liquid nitrogen toinduce freezing and maintaining a negativetemperature in the part processing area.

To determine the partial solutions at conceptuallevel, an external research was prepared mainlyconsisting in the study of several patents,consulting several experts in this domain, andinterviewing top users.

Following the sorting of concepts, two of themwere retained, which are shown in figures 2 and 3.

Figure 2: Concept no. 1


103

Figure 3: Concept no. 2

Applying the multi-criteria method of analysis,Analytically Hierarchy Process (AHP) – weselected concept no. 1, to be further examined atthe detailed design phase.

4. Detailed design of the ice parts rapidprototyping equipment

At this stage, the detailed design was preparedof the product, considering both the total assemblyand its component parts, to provide the prototypedocumentation.

In the framework of the constructive design ofthe total assembly, the product architecture wasprepared, the configuration of parts andsubassemblies, as well as their preliminarydimensioning, based on their functional role.

Design was carried out, while taking intoconsideration a series of criteria in accordance withthe technical service life cycle of the product, suchas its functionality, usefulness, ergonomy, labourprotection, ecology, aesthetics, manufacturing,product disposal and minimal cost.

For the selected product concept, the productarchitecture was determined in several stages, asfollows.

To better observe the interfaces between theproduct blocks, an approximate constructiveproduct plan was first created comprising fivelarge blocks – the system of transmissions,actuation, the item corpus, the water circulationsystem and the liquid nitrogen circulation system.Next, we developed the detailed constructive planof the part, as indicated in figure 4 and weidentified the fundamental and incidentalinteractions.

Figure 4: Detailed constructive plan of the product

Based on this approach, the authors preparedthe final block diagram, stating the constructivesolutions for the principal interactions (figure 5).

Figure 5: Final block diagram

Based on the above mentioned, two alternativesolutions were prepared of the rapid prototypingequipment.

In the first solution, the horizontal movement isperformed by the working head, while the verticaldisplacement by the mass of the equipment. In thesecond alternative solution, the displacements onthe three axes are provided by the working head.

Figures 6, 7, 8 and 9 show the constructivesolutions adopted under the first alternativesolution.

The shape of the working head is shown infigure 6. The water is pumped by the cylinder andpiston unit upon disconnecting of power magnet,then water is injected through nozzle, when thepower magnet is actuated through computercontrol. Injected water comes in contact with thecold environment produced by liquid nitrogen andfreezes, getting thus attached to the deposition icelayers. This shape of the working head leads to thesignificant decrease of the ice layer deposition


104

time, as compared to the working head with waterdropping on the working surface.

Water injection nozzle

Valves

Spring

Electromagnet

Piston

N2

N2 injection nozzleWater

Figure 6: Construction of the working head

Figure 7 shows the working head movementprinciple. Movement is conveyed through motorV-belt gear assisted by planetary rotary speedreducer. The movement of the working head isachieved by linear transducer (with photodiodes)and calibrated ruler. The intention was to improvethe positioning precision in the case of higherworking speed.

Motor

Columndirections

Dewar tank

Displacementtraducer

Working head

Trapezoidal toothedbelt drive

Scale rule to measuredisplacement

Figure 7: Working nozzle movement principle

The less sophisticated shape was selected forthe chamber of ice part prototyping and constanttemperature maintenance, to minimize the cost ofpaying off the prototyping equipment and the icepart prototypes, respectively (figure 8).

WorkingTable

Ball screw nutmechanism

Refrigerationsystem

Worm and wheelelevator

Figure 8: Ice part prototyping and temperaturemaintenance chamber

Figure 9 shows a general view of the firstalternative solution of ice parts rapid prototypingequipment.

Figure 9: General view of RFP equipment firstalternative solution

The constructive solution for the secondalternative was designed to provide the movementof working nozzle along the three axes. Figure 10shows the general view of the RFP equipment inthe second alternative solution.


105

Figure 10: General view of RFP equipment secondalternative solution

The principal constructive elements of the unithave been selected, or, from case to case, designed,as described below.

For the selection of motors, we examinedseveral alternatives from the products ofmanufacturers available on the specialised market,considering DC motors in miniature variant, withor without reducer. The authors considered linearmotors, servomotors and step by step motors.

Linear motors have the advantage of highprecision and reliability, high feeding speed andacceleration and relatively high cost with respect tothe cost objectives envisaged for the RFPequipment. Other shortcoming of this item is that itneeds a source of compressed air to make itpossible to maintain the distance between rotor andstator, which makes the equipment a rathercomplicated construction.

Servo motors are direct motors which requireposition actuators (encoders), which increasesignificantly the cost of equipment. They arehowever less expensive than linear motors, butshow lower repeatability performance.

Step by step motors rotate at a specified angularstep and require no position actuator. They are notexpensive and require no sophisticated electroniccontrol units. Other advantage is their rotation by1/2 to 1/32 of the normal step, thus increasing thepositioning precision, yet at the expense of smallerfeed speed. The principal disadvantage of these

types of motors is the so-called phenomenon ofelectro-magnetic skidding at high speed, that is themotor skips over rotation steps. This would resultin low feeding speed, which is a minordisadvantage, considering the smaller CNCoperation travel (150x200X350mm) in the case ofthis construction of the authors. To prevent electro-magnetic skidding of motors, the feed speed hadbeen limited to 500 mm/min. maximum, whichmeans smaller equipment output, still better qualityof the machined surface.

Considering the above-mentioned, the authorsselected step by step three-motors manufactured byNEMA.

Also, due attention was paid to transmissionscrews: we selected ball screws produced by ISELGmbH, made of special hardened steel CIF at60HRC, which ensures high wear resistance andcore elasticity.

The screw thread flank is obtained throughrolling, which leads to high precision ofpositioning of the nut, with respect to the leadingscrew. The ball nut also from the producer ISELGmbH made of steel 16MnCr5 hardened andnitrated in gas atmosphere, which provides lowwear percentage of the rolling flank.

The authors selected a free software equipmentthat can be purchased via Internet, with the resultthat the cost of this component part of the RFPequipment is zero.

The performances of this software arecomparable with those of genuine softwareequipment, its technical characteristics beingtotally compatible with the design RFP equipmentof the authors.

This software makes the translation frommachine code ‘’G-code’’ into one decodable by theequipment controller, thus enabling the executionof rapid working feed movements, rotary speedand other auxiliary movements. This codeconversion is made by the so-called‘’Postprocessor’’ specific to the RFP equipmentjointly working with the electronic controller.

Generally, this software is used with industrialPCs integrated in the equipment construction, butin the case of the equipment conceived by theauthors, this software will be installed on a laptopcontrolling the equipment.

Two software items, Mach3 and K-CAM, havebeen examined, the second being selected, sinceMach3 requires a stronger computer for itsoptimum operation and in addition, it was noticedthat this software evinces a delay between the


106

computerised control and the real control responsefrom the equipment engines.

The electronic controller provides theconnection between software and hardware, beingcompulsory for the control function of the RFPequipment, which is feasible in two alternativesolutions:

1. The G code is hardware-integrated incontroller, which requires more complexelectronics and is more expensive.(microprocessors and ROM memory).

2. Code G is software-generated and requireslower costs.

The power supply for the switch unit is highoutput, which is specific for these types of powersupply systems and is available for acquisition atConex Electronic Bucharest.

For water supply, the authors found the solutionof customizing a pump used for car wash, namelyfront screen wash. This type of pump is computercontrolled in accordance with the slicing section ofthe CAD model, water being pushed towards thespraying system above the part.

The equipment is endowed with a liquidnitrogen storage and circulation unit comprisingDewar tank, insulated pipes, valves, electric fusefor the tank pressurisation, etc. By the heating ofthe electric fuse fixed on the pipe for nitrogenextraction from the tank, part of the liquid nitrogengets evaporated and pressurises the tank. Thus, thenitrogen is forced up the discharge pipe andreaches the working zone through opening a valveactuated in correlation with the way in which wateris sprayed.

The prototype of second alternative solutionwas produced, tested and standardised in the Dr.Kocher Center inside the POLITEHNICAUniversity of Bucharest.

5. ConclusionsBased on the above-mentioned descriptions and

definitions made by the authors, severalconclusions can be stated, as follows:

1. The authors proposed a solution of rapidprototyping cost reduction consisting in theutilization of less expensive raw material in theprocess. Ice parts rapid prototyping involves theutilisation of a very cost-efficient material, whichis water.

2. Conceptual design of RFP equipment isbased on the methodology developed duringseveral years of research by the authors within the

TCM Department of POLITEHNICA University ofBucharest.

3. Two alternative solutions were prepared ofthe RFP equipment. The first solution containshorizontal movement by aid of working head, andvertical movement by aid of the equipment table.In the second alternative solutions, the movementon the three axes are performed by the workinghead.

4. Both solutions are based on the sameoperation principle of selective freezing usingliquid nitrogen, of the sprayed water on theworking surface, in conformity with the slicingsection of the CAD model.

References[1] Beeley P. R. and Smart R. F., Investment

Casting, 1st Edition, The University Press,Cambridge, UK, 1995.

[2] Leu M. C., Zhang W., and Sui G., Anexperimental and analytical study of ice partfabrication with rapid freeze prototyping,CIRP Annals, 49(1), pp.147-150, 2000.

[3] Liu Q., Sui G., and Leu M. C., Experimentalstudy on the ice pattern fabrication for theinvestment casting by rapid freeze prototyping(RFP), Computers in Industry, 48(3), August,2002.

[4] Peters D. M. Patterns in ice, FoundryManagement & Technology, 123(8), 1995.

[5] Yodice A., Freeze cast process, US Patent5,072,770, 1991.

[6] Yodice A., Freeze process cuts casting costs,Advanced Materials and Processes, 155 (4)(1999) 35-36.

[7] Wohlers Terry, Rapid Prototyping Mouldsthe Future of Manufacturing Available from,http://www.wohlersassociates.com.

[8] Wu X., Study on ceramic mould investmentcasting based on ice patterns made by rapidprototyping method, Bachelor Degree Thesis,Tsinghua University, Beijing, China,1997.

http://www.wohlersassociates.com./


107

COMPOSITE SHAPE MEMORY ALLOYS USED IN ENERGYDISSIPATION APPLICATIONS

I. Cimpoe u*, S. Stanciu, A. Enache, D. Dan , V. Manole, P. Paraschiv

The ”Gh. Asachi” Technical University of Iasi Faculty of Materials Science and Engineering, Prof.dr. docent Dimitrie Mangeron Rd., 61 A, 700050, Ia i, Romania

Abstract: Shape memory alloys exhibit through them martensitic transformation a high dampingcapacity that can be used in many applications. Microstructural and chemical analyses were madeconcerning a shape memory alloy based on copper that can be use as matrix material in a compositematerial for mechanic energy dissipation. Scanning electrons microscope and X ray analysis equipmentwere used to characterize some material properties.

Keywords: shape memory alloys, damping capacity, martensite transformation

1. Introduction

Shape memory alloys, because of theirunique thermo-mechanical behavior, havedeveloped considerable works towards theperformance of intelligent materials andstructures. Indeed, these alloys are particularlyuseful when large deformation and recovery ofthe shape is observed under a small rate ofstress or temperature. These properties areattributed to the ability to undergo martensiticphase transformation. This approach presentsthe advantages to be easily incorporated in astructural computation. Shape memory alloyshave ability to modify both their shape andproperties in response to the thermo-mechanical environment. They may beembedded into an inactive material totransform it into an active composite: by asuitable choice of their volume fraction, shapeand alignments, they accomplish a specificfunction. Many applications can be profitablyexplored in aerospace structures, vibrationcontrol among others.

Actuation by shape memory alloys (SMAs)has generated considerable interest and someapplications have been developed across awide range of industries. Successfulapplications are still largely restricted to nichemarkets, although the extent of theirapplications is growing rapidly. Two factorsseem to have restricted growth into othermarkets. These are the early dominance by themedical market, where small sections of low

temperature materials have been developed.This has partially led to the second limitingfactor, which is the lack of availability oflarger engineering sections and techniques.This second limitation is partially due totechnical difficulties in scaling up of thematerials and techniques, plus a lack ofvisibility of the benefits.

Nonetheless, impressive demonstrations oflarger scale advanced actuator applicationshave been achieved, particularly in theaerospace industry. Applications within a gasturbine are an order of magnitude moredifficult due to the more robust requirementsand higher temperatures encountered [1].Many research grants are in hand to investigatethe opportunities for smart technologies to gasturbines, with SMAs attracting considerableinterest. The opportunities for enhancedperformance and cost reduction are generallyobvious within the industry, but this must beachieved with realistic cost and reliability fromthe technologies.

In general, a high integrity, robust and stiffstructure must be achieved. This has led toseveral programme investigating largerstructures and supporting technologies such asjoining techniques. The control of noisepollution has become increasingly important inthe aerospace industry. The generation andradiation of noise from aero-engines poses aproblem of immense complexity. Over recentyears, noise reduction has become a prime


108

requirement in the specification for next-generation propulsion systems.

Engine manufacturers are constantlyinvestigating ways that will reduce the noisecreated by gas turbine engines. CommercialSMAs are usually available ‘off the shelf’ inthe form of wires and thin strips. The SMAbusiness is very much ‘supply driven’ andwithout major demands for large sectionSMAs, suppliers are reluctant to invest anddevelop the process technologies required tomanufacture them. This has been a majorfactor hindering the advances in the alloydevelopment and process technology of SMAs[2-4].

In this study a composite matrix materialmade of shape memory alloy based on copperis analyzed in martensitic state bymicrostructure and chemical compositionpoints of view. Number of martensitic variantsand grains dimensions influence the materialsdamping capacity through them boundariesdimensions and number so is important toanalyze them characteristics [5-7].

2. Experimental details

To obtain the alloy was used a laboratoryfurnace with graphite crucible using copper,zinc and aluminum high purity materials withreduce percentage of iron [8]. The heattreatments were realized on a laboratory

Vulcan furnace with controlled temperaturevariation [9].

Chemical composition was determinedthrough spark spectrometry analysis usingFoundry Master equipment (for matrix andreinforcement elements chemical analysis) andEDAX analysis as well for matrix study. Inthis study different EDAX softwareapplications were used to determine thechemical variation of the elements on LinePoint mode with automatic or element listconsiderations. Microstructures of thecomposite matrix were obtained with ascanning electron microscope (SEM) LMH IIby Vega Tescan brand using a secondaryelectrons (SE) detector to characterize thematrix in martensitic state details.


Microstructural and chemical investigationswere carried out to analyze a shape memoryalloy based on copper used as matrix materialin a composite. SEM microstructure of thecomposite surface is presented in figure 1 at a500x image amplification. Composite materialis made of a shape memory alloy as matrix andan arch material (steel) reinforced in metallicmatrix. Analyses concerning some interfacephenomena were done [10] presenting a nicecombination of the materials and a diffusionmade layer between SMA and steel materials.

Figure 1: SEM microscopy of a composite material with a shape memory alloy matrix


109

Microstructure of a shape memory alloyused as matrix material in a metal-metalcomposite material is analyzed in figure 2.SEM microscopy presented in figure 2 wererealized with a SE (secondary electrons)detector at a 30 kV alimentation tension lampand different image amplifications like 500x in

a), 1000x in b) and 10000x in c). Using theTescan Software 3D analyzes of the martensitevariants was done based on the electronsanalyzed from de sample.

a) b)

c) d)

Figure 2: Microstructural analyze of a shape memory alloy based on copper used as matrix in a compositematerial in martensite state a) grains intersection b) detail of grains intersection from a) with martensite

variants type c) geometrical and dimensional characteristics of martensite variants d) 3D image of area selectedin c) for martensitic variants characterization

Martensite microstructure of the copperbased shape memory alloy propose for thematrix material has different types of variantslike plates or V shape with different kind ofdimensions between 2 or 3 µm to 50 to 100nm. Secondary martensite variants appear as

well in the microstructure, marked with redline in figure 2 c), having nanometricdimensions and being formed between bigmartensite variants. Different grains containdifferent orientation variants of martensite,presented in figure 2 c), parts of them


110

orientated at 90 degree. The 3 D representationof figure 2 c) present three different types ofvariants not only as shape and orientation butas high as well.

Chemical investigations of the matrix shapememory alloys were done on different areas of4 mm2 from the material surface and chemical

composition results (ten different areas) beingaveraged in table 1.

Spectral energies characteristics to shapememory alloy analyzed are presented in figure3 with two energy connections for copper andzinc and a single characteristic energy foraluminum.

Figure 3: Spectral energy distributions of elements from composite material matrix

In table 1 weight and atomic percentages ofthe component elements are presented having a72.84 % wt of copper, 20.42 % wt of zinc and6.73 % wt aluminum. Atomic percentages

present a 67.11 contribution of copper, 18.29% zinc and 14.597% aluminum that arecombined as CuZn, Cu5Al9 and Cu5Zn8formations [11].

Table 1 Chemical composition of matrix material consisting of a shape memory alloy

Element AN series Net [wt.%] [norm. wt.%] [norm. at.%] Error in %Copper 29 K-series 77223 64,25326 72,84393 67,11193 1,617906

Zinc 30 K-series 18388 18,01927 20,42844 18,2902 0,487459Aluminum 13 K-series 3272 5,934216 6,727622 14,59787 0,360829

Sum: 88,20674 100 100

In figure 4 distributions of chemicalelements copper, zinc and aluminum on a 100µm line are presented and can be consideredthat the material homogeneity id very goodexhibiting nice properties on entire materialmass. Following the diagram from figure 4 canbe observe a smoother behavior distribution ofcopper on the smaller martensite variants andbigger variations of all elements in the other

parts of the material. Having a smootherhomogeneity shape memory alloys can bebetter controlled in practical applications.Shape memory alloys exhibit a nice dampingcapacity due trough martensitic transformationthat can be used in dissipation applications.Reinforcement of a shape memory alloy toincrease them strength and helping the shape


111

memory recovery represent a goal of the researchers in this field.

Figure 4: Chemical elements distribution on a 100 µm line of a shape memory alloy based on copper

Energy dissipation in shape memory alloysfrom mechanical to thermal is made especially onboundaries matrix between grains and martensiticvariants. Modifying the grains number anddimensions and the variants shape, dimension andnumber can improve damping capacity of thematerial.

4. Conclusions

A shape memory alloy based on copper wasobtained through classical melting methods andanalyze by Microstructural and chemical points ofview.

The material is proposed for a dampingcomposite having a high dissipation capacity inmartensitic transformation domain of mechanicalenergy in thermal energy. Grain boundaries andmartensite variants support most of the energydissipation so is very important to control thematerial microstructure.

Showing different martensite variants andvarious dimensions the material present a goodpotential for damping composite metallicmaterials.

AcknowledgementThis paper was realised with the support of

POSDRU CUANTUMDOC “DOCTORALSTUDIES FOR EUROPEAN PERFORMANCESIN RESEARCH AND INOVATION” ID79407

project funded by the European Social Found andRomanian Government.

References

[1] Webster, J., Proceedings of the InternationalSociety of Air Breathing Engines Conference,Bangalore, India, September, 2001.

[2] Caicedo, M., Jacobs, J.J., Reddy, A., Hallab,N.J., J. Biomed. Mater. Res. A 86 (2007) 905.

[3] Shabalovskaya, S., Anderegg, J., VanHumbeeck, J., Acta Biomater. 4 (2008) 447.

[4] Sun, T., Wang, M., Surf. Coat. Technol. 205(2010) 92.

[5] Cimpoe u, N., Stanciu, S., Meyer, M., Ioni , I.,Cimpoe u Hanu, R., Effect of stress on dampingcapacity of a shape memory alloy CuZnAl, ,Journal of Optoelectronics and AdvancedMaterials, Vol. 12, Issue 2, pg. 386-391, IDS570DE, ISSD 1454-4164, 2010.

[6] Paun, V.-P., Cimpoesu, N., Cimpoesu Hanu, R.,Munceleanu, G. V., Forna, N., Agop M., On theEnergy Dissipation Capacity and the ShapeMemory. A Comparative Study betweenPolymer Composites and Alloys, MaterialePlastice, vol. 47, nr. 2 , pg. 158-163, 2010.

[7] Cimpoe u, N., Cimpoe u Hanu, R,, Vizureanu,P., Ioni , I., Agop, M., Experimental andtheoretical results concerning internal frictioninvestigation of a shape memory alloy based oncopper, Metalurgia Interna ional, vol. XV, nr.12, pg. 48-59, ISSN 1582 – 2214, 2010.


112

[8] Mares, M A, Carcea, I, Chelariu, R, Roman, C,Mechanical and tribological properties ofaluminium matrix composites after ageingtreatments, Euromat 97 - proceedings of the 5theuropean conference on advanced materials andprocesses and applications: materials,functionality & design, vol. 1 - metals andcomposites pg. 423-426, 1997.

[9] Cimpoe u, N., Enache, A., Stanciu, S.,Nejneru, C., Achi ei, D., Hopulele, I.,Composite shape memory materials obtainingmethods, Simpozionul international Artcast2008, Galati, pg. 293-296 ISBN 978-973-7845-94-8, 2008.

[10] Enache, A., Cimpoesu, I., Stanciu, S.,Hopulele, I., Axinte, M., Cimpoesu, N.,Interface analyze of a composite material withshape memory alloy matrix, Bul. Inst. Polit.Iasi, t. LVI (LX), f. 1, 2011.

[11] Morton, A.J., The -phase regions of the Cu-Zn, Ni-Zn and Pd-Zn binary systems, Actametall., 27, , 863-8,1979.


113

IRESEARCH ON THE CORROSION OF CONSTRUCTION MATERIALSFISSURE MACHINERY HYDROCARBONS IN CONTAMINATED

ENVIRONMENTS WITH STAFF FROM CORROSIVE BY CRUDE OILDISTILATION

Moro anu Marius 1, Ovidiu Georgescu2

[email protected].

Abstract: In the atmospheric distillation of crude oil facilities and in vacuum of the fuel oil werehighlighted in uniform and uneven general corrosion, local corrosion in the form of points, spots, cavesand cracking, selective, crevice, etc. .. All these forms of corrosion are dangerons, corrosion cracking,however, occupies a special place, because in a short time can switches off the machine or plant with hardconsequences sometimes to assess. If sulfur crude oil processing appeared to corrosion cracking ofaustenitic stainless steel scales of DA-C1 column at the top, after 1.5 years of operation. Also, heatexchangers at the top of the columns DA and DV, with monel or carbon steel pipe. Hydrocarbonenvironments in the presence of additional factors, are potentially active for the emergence anddevelopment of corrosion cracking. The paper presents some aspects of metal corrosion crack to metalequipments in the crude oil distillation plants and fuel oil vacuum

Keywords: stress corrosion, load test, time fractional (max. 5)

1. Aims and background

Corrosion cracking under stress is one of themost dangerous forms, given the destruction that itcauses to metal materials used in manufacturingindustrial machinery. This process is manifested byintergranular cracking and transgranular,accompanied by a reduction in strength propertiesand finally, the fracture mechanics of materials inthe cracks.

Environments containing hydrochloric acid,hydrogen sulfide, naphthenic acids, cyanides, etc..are specific environments that cause cracking andultimately fracture of steels subjected to corrosioncracking under stress in the normal operation of theplant from atmospheric distillation of crude oil andfuel oil vacuum. Sulfur compounds, chlorine,nitrogen and oxygen are converted to hydrogensulphide, hydrochloric acid, ammonia andrespectively water [1-3].

The compounds of conversion in the absence ofcorrosion inhibitors have a corrosive action onmetal equipment from carbon steel, stainless steelor alloys CuZn, Cu Ni, itself and their interactionproducts (ammonium chloride, ammonium sulfide,ferric chloride, etc.. ). These building materials inspecific environment conditions, are susceptible tocorrosion cracking - [4-6].

2. ExperimentalThe paper presents experimental data obtained from

tests performed by the conventional method ofdetermining the time of failure and the propagation ofcracks in the corrosion of metal materials.

Materials used are carbon steel S235JR , stainlessferrites steel 7AlCr130 and austenitic stainless steelsX6CrNiTi 18-10 and 17-12-2 X6CrNiMoTi of highmechanical strength, whose composition is shown inTable 2.1.

Tabel 2.1.- Chemical composition of studied steel

Also, test samples were stabilized brass (70%Cu, 29% Zn, 1% Sn) and Monel W2.4230 (65.9%Ni, 32.5% Cu, 0.03% Al, 1, 57% Fe).

Steel Components , %C M

nSi

Cr Ni Ti Mo

P S

S235JR 0,18 0,53

0,30

- - - - 0,04

0,04

X6CrAl13

0,08 0,60

0,60

13,2

- 0,1 - 0,03

0,015

X6CrNiTi

18-10

0,06 1,8

0,7

18,5

9,5 0,51 - 0,03

0,014

X6CrNiMoTi

17-12-2

0,07 1,7

0,8

18,3

11,2

0,52 2,4 0,04

0,015

mailto:marius.morosanu2003:@gmail.com


114

The specimens used in corrosion cracking testsare of cylindrical shape with a diameter of 3 mm,threaded at both ends to be caught in fast-tensioning device to ECO Before eachdetermination, the metal specimens were machinedand polished with metallographic paper to 600 grit,then degreased in acetone.

Tensioning specimens was performed under aconstant load voltage to a unidirectional, its valueto 50%, 70% and 85% of breaking strength of thematerial tested.

Tests were conducted at temperature of 200C incondensate aqueous collected from the top of theatmospheric distillation column column AD-C1-C5VD vacuum distillation, samples weher pH wasmaintained in the 6.0 to 6.5 and the chloridecontent was 30-60 ppm.

The samples tested were mounted in thecontainer unit composition aggressiveenvironment. Lever was loaded with weights untilthe tensioning force of the specimen (Figure 2. 1).Test time is measured by timer with digitalreadout, and the cracking sound samples isreported. The timer stops to the break, if thesample and indicate the number of hours during thesample remains under load.

1– cell solution;

2 – heating jacket;3 – thermometer4 – corrosion

specimen;5 – refrigerant;6 – hydraulic

closure;7 – sealing;8 – lever;9 – greut i;10 – dispozitiv de semnalizare a ruperii.Figure.2.1: Divice for determining a corrosion

crack with unidirectional voltage

Also, the tendency to corrosion cracking ofsteels investigated was determined according to SRSR ISO 7539-2:1994/A99:2002. The methodinvolves immersing for a specified duration ofepruvette, previously deformed in cold and tenseelastic device shown in Figure 2.2

Figure. 2.2 : Sample tensioning device

Test environments at the top of the columnswere maintained at 80 temperature and thecontinuous refluxing and extensive examination ofthe specimen surface was made after its removalfrom solution. The samples were processed only tothe interior, which comes into contact with themandrel. The specimens were mounted in thedevice so that the processed surface to be placed inthe compression and tension by tightening thescrews Co up to a distance of 2 mm between thespecimen cap and base plate of the device.Specimens caught in the device tensions weredegreased with acetone and were placed in acorrosive environment, which should completelycover the specimens, the duration of the test.

The specimens were examined visually forcracks control. The examination was conducted bytaking samples of the solution, washing with hotdistilled water and observing with the naked eyeand with a 10x zoom lens.

3 Results and discussion

Composition of condensate aqueouscollected from the top of the atmospheric crudedistillation column AD-C1 (Figure 3.1) andvacuum distillation of fuel oil VD-C5 (Figure 3.2)are presented in Table 3.1, and stress corrosioncracking test results are presented in table 3.2.

Figure 3.1: The technological scheme foratmospheric distillation plant (AD-C1)


115

Fig. 3.2: The technological scheme for viddistillation plant (VD-C5)

Table 3.1 Quality of aqueous condensate testNo Qualitative

CharacteristicsUM Working place

Columnrefluxtank AD-V1

ColumrefluxrecipientDv-V100

1 pH upH 6,4 6,7

2 Chlorides ppm 42 31

3 Sulfur ppm 52 45

4 Iron ppm 0,1 0,2

Table 3.2: Values for fracture time of the testedmaterials at temperatures of 200C and 800C in thealquons condensate from the top of the column AD-C1,pH=6,0

It appears that after 1000 hours of testing in theaqueous condensate from the top of the columnAD-C1 did not show cracks on the surface ofspecimens under unidirectional tension of 50% ofthe breaking strength of the material. At test loadsof 85% from the tensile strength, fracture timedecreased. This decrease becomes morepronounced by increasing the operatingtemperature of 200C to 800C.Also tested were the same types of materials into

the tension that in Figure 1. Lack of large crackson the (upper) of the specimen showed resistanceto corrosion cracking under stress of the testedmaterials (Table 3.3).

Specimens of carbon steel S235JR hadcorrosion rates of 0.118 mm / year at the top of thecolumn DA-C1 and 0.130 mm / year at the top ofthe column C5 DV-corrosion showed punctuate onthe whole surface. Austenitic stainless steels havemuch lower corrosion rates and their surface is

uniform general corrosion. Furthermore, brassCuZn 28 and Monel W24230 SN1 have corrosionrates of 0,012 mm / year respectively 0.0103 mm /year.

No

Mat

erial

Brea

king

resu

lts R

mN

/mm

2%

from

Rm

N/m

m2

Task

test

N/m

m2

Tem

pera

ture

s0C Fo

rFa

ctur

e tim

e

Obs

erva

tion

1

S235

JR

420

50 210

20 >1000 Nu s-arupt

80 >1000 Nu s-arupt

70 294

20 >1000 Nu s-arupt

80 750 S-a rupt

85 357

20 390 S-a rupt80 105 S-a rupt

2X

6CrA

l13

560 70 392

20 >1000 Nu s-arupt

80 >1000 Nu s-arupt

3

X6C

rNiT

i 18-

10

550

50 275

20 >1000 Nu s-arupt

80 >1000 Nu s-arupt

70 385

20 >1000 Nu s-arupt

80 >1000 Nu s-arupt

85 467.5

20 >1000 Nu s-arupt

80 >1000 Nu s-arupt

4

X6C

rNiM

oTi1

7-12

-2

540

50 275

20 >1000 Nu s-arupt

80 >1000 Nu s-arupt

70 385

20 >1000 Nu s-arupt

80 >1000 Nu s-arupt

85 467.5

20 >1000 Nu s-arupt

80 >1000 Nu s-arupt

5 Bras

s

350 70 245

20 >1000 Nu s-arupt

80 >1000 Nu s-arupt

6

Mon

el

580 70 406

20 >1000 Nu s-arupt

80 >1000 Nu s-arupt


116

Tabel 3.3: Test results of the tested materials corrosioncracking at 80 temperature for 96 hours

On visual inspection of samples there were notfound am cracks. The values of corrosion ratesranged in the materials stability group metals. Thedata presented show that the methods used thestudied materials fracturing time is high.

4 Conclusion

1. Corrosion cracking tests were performedwith samples under a unidirectional tensin at aconstant load and with samples deformed at coldand elastic tensied.

2. After 1000 hours of testing in the aqueouscondensate from the top of the column DA-C1 didnot show cracks on the surface of specimens underunidirectional tension of 50% and 70% of breakingstrength and break more quickly specimens at atemperature of 800C.

3. By the cold deformation and elastic tension,the samples exposed in the media at the top of thecolumn DA-C1-C5 DV, at boiling temperature,were not affected by cracks.

4. The presence of chloride with hydrogensulfide accelerates corrosion phenomena,compared with the corrosions processed only byhydrogen sulfide. Hydrochloric acid aqueoussolutions under high unit action efforts, cause

corrosion cracking of austenitic stainless steels.5. Low concentrations of chlorides, ppm order.

produce austenitic cracks when water with salts isconcentrated by boiling. The tendency to crackingincreases with the temperature increasing at aconstant voltage and with increasing salt content.

6. Variable-valence metal chlorides Fe2 + / Fe3+, Cu2 + / Cu + give rise to pitting corrosion,which may constitue germs of corrosion cracking.

References

[1] Shalaby H.M., Refining of Kuwait’ s heavycrude oil: materials challenages, Workshop oncorrosion and protection of metals, Arab schoolfor science and tehnology, December 3-7Kuwait

[2] Antonescu N.N., Moro anu M, Petrescu M.G.,Niculae A.D., Georgescu L., Georgescu Ov.,Issues on corrosion equipment in instalation bydistribution DAV naphthenic acids from crude oil,Poster WP15- corrosion in Refinery Industry,EUROCORR 2010, The European CorrosionCongres, European Federation of Corrosion 13-17 September 2010.

[3] Burlov V.V, Altsybeeva A.I., Kuzinova T.M.,Sokolov V.L, Corrosion problems of modernpetroleum refinery, Poster WP15- Corrosion inrefinery industry, EUROCORR 2010, TheEuropean Corrosion Congres, EuropeanFederation of Corrosion, 13-17 september 2010.

[4] Ropital F. Mecanismes et mecanique desinteractions plastic envirovmements. Casd’industrie petroliere, Plast Ox 2007 (2009),p265-275., EDP Sciences 2009

[5] Lobley G.R., Stress corrosion cracking: caseSaudies in refinery equipment, The 6 th sandiEngineering Conference, KFUPM, Dhahran,December 2002, vol.5, 2002, p17-26.

[6] Feaugas X., Creus. i., Sabot R., El Alami H.,Lange D., Sahal M., Savall C., Huvier C.,Influence d’un etat mecanique sur la reactivitede surface en milieux aqueux des metaux c.f.c,Plasto x 2007 (2009), p221-236, EDP Sciences2009.

No WorkEnvironment

NatureMaterials

Timeto

breakafter140

hours

The corrosionrate

Kg,g/m2

h

P,mm/an

1 Condensation from thetop ofdistillationcolumn AD-C1

S235JR Nocracks

0,107 0,118

X6CrAl13 Nocracks

0,022 0,024

X6CrNiTi18-10

Nocracks

0,014 0,016

X6CrNiMoTi17-12-2

Nocracks

0,0044 0,0049

Brass Nocracks

0,011 0,012

Monel Nocracks

0,0093 0,0203

2 Condensation from thetop ofdistillationcolumn VD-C5

S235JR Nocracks

0,12 0,13

X6CrAl13 Nocracks

0,024 0,027

X6CrNiTi18-10

Nocracks

0,019 0,021

X6CrNiMoTi17-12-2

Nocracks

0,0052 0,0058


117

SOURCES OF POLLUTION IN SIDERURGY AND TECHNIQUES FORREDUCING NOXIOUS EMISSIONS

Elisabeta Vasilescu , Ana Doniga , Marian Neacsu

1“Dun rea de Jos” University of Galati , [email protected],3 Dun rea de Jos” University of Galati , [email protected].

Abstract: The systematic identification of the various sources of pollution creates the necessarybackground for organizing actions in order to prevent and fight against pollution. The diversity of thesources and forms of pollution are characteristic of industrial activities, especially of those in thechemical industry, metallurgical engineering, transportation industry, generally, industrial sectors dealingwith raw material manufacturing.

In the metallurgical industry, the air pollution takes place especially on the basis of the followingcategories of polluting agents: solid particles resulted from powders, particles drawn by combustiongases, particles of coal, ore, coke and fusing agents, slag, iron oxides, sulphur oxides (SO2 and SO3),nitrogen oxides (NO2 and NO3) and, although in smaller quantities, by tars, hydrocarbons, soot, organicacids and others.

Being familiar with the sources of pollution, their monitoring and control represent a first steptowards reducing the quantity and the toxicity of all emissions, focusing on applying a “cleaner”production in the industry of elaborating and processing metallic materials, too.

This paper succinctly presents the main sources of pollution in the primary siderurgical sectors, aswell as a series of techniques to reduce pollution, especially in the process of steel elaboration.

Keywords: pollution, noxious emissions, siderurgy

1. Introduction

During the period between1950 and 1960, steelindustry represented a major source of pollution,especially air pollution. The problem of reducingair pollution was raised at once, which made thespecialists focus their attention, initially, onmodernizing the installations for gases and dustcollection, but this approach was replaced by theintroduction of new installations designedaccording to the concept of “cleaner” production.A “cleaner” technology is based on less pollutingproduction procedures or procedures which userecycling technologies, reintroducing in themanufacturing process the waste material resultedfrom a certain technological phase, or procedureswhich revaluate waste material turning it into rawmaterial for a secondary production. In a lesspolluting procedure, investments becomeproductive and the expenses implied bydecontamination are an integral part of themanufacturing process.

From the economical point of view, “clean”technologies, by fighting against wasting, allow

achieving important energy and raw materialsavings, which lead to cutting down the terms forrecovering the investment expenses.

2. Sources of pollution and techniques forreducing pollution in the primary siderurgicalsectors.

In coking plants, agglomeration andelaboration sectors (cast iron-steel), air pollution isthe most important environmental problem anddiminishing or completely eradicating noxiousemissions in the atmosphere is connected with thepreservation of energy and resources available(fig.1).

Thus, residual gases represent a valuablesource of fuel, at the same time being an importantlink in the energetic chain of a plant.

In the case of coking plants, the noxiousemissions in the air consist mainly in powdersresulted from coal supplying, coal processing and

mailto:elisabeta.vasilescu:@yahoo.com

mailto:uscaeni:@yahoo.com


118

coke sorting. Burnt gases from coking batteries,coke quenching and cooling are also released in theair. Generally, before being used, the resultedgases are carried through different cleaningsystems, the residual water resulted containingsolid particles, heavy metals (in the case of furnacegas and converter gas cleaning) or organicpolluting agents (cyanides, ammonia and others) in

the case of coking gas cleaning. Besides, powdersand silt containing iron can also be collected. Thus,by reducing the emissions of secondary gases inatmosphere, an ironworks can benefit from theexistence of a collected power source, reducing theother expenditures, and as a result of recyclingmaterials containing iron, the consumption of theavailable resources is diminished.

Figure 1 Relative emissions in the air originating in the primary siderurgical sectors

Temporary gas emissions appear during theoperations of liquid pig iron transport, preliminarytreatment and operations of loading, charging andevacuation of the LD converter. The secondarystirring installation and the continuous casting oneare less important sources of noxious emissions inthe air.

An inert gas can be used to reduce the ironoxides emissions during the operations of liquidpig iron transport and of loading the converter, andthe emissions generated during loading can bereduced by optimizing the loading technologies aswell as the characteristics of scrap iron (quality,distribution in furnace).

In the case of steel works, there are emissionsresulted from secondary gas burning, temporaryemissions from the steel works, flashing emissionsand the cleaning of the secondary extraction gas.

The secondary emissions during charging canbe controlled by providing a tight adjustingbetween the converter opening and the seal boot of

the primary residual gas system, but this can oftenbe prevented by the drops of slag and metalproduced during the charging operation.

The secondary gas from the LD converter canbe collected and recycled as fuel, although a part ofit has to be burnt every time at the beginning andin the end of each charge when the COconcentration is low. Thus, the efficiency of thegas cleaning installation as well as of the gascollecting will have an important effect on the totalquantity of the emissions in the air. Technologiesapply computerised control systems in order toincrease the efficiency of gas collecting.

The converter gas used as fuel has a caloricpower of about 9 MJ/m3, thus saving 0,7 GJ/t ofraw steel. In Germany, for instance, the recyclingof the converter gas saves the equivalent of 300Nm3 of natural gas.


119

The system works by carrying the hot gasthrough a recuperator where the heat is used togenerate steam (fig.2). Then, the gas goes througha purifying system or through a wet cleaningsystem, and, when a certain value of the COconcentration is reached, the gas is collected,

cooled and stored. The electric filter purifyingsystem has lower power consumption as comparedto the corresponding water systems and it isinstalled in the place where the gas is collectedfrom the converter.

Figure 2 The purifying and recycling system for LD converter gases

In electric steel works, a primary extractionsystem can be used in order to directly collect theemissions generated in the furnace during themelting process and a secondary system can beused to collect the temporary emissions in the steelworks during loading, melting and evacuation(fig.3). The primary gas is mainly filtered in orderto remove the particles before their being releasedin the atmosphere, but this system may not beequally efficient in lowering the level of the

volatile compounds, of the smoke resulted and ofthe traces of organic compounds, of volatile metals(e.g. mercury) etc. The level of these compoundscan be controlled to a certain extent by sorting andpreparing the scrap iron. Another problem in thecase of electric furnaces is represented by thedioxin – type toxic gas emissions, for which aseries of researches have been made in order todiminish or completely eliminate them.

Figure 3 The situation of the powders collected from the primary and secondary gases evacuated from theelectric arc furnaces (data processed from 67 installations)


120

3. Modern technologies used to reduce thedioxins emissions in the electric arc furnaces

Among the polluting substances emittedduring the metallurgical processes, dioxins are themost toxic.

The problem of dioxins emissions is moreacutely raised in the case of electric arc elaborationfurnaces, where important quantities of scrap iron(along with the organic and chloride constituentswhich they inevitably contain) are melted in orderto obtain steel.

Dioxins define a group of compounds whichconsists of about 75 polychlorodibenzodioxins(PCDD) and 135 polychlorodibenzofurans(PCDF).

At high temperatures, dioxins and furansdecompose.

In the gas cooling phase, they are remade bythe chemical reactions between the organiccompounds and the chlorides present in the gas andpowder emissions, at temperatures of 200 – 600°C.This “novo” synthesis reaches the maximumintensity between 250 and 400°C. In fig.4, it is

schematically presented the variation of thereaction speed in the process of dioxins formingand decomposing, function of temperature raise.

Dioxins remaking depends on the gaseskinetics, especially during the above mentionedtemperature interval. Thus, in order to reduce theseemissions, it is necessary a rapid cooling of thegases resulted in the process, at temperaturesbellow 200°C to prevent dioxins remaking.

The strategy for reducing the concentrationof dioxins in the electric furnaces emissions isfocused on three aspects:- thermic decomposing as advanced as possible ofthe dioxins contained in emissions by an extensiveexposure of these compounds at temperatures over850°C;- prevention of their remaking (by “novosynthesis”) during the gas cooling phase byaccelerating the process within the temperaturedomain 600 - 200°C:- separating the residual dioxins from the emissionflow, by injecting an absorbent substance, followedby a filtering operation.

Figure 4 Dioxins forming function of temperature

To reduce the dioxin emissions in atmosphere,a modular system for treating emissions wasdeveloped, which takes into account the specificconditions of exploitation and the type of electricfurnace (classical or tub furnaces).

Each part of the system (equipment)corresponding to a technological treatment stagewas optimised in industrial conditions (fig.5).


121

Figure 5 The efficiency of absorption depending on the filter temperature and the injecting of the absorbentsubstance

In the case of classical electric furnaces, aseparator was installed, which captures the dustparticles and the drops of liquid slag. The hot blast(which includes gases and solid particles) canreach 1000 °C and contains an important quantityof gaseous CO which burns in contact with theinjected air.

Simultaneously, the volatile organiccompounds contained in the gaseous mixture,together with the dioxins resulted from the organicsubstances present in the scrap iron and therecycled materials in the charge, are decomposed.

In the case of the “shaft” furnaces, a quantityof air and liquid fuel are injected and burnt in acombustion chamber located at the outlet port ofthe charge preheating tub in order to achieve atemperature high enough (over 850°C) to assurethe decomposing of the organic materials and ofthe dioxins contained in the mixture of gases andpowders evacuated.

The dense particles of powders and the slagdrops are evacuated quite efficiently through thefilter made by the charge column situated in thepreheating room.

Gas coolingAfter combustion, hot gases are cooled to

approximately 600°C in a pipe cooled by water,and then they are passed either in a mist cooler orin a gas-gas cooler with forced draught to reach atemperature of approximately 200°C. Theadvantages of mist cooling are: the high coolingspeed and, due to a small pressure lowering, theeasier circulation through the system.

The forced draught cooling has the advantagethat it can function without water (the one whichmakes the mist), which simplifies the exploitation.

Dioxins absorptionThe residual dioxins contained in the gaseous

mixture are eliminated by condensation and / orabsorption followed by a filtering stage. Theefficiency of the process depends on ahomogeneous dispersion of the particles having acertain size and shape (with large specific surface)in the emissions mixture to obtain a maximumdegree of compression and absorption of dioxins.For the purpose of assuring homogeneity, activecarbon (coke) or active carbon mixed with limewere used. The quantity and the quality of theabsorption mixture are important factors whichassure the security of the installations within theaggregate, with a view to reduce the danger oftaking fire and explosion inside the filteringchamber.

The injection system of the absorbentsubstance is made up of a bin (storage chamber)for the absorbent substance, a (“spider”)distribution aggregate and the proper infusionsystem.

The particles enriched with dioxin areseparated from the emission flow in a filteringchamber containing sack filters and uncloggingdevices with counter-current air. Powders arecollected in a bin, and then they are agglomeratedto be recycled or evacuated in the end.

Industrial uses1. System for a tub furnace.


122

The system of dioxin evacuation has acombustion chamber equipped with CH4 burners.After leaving this chamber, the exhaust gas andpowder mixture goes through a mist cooler. Anabsorbent infusion device is installed next to thechamber where primary and secondary gases aremixed. The purifying of the dioxin loaded gases isachieved in an air filter.

2. System for classical electric furnaces.The burnt gases go into a separator where thick

particles are eliminated either by the emission fluxor by a mist cooler or a forced draught air cooler.Gases go then through a fine filter where thesmallest particles are eliminated.

To sum up, we can make the followingobservations:- It was noticed that the efficiency of the systemsfor the removing of dioxins in the electric ovensemissions has been substantially improved byintroducing absorbent substances infusion systems;- The dioxins tendency of concentrating, byabsorption and condensation, on the particles ofvery fine powders (which facilitates theirelimination in sack filters) becomes moreaccentuated as the temperature of the gases at thefilter entrance goes lower. This temperature shouldbe under 80°C to assure a removal of the dioxinsas advanced as possible.

In the case of the system treating the emissionsfrom an electric furnace, the efficiency of purifyingdepends on the installation building as well as onthe exploitation technology of the equipment.

ConclusionsThe dioxins concentration can be reduced at

much lower levels than those imposed by theenvironmental legislation in force, due to thecombination between the thermic treatment of theprimary emissions (combustion gases), with rapidcooling and gas filtering.

The solutions proposed and experimentedcorrespond to the most advanced techniquesconcerning environmental protection theintroduction of some “clean” technologies inmetallurgical engineering.

References[1] Poluanti organici persistenti în mediul înconjur tor.

Brosura informativa a Asociatiei expertilor demediu – 2003

[2] Jiroveanu, M., Oprescu St. – Captarea si epurareagazelor în industria metalurgic prelucratoare. –EDP, Bucuresti.

[3] Pumnea C.,Ionit N.- Tehnologii în industriaprelucr toare. EDP – Bucuresti

[4] Gheorghe N.,Gheorghe D. - Poluarea în industriametalurgic si chimic – Editura Performantica,Iasi 1997.

[5] Voicu V. – Combaterea noxelor în industrie ,Editura Tehnic , Bucuresti 2002

[6] * * * Technical and Management Issues, pag.65[7] Lehner J., Friedacher A., Gould L., Fingerhut W. –

Des solutions economiques pour l’elimination desdioxines dans les fumees des fours electriques a arc– La Revue de Metallurgie nr.1-2004


123

THE TEORETICAL CONSIDERATIONS ABOUT THE PRECISIONINCREASE OF CONICAL DRAWING PIECES

Lucian V. Severin1, Traian Lucian Severin2

1University of Suceava, Romania, [email protected] of Suceava, [email protected]

Abstract: The paper presents same theoretical considerations about the precision increase of conicalpieces drawing, among other things the angle calculus work zone of tools. This is influenced byinitial dimensions and the initial thickness of the work piece and the size of deformations.

Keywords: conical pieces, precision, drawing, deformations, angle.

1. Introduction.The process of drawing conical pieces with

rigid tools is developing under more difficultconditions compared with the cylindricalparts. In this mode the deformation processingis transmitted from the punch to semi-productthrough a smaller area with the same degreeof deformation and therefore drawing stresshave greater values.

Also, in comparison with the cylindricaldrawing, the interstice between deformation tools,at the beginning process has great relativelyvalues, what determined taking technological andconstructional solutions in technologicalequipment conceiving, corresponding the relativesheet metal thickness [1].

Comparative analysis of mechanicaldeformation material scheme shows that these aredifferent in walls pieces (fig.1). Thus, in conicalpieces case the stress state is plane and thedeformation state is spatial.

The low and middle conical pieces have a littledeformation degree of the material and thesedetermined elastically arching when the pieces areremoved from the die, particularly at reducedrelative thickness of sheet metal. This difficulty iseliminated if use at drawing processes the mouldswith special construction (moulds with strongpressing of semi-product by blank-holder mouldswith redrawing rib and so on) [1, 2].

Low and middle conical pieces drawingprocess, with a high precision realized in a singleoperation, in many situation is needed to berealized the drawing pieces with the dimensionalcalibration. For that it must angle work zonetool’s adequate determination. If these are realizedat adequate value from constructive referencematerial, is not adequate realized the dimensionalcalibration, or it must great deformation forces byvolumetrically calibration of conical piece wall(fig.3).

1

3

23

2

1

1

3

1

3

3

1

2

RM

Rp

Figure 1 Mechanical scheme ofcylindrical pieces. drawing.

mailto:severin:@fim.usv.ro

mailto:severin.traian:@fim.usv.ro


124

It is not produced a good calibration of piece,because so that is showed by the deformation stateof semi-product in upper conical piece wall zone,the thickness increase, from initial semi-productthickness s0, to thickness s, at upper conical piecewall.

The question is how to determine theangles work areas of active elements becauseat the end operation processing calibrationtrack its makeover on entire surface

2. Upper piece wall thickness calculusIt is considered a semi-product with the

thickness s0 and radius R0 (fig. 4). On this semi-product is considered a point M placed on a circlewith radius . After conical drawing of semi-product, this point take position M1 on piece wall,at distance r to piece axis, where wall piecethickness is s1

If it is considered a reference system incylindrical coordinate ( , z) orientated on themain directions of stress and deformations can bewriting:

30

12 s

sln;rln z (1)

where i z are the real main deformations inpoint M1.

Writing Hooke relations between stress anddeformations for a plain stress state in this point,and the constance volume relations in plasticdeformations, result:

0zz

z ; (2)

If it is note ,a can be writing:

aa

z 21 (3)

Using relations (1), relation (3) become:rln

aa

ssln

21

0

1

With this relation can be calculated thethickness s1 by relation:

aa

rss21

01(4)

In upper conical piece wall the thickness s isdetermined with relation (4) where is substitutedwith R0, r with R and a with zero value, becausethe stress has zero value in point A. Therefore itresulted:

21

00 R

Rss (5)

If note the report R/R0 with m, report whatdefines piece drawing coefficient at the upperwall, relation (5) can be write with relation:

mss 0 (6)

Suitable relations (4) along conical wall exista position of the point M1 where z=0.This position is obtained if a=-1 .Thus can bewriting [2]:

RMRp

3

2

1

2

3

2

1

1

2

2

13

3

2

1

3

Figure 2 Mechanical scheme of conicalpieces drawing

Figure 3. Drawing scheme with piececalibration

RMRp s 0

s


125

11 0

0

Rln

Rlna

c

c

and by solving is

obtained:

00 6070 R,e

R(7)

This is the point M position on semi-productwhat after drawing arrives on piece in the pointM1, without the change of material thickness.

3. Establish the angles of active zone tolls.Because of the deformation state described

by the relation (4), the thickness conical wall ofdrawing piece is variable.

The greatest thickness value is calculated byrelation (6) and this is in point A1, because thematerial deformation degree is greatest in thispoint.

The drawing process with calibration piecedimensions must to be making on the mould withcorrected active zone tools with angle (fig.3).

If piece angle is quoted on middlethickness, then can be writing:

p and M (8)If piece angle is quoted at piece inside, the

work zone tools will have angles:

p and 2M (9)and it is quoted at piece outside

2p and M (10)

In relations (8), 9) i (10) it is noted:piece angle- ; punch angle- p; die angle- M; anglebetween inside generatrix of conical surface andthe outside generatrix surface - .

Angle size can be established with theposition of point M defined by relation (7),tacking account the semi-product thickness is notmodified in piece processed. Wall thickness inpoint A1 is determined by relation (5). If it is notedgeneratrix length between point A1 and M1 with L,can be writing:

Lssarctg

20 (11)

where is the angle between conical surface ofinside generatrix or conical surface of outsidegeneratrix and middle conical surface generatrix.

The size L can be determinate with relationdetermined by material volume equality includedin plane circular crown with 0,393R0 width withthe material volume included in conical piece wallwith middle generatrix by length L.

This volume can be calculated with thesecond Pappus-Guldin theorem, determined withrelation [3]:

cp xLssV 22

0 (12)

where, figure 3 corespondly, xc has the value:

D=2Rs

L

s 0

0,607R0 D0=2R0

Rp

D=2R

L

yc

xc

s

s0

Figure 4. Wall thickness change of drawing conicalpiece

A

A1

M

M1

A1

C

R

R0

r

z

A

A1

M

M1


126

sinyRx cc (13)where :

0

023 ss

ssLyc (14)

By yc substitution in relation (13) with thevalue determined by relation it is obtained:

sinssssLRxc0

023

(15)

Material volume of semi-product what isprocessed in piece wall with length L is obtainedby relation:

2006320 Rs,Vs (16)

By equalization the relation (12) with (17)and taking account by relation (6) and (15)results:

sinm

sinmm,mmmmmRL

21

2842012

322

0

(17)Substituting the length L determined with

expression (17) in relation (11)and taking accountby relation (6) it is obtained:

sin2mm0,842m1mmmm

sinm21m1m3R

s

arctg22

0

0

(18)

With the angle size obtained with relation(18) substituting in relations (8), (9) or (10) and itare determined the work zone tools of the conicalpiece with calibration drawing.

4. ConclusionsThe thickening angle of the conical piece

drawing wall depend on initial semi-productdimensions used, the initial thickness anddrawing coefficient in piece processingdetermined on greatest diameter of this.

The angle size increases with growth ofinitial semi-product thickness, with increasingthe degree deformation and decreases withincreasing semi-product deformed radius atthe same degree of deformation and the sameinitial thickness.

4. References[1] Romanovski, V., P., tan area i matri area larece. Editua tehnic , Bucure ti, 1970;[2] Teodorescu, M., Al., Tehnologia pres rii larece. Editura didactic i pedagogic , Bucure ti,1980;[3] Bachman, K., H., Mic enciclopediematematic . Editura tehnic , Bucure ti, 1980


127

METHOD FOR THE CONTROL OF THE GEOMETRIC ANDCONSTRUCTIVE ELEMENTS OF THE SPIRAL-FACE AND END

MILLING CUTTER

Str jescu Eugen1, Pavlov Olimpia2, Dogariu Constantin3

1,3Politehnica" University of Bucharest-Romania,[email protected],3 S.C. Munplast S.A. [email protected].

Abstract: In the paper are presented the bases of a methodology for the determination of all thegeometric and constructive elements of the cutting tools for the milling machines, starting from thegetting of a solid model 3D of the measured tool. There are shown the possibilities to obtain a 3D modeland the graphic model of the geometric parameters control.

Keywords: Control, Cutting Tools, 3D Model, Geometry.

1. General Introduction

Unprecedented development of materialmanufacture and technical orientation towardhigher accuracy of products required developmentof the cutting tools' construction, made by usingnew materials to increase the precision of theirexecution. Increase bad-dimensional and geometricprecision of the tools required technologicalprocesses and new machinery and apparatus at thehighest level.

The necessity to exactly endeavor thedimensions and cutting tools' geometry, especiallycomplex, assembled or curved cutting edges andordered development of techniques and differenttechnologies, which would shorten the time neededto control and would permit to obtain endeavorparameters at all cutting edges, to verify if all thegeometric parameters are real and includedbetween acceptable limits, more or less close to theoptimum.

To control the cutting tools' geometry there areimagined, developed and built a large variety ofdevices and control instruments, from simpletemplates, rulers, sublime and reporting workshopmicroscopes to larger or smaller cars and thecoordinated measures [Minciu et al. 1995].

Currently, the explosive expansion ofcomputer's use and devices permitted a newapproach of the problems of the cutting tool'scontrol, based on other principles, that we did notfind in the autochthon literature or in the strangeliterature [Strajescu 1989].

The original idea developed further make useone of innumerable programs (software) made byfirms specialized in CAD, CAM (3D model-horizon), in order to determine all the geometricand constructive elements of cutting tool using 3Dsolid model of the studied cutting tool obtained inadvance.

2. Obtaining 3D Solid Model Of CuttingTools

The methodology that we propose is based onthe getting of a 3D solid model of the controlledcutting tool [Enache et al. 1988]. As is easilyintuited, if the project of the tool is made with acomputer using software CATIA, Solid Works,Solid CONCEPT, the 3D model exists, becausethat model is generated by the software. It isknown that at the design of the cutting tool, thereare some preconditions which allow theconstruction of integral tool. Using 3D modelmade on this route would allow the designervisualize all geometrical elements at all edges ofthe cutting edge which classical designmethodologies not allow. Other two possibilities toobtain 3D model are taken into considerationtoward the authors of this article. This 3D model isobtained by scanning the cutting tool (modernscanners will produce geometric precision and highdimensional, maximum deviations hundredth of amillimeter is now toward normal), and by

mailto:S.A.ostefu:@yahoo.com


128

obtaining the 3D model using photography, photos,and treatment with specialized software.

2.1. Obtaining 3D model of the phase ofprojects hard.

The trend of placing in the cutting tools' currentdesign of the technical computing, imposed anddeveloped by the personal computers requires acomplete description of the geometric shape anddimensions of tools.

The assisted design of the cutting tools can beoptimized by programs.

For the cutting tools of any type, designsubsystems are built in the next order:

1. elaboration of the principles for thedevelopment projects;

2. parameters' description of the processedparts;

3. implementation of the model for calculatingthe geometric and constructive elements of thecutting tools;

4. elaboration of the block scheme ofcalculation;

5. developing computer programs.In order to realize an automatic design of the

cutting tool with efficiency, it would essentiallydetermine the correct informational structure of thecutting tool. In this respect, it should be clarifiedthe following connections:

a) spatial, which determine place andsequence of placement;

b) function, which determine the parameters'size (e.g. parameters of the constructive elementswhich determine strength and rigidity);

c) external, upon the interaction's characterand conditions with the manufactured piece.

The external connections determine the initialdata formation, the external factors being verynumerous. The multiple integration, multidirectional,programs from the computer-assisted sphere,specialized programs on the problems ofgenerating geometry (3D drawing), and simulationand calculation of static and dynamic behaviorpresents a lot of advantages.

We gate away from the pre-project until toreach the final product.

So, the successive stages are: specification ofthe project's theme, the preliminary drafts,realization of the pre-project, the model calculationand analysis with logical for cinematic anddynamic simulation, the results' getting, thedecision on the accuracy, the finite element

analysis, decision on the analysis' results, therealization of the prototype or experimentalproduct, experimental research, simulation andoptimization through testing and confirmation ofthe procedures.

There is currently a tendency to extend the useof the solid model geometry, which is ageneralization of the geometry in space.

A solid model provided a complete andconsistent definition of the piece geometry.Principles of the edition remain available in three-dimensional space, with the remark that changesaffect mass properties of the ensemble (center ofgravity, moment of inertia) that can be calculatedautomatically. The parametric design of the simple parts orassemblies is actually an effective concept andrepresents the way to suit for the purpose of thispaper. This concept design integrates design andanalysis steps.

Using generalized cutting tools on differenttypes of tools are included in the concept ofparametric design.

In this case are necessary complex calculationsand the project no longer analyses directly thepiece geometry. Analysis is made on a generalizedor idealized model to be complete in a timely andwith an acceptable coast. Switching to an idealgeometry ask for a fundamental change in theinformation model. Idealized model may bedifferent depending on the type of analysis (e.g.static or dynamic, structural or thermal, linear ornonlinear). A simulated event is equivalent to asingle running of an experiment. Therefore it mustperform a statistical analysis of a set ofexperimental results. The accuracy is higher andthe confidence in the results is greater when the setof runs is greater. It is possible to run multiple testsunder different conditions. The results can bestatistically processed.

Following such a process to obtain highaccuracy with a 3D solid model of cutting tool thatcan be cut with all the desired measuring planes,the programs used for this purpose may put up allsearched angles, without calculations or otheroperations.

2.2. Obtaining the 3D model by photography.An absolute new path in the field, usable to

obtain a 3D model can be visualized today by theexistence of some computer programs that allowreconstruction 3D model from photos made infixed positions that encircle the subject of the


129

wanted model. Initially such programs weredeveloped for the architecture, but appear variantswith net superior precision that can be applied tothe cutting tools. Software developments in thisarea will bring increased safety with precision anddimensional accuracy. We must say that theseprograms function today, but the necessities for theobtained images' processing are great timeconsuming.

The D Sculptor computer program permits tocreate computer photo-realistic models of a largedomain of objects, using common pictures ofobjects in a relatively easy mode and quite fast. Noneed for special hardware items, but only acomputer and a camera. It is recommended adigital camera, but it is possible to use scannedimages.

Figure 1: The main screen of D Sculptor software.

The basic processes for the composition of thebase model are the next:

- it is placed the object for which we need themodel on the calibration plane (fig. 1);

- the object is photographed from multipleangles;

- the pictures are imported in the D Sculptorsoft;

- D Sculptor automatically detects the model;- the outside lines of the object in each

photography are marked, using the tools formasking from the D Sculptor menu;

- D Sculptor computes the three-dimensionalmodel;

- the three-dimensional model is exported foruse with other software (as it is necessary shown).

While D Sculptor 2.0 brought improvementsboth at the technical level and in terms of interface.You can create models faster than before, and DSculptor 2.0 Professional last version has a highaccuracy.

2.2. Obtaining 3D model by scanningA convenient way to achieve easier a precise

3D model of a cutting tool is by scanning andreconstruction model. The method is really simple,but its application is limited by the high coast ofthe scanner device and, some time, by the necessityof an ulterior addition of the model. This lastaction can be complicate if the scanned object iscomplex and have many concave surfaces ormasked by other parts surfaces.

Figure 2: The solid model for a lathe cutting tool. Themain screen of D Sculptor software.

For the present paper the solid model wasobtained using a scanner 3D (Model 2020iDesktop 3D Scanner) that gave very good results.

In conclusion at the chapter 2, it is possible totell that the use of the 3D model obtained in theproject phase permits the integral determination ofthe geometric and constructive elements in everypoint of the active surfaces or cutting edges,avoiding for the designer ulterior surprises.

The getting of the model by the two other ways(photographing and scanning) permit the completeand precise determination of the geometric andconstructive elements of the studied cutting tools.

3. The Use of the Solid 3D Model for theControl of the Cutting Tools' Geometry

Starting from the solid model obtained in one ofthe ways described above (Fig. 2) and using adesign mean as CATIA or Solid Works, by simplecommands consecutive to the points assessment inwhich the geometry is measured it is possible todraw planes considered as main planes: theorthogonal plane PO, the tangent plane (actually theplain of the edge projection on Pr) PT,, the normalPn (less important for practice, but extremelyuseful in complex geometric calculations), Pffrontal plane and the posterior plane Pp.

4. Application of the Use of the Solid 3DModel for the Control of the spiral-face and endmilling cutter


130

This kind of tool work simultaneously twoperpendicular surfaces, having teeth on thecylindrical and frontal part.

The exterior diameter is established in functionof the technological needs and of the tool type,being content generally between 1 and 160 mm.

For the diameters between 1 and 75 mm thereare preferred the constructions with a cylindrical orconic tail, and for the bigger ones there are usedconstructions with circular holes.

Figure 3: Intersectii cu plane normale frontale.

Figure 4: Intersections with planes perpendicular onthe cylindrical helix

The used plans (perpendicular on the cuttingedges, named in ISO Standards orthogonal planesPO) for the frontal or cylindrical teeth are shown inthe fig 3 and 4. Using the program possibilities, thevalues of the cutting angles n and n are precisefor different points of the active edge.

0 0 0 0

-3.99 -3.99 -3.99 -3.99

-5

-4

-3

-2

-1

0

n

n

n 0 0 0 0

n -3.99 -3.99 - 3.99 -3.99

M1 M2 M3 M4

Figure 5: The graphic representation of thevariation of the angles’ value for 4 point.

Figure 6: Intersections with orthogonal planes.

69.38 67.9575.33 73.32

-4 -4 -4 -4-10

0

10

20

30

40

50

60

70

80

M1 M2 M3 M4

0

0

Figure 7: The representation of the section planesin which are determined the O and O angles and thegraphic representation of the variation of the twoangles’ value for 4 points.


131

4.1. Intersections with tangential planesAfter the intersection of the tangential plane

with the reference plane, it was obtained the angleKr, shown in fig. 8.

The method is especially useful for complextools, small, with the active surfaces and curvedcutting edges, to which access control with currentinstruments is very slowness or impossible andwhere the definition of theoretical planes controlangles is difficult to apply.

Figure 8: The tool sectioned with different planes inwhich is shown the angle r.

66.55 66.33 66.21 66.09

0 0 0 00

10

20

30

40

50

60

70

kk

k 66.55 66.33 66.21 66.09

k 0 0 0 0

M1 M2 M3 M4

Figure 9: The Graphic representation of the rangles’ variation..

10. Conclusions

A proposal for a complete control of cuttingtools made on the basis on the getting of a 3D solidmodel is new and brings the first advantage that itis possible to obtain all the angles searched at anypoint of the cutting edges or active planes. In thesame time, the solid model permits at any time toobtain again the lost values. Our research led to thestroke for the control of a milling tool very low,but getting a solid model consumed time andresources, especially to correct the model obtainedby scanning, correcting imperfections due to theirsoftware interpreting the results of the scan.

The other method of obtaining the 3D modelthrough photography has its limits, given by theprecision low shape recovery, but the rate ofincrease in performance software creates goodpremises for future use. An important observationconcerns the tools with edges and surfacescurvilinear, at which we can determine all thegeometric and constructive parameters, achievingwith the help of additional software included in thesoftware graphics of the parameter's values of alledge points. Further research will developmethodologies for control of complex tools forvery large or very small, with curved surfaces,methodologies able to change the angles of thelong edges or surfaces with steps as small, basedon programs developed independent and includedin the design software used for analyzing 3Dmodel.

11. References[1] Minciu C., Str jescu E., .a., (1995) Scule

chietoare, Îndrumar de proiectare. EdituraTehnic , Bucure ti, 1995.

[2] * * * Standardul ISO nr. 6599/2/1982chiere i scule a chietoare /2. ISR

Bucure ti.[3] Str jescu, E., (1984) Proiectarea sculelor

aschietoare, Litografia I.P.Bucure ti, l984.[4] Enache, t., Str jescu, E., Minciu, C., (1988),

Metode i programe pentru proiectareaasistat a sculelor a chietoare. LitografiaI.P.Bucure ti.

[5] * * * Standardul ISO nr. 6599/2/1982chiere i scule a chietoare /2. ISR

Bucure ti.[6] x x x Scule a chietoare i portscule pentru

chierea metalelor. (1988 - 1989), Colec ieSTAS, vol. 1 i 2, Editura Tehnic Bucure tti..

[7] Duca Z., (1967), Teoria sculelorchietoare. Editura Tehnic , Bucure ti.


132


133

ZINC COATINGS ON STEEL SUBSTRATE ATTAINED BY DIFFERENTELEMENTS ADDED

Radu Tamara, Vlad Maria

University “Dunarea de Jos” of Galati Romania

Abstract: The purpose of this work was to identify the influence of different process variables, suchas bath temperature, immersion time and bath alloy additions, on the morphology and coating thicknessattained using innovative alloys systems containing different percentages of Ni, Pb, Sn and Bi. Steelsamples were galvanized by the hot-dip method in micro alloyed zinc baths at different temps andtemperatures Experiments were aimed to obtain zinc coatings with increased resistance to corrosion,which adheres very well to steel support, not conducive Fe-Zn reactions, not have negative environmentalimpacts and presents uniformly dispersed phase structure in zinc matrix. Layers obtained in microalloyed zinc with nickel, bismuth, tin, lead are analyzed.

Keywords: galvanizing, Ni, Pb, Sn, Bi, thick layer, corrosion.

1. Introduction

Hot dip galvanised coatings are widely used inindustry for corrosion protection of steels. Thesteel is protected by the Zn coating through abarrier effect and a galvanic effect, in which Znacts as the sacrificial anode while steel acts as thecathode. Although Zn has high corrosion resistancein most non-aggressive atmospheric environments,corrosion problems of Zn coatings have been oftenfound in aggressive atmospheric environmentswhere salt and sulphur dioxide are present. Toovercome this problem, new types of Zn coatingswith higher corrosion resistance are needed.

To increase the resistance to corrosion,adherence, limiting pollution and zinc-ironreaction, its use more and more alloyed zinc melts.Alloyed zinc melts in galvanising is ollsoimportant in the case of Sandelin steel galvanising.It is characterized by the formation of thick layersand which can be hot - dip galvanised only inmicro alloyed zinc melts. There is a great interestin carrying out research which explores theinfluence of the alloying elements in zinc melts oncoating layers characteristics and thecharacteristics of the melts [1, 2, 3, and 4].

The paper aims to analyse the effect of smallquantities of nickel, tin, bismuth and lead on thezinc melts.

Lead has an important role in zinc baths [2] thatmust be taken over by another micro-alloying

element and the research in this area suggest the useof bismuth. Bismuth isn’t toxic [3] and like lead ithas the same fluidization effect on the melt alsodecreasing the superficial force [4; 5].

There is a great interest in carrying out research[1], which explores the influence of the alloyingelements in zinc melts on coating layerscharacteristics and the characteristics of the melts.Thick layer of hot dip galvanizing process varieswith temperature and immersion time and dependson the chemical composition of melt. Elements ofmicro alloying change the fluidity and surfacetension of zinc melt and consequently drainage ofzinc when support is extract from the melt [2, 3].Immersion temperature affects both quality andquantity of zinc deposit on the surface of bandsand produce ash, slag and dross. The usualworking temperature in galvanizing process ischosen between 450-460oC. At lower temperaturesis more difficult adhesion between zinc and steeland at higher temperatures (over 470oC) when thelayer thickness starts to decrease, Fe-Zn alloysbegin to break passing interface in the melt andform a large amount of dross. Time of immersionis determined by the thickness of steel sheet anddesired thickness of the layer. For a giventemperature, a given composition to melt and thesame work speed, increasing the duration ofimmersion leads to a corresponding increase indeposit weight. The optimal duration is determinedby technology. Times too small leading to


134

defective adhesion and uniformity, and too largeleading a strong attack, large layer of alloy Zn-Feand the emergence of such compounds in the melt(matt, ash), [5].

2. Experimental research

In the framework of the research were analyzedfour different zinc melts micro alloyed with nickel,tin, bismuth, lead whose chemical composition ispresented in Table 1.

It has galvanized steel sheet to chemicalcomposition according to Table 2.

Table 1. The chemical composition of micro alloyed Znmelts

Alloy Ni[%]

Bi[%]

Sn[%]

Pb[%]

Zn[%]

Zn-Ni-Bi-Sn 0.16 0.71 2.95 - 96.18

Zn-Ni-Pb-Sn 0.16 - 2.88 0.72 96.24

Zn-Ni-Pb-Bi-Sn 0.16 0.41 3.49 0.43 95.51

Table 2. The chemical composition of steel support,in %

C S Mn P S0,030 0,030 0.300 0,015 0,010

Al Ti V Ni Cr0.046 0.002 0.001 0.008 0.025

For the preparation of the zinc bath, pure zinc(SHG) was used.

Micro alloying with nickel, bismuth, tin, andlead was made directly, using metallic elements,finely crushed, followed by mechanical mixing.Laboratory experiments at the micro alloying zincmelts were performed in the temperature rangetypical galvanizing processes, working at 450-4600C.

Experimental immersion times were 3, 5, 8, 15seconds. In the experiments did not apply anycontrol and uniformity process, coating thicknessresulting in free flow of zinc from the sample.

According to Zn-Ni phase diagram (Fig.1),nickel and zinc forms intermetallic compounds andsolid solutions based compounds. Nickel isinsoluble in zinc and at the 418oC and 94.8% Znforms the ( + Zn) eutectic.

Alloying with Ni was made directly through theuse of finely crushed Ni, (fine pieces < 1mm) andmechanical stirring. Micro alloyed process is

longer, being at least one hour at the temperatureof 7000C.

Assimilation efficiency is low because nickelwas lost in slag and dross (Table 3 and Table 4).

Figure 1: The Zn - Ni phase diagram [6]

Table 3. The chemical composition of dross

seample 1 2 3 4 5

Ni, % 0,024 0,021 0,034 0,022 0,018

Fe, % 0,024 0,024 0,021 0,025 0,024

Zn,% diff. diff. diff. diff. diff.

Table 4. The chemical composition of slag

seample 1 2 3 4 5

Ni, % 0,019 0,021 0,024 0,016 0,018

Fe, % 0,016 0,020 0,021 0,020 0,020

Zn,% diff. diff. diff. diff. diff.

Assimilation efficiency of the direct alloyingexperiments was found in 77%, measured threehours after the introduction of nickel (to calculate aconcentration of 0.21% Ni and 0.16% Ni wasobtained). In the literature it is recommended bothdirectly alloying with metallic nickel and the use ofalloys with a maximum of 5% nickel [5].

In the equilibrium diagrams of Zn-Bi (Fig.2),there is an insolubility of the two metals whoforms a eutectic at 97.3% Bi and 254.5 0C. Giventhe low melting temperature of bismuth (2710C)micro alloying was made with metallic bismuth,grinding and mixing in the melt by mechanicalstirring.


135

Assimilation process of bismuth in the melt wasstable, maximum efficiency is obtained. Bismuth isa micro alloying element used to replace lead, thesame effect of melt fluidity and reduction ofsurface tension without being toxic.

Micro alloying the melt with 0.1% Bi,influences surface tension and fluidity similarlyusing a Pb content of ~ 1% [5,7]. Althoughbismuth is more expensive, the quantity needed foralloying bath is much lower, costs arecompensated. Bismuth is also very stable in themelt and requires replenishment just proportionaladded zinc.

Figure 2: The Zn - Bi phase diagram [6]

In the experiments, micro alloying with tin wasmade with metallic tin. Sn-Zn equilibrium diagram(Fig.3) shows a total insolubility of this element inzinc with the formation of a eutectic at 198.5 0Cand 91.2% Sn. Tin, like most analyzed elementsform intermetallic compounds with nickel.

Figure 3: The Zn – Sn phase diagram [6]

Microstructure of coatings obtained is presentedin Fig. 4. It shows a thin layer of intermetalliccompounds Zn-Fe and metallic layer consisting ofintermetallic compounds (Zn-Ni, Zn-Ni-Sn, Ni-Sn,Ni-Bi) finely dispersed in zinc matrix.

a) Ni-Sn- Bi

b) Ni-Sn- Pb

c) Ni-Sn- Pb-Bi

Figure 4: Microstructure of coatings obtained inzinc alloyed

Analysing changes in layer thickness ofcoatings obtained from micro alloying with nickel(and elements for improving fluidity and structure)shown in Figs 5, 6 and 7. At 4500C, is observedpoor uniformity and high values of thickness,compared with 4600C operating temperature ofmelt. After the trials for these types of coatings is

90 m

90 m

90 m


136

proposed technological temperature of 4600C andmaintenance period of 3-5 seconds. The elementsof micro alloying used can enhance the surfacequality of the zincked steel by: uniformity, texture,and luminosity. This surface layer is dependent onthe melting fluidity, its superficial tension and thesolidification characteristics.

48

3935.5

43.3

38.5

3432.537

30

32

34

36

38

40

42

44

46

48

50

3 5 8 15

Immersion time, [sec]

Laye

r thi

ckne

ss, [

m] T = 450 0C

T = 460 0C

Figure 5: Layer thickness variation depending ontemperature and duration of immersion, ( alloyed Zn

with Ni-Bi-Sn)

43.5

52

48

46

51.5

46

4849

38

40

42

44

46

48

50

52

54

3 5 8 15


Laye

r thi

ckne

ss, [

m]

T = 450 0C

T = 460 0C


with Ni-Pb-Sn)

63

39.5

61.5

57.555

3836.5

33.5

30

35

40

45

50

55

60

65

3 5 8 15


Laye

r thi

ckne

ss, [

m]

T = 450 0C

T = 460 0C


with Ni-Pb-Bi-Sn)

According to the combination of the microalloying elements the surface can have differenttypes of metallic layer which can have animportant effect on the coating aspect.

One can notice the changing of the surfaces’morphology according to the micro alloyingelements as compared to surface morphologyresulted at the coating with pure zinc; thus, atmicro alloying with Ni-Sn-Pb, the surfacemorphology reveals the forming of some bigcrystals with a fan aspect (Fig. 8). If a part of Pb ischanged to Bi the crystals are significantly gettingfinished (Fig. 9).

Figure 8: The coating morphology with Zn-Ni-Sn -Pb,x50

Figure 9: The coating morphology with Zn-Ni-Sn-

-Pb-Bi, x50For measuring the corrosion resistance, it was useda potentiostate PGP type 201. A calomel saturatedelectrode was used as reference electrode and aplatinum wire electrode as an auxiliary one. Thecorrosive environment used for electrochemicaltests was a solution of 3% NaCl at roomtemperature.The samples were prepared for analysis by beingdegreased with acetone, washed and dried [5].


137

From the experimental data obtained frommeasurements we chose representations in theform of Tafel, lg. Icor - f(Ecor) polarizationcurves. The analysis of the graphicalrepresentations (Figs. 10-12) made it possibleto determine the characteristic quantities ofcorrosion: corrosion current intensity Icor,corrosion potential Ecor, corrosion current densityicor, corrosion speed vcor and penetration index p isshow in Table 5. Corrosion behavior in seawaterappreciated by electrochemical tests show a lowerresistance to corrosion coating with Zn-Ni-Pb-Bi-Sn. Comparing the chemical composition of the

threetypes of coatings is a big difference incontent of tin (more than 3% in sample no.3).

Table 5. Values of the corrosion process

Tip of coatingicor

[A/m2]vcor

[g/m2h]p

[mm/an]1. Zn- Ni-Bi-Sn-Cd 8.6230 0.0253 0.0012

2. Zn- Ni-Pb-Sn 9.2242 0.0271 0.00133. Zn- Ni- Pb-Bi-Sn 13.8614 0.0405 0.0020

Figure 10: Tafel curve for Zn-Ni-Bi-Sn

Figure 11: Tafel curve for Zn-Ni-Pb-Sn


138

Figure 12: Tafel curve for Zn-Ni-Pb-Bi-Sn

Conclusions

- coating thickness is less at the same time addinglead or lead and bismuth show a higher fluidityfor zinc melt.- Microstructure shows a thin layer ofintermetallic compounds Zn-Fe and metallic layerconsisting of intermetallic compounds (Zn-Ni,Zn-Ni-Sn, Ni-Sn, Ni-Bi) finely dispersed in zincmatrix.- At alloying with Ni-Sn-Pb surface morphologyreveals the forming of some big crystals with afan aspect. If a part of Pb is changed to Bi thecrystals are significantly getting finished.- Corrosion behavior in seawater appreciated byelectrochemical tests show a lower resistance tocorrosion coating with Zn- Ni- Pb-Bi-Sn.

References

[1] John Zervoudis and Graeme Anderson, p. 4, AReview of Bath Alloy Additives and theirImpact on the Quality of the Galvanized

Coating, Teck Cominco Metals, Toronto,Ontario, Canada 2007.

[2] Galvanizing Reactive Steels, a guide forgalvanizers and specifies,” InternationalLead Zinc Research Organization, 2009.

[3] Gagne, M., Zinc Bismuth Alloys for AfterFabrication Hot Dip Galvanizing, AmericanGalvanizers Association 1997, Houston, TX.

[4] Krepski, R.P., the Influence of Lead in After-Fabrication Hot Dip Galvanizing,14thInternational Galvanizing Conference(Intergalva’85), Munich.

[5] Beguin, Ph., Bosschaerts, M., Dhaussey, D.,Pankert, R., and Gilles, M., “Galveco©, ASolution for Galvanizing Reactive Steel,”Intergalva 2000, Amsterdam.

[6] ASM Metals Handbook, Tenth Edition,Volume 3, “Alloy Phasse Diagrams,” ASMInternational, Metals Park OH.

[7] Tamara RADU, Florentina POTECASU,Maria VLAD, Viorel DRAGAN , Researchon obtaining and characterization of zincmicro-alloyed with bismuth coatings, Metal2010.


139

ACCELERATION TEST MACHINE

Marc Juwet1, Koert Bruggeman2, Filip De Bal3

1KAHO Sint-Lieven, [email protected] Sint-Lieven, [email protected]

3KAHO Sint-Lieven, [email protected]

Abstract: Road transport of cargo is increasing year by year. The European action program ontransport safety aims at an increased safety level. One important aspect of transport safety is loadsecuring: in normal road conditions - including sudden change of lanes and emergency braking - thecargo should not move on the vehicle. In this text a very simple and efficient method for load securing isdescribed. The method is based on the idea of “rigid load units”. Also a test machine to certify andoptimize the rigidity of a load unit is described.

Keywords: road transport, cargo safety, acceleration test.

1 Cargo Securing MethodsThe idea behind cargo securing is that the cargo

should not slide, tilt, wander nor deform on thetruck. Sliding and tilting can easily be understood.Inertia forces that can cause sliding or tilting areknown very well: 0,8g in the forward direction,0,5g in the sideward and backward directions.Wandering is a phenomenon related to vibration ofthe vehicle. At inspection it is found that the cargoor part of the cargo has moved several cm or evenmore in a random direction. Also deformation ofthe load is unacceptable. Deforming load units cancause unacceptable inertia forces and can endangerthe stability of the vehicle, e.g. a heavy piece ofmetal sliding sideward in a rigid crate, can cause atruck to turn over.

Cargo securing can be done based on 3different principles, as outlined below.

1.1 Top Over Lashing – Friction Lashing

Friction between the loading platform of thevehicle and the bottom surface of the cargo shouldprevent sliding. Friction is increased by top overlashing: some lashings over the load pull down theload and increase friction between the loadplatform and the cargo. At the same time, frictionbetween several layers of the load is increased,thus preventing sliding between layers. Top overlashings at the same time prevent tilting andwandering.

Top over lashing is very well known by almostall truck drivers. On the other hand, it has threemajor disadvantages:

1. In most cases the number of top over lashingsis very high. On average for a full truck of 25tons up to 80 top over lashings are necessary(depending on the type of lashings used,depending on the materials and the position oflashing points). Fixing and removing thenecessary lashings can take up to 2 hours,causing an unacceptable workload for thedriver and an unacceptable cost for thetransport company.

2. The top over lashings have to be tensioned, inmost cases up to a tension force of 200 to500daN depending on the design of thetensioning device. A tensioned lashing candamage the cargo unless some cornerprotection is used. This will increase cost,weight and material usage.

3. Very often the tension in a top over lashingdecreases during transport due to vibrationsand/or small movements of the cargo. In thatcase, load securing is inadequate.

1.2 Direct LashingLoad securing devices such as lashings, steel

wire ropes or chains are fixed between the load andlashing points on the vehicle superstructure. Forun-deformable loads such as heavy constructionmachinery, … four different tensioning devicesare used to compensate for the inertia forces. Forpalletized goods, one lashing device for eachdirection is used: e.g. to compensate for the forcesduring braking, a lashing is put horizontally in

mailto:Marc.Juwet:@kahosl.be

mailto:Koert.Bruggeman:@kahosl.be

mailto:Filip.DeBal:@kahosl.be


140

front of the load and tensioned backwards at eachside of the load. Typically the actual tension forcein the tensioning device during transport is notimportant. The tension will increase if the loadtends to move. This means that in most cases, theload will not be damaged by the tensioning deviceand that no repetitive tensioning during transport isnecessary. For palletized goods direct lashing canbe an economical solution if the load units do nottend to tilt nor deform.

1.3 BlockingThe load is blocked against a certified part of

the vehicle or another part of the load. In this casea so called XL-coded vehicle is very interesting:the European standard EN12642:2006, code XL,guarantees that the walls of a vehiclesuperstructure will resist the forces from the load ifthis load is “un-deformable” and put against thewall without a gap. The walls of an XL codedvehicle type are tested statically or dynamically.The front wall resists a distributed load of 50% ofthe loading capacity of the vehicle, sidewalls resist40% and the back wall resists 30% of the loadingcapacity. Such rigid walls and the friction betweenload platform and load compensate for the inertiaforces. Gaps and deformation of load units areunacceptable since these can cause impact forcesthat are much higher than the inertia forces. Only asmall gap of some cm between the pallets and thewalls is generally accepted: 2 pallets of 120cm or 3pallets of 80cm can be put between two sidewallsof a standard XL coded vehicle without furtherload securing. In many cases the loaded productsare somehow smaller than the pallet. Such a gapbetween the products on neighbouring pallets is noproblem if the load units are really rigid.

Blocking is the most efficient load securingmethod for palletized products. The load securingefficiency does not depend on the efforts of thedriver. If the products are fixed rigidly on the palletand the XL coded vehicle is loaded without gapsbetween pallets and vehicle walls, no additionalload securing action is required. A new XL codedvehicle should not be more expensive than anothervehicle. No time is lost for load securing. The mainuncertainty is the rigidity of the load units.

2 Test Methods for Load Unit RigidityA load unit on a pallet that is blocked, has to

resist inertia forces of 0,5g or 0,8g depending on

the direction. The rigidity depends on severalparameters such as the product type (e.g. fertilizergrains, liquid washing powder, concrete tiles, …),the primary packaging (paper or PE bags, glass orPE or PET bottles, carton boxes, …), thesecondary packaging (tray, American box, shrinkfilm, …) , the stacking pattern (interlocked, …),the transport packaging (wrapped foil, stretchhead, tie sheets, corner protection, straps, …). It isnot possible to predict rigidity at a glance.Therefore a test method has been developed.

2.1 Inclination Test

In the transport community, an inclination testis well known. According to Newton’s law aninclination of 26,7° is supposed to be equivalent toan acceleration of 0,5g. As Newton’s law is notvalid for deformable goods, this test is not valid totest the rigidity of palletized goods. Even for manyun-deformable goods this test is not applicable, e.g.a concrete prefab wall of 10cm thickness and 2mheight will tilt if inclined under 26,7°.

2.2 Vibration Test, Impact Test

In the packaging community these two types oftests are very well known. Vibration tests are beingused to simulate transport conditions in laboratorycircumstances. The deformation of the load unitstrongly depends on the frequency of the vibration.Standardized frequency spectra or measuredfrequency spectra can be used. In most cases avibration test will predict problems related toprimary and secondary packaging during roadtransport, but it does not lead to an unambiguousconclusion on transport packaging efficiency.

To test the transport packaging an impact test isvery well known. The load unit hits a wall with acertain velocity. During such an impact, thedeceleration forces can rise up to 3 or 15gdepending on the damping material that is beingused. However the duration of these forces is veryshort: 10 up to 80 ms. The effect of an externalforce on a deformable product is a deformationvague going through the product. The effect of theexternal force will depend on the duration of thatforce. During road transport, the inertia forces canoccur for seconds. Therefore in most cases animpact test does not predict the effect of realinertia forces in a truck.


141

2.3 Real Truck TestOf course a real truck test is a relevant test to

certify the rigidity of a load unit. In practice acertification test on a vehicle is very expensive anddifficult to do. The load unit is put on the loadingplatform of a vehicle and sliding is prevented, e.g.by putting a high friction material in between theplatform and the load unit. The vehicle is equippedwith some measuring system in order to measureaccelerations in forward and sideward directions.A series of movements are performed withincreasing inertia forces. The load unit is inspectedpermanently during this test.

2.4 Laboratory Acceleration Test

An acceleration test that can be performed inlaboratory circumstances appeared to be necessary.Therefore an acceleration test machine has beendeveloped (Fig. 1). A table is put on slidingbearing and can move along a horizontal trail. Avery high torque servo motor has been developedto move this table. The acceleration of the table iscontrolled based on a feedback loop in which theactual position of the table is measuredpermanently. The acceleration of the table can bechosen between 1 and 10m/s². A load unit is put onthe table and is subjected to the chosenacceleration in the same way as in a vehicle. Thedeformation of the load unit during theacceleration can be monitored by means of a highspeed video. The permanent deformation of theload unit after the test can be measured in detailafter the test.

Figure 1: The acceleration test machine afteran acceleration test, with a failed load unit.

Up to 1200 tests with different load units havebeen done until now. The influence of severalparameters of the test, has been examined.

2.4.1 Acceleration or Deceleration.The acceleration test machine can be used in

two possible ways: the load unit can be acceleratedslowly (eg.0,2g) up to a certain speed anddecelerated with an accurately controlleddeceleration (eg.0,5g). Alternatively the load unitcan be accelerated in a controlled way (e.g. 0,5g)in order to check the rigidity and deceleratedslowly (e.g. at 0,2g). The results of both alternativetests should be the same for an un-deformable loadunit.

A real load unit of palletized goods is somehowdeformable and the results of both types of testsare in most cases not identical: an acceleration willdeform the products in the backward direction.During braking the deformed products are forcedto move back in the forward direction starting froma deformed shape. In the second type of test, theslow acceleration does not cause any deformation.The deformation during braking starts from un-deformed products. Therefore the acceleration -deceleration sequence is the most demanding andis chosen as the basis for a test standard.

2.4.2 Duration of the Acceleration.The elastic and permanent deformation of a

load unit depends on the duration of theacceleration. Several experiments have been donewith four types of load units: palletized bricks,palletized bags with PE granulate, palletizedlaminate planks, palletized carton boxes. About 10identically packed pallets of each type have beenaccelerated for different durations. Increasedduration of the acceleration results in increaseddeformation up to a certain acceleration duration.For palletized bricks a duration longer than 0,12sdoes not result in more deformation. For palletizedbags with PE granulate, an acceleration over 0,25sdoes not result in more deformation. Therefore anacceleration duration of 0,275s is put forward as abasis for a test standard.

2.4.3 Acceptable Deformation.Almost all palletized load units will deform

under inertia forces, most often with somepermanent deformation. No scientific basis wasfound to establish criteria for an acceptablepermanent deformation. Therefore several tests


142

have been done with different types of products ondifferent pallet types. Most industrially palletizedproducts are somehow understacked, meaning thatthe product on the pallet is smaller than the pallet.Around the pallet some cm of the pallet is notused. On the other hand, some products areoverstacked, meaning that the product is largerthan the pallet.

Most often understacking is preferred since itprotects the products during manipulation.Understacking and overstacking are veryunfavourable for the rigidity of a load unit, e.g. anunderstacked load unit that is wrapped with foilwill tend to slide on the pallet if the accelerationforce is higher than the friction between productsand pallet. The wrapped foil will not prevent orreduce sliding until the border of the product isaligned with the border of the pallet. On the otherhand, sliding of the product on the pallet is notharmful as long as the product does not continue toslide over the border of the pallet. Therefore, acriterion for an acceptable deformation mustdistinguish between sliding on the pallet anddeformation of the palletized products. Severaltested products have been categorized by 20transport or packaging experts in acceptable andunacceptable permanent deformation.Unacceptable deformation can be described as:displacement before and after the test measured inany horizontal plane is smaller than 6cm andsmaller than 4% of the height of the pallet.

3 Additional Advantages of Load UnitRigidity

The initial purpose of an acceleration test is theverification of the rigidity of a load unit for reasonsof load securing. A rigid load unit in an XLcertified vehicle does not require further lashing ifgaps can be avoided. Some multinationalcompanies systematically implemented thisstrategy for load securing. Unexpectedly this led totwo additional advantages.

3.1 High Speed Camera RecordingWhen doing an acceleration test with a

conventionally packed load unit, the deformation isfilmed with a high speed camera. In most cases thedeformation is unacceptable and in some cases theload unit collapses. The deformed or collapsedload units are very well known from pictures ofpalletized goods arriving at their destination.However, these pictures do not give a lead for

improvement of the packaging. This problem isovercome by the videos. The cause of the failurecan be found, e.g. the damage starts withinclination of the boxes on the second layer,followed by tearing up the foil at the foot of thepallet. The cause of the damage is not a lack ofrigidity of the foil at the foot of the pallet, but anun-sufficient stiffness of the foil at the level of thesecond layer. The transport packaging can beimproved systematically by correcting the primaryfailure modes, in the above example by increasingthe pre-stretch of the wrapped foil at the level ofthe second layer. At the same time packagingmaterial that does not contribute to the rigidity canbe removed. In many cases wrapped foil in theupper half of a pallet can be decreased, the numberof tie sheets can be decreased, …

The videos can provide information related tothe stacking pattern also. In most cases somemixture of stacking strategy turns out to be mostefficient. Columnar stacking (one box on top of theother with the same orientation) can be preferablefor the lower layers since this type of stacking cancarry higher loads. The stability of the load unitcan be increased by interlocked stacking (otherposition or other orientation of the boxes indifferent layers) in the top layers. The use ofinterlayer sheets (paper, corrugated board, highfriction paper, …) strongly influences the optimalstacking pattern. In case of a secondary packingincluding corrugated board (trays, boxes, displayboxes, …), a global optimization of secondarypackaging, interlayer sheets, stacking pattern andwrapped foil can give rise to considerable savings,e.g. up to €20 per pallet.

3.2 Cost Reduction

Two independent large scale studies in Europeconclude that approximately 4% of products aredamaged upon arrival at their destination. Probablythis percentage is even higher since most (small)companies do not keep records of this type oflosses and they do not like to publish this kind offigures. Insurance companies pretend that in someindustry sectors the percentage is about 6%. Theadditional costs for the shipping company are evenmore considerable (disturbed customer relation,emergency delivery, claims, processing ofdamaged products, …) The real cause of thedamage is in most cases not known: damagerelated to manipulation, to forces occurring duringtransport, to environmental conditions, …. In the


143

past months the transport packaging of several loadunits has been optimized and a rigidity up to 0,5ghas been certified. In at least 5 different cases thefigures of damaged products before and after theoptimization are available. On average for the 5different cases, the percentage of damagedproducts has decreased by 50%. Two types ofproducts are delivered as full truck loads directly tothe customer. At loading time in the productionplant they are most likely not damaged. They aretransported by road and inspected upon arrival.Damaged products are in many cases not unloaded.Damage is thus most likely related to forcesoccurring during road transport. In both cases thedamage percentage was reduced by 80%.

4 Conclusion

Load unit rigidity is very important for efficientload securing of goods during road transport. Thisrigidity cannot be quantified with conventionalstandard tests available in the packaging sector.Therefore an acceleration test machine has beendeveloped and a new test standard is proposed.

The load unit is accelerated for at least 0,275sand no significant permanent deformation isallowed. This acceleration test does not only allowfor certification of load unit rigidity, it also allowsto optimize transport packaging in a systematicway. In almost all cases this leads to a reduction inpackaging material and a reduction in damagedproducts upon arrival. In many cases this evenleads to a reduction of the total packaging cost.

5 References

5.1 Standards Related to Packaging Tests

[1] DIN 55440-1, Packmittelprüfung;Stauchprüfung; Prüfung mit konstanterVorschubgeschwindigkeit

[2] ISO 12048, Packaging - Complete, filledtransport packages - Compression andstacking tests using a compression tester

[3] ASTM D642-00, Standard Test Method forDetermining Compressive Resistance ofShipping Containers, Components, and UnitLoads

[4] ISO 527-3, Plastics - Determination oftensile properties - Part 3: Test conditionsfor films and sheets

[5] ISO 1924-2, Paper and board -Determination of tensile properties - Part 2:Constant rate of elongation method (20mm/min)

[6] EN 60068-2-1, Environmental testing - Part2-1 : Tests - Test A : Cold

[7] EN 60068-2-2, Environmental testing - Part2-2 : Tests - Test B : Dry heat

[8] EN 60068-2-6, Environmental testing -Part2 : Tests - Test Fc : Vibration (sinusoidal)

[9] EN 60068-2-27, Environmental testing -Part 2-27 : Tests - Test Ea and guidance :Shock

[10] EN 60068-2-64, Environmental testing -Part 2-64: Tests - Test Fh: Vibration,broadband random and guidance

[11] ASTM D4728, Standard Test Method forRandom Vibration Testing of ShippingContainers

[12] EN 60068-2-30, Environmental testing -Part 2-30 : Tests - Test Db : Damp heat,cyclic (12 h + 12 h cycle)

[13] EN 60068-2-80, Environmental testing -Part 2-80 : Tests - Test Fi : Vibration -Mixed mode

[14] EN 60068-2-81, Environmental testing -Part 2-81 : Tests - Test Ei : Shock - Shockresponse spectrum synthesis

5.2 Books

[15] The ISTA 2011 resource book, InternationalSafe Transit Association, January 2011

[16] M. Juwet, Basisregels ladingzekering voorwegvervoer, 92p, Tandem, 2010


144

[17] M. Juwet, Règles de base – arrimage descharges pour transport routier, 92p,Tandem 2010

[18] M. Juwet, Basic principles of load securingfor road transport, 86p, Tandem 2011-03-28

5.3 Conference Papers

[19] M. Juwet, K. Bruggeman, Test method forstability of load units, Dekra InternationalConference on Load Securing,Eurospeedway Lausitz, Duitsland, 8-9October 09

[20] M. Juwet, K. Bruggeman, Opportunities fortransport packaging optimization, ISTAEurope international conference, Valencia,Spain, 26-28 October 2010

[21] M. Juwet, K. Bruggeman, Verpakken enzekeren van lading bij wegvervoer , EVOconference, Asten, Nederland, 18 november2010

[22] M. Juwet, J. Dendauw, Methods to quantifyload unit rigidity, ITW seminar, Virton,Belgium, 15 February 2011

[23] M. Juwet, Nieuwe Europese regelgevingvoor transportverpakking, Empackconference, ’s Hertoghenbosch, Nederland,15 maart 2011

5.4 Articles

[24] Zeng, F., Le Grognec, P., Lacrampe, M-F.,Krawczak, P., A constitutive model forsemicrystalline polymers at hightemperature and finite plastic strain:Application to PA6 and PE biaxialstretching, Mechanics of Materials, 2010

[25] D. Briassoulis, A. Aristopoulou, Adaptationand harmonisation of standard testingmethods for mechanical properties of low-density polyethylene (LDPE) films,Department of Agricultural Engineering,Agricultural University of Athens, Athens,Greece, November 2000

[26] Sangkeun Rheea, James L. Whiteb, Crystalstructure, morphology, orientation, andmechanical properties of biaxially orientedpolyamide 6 films, Institute of PolymerEngineering, The University of Akron,Akron, USA, 8 July 2002

[27] Naresh Bhatnagara,_, Rakesh Bhardwaja,Palani Selvakumara, Mathias Brieub,Development of a biaxial tensile test fixturefor reinforced thermoplastic composites,Department of Mechanical Engineering,Indian Institute of Technology, New Delhi,India, 6 September 2006

[28] Arthur Bobovitcha,b, Emmanuel M.Gutmanb, Sven Henningc, Goerg H.Michlerc, Morphology and stress-relaxationof biaxially oriented cross-linkedpolyethylene films, Department ofEngineering, Martin-Luther University,Germany, 17 September 2002

5.5 Patents

[29] Test machine for palletized products,application 2010/0155, March 9, 2010


145

DEVELOPMENT OF WASTE MANAGEMENT SYSTEMS IN ANINTEGRATED SHIPYARD

Daniela Buruiana

1 „Dun rea de Jos” University of Galati, [email protected]

Abstract. This paper aims to present a waste management system in the shipyard industrial areaand the solutions adopted to reduce the environmental impact caused by wastes from ship repair andmaintenance work. The waste impact (material resulting from the hull blasting , oil and paint wastes )implies the occupation of the land in the vicinity of the shipyard studied, soil and water contaminationand landscape change. Even if some of them are inert, solid naval wastes cause storage problems ,primarily because of the large quantities resulted . Identifying solutions for reuse of solid waste leads toreduced costs and superior capitalization of those materials. The waste management policies to beadopted by a shipyard must find solutions to their very source involving both reduction and recycling.It is also necessary to limit the negative environmental impact by greening works in the already aaffected areas by the uncontrolled deposits

Keywords: management, wastes , recycling, environment, efficiency

1. General

An effective integrated waste managementmust focus on the following purposes: identifyingthe sources/activities generating emissions,emissions monitoring and management, measuresto reduce emissions through the application of thebest practices in the sector concerned

Waste elimination and/or minimization shouldbe a priority in any integrated environmentalmanagement plan. Full or partial recycling leads toreducing the amount of waste being deposited, aswell as reducing the amount of raw materials usedin any technological processes

When referring to an efficient solid wastesmanagement in shipyards, it is necessary to firstconduct a thorough analysis of the processes ofconstruction, repair and maintenance of vesselsand the volume of materials used in shipbuildingindustry. The results of such an analysis wouldprovide for the knowledge, appreciation andapplication of the best integrated wastemanagement solutions. For most shipyards, thesesolutions would allow the recovery of waste inother applications, cleaning the areas affected byuncontrolled waste disposal .

The evolution of marine industrial systems inthe context of the modern concepts of total qualityand sustainable development promoted by anyenvironmental policy is based primarily on the

understanding of technological processes related tomarine environmental problems and restoring thebalance, particularly through recycling in therelationship repair / maintenance- vessels-environment – recycling- material – energy.

An issue of increased interest in the worldtoday, also supported by environmental policyapplied to a shipyard, is represented by wastesfrom both an environmental and economic pointsof viewEnvironmental concerns in the strategies pursuedand applied to shipyards around the world fall intotwo directions:- development of advanced technologies thatsignificantly reduce emissions;- increasing waste recycling and recovery outputto levels close to 100%

2. Defining and identifying pollution sourcesin a shipyard

Shipyards all over the world are faced withthe need to restructure their work so as to meet thesustainable development requirement byprotecting the environment as efficiently aspossible since the activity in this sector requires amajor consumption of raw material and energywhile generating considerable amounts of pollutingemissions such as gas, solid waste and wastewaters. Compared with the practice and global

mailto:dnegoita:@ugal.ro


146

trends, Romanian shipbuilding industry featuresdeficiencies both in the collection, transport andstorage of all categories of waste and wasterecycling and /or reuse. The tendency to dismantlehundreds of ships each year occurs continuously.This is a consequence of rapid development in thefield of naval technology, the competition betweenthe carriers and the current crisis. We can give asan example the situation in some Asian countrieswhere more than 80% of the ships are dismantledin shipyards located on beaches in India,Bangladesh, Pakistan and Turkey. Bangladesh isthe country where most ships are dismantled.

Most shipyards use the cheapest method, butalso the most harmful to the environment: "failedon the beach" for dismantling ships. This methodhas a high price in human lives and causes manydiseases due to exposure of workers to toxicaction. The European Economic and SocialCommittee is aware that in the foreseeable future,the failure of ships for dismantling on the beachwill remain the preferred method. Therefore,current conditions in shipyards must be improvedso that they operate in a safe and ecologicalenvironment

Figure 1: Technological processes specific to ship repair and maintenance sector and specific pollutant emissions

Blasting /first treatment

Ship dismantling

Blasting and paintingon different portions

BLASTING ANDPAINTING HALL

Cleaning withSolvents; chemical

Degreasing, removal

Assembling theship by welding

Maintenance

Painting

Sand, slag, VOCs

WastewaterSlag and smog

Sand, slag,VOCs

SolventVOC emissions

Metal surfacecleaning, lubricating

products

Batteries, waste fromoil, VOC emissions


147

Conditions relating to environmental and socialprotection are disastrous in South Asian shipyards.

They result in unfair competition against theirEuropean counterparts. Moreover, high localdemand for recycled steel is an additional problemfor European competitiveness [2].

Worldwide, the shipbuilding technique in thelast decade has reached a high level ofperformance, demonstrating a great capacity toadapt to changes brought in by raw materials andenergy conditions, the need to increaseproductivity, reduce consumption and comply withthe environment legislation that is becoming moresevere.

It can be considered the most representativetechnological processes generating pollutantemissions from a shipyard are those in the sectorsof dismantling, repair and maintenance of ships.Pollutant emissions generated by the blast cleaningof metal surfaces as well as the process of paintingor welding of metal components are the mostimportant Figure 1 illustrates the technological

processes in a sector of ship repair andmaintenance along with their pollutant emissionsand specific wastes.

Monitoring pollutant emissions in conjunctionwith the process that generate them calls for, andsimultaneously allows for, the implementation ofan integrated environmental management in thesesectors.. Today it is accepted that, at the currentlevel of knowledge, is not yet possible to establisha balance of environmental pollutants on globaldevelopment in the shipping industry. This themain reason why, in their analysis on thecorrelation energy-environment-waste orrecycling-environment, they still use energybalances, material balances respectively, for theprocess being analyzed

Material balances are made in companies inorder to be able to monitor the emissions based onmaterial inputs (Table 1). This model is promotedby the shipyards in Japan, which managed to reachwhat is found in the literature as " waste-freeindustry" or "zero waste”:

Table 1: Balance sheet materials – emission and waste

Process Rawmaterials/materials

Emissionsemissionsto air

wastewater

solid waste

Shipdismantling –assembling by

welding

Electrodes SmokeCO2

- Slag fromwelding electrodes

SurfacePreparation

Sandblast GritSolvents

Degreasing Alkalineand acid agents

SandblastGrit

SolventsDust

emissionsfrom thecleaningprocess

Solventsand acid

water

Recyclablewaste grit

Slag Residuesfrom cleaning areas

SurfaceFinishing

Sandblast GritSolvents Paint Primer

- - Waste WaterContainers of paint,

brushes

Reducing water pollution by increasing therecovery of all categories of waste and theircontrolled storage is a priority in environmentalpolicy analysis proposed in a yard analysis. As afirst step, it should be made rigorous managementof the waste generated at source with particularfocus on reducing the quantity and harmfulness of

waste, advanced recycling of waste byreintroducing them into various stages of the sametechnological flow

This ensures protection of natural resources,increase reuse of waste materials by convertingthem into raw materials for their own flows or of


148

other industries, controlled storage of all categoriesof waste.All these objectives are proposing an action ofinventory and management of sources generatingwastes and monitoring the pollution they cause.Monitoring means the acquisition, storage andprocessing input data to assess some processeswithin the shipyard. This would allow makingcorrect decisions, argued on reconsideration oftechnology and design in relation to each part ofthe technological process

3. Peculiarities of waste management in ashipyard

Waste management is a complex issueconsisting of a multitude of components. Inparticular, waste management in a shipyard in theEuropean Union represents a particular challengein that it should not disrupt the European market ofship construction and repair. Although there is nosingle recipe that can be applied in all cases, theEuropean Union has firm principles underlying thedifferent levels of waste management systems [3]:-principle of prevention - the production of wasteshould be minimized and avoided if possible;- principle of responsibility for pollutinggeneration and payment – he who produces wasteor contaminates the environment must pay the costof his actions; precautionary principle- potential problems must be anticipated proximityprinciple - waste should be stored as close aspossible to the place where they were produced

The Romanian legislation stimulates the“polluter pays” principle (Article 3 of Law137/1997). The Ministry of Environment andWater has transposed into national legislation onthe environment, which fully complies with thisprinciple. The basic principle of the Directive96/61/EC on integrated pollution prevention andcontrol is best reflected in the concept ofsustainable development, which requires measuresto prevent. The concept of best available techniqueinvolves taking all measures to prevent pollutionfrom the plant design phase and continue in theconstruction, operation and decommissioningstages. Also, the holder of activity is directlyresponsible for any environmental damage he maycause

In an environmental policy, waste managementis an essential component due to increasedproduction and subsequent increase in the wasteswith negative impact on the environment. Thenegative effects of waste on the environment andon human health are well known and publicized

throughout the world, and failure to adequatelyapproach this issue can have catastrophicconsequences

Therefore solutions must be found to reducethe amount of waste generated and to limit theirnegative effects, by reusing the recyclable wastesin the production processes as secondary rawmaterials (such as those resulting from the gritblasting) or by neutralization and final controlleddisposal of those who can’t be recovered. Wastemanagement is the model of industrializationwhich allows increasing the contribution ofindustry to achieve economic and social benefitsfor the present and future generation’s resourceswithout damaging the environment.

In most cases, the dynamics of industrial wasteamount generated in a shipyard can’t be preciselydetermined due to lack of a consistent database andchange of the definitions. This problem isparticularly serious, because to find the totalvolume of industrial waste generated, we shouldknow exactly correlated with the waste evolutioncorrelated with naval dynamics, then to develop acoherent policy with objectives, criteria, measuresand investment costs.

The importance of cost / benefit analysis todetermine the overall economic efficiency of theinvestment consider the direct economic efforts(investment costs) and indirect economic efforts(for example, pollution costs, costs allocated foraffected human health, environmentalrehabilitation costs , costs that are usuallyoverlooked or minimized).

4. ConclusionsA special chapter in the world today and

supported by the current environment policy of ashipyard is represented by the wastes, both interms of environmental and economic points ofview. A successful management will lead tofinding the best solutions for the recovery of wastegenerated and thus both costs and environmentalimpact can be reduced.

The Concern for compliance with legislativerequirements on environmental protection and theneed for harmonization of the economic progresswith rational management of material and energyresources should lead to the recovery of waste bytechnologies that offer both economically andecologically optimal solution:-rigorous management of waste and reduction ofwaste quantity at source;- controlled storage of all categories of waste


149

generated on the platform of a shipyard;- advanced recycling of the waste produced bytheir reintroduction into the various stages of theprocess flow, thus ensuring protection of thenatural resources of raw materials;- increased use of wastes by converting them intoraw materials for other industries.

We believe that a focus on a strongmanagement for the environmental protection aspart of the administration and operation of ashipyard will enhance the environmental qualityperformance. The evolution of marine industrialsystems in the context of modern concepts of totalquality and sustainable development promoted bythe proposed environmental policy is basedprimarily on the understanding of specifictechnological processes in conjunction withenvironmental issues focused on restoring thebalance, particularly through recycling: process-environment-recycling-energy. Environmentalexcellence, incorporated into all processingactivities, can be promoted through the principlessupported by:

- continuous improvement of theenvironmental performance through systematicmonitoring of the environment factors, minimizingthe environmental impact and using the principleof pollution prevention;

- reducing waste at source, and where theyinevitably occur, they should be capitalized or ifthis is not technically and economically feasible,they must be controlled stored while avoiding orreducing any impact on the environment;

- implementing clean technologies and bestpractices in the field by investments leading toenvironmental protection and savings of bothenergy and raw materials.

References

[1] Negoi D, Contribu ii la reducerea polu riiprovocate prin func ionarea unui antiernaval, Teza doctorat 2007;

[2] Avizul Comitetului Economic i SocialEuropean privind Comunicarea Comisiei c treParlamentul European – O strategie a UEpentru ameliorarea practicilor de dezmembrarea navelor, http://www.eurlex.europa.eu;

[3] Recycling Spent Sandblasting Grit and SimilarWastes as Aggregate in Asphaltic ConcreteNaval Facilities Engineering Service CenterPort Hueneme, California 93043-4370.

[4] Rojanschi ,V, Bran, F., Politici i strategii demediu, Editura Economic , Bucure ti, 2002.

[5] Negoi D., Dispersion of the atmosphericpollutants on the platform of a shipyard,published at the Sustainability for Humanity &Environment in the Extended- ConnectionField Science-Economy Policy, Timi oara, 24-25 Febr. 2005, pp. 115-119.

http://www.eurlex.europa.eu;/


150


151

METHOD FOR DETERMINATION OF THE DISTANCE BETWEENELECTRODES TO ELECTROHYDEAULIC UNDERWATER THROUGH

SCINTILLAS DISCHARGE

Dumitru IACOBtefan cel Mare” University of Suceava

Abstract: In this paper to present a method for fast-track determination of the distance between electrodes tothe discharge of the energy from a condensator batherry for a realize electrohysraulic effect, in function of the principalwork parameters of the installation, tension of loading U and capacity C, as well as maximal work tension of thebatherry maxU .

Key words: electrical, energy, vapacitor, tension, capacity, electrohydraulic, effect.

1. INTRODUCTION

It is knuw the fact that theelectrohydraylic effect to generate to thesudden perforation of the electrolyte betweenof the electrofs from the electrohydraulicchamber, under electrical high tension, withwhich is loaded yhe capacitor bathery, to thesudden discharge of the electrical loadedenergy in the capacitor energy. To the suddendischarge energy, to generate a pressionimpuls which can be utilize in technologicalpurposes.

2. CONDITIONS FOR THEREALIZATION ELECTROHYDRAULICEFFECT

For the realization electrohydeaulicunderwater through scintillas discharge isnecessarytheelectrical perforation betweenelectrodes. There are following situationsrespecting electrical perforation of the intervalfrom the two electrodes and namely>

If rhe intensity of the electrical fieldbetween electrods m

KVE 360 to

realise sudden perforation, which isfavourable for execution of thetechnological operations, because in

this case to generate the pressionimpulse;If rhe intensity of the electrical fieldbetwe enelectrods

mKVEm

KV 36030 to realise

thermal perforation, with heating andplace ionization of the liquide with torealise the pression impulse;Clearly,that this situation can not utilise integhnological purpose;If rhe intensity of the electrical fieldbetween electrods m

KVE 30 the

perforation of the liquide is notpossible.

In other papers, to recommend that for thedetermination of the maximum distancebetween electrods which produce suddendischarge, to utilise the equation:

mmUKUCE2

123

max 06,0(1)

in which:U - the tension of loading pf the capacitorbatherry in KV’;C - the electrical capacity of the batherryin F . From the first condition on theelectrical field from electrods, resultequations (2)and (3):


152

mmUUK

mmUmUC

777,2360

1000360

1

2

max

(2)

mmUUK

mmUmU

333,3330

100030

1

3

min

(3)where:

naxC - the maximum distance for torealise sudden discharge;

min - the minimum distance for whichthe perforation of the liquide is notpossible.

The angles of inclination pf the rights,in comparations wirh abscissa asis, are:

For Cmax :0

2 20,70777,2arctgarctgK(4)For :min

03 25,88333,33arctgarctgK

(5)

From the equality equation (1) and (2) result:2

123

2 06,0 UKUCUK(6)

1

2

KKU e

(7)where eU is tension for which:

EC maxmax

(8)The value 777,22K and the value

1K for the capacity of the instalation fromlaboratoy FC 70 is:

247,07006,0 31K

(9)For this case result:

KVKKUe 2,11

247,0777,2

1

2 (10)

The maximum distance between electrods inthis case is:

mmUKe 2,3124,11777,22max

(11)In figure 3 are represented the two curves:

mmUUKC 777,22max (12)mmUUKE

221max 247,0

(13)To find that for eUU , the condition

(15) is principalbecause if this isaccomplidhed, result that and the thecondition (14) is accomplidhed.

For eUU the condition (14) isprincipal, because if this is accomplidhed,result that and the the condition (15) isaccomplidhed.

If the capacity of the bathery increase,then the value of the tension eU desrease. Forexample, if FC 350 , the value

422,035006,0 31K and

KVUe 6,6422,0777,2

1 .


153

So, with the increase of capacity, decreasethe interval for which the condition (15) isprincipal. For the capacities very large, themaximum distance between electrods, isdetermined of the condition (14).

The optimal distance is:

2max

o (14)

because the variation of the electrical currentin the interval max,0 is about that in figure4.

The value Emaxmax , if maximum worktension of the insyalation eUUmax and

Cmaxmax

if maximum work tension of the insyalationeUUmax

CI - the intensity of the current determined invery large part of electronic conductedcurrent: ;

TI - the intensity of the current determined invery large part of thermal ionization.:

For the capacity of the instalationfrom laboratory FC 70 , the tension

KVU e 2,11 and maximum work tensionKVUKVU e 2,116max . So, in this case

mmmmR 9,8892,8max . The optimaumvalue result:

mmmmE 45.4446.42892,8

2max

0

(15)vvalue with wich have work fortexperimentation.

The experimental determinations confirmfull previous assumpsions.The technological work comain is undercurve OABC from the figure 3.For the tensions KVU 6max , proper for theinstallation from plastic deformation


154

laboratoryof the mechanical engineeringfaculty, maximum and the optimal values to

present in the table

Tabel 1.

3. CONCLUSIONSVirtually, for the determination of the

maximum and optimum distance fromelectrodes, to underwater scintillas discharge,to propose following methpdology:

Function of the capacity of thecapacitor bathery, to determine thevalue 2KTo determine the balue of the yension

eU for which the two distances isrqual:

KVKKU e

1

2 (16)

To compare the maximum worktension of the installation maxU withthe value eUIf eUUmax the maximum distancebetween electrodes is naxE ;If eUUmax the maximum distancebetween electrodes is naxC ;The optimal distance from electrodesid:

2max

o (17)

The value max from the equation (19)is equal with naxE , or with naxC ,function of

reporteU

Umax . Clear, that for eUUmax , is

whatever which from the two equation toutilise, because to obtain same result.

REFERENCES[1]. IACOB DUMITRU, Cercet ri teoretice i

experimentale privind defprmarea plastic la rece asemifabricatelor metalice utilizând energia acumulatîn condensatoare, Tez de doctorat, UniversitateaTehnic „Gh. Asachi” Ia i, 1994.

[2]. IACOB DUMITRU, Braha V., Rusu B.,Cercet ri experimentale privind perforarea idecuparea electro-hidraulic ; Universitatea Maribor,Slovenia, 27-29 octombrie 1994.

[3]. IACOB DUMITRU, Braha Vasile, Luca Liviu,Cercet ri experimentale privind razele de racordare încazul energiei minime ob inute la perforarea idecuparea electro-hidraulic ; Revista Deform riPlastice (Journal of Plastic Deformation) Centrul deStudii si Cercet ri pentru Deform ri Plastice,Universitatea "Lucian Blaga" Sibiu, vol.1(1994)Nr.2, pag.22-26, ISSN 1222-605X.

[4]. IACOB DUMITRU, Braha Vasile, La pressionmaxime dans la chambre de travail a la deformationpar des chocs mecano-hydrauliqes; Buletinul tiin ifical I.P. Ia i, TOM. XLII (46) Fascicola 1-2, Sec ia V,Construc ii de ma ini, 1996, pag. 136-140.

[5]. IACOB DUMITRU, Mironeasa Costel,Deformarea electro-hidraulic , Editura Matrix,Bucure ti, 2008, ISBN 978-973-755-318-8.

[6]. IACOB DUMITRU, Mironeasa Costel, Camerelectrohidraulic cu electrod mobil în timpuldesc rc rii, Brevet de inven ie nr. 123073 din30.09.2010.

Tension U, KV 1 2 3 4 5 6Maximum distance Emax , mm

222max 247,0 UUKE

0,247 0,988 2,223 3,952 6,175 8,892

Optimal distance 0 , mm

2max

0E

0,123

0,1

0,494

0,5

1,111

1,1

1,976

2,0

3,087

3,1

4,446

4,5


155

SUPERIOR WASTE RECOVERY IN THE METALLURGICALINDUSTRY

Elisabeta Vasilescu , Ana Doniga, Alexandru Chiriac

1“Dun rea de Jos” University of Galati , [email protected],3 Dun rea de Jos” University of Galati , [email protected].

Abstract: On the global scale the generating rhythm of the scraps exceeds the natural assimilatingcapacity of the most regional ecosystems.

The numerous studies mode about the good “practices” spotlighted the succe of the ecologicaltechnologie, but expoed the exitence of many indutrial ineficient, polluant and dangerous installation too,wich are, still, into operation. To the last evaluation, the industrialized countries should cut down tominimum four times the row material consumation, so that the considerable effort are necessarytodecrease the quantities of scraps resulted in the industrial processes and increase their recirculation.

A “clean” technology could rely on a new, less polluted production procedure, that recirculates theresulted scraps in a certain phase of fabrication or, on the utilization procedure of the scraps becomingraw material for a secondary production.

The conception of the scraps recovery and recirculation should be brought up to date emphasizing themanagement and the marketing activities to find the utilization fields of the accumulated scraps in theindustrial dump – yards.

Keywords: scraps, metallurgical sources, metallic slags, metallic powders

1. Scraps Sources from the MetallurgicalPlants

In the frame work of the technological flours,typical to the iron and steel industries, besidesmain products the important quantities of the by –products and scraps outcome. In generally, the by –products are re – introduced in the technologicalflow, boving high degree of utilization.

The scraps are stored on the particularly fitedout surfaces but, at present, the recirculating ofsuch materials is pursued by the nonpollutedtechnologies as possible.

A classification of such scraps has in view thefollowing criteria:

-the content of the certain chemical elements,depending on the technological process from

which they proceeded;-risk degree related to the environment and

health impact;-aggregation state at generation moment ; the

intinsic value of the scrap and its utilizationpossibility;

-generation sources (technological flow andoperation which generated the scraps).The mainscrap generating sources from metallurgical fieldsare: coke – chemical plants, sintering, Blast –furnaces plants, Steelworks, Rolling Mils, Forging– plants, Thermal – treatments shops, Mechanicalprocessing – works and their types of scraps are:powders, sludge (slurry), slags, refractory debrischips (splinters), scales, etc. For example, fromrolling mills the scraps which the followingcharacteristics resulted (tables 1, 2):

mailto:elisabeta.vasilescu:@yahoo.com

mailto:sandu.chiriaci:@yahoo.com


156

Table 1 Average grain – size of the scale, slag and metallic melting.

Table 2 Chemical composition of the metallic scrapsScrap type Chemical composition (%)

Fe SiO2 CaO MnScale (%) 66.97 – 71.23 0.70 – 1.97 0.36 – 1.47 -

Metallic slag (%) 63.66 – 70.16 0.11 – 1.46 1.14 – 1.52 0.42 – 0.8Metallic melting(%)

74.48 – 78.37 0.30 – 0.45 1.08 – 1.35 0.43 – 1.13

Having in view the high iron quantity of suchscraps, their higher valorification was layed on thetable, aiming to the iron extraction and its use inother fields.

One of the up-to-date valorification method offerrous scraps is the powder getting used for smallparts manufacturing, with complex geometry, ofwhich fabrication, by casting, isn’t profitable fromtechnical and economical point of view, or partsmanufacturing which, by their operation nature,should have a porous structure (filters, self-lubrication bearings, porous plates foraccumulators etc.).

At present, the utilization methods of themetallic powder-scraps should answer to thefollowing questions:

-getting of the acceptable purity powders forthose fields where they will be used;

-achievement of the wished phisical (shape,grain-size etc) and technological characteristics(compression, sintering capacity)

-acceptable cast of the row materials; resourcesavailability; value of the processing investments.

A large range of the productsmanufactured by the typical methods of thepowders metallurgy is that of the iron and steelproducts with higher characteristics, used formechanical engineering, welding rods (electrodes),electromagnetical couplings, sintered bearings, et.

Generally, the powders, necessary to theseitems manufacturing, are gotten from the oxidesresulted from metallurgical and chemical process,and they are pure products; their purity degree andshape and grain-size depending on the fabricationmethod.

The scraps resulted from the parts and toolsmanufacturing as well as scraps of the plasticdeformation are recovered by the alloyed meltingspraying (pulverization) or by the mixing of thepure metal powders in the given ratio.

First procedure shows the draw hack of thespherical grain-size powder getting good for the

cases where the required porosity parts arenecessary only and that of the contamination by themelt contact to the furnace refractories.

The second procedure supposes the puremetallic powder manufacturing, powders, that, notin all cases, could be gotten with the necessarycharacteristics and either in convenient economicalconditions.

Manufacturing of the alloyed powders, in theadvanced homogeneity and convenient economicalconditions, on the electrochemical way, removesthe above mentioned, drawbacks.

Complying with this technology, the anodicdissolution of the respective alloy takes place fromsemi-finished or worn out parts, in an electrolyte ofsodium or amonium chloride aqueous solution,having concentration of 20%, voltage 12-24V,current density 0.01A/dm2

Formation of the metallic hydroxides takesplace, which are washed, filtered and dewatered inair or nitrogen.

The gotten metallic oxides are reduced byhydrogen at temperature between 500-1000 Cwith or without final crushing.

To get powders from high alloyed steels, theanode is made of the worn out tools or scrapsresulted from the respective steel-grade toolsfabrication. The cathode is made of copper, nickelor graphite. The gotten powder is screened, graded

Scrap type Grain size (mm)0…1 1…5 5…10 1…15 15…20 >20

Scale (%) 8.24 25.06 28.72 32.78 4.80 -Metallic slag (%) 7.14 9.22 7.15 9.50 31.85 33.30Metallic melting (%) - 3.6 4.72 11.78 10.62 68.78Sludge scale (%) 71.4 28.50 - - - -


157

and mixed up with graphite powder additions toachieve, finally, the necessary carbon content.

The powder is pressed in plates and sintered,afterwards.

By same technology the non-alloyed metallicpowders could be gotten using pure metallicanodes and electrolyte consisting of the aqueoussolution of pure amonium chloride, the powderpurity being some to that of the anode (purity).

The cathode could be same metal (alloy) as theanode: copper or graphite.

Technological steps :

The chemical composition of the gotten powderis same as that of the anode and the grain –size ispolyhedronical shape. Grain-size is direct variableto the current density.

By these technologies the followings are gotten 1) The alloyed powders, for every part of

the mechanical engineering, made in thetechnological flaw: pressing, sintering; as well asevery tool, using the worn-out parts and tools and,also, by the valorification of the scraps resultedfrom the fabrication process of the parts, tools,

rolled, forged, extruded – and cast semi-finishedproducts.

2) The non-alloyed powders of high puritymetals, the powder purity being same as to that ofthe anode.

3) The scraps, as stainless steel chips,could be changed in powders using theintercrystalline corrosion method. For that itshould be started from the material with highercarbides precipitated on the crystalline grain-sizeborders, in order that the intercrystalline corrosionprocess to be developed quickly. This method usesthe austenitic stainless steel chips proceeded fromthe mechanical processing.

Several thermal treatments are used toprecipitate the carbides of the grain size bordersand to sensitive the steel at the intercrystallinecorrosion:

- first treatment is quenching in solutionconsisting in steel heating at temperatures rangedin the stability zone of the solid solution, enoughmaintained to dissolve, partially or totally, thesecondary phase grain size (carbides) in matrix(austenite) followed by the high speed cooling, toprevent the precipitation of the secondary phaseand to maintain the solid solution in solved state,thus under-cooled and supersaturated.

After treatment a homogeneous austenite wasgotten where the carbides of steel were dissolved.

- the second treatment is named sensitivenessand it consist of the heating between 650-680 C, with maintain at heat 1 – 2 hours for amassive precipitation of the carbides at grain sizeborders.

- thus treated the steel is corroded by boiling inan aqueous solution of 10% H2SO4 + 11% CuSO4for 48 hours. Steel becomes breakable, so that,after this operation, by mortaring, stainless steelfine powder is gotten.

In the engineering industry, in the hot-rollingdepartments, the surface laminates is covered witha layer of dross (oxide) to be removed by differentmethods work because the damage and harm thequality of equipment processed material. Skimslayer consists of three types of oxides: FeO(wüstite), Fe3O4 (magnetite), Fe2O3 (hematite).

In cold rolling departments, the hot rolled bandis cleansed of dross layer by chemical etching withacid solutions (HCl or H2SO4).

In the technology of the cold strip rolling, animportant operation is the pickling of the semi-finish (hot rolled strip) to remove the oxides and toclean the strip surface.

Electrolyses

Filtering

Metallic hydroxides

Drying

Calcination

Metallic oxides

Processing with H2

Screening

Grading

Metallic powders


158

A solution of H2SO4 is used, having theconcentration of 15 – 20% in water at temperatureof 80 - 90 C. The following chemical reaction takeplace:Fe2O3 + 3H2SO4 = Fe2(SO4)3 + 3H2O

Fe3O4 + 4H2SO4 = Fe2(SO4)3 + FeSO4 + 4H2O

FeO + 4H2SO4 = FeSO4 + H2O

Fe + H2SO4 = FeSO4 + H2

Also, in the bath, the reduction of the ferricsalts to the ferrous salts takes place in the base ofreaction:

Fe2(SO4)3 + H2 = 2FeSO4 + H2SO4

After pickling, an important quantity of thecrystalline light-green ferrous sulphate FeSO47H2O (vitriol iron sulphate) is gotten and, which iseasy oxidated in contact with the oxygen of the air.

Ferrous sulfate is practically a waste resultingfrom the etching. It can be sold or used forindustrial water treatment industry dyes,medicines, agriculture.

The crystallized ferrous sulphate FeSO4 7H2O,and wastewater of the sulphuric acid picklinginstallation could be used, with good results, to get

some aqueous or non-aqueous systems, whichcontains the iron oxides or hydrated oxides, incolloidal state or in suspension.

Finally it seeks magnetite (Fe3O4), a powderwith good magnetic properties and different uses.A simple method for obtaining Elmore Fe3O4 wasexperienced since 1938 and consists of magnetiteprecipitation from solutions of salts of bi-andtrivalent iron, by the action of excess sodiumhydroxide.

This method was applied based on the solutionof FeSO4 • 7H2O Etching bands resulting from theCold strip mill at Mittal Steel Galati. It is usuallyadded and a solution of FeCl3 • 6H2O.

This method used the chemical reaction:

Fe(aq)+2 + 2Fe(aq)

+3 + 4HO - = Fe3O4 + 4H+

Fe3O4 is gotten in suspension or colloidalsolution in various aqueous or non aqueousmediums.

In order not to cause a rapid decrease in pH -his alkaline solution is added in excess (10%NaOH in water). After precipitation of Fe3O4 oxidation processcan proceed both in solution and in the suspensionduring filtration, washing and milling.

Technology of getting Fe3O4 from solution,which contains Fe II and Fe III atoms from thesystem, the following operations are presented infigure 1.

Figure 1 The main operations of getting Fe3O4 from solution

Dissolution is made in distilled water to preventfurther oxidation and eliminate some of thephysically dissolved oxygen.

The mixing done so that the oxygen ofatmosphere is not driven in solution rather than insmall extent.

The mixed oxides of Fe II and Fe III is knownas magnetite (ferroferric oxides) in technic; it isspinel type (MIIO H2

IIO3) with magneticalcharacteristics. The numerous utilizations it has,both as the colloidal state (magnetical fluid) and inthe shape of the dried powder in electronic andelectrical engineering, radio-broadcasting, non-distructive control and electro-domesticequipments. Filtering is made by the under-

DissolutionFeCl3

Mixing Precipitation Filtering

DissolutionFeSO4 + 7H2O

Washing Drying Crushing Grading Fe3O4 grade


159

pressure filters, maintaining a liquid cushion, aslong as possible, on the precipitation surface. In thesame way running and washing.

The drying is made in vacuum drying stove at105 C to avoid oxygen action on the precipitate.Depending on how to exploit Fe3O4 obtainedthrough this process, choose the appropriate sizerange. Sorting can be done on an assortment ofvibrant site. If they need a very fine grained (<1 µm) provides a wet or dry milling. An optimumgrain can be obtained by passing an appropriateaqueous phase precipitated in the organic phase.Keeping quasicoloidale very fine particles insolutions can be achieved with surface-activesystems. Oleic acid precipitate moisten and mixone more time, so that each particle to be coveredwith a thin film of oleic acid, which will prevent

congestion. For an advanced mixing colloid millcan be used according to the method Bibik [..]

Finally you can add a quantity of toluene for avariable density, depending on the purpose. Youget a ferrofluid with magnetic properties withmultiple uses:- ferrofluid seal;- ferrofluid bearings;- the effects of levitation applications;- ferrofluid systems of writing and posting.

The powder obtained by precipitation method isblack, very fine, the reason is not flowing, formsmall clusters and is magnetic. Density is about.0.33 g/cm3. Particle size, d = max.3, 7 µmDiffractometer analysis showed a high purity ofthe powders (Fig. 2)

Figure 2 Fe3O4 powder diffractometry

2. Conclusions

To get high purity metallic powders from themetallic or non-metallic scraps is one of the up todate method of higher valorification of the scraps,from the metallurgical plants.Non-polluted “clean-technologies” are used,generally, with low consumption of energy andmaterials.

Technologies require simple installation withhigh performances

By a good management of the technologicaloperation the wished purity of powder could be

gotten, into a large range of grain size, grainshapes and with multiple utilisations.

References

[1] Negulescu, M; .a. Protec ia mediuluiînconjur tor, Editura Tehnic , Bucure ti,1995

[2] Rojanschi,V; .a. Protec ia i ingineriamediului, Editura Economic , Bucure ti 2001

[3]*** Gospod rirea de eurilor-Buletin deinformare INFOTERA, Bucure ti,1998.


160

[4] *** Unser Metall Recycling zu 100%,nr.6/1997, pag.398.

[5] *** De eurile-restric ie sau stimulent aldezvolt rii- Simpozionul CMRM;Bucure ti,1993

[6] E. Vasilescu; A. Doniga Valorificareasuperioara a deseurilor la pulberi micronicecu caracteristici fizico-chimice impuse,Contract de cercetare nr.262/ 2000

[7] M.Vlad, E. Vasilescu, A. Doniga Cercetariprivind valorificarea deseurilor identificatein actiunea de implementare a sistemului demanagement de mediu, Contract cercetarenr.345/2001

[8] V.Popa, C.Caraci Valorificarea deseurilormetalice Editura Tehnica Bucuresti1973

[9] V. Constantinescu, R.L Orban s.a. Aspectetehnico-economice ale elaborarii pulberilorprin valorificarea superioara a materialelormetalice scoase din uz, Conferinta demetalurgia pulberilor, 1983, vol. 12, p.119.

[10] Gh.Facsko, C.Radovan, C DaminescuObtinerea pulberilor de nichel pe caleelectrolitica, Conferinta de metalurgiapulberilor, 1975, vol. 1, p.19.

[11] F.Oprea , I. Constantin, P. Nicolae, C.Rohr Procedeu semicontinuu de valorificare

complexa a materialelor refolosibile dincarburi metalice sinterizate, Conferinta demetalurgia pulberilor, Cluj-Napoca 1983, vol.2, p.125


161

IMPACT OF STREET’S LANDSCAPE ON PEOPLE ORIENTATIONAND ROAD SAFETY IN THE CITY

A. Polyakov, V. Shvets, O. Veremiy, M.Grabenko

State Technical University Vinnitsa, Ukraine, [email protected]

Abstract: in this work is shown a sight of the effect of road’s gardening on its aesthetic perceptionand on road safety in cities. Planting of trees are described and the principles of their use as well as theclassification of trees by psychological influence on people.

Keywords: gardening, orientation, street, traffic safety.

Introduction

The most progressive type of humansettlements development is the creation of urbanagglomerations, which allows people living inlarge areas to get to administrative and industrialcenters of cities and consequently get higherpaying jobs and improve their standard of living.

With the growth of cities, its industrialdevelopment is becoming a more difficultproblem for the environment and the creation ofgood conditions for human recreation. Intensivedevelopment of the industrial economy isaccompanied by significant violations of thesurrounding environment.

Gardening of the cities of Ukraine, during theyears of economic crisis, was increasinglyabandoned, there is a lack of recreational, arts andartistic features, the existing do not carry anyinformation and meaning.

With the building of urban agglomerationsincreases the number of transport’s means,especially individual, the number and length ofhighway between the cities ofagglomeration. There is therefore a significantcomplication in road markings, making it difficultto orientate drivers on the road and usually leadsto accidental road impacts.

To solve the above problems, we propose toapproach the landscaping of the streets not onlyby improving the ecological condition of

decorative and aesthetic appearance, but also bypaying attention to the informative planting oftrees and their psychological impact on drivers.

Main part

Road-maintenance organizations involved ingardening of roads mostly carry out monotonousordinary planting that did not change during anextended period of time. This is not the bestsolution. There is a clear relationship between theaesthetic qualities of highway and movementsafety (in primitive solution, monotonous roadsplanting, traffic accidents are much higher). Thatis why it is often said: "The issue is not about howmuch would cost the construction a city’sbeautiful highway, but about what will cost itsimperfection.

Visually smoother road is essential for safeand confident driving. This option does not onlyrequires the usual driver's visual reference points(such as the edge of the carriageway, brow subgrade, concrete pavement joint axis), but alsoadditional information about the direction of themovement (such as the use of contrasting linesmarking, boundary lines, columns and bar guides'interiors, vegetable crops).

It should be noted the ability of shrub plantingto soften the consequences of a car coming of theroad for any reason. For example, a shrub of 30-

mailto:elena.veremiy:@mail.ru


162

40% provides the deceleration of a car (whichcame at a speed of 90 km/h on the road and at anangle of 30 ° to the axis of the road) on only 3-4m.

In lowland areas it is necessary to avoidmonotonous planting. Longitudinal planting onurban roads was formerly widespread in WesternEurope, directly and not on too long plots. Theyare good in terms of traffic safety in fog or snow,but under sunlight they give some alternation ofshadow bands, that tend to tire the drivers (themost dangerous frequency of alternatingillumination and shaded areas are 10-15 persecond, which corresponds with a speed of 80-100km/h and the distance between trees of 2-3 m). Inaddition, longitudinal plantations create a corridorand close views of the surrounding landscape andbuilding. Application of longitudinal planting inreconstruction and construction of new urbanroads is appropriate only for high slopes andembankments along the banks of rivers.

The practice of decorative landscapingorientation of roads provides a visual plantingguidance, which can be divided into 4 groups:

1) Direct planting (linear reception) –indicates a change in the movement, from farprompts the driver about the degree of aturn. They can only be linear, parallel to the axisof driving, beyond the roadbed. Their lengthdepends mainly on the radius of the turn, and theirline should visually cover the entire width oflanes, a look at the turns at their the approaches(Fig. 1);

Figure 1: Direct planting on a turn shown in aplanar view:

a) Turn with a small angle and a large radius;b) For small radius turn

2) Barrier planting - prompt the driver aboutthe inability to continue moving in its direction,

and at the same time is the visual "reflector"view that forces to move to the rightdirection. They are planted on the same principleas direct planting: they are needed mostly atintersections, bus stops, traffic intersection, butcan be used at venues and leisure complexes andmaintenance places (Fig. 2);

Figure 2: Examples of barrier planting onadjacent roads:

a) Opposite the driving direction b) at the end of atransitional speed zone;

c) At a convex vertical turn

3) Emphatic decorating planting – areintended to prevent driver distraction from themost important or potentially dangerous parts ofthe road (decoratives) or, conversely, to draw itsattention or its focus on points of importance or ofsafety, or for architectural reason of the road (e.g.,separation of pools). An example of newapproaches and planting can be a gate made by afractured convex longitudinal profile (Fig. 3);

Figure 3: Pairedplanting trees on a

convex fracture of a longitudinal profile:a) In an enclosed area, b) in open area

4) Safety (ordinary) planting - the protection ofdrivers from blinding headlights of oncomingcars; protection from side wind by ordinary trees,shrub planting arrangement that delay car that areleaving the road (Fig. 4);


163

Figure 4: Example of mixed mode of decorative landscaping of the road:1 – snow protection band, 2 – regular tree planting, 3 – landscape planting

To ensure visibility at intersections andhighways adjacent to a level, planting strips arearranged according to figure 5. Estimated distancevisibility road surface (La, LB) should beequivalent to the average speeds on the roads andtaken in table. 1, A width of the strip adjacent tothe road, provides side visibility (L ) should be25 m (from edge of carriageway for) roads of I-IIIrd categories and 15 m for roads of IV and Vcategories.

Table 1: Estimated distance for visibility ofroad’s surface (La, LB), m

Des

ign

spee

d,km

/ h

150

Estim

ated

dist

ance

for

visi

bilit

y, m

250120 175100 14080 10060 7550 60

40 5030 40

It is also important when selecting trees totake into account the nature of their psychologicalimpact on humans (Table 2).

Figure 5: Scheme of planting strips for visibilityat intersections of roads

Table 2: Classification of trees by psychological impact on humans

Form type sketch Scope of use1. Annoying Used for drawing the attention of drivers

1.1 Widely sprawling Create snow protection strips; design border of country,region, district; fulfill the conditions of the visual orientationat crossroads, at plans with turn and areas with limitedvisibility

1.2Narrowlysprawling

1.3 Conical Made for the visits of city, entrances of memorial space,creating dominant zones


164

1.4 Column form

2. Decelerating Used to create places of rest and quiet and slow driving rate2.1 Oval form planting for recreation, special areas, decorative plantings on

Historic streets

2.2 Spherical form

2.3 Umbrella form Planting for recreation, create forms on which are placedinformation signs

2.4 Tear-drop form

Conclusions

– We have shown the clear relationship betweenthe aesthetic qualities of highway and trafficsafety. To ensure the visual smoothness of theroad it was proposed to use decorativelandscaping orientating roads, namely: direct,decorative, emphatic and protective planting.– To ensure traffic safety at intersections andhighways adjacent to a level, planting strips areplaced according to the dependence between theestimated speed and estimated distance forvisibility of road surface.– When choosing trees, it is necessary to considerthe nature of their psychological impact onpeople. Thus, we distinguished class of annoyingand decelerating plantings. We should also keepin mind that the form of an object is perceiveswell only under the comparison with other similarobjects, that is why it is necessary to consider thenature of the combination of two volumes ofcrowns of different architectural structure.

References

[1] :

. , 2008 . / .. . – : , 2008 –

. 195-199.[2] [

]: / . 1960

. : http://landshaft-m.at.ua/publ/klassifikacija_reguljarnykh_form_drevesnykh_rastenij/1-1-0-12.

[3] :

18-84 – [01.01.1986]. – .: ,1984 . – 74 .


165

IMPLEMENTING A SPC INTEGRATED SYSTEM TO IMPROVEMANUFACTURING PROCESSES IN AUTOMOTIVE INDUSTRY

L., Lobon 1, C. V., Kifor2, C., Oprean3 and O., Suciu4

1Lucian Blaga University of Sibiu, [email protected] Blaga University of Sibiu, [email protected]

3 Lucian Blaga University of Sibiu, [email protected] Compa SA Sibiu, [email protected]

Abstract: Nowadays challenges emphasize greatly on the accomplishment of a high level of qualityin manufacturing processes, especially in automotive industry. Various methods have been used in orderto find ways to organize monitor and control the manufacturing processes. A method used widely toachieve these requirements is statistical process control (SPC). The objective of this study is to present asystem which integrates SPC with the results from 3D measuring systems in order to calibrate andmonitor manufacturing processes from an automotive industry organization. The implementation of theintegrated system has resulted in improvement of the manufacturing processes and, consequently, to abetter quality of the final products.

Keywords: measuring systems, 3D, SPC

1. Introduction

The challenges in nowadays industry are verystrong, the emphasis stressing the need foraccomplishing a high level of quality duringmanufacturing processes. Various methods havebeen used, especially to find ways to control themanufacturing processes. A method used widely isstatistical process control (SPC). Statistical processcontrol is recognized as a technique to achievecost-effective quality control through continuousmanufacturing process improvement [1]. StatisticalProcess Control (SPC) is a statistical basedapproach able to determine whether a process isstable or not by discriminating between thepresence of common cause variation andassignable cause variation. It is a well-establishedtechnique, which has shown to be effective inmanufacturing processes.[2]

It is impractical to inspect quality into aproduct; the product must be built right the firsttime. The manufacturing process must therefore bestable or repeatable and capable of operating withlittle variability around the target or nominaldimension. Online statistical process control is apowerful tool for achieving process stability andimproving capability through the reduction ofvariability.[3]

The statistical process control use differentdevices and instruments for initial acquisition ofdata. The evolution of SPC can be seen by thelarge number of specialised hardware andsoftware. In its earlier time SPC means only dataanalysis from manufacturing processes, today SPCmeans informatic systems with complex analysisfunctions, decisions and management. [4].

A widely used SPC system configuration isshown in figure 1.

SPC systems have been the central point ofinterest in various researches and have knowndifferent approaches according to the specificobjectives of each study. Following we presentsome of these approaches.

Systems that use previous collected data toimprove the quality of the manufacturingsystems. In one of the studies adopting thisapproach [5] the authors developed aframework that makes use of self-learningalgorithms that enable the manufacturingsystem to learn from previous data andresults in eliminating the errors andconsistently producing quality products.The framework relies on knowledgediscovery methods such as data miningencapsulated in a process analyzer toderive rules for corrective measures tocontrol the manufacturing process.

mailto:lucian.lobont:@ulbsibiu.ro

mailto:claudiu.kifor:@ulbsibiu.ro

mailto:rector:@ulbsibiu.ro

mailto:octavian.suciu:@compa.ro


166

Analysis of statistical instruments used inSPC, such as control charts with theobjective of identifying the real time of theprocess change. [6]. The study presented amethod based on the concepts of fuzzyclustering and statistical methods anddeveloped a novel hybrid approach whichis able to effectively estimate change-points in processes with either fixed orvariable sample size. The proposedapproach can be employed for processeswith either normal or non-normaldistributions. Also, it can estimate the truevalues of both in- and out-of-controlstates’ parameters.Decision support system used to makedecisions after analyzing statistical datafrom SPC [7]. There is presented anadvisory decision support system. Thissystem helps in collecting statistical dataand thereafter analyzes the enormousvolume of data and aids in making qualityrelated decisions. Unlike the conventionalSPC applications where the analyzedresults have to be interpreted by qualitycontrol specialists, Manufacturingexecution system (MES) based unmannedmanufacturing environments requireautomation of the interpretation process.The developed advisory system helps inselecting and designing control charts

based on various cost, rule or heuristicsmodels. The system also providesinterpretation expertise by configuring andapplying various rule sets. On violation ofthese rules, signals are generated by thesystem and the expert system advices forappropriate remedial actions. Thus thesystem acts as an advisory support system.Analysis of production system architecturewhich use SPC systems and techniques.The authors [8] presented a new analyticalmethod for evaluating the performance ofproduction systems in which statisticalprocess control (SPC) techniques areimplemented. Machines’ behaviour ismonitored by measuring qualitycharacteristics of the produced partsthrough off-line inspection devices andsampling inspections. The numericalresults show the good accuracy of theproposed method, provide new insight inthe relations among the two areas and pavethe way to the joint design of productionlogistics and quality control systems.

In concordance with the recent researchesregarding SPC used in manufacturing, theobjective of this study is to present a system whichintegrates SPC with the results from 3D measuringsystems in order to calibrate and monitormanufacturing processes from an automotiveindustry organization.

Figure 1: Generic SPC system.2. System development 2.1. Requirements


167

The development of the system started from aset of specific requirements:

- monitoring workstations in order to rapidlytake action when the process is out oflimits;

- validation of settings following the changeof part, tools or if the process is out ofcontrol limits;

- statistical sampling of products key featureswhich cannot be directly monitored at theworkstation because of the complexity ofthe measuring device;

- setting and monitoring of a large number ofworkstations that otherwise would have

needed a numerous and sophisticatedmeasurement devices;

- reduction of cost related to calibration andmaintenance of a large number ofmeasurement devices.

2.2. Procedures

The design and development of the system isbased on a series of quality procedures developedwithin the organization. Such procedures are: theprocedure for process statistical analysis (fig. 2),process monitoring, process validation, processcontrol through SPC, etc.

Figure 2: SPC procedure.

2.3. SPC conceptual framework system

Confirmedcapability

?

There are previousdata?

Control planI-QS-….

CPPM

SPC process plan

Quality engineer

SPC Team

SPC characteristics analysis

Previous dataTechnological documents

CPPM SPC Team

Acquiring process information

I-QS-……

SPC TeamSPC Team

Establishing sample size and sample collectionfrequency

SPC process planControl chart

Quality engineer

Control chart

Quality engineerOperator

Measuring and data collectionControl chart

Quality engineer

Control chart

Quality engineerSPC Team

Perform calculations and analysis resultsPlan process SPC

Quality engineer

Control chart

Quality engineerSPC Team

Variation causes analysis and their elimination Action plan

Quality engineer

YES

YES


168

The SPC system framework (fig. 3) wasdeveloped after analyzing the requirements and theprocedures. The system is composed of four mainelements: 1. the workstations – where theinformation comes from; 2. coordinate 3Dmeasuring system; 3. the monitoring room – where

the process parameters are assessed and visualized;4. SPC system which integrates the informationand transmits the results to available displays forvalidation and monitoring in the workstations areaand the monitoring room.

Figure 3: Conceptual structure of integrated SPC system.

2.4. System realization

The real system has two main components: thehardware – all the physical equipments that formthe SPC system and the software – whichcomprises the SPC analysis and monitoringapplication and ensures the information transfer

within the system. A detailed presentation of thesecomponents follows:

The hardwareCoordinate 3D measuring system (fig. 4). It is usedfor complex and precise measurements of thecollected samples from the workstations.

Figure 4: 3D coordinate measuring system.Computers, displays, transferring data systems etc.(fig. 5). All these components are used for datacollection, analysis, transfer and visualisation.

They assure the correct flow of information and arepositioned near the workstations, the measuringsystems and the monitoring room.

information flow

information flow

information flowSPC SYSTEM

hardware+software

3D coordinatemeasuring system

workstation 1 workstation 2 workstation 3

display display display

Global view displayMonitoring room


169

Figure 5: Components of the SPC system.

The software1. The SPC analysis and monitoring application

is developed and implemented within theorganization allowing:

a) for SPC system:-data collection from measuring devices;-data processing and calculation;-generation of specific SPC graphical and

numerical reports (fig. 6);

-process capability study;-export of reports data;

b) for monitoring (fig. 7):-support for generating organization layout;-establishing the connections between

system elements;-real time graphical representation of

problems encountered in fabricationprocess.

Figure 6: SPC graphical report.


170

Figure 7: SPC monitoring layout.

3. Results following the system’simplementation

The system was implemented in an automotiveindustry organization and has been functional fortwo years. The results obtained in this period oftime are reflected in several areas:

-the reduction of time with 5%-10% forpreliminary and ongoing settingsrequired for the manufacturingprocesses;

-ppm reduction with 5%-20%;-other costs reduction 1,5%-3% (example:

costs related to calibration andmaintenance of a large number ofmeasurement devices);

-raising the awareness of the workers aboutthe functioning of the entiremanufacturing system and their role asa part of it;

4. Conclusions

This study has presented a system whichintegrates SPC with the results from 3D measuringsystems in order to calibrate and monitormanufacturing processes from an automotiveindustry organization. The development of thesystem had followed certain steps. First there were

identified the specific requirements in theorganization where the system was created. Then,the design and development of the system followeda series of quality procedures developed within theorganization. The next step was to conceptualize aframework for integrating the main components ofthe system. Based on this framework the physicalsystem was developed combining hardware andsoftware elements. All the software componentswere realized within the organization so as to fitbest for the purpose intended.

It’s important to notice that, although thesystem was created for an organization fromautomotive industry, it can also be adjusted withminimum changes for manufacturing processesfrom other industries.

5. References1. Zhu, Y. D., Wong, Y. S. and Lee, K. S.

Framework of a computer-aided short-run SPCplanning system, The International Journal ofAdvanced Manufacturing Technology,Springer, London, 2006.

2. Baldassarre, T., Boffoli, N., Caivano, D. andVisaggio, G., Managing Software ProcessImprovement (SPI) through Statistical Process


171

Control (SPC), Springer, Berlin Heidelberg,2004.

3. Montgomery, D. C., and Runger, G. C., AppliedStatistics and Probability for Engineers - ThirdEdition, John Wiley & Sons, Inc., ArizonaState, 2003.

4. Oprean, C., Lobon , L. and Kifor, C. V.,Integrated statistical process control system formanufacturing processes, Academic Journal ofManufacturing Engineering, Vol. 5, No. 2,2007.

5. Kumara, S., Nassehia, A., Newmana, S. T.,Process control in CNC manufacturing fordiscrete components: A STEP-NC compliantframework, Robotics and Computer-IntegratedManufacturing 23, pp. 667–676, 2007.

6. Alaeddini, A., Ghazanfari, M., Nayeri, M. A., Ahybrid fuzzy-statistical clustering approach forestimating the time of changes in fixed andvariable sampling control charts, InformationSciences 179, pp. 1769–1784, 2009.

7. Chakraborty, S., and Tah, D., Real timestatistical process advisor for effective qualitycontrol, Decision Support Systems 42, pp. 700–711, 2006.

8. Colledani, M., and Tolio, T., Performanceevaluation of production systems monitored bystatistical process control and off-lineinspections, Int. J.Production Economics 120,pp. 348–367, 2009.


172


173

A NEW APPROACH OF THEMAIN LANDING GEAR EQUATIONS

Daniel BOSNICEANU1

1 Military Technical Academy, Bucharest, Romania [email protected]

Abstract: In this paper an approach for modeling landing gear systems is presented. Specifically, anonlinear model of an main landing gear is developed. This model includes nonlinear effects such as apolytropic gas law, velocity squared damping, a geometry governed model for the discharge coefficients,stick-slip friction effects and a nonlinear tire spring and damping model. An initial model was developed thatonly included the air-spring above the fluid, fluid dynamics through a fixed orifice, and a linear tire springterm.

Key words: main landing gear, hydraulic spring, shock damper, linear model of equations

1. Initial landing gear investigation

This chapter is intended to familiarize the readerwith landing gear terminology and to demonstratea mathematical development of the equations ofmotion for a telescoping landing gear. Figure 2-1 isintended to acquaint the reader with basic landinggear components. It shows the simplifiedcomponents of a telescoping, main landing gear(as opposed to a nose gear).

Point 1 on the figure is a rigid bodyrepresentation of the aircraft fuselage. Point 2 is achamber containing compressed nitrogen whichserves as a spring that carries the weight of theplane in ground operations. Point 3 refers to themain, upper cylinder which houses thecompressed gas, hydraulic fluid, and within whichthe piston slides. Point 5 is the orifice plate. It isessentially a circular plate with a hole in the centerthrough which the hydraulic fluid flows when thestrut is stroking. It, along with the metering pin,point 6, controls the damping characteristics ofthe gear. Point 7 locates one of many rebound orsnubber orifices. These holes lead into a smallvolume on the backside of the piston head (point8) called the rebound or snubber chamber. Thepurpose of the snubber is to provide dampingwhen the strut extends. The Point 9 is the piston.

Figure 1. Schematic of typical telescoping mainlanding gear studied

It houses the metering pin and is also the rigidconnection of the wheel axle. Finally, point 10 isthe tire. This element of the gear adds both springand damping characteristics to the overallperformance of the gear, and is selected carefullyfor various applications.

mailto:bosniceanud:@yahoo.com


174

2. Nonlinear model development

This research discusses an independentdevelopment of a mathematical model of a mainlanding gear with all the relevant physicalparameters included. The nonlinear equations ofmotion are developed for a telescoping main gear.

An initial model was developed that onlyincluded the air-spring above the fluid, fluiddynamics through a fixed orifice, and a linear tirespring term. This simple model allowed sometrend comparison between the results of this modeland the early results of the linearized gear. Ametering pin was then added to change the mainorifice effective diameter as a function of stroke.Another variation was the addition of a snubber,or rebound chamber. This feature providesdamping while the gear is extending. The modelincludes constant seal friction as well as a variablefriction that is a function of stroke. In a furthereffort to be realistic, a nonlinear tire model wasadded. This tire model has a spring rate that is afunction of tire deflection and dampingproportional to compression rate. In the equationsdeveloped below, the spring and dampingcoefficient are used as if they were constant. Thenonlinear characteristics of each of these terms isincluded in the equations of motion that areactually integrated.

Figure 2 is a schematic of the gear used in thedevelopment of the equations of motion. Thisschematic is representative of a generaltelescoping-type main landing gear. It includes theaerodynamic lift on the plane, Lift, the upper mass(of the plane's fuselage) and the mass of the maincylinder lumped together as a rigid mass, M,, andthe mass of the piston and the mass of the tire, alsolumped together as ML. The inertial coordinate ofthe upper mass is Xwg. The zero value for Xwg is when thegear is fully extended with the tire just touching theground. From this same gear configuration, Xa, thecoordinate of the lower mass, is taken as zero at theaxle of the tire. Therefore, when the gear is in somecompressed state, Xa measures the deflection of the tirewhen the ground input, U(t), is zero.

In the compressed nitrogen chamber (uppercylinder) with cross sectional area of A u the pressure isPu. Likewise, in the lower chamber with cross sectionalarea of AL, there is a pressure of PL. In the snubberchamber, with annulus area of AR, the pressure isdefined to be Ps. The orifice plate has a hole of

diameter D op through which the metering pin, withvariable diameter Dpin moves. Fluid reaches the snubberchamber through the orifices ds

c and dsE, where the

superscripts represent either the compression mode orextension mode respectively. The diameter of the piston,Dpi, is used to calculate AR. Simply subtract the area ofthe piston shaft from that of the lower cylinder to getAR. The tire is also shown in Figure 2 with a distinctionof pointing out that the tire spring and dampingcoefficients, K t and C t are nonlinear and contribute tothe calculation of the tire force Ft.

Figure 3 shows the forces acting on the uppermass. Balancing the forces on the upper massgives the following equation:

fAPAAPAPLgMXM RSOLLUUUwgU )(..

(01)

The term on the left hand side of Eq. (01) is theinertial motion term, g is the gravitationalacceleration, f is the friction present in the gear,and all other terms are as described previously.This equation assumes that the fluid pressure in theupper cylinder is identical to the pneumaticpressure.

Figure 2 - Schematic of telescopinc landing gear


175

Figure 3 - Schematic of upper mass and main cylinder

In this area, reflects the fact that the meteringpin is included, i.e. it is a variable cross-sectional

area depending on stroke.

Figure 4 - Schematic of lower mass

Figure 4 shows the forces acting on the piston.Summing the forces on the lower mass (piston)the force balance equation is:

fFAAPAAPgMXM tSRSSLLLaL )()(..

(02)

Where the left hand side of Eq. (02) is theinertial motion of the lower mass and As is the areaof the snubber orifice. Ft is the force that istransmitted through the tire from the ground and

has the form:

)()(..

UXCUXKF Utatt (03)

where the tire force is a function of a nonlinear tirestiffness and a damping force that is composed of adamping coefficient that is proportional to the tirestiffness and the time rate of change of the tiredeflection.

3. Relation of pressures to stroke position andstroke rate

The pressure terms in Eqs. (01) and (02)are as yet unknown and need to be related to thepositional variables Xwg and Xa or their derivatives.The pressure of the compressed nitrogen in theupper cylinder can be described by the polytropicgas law for a closed system as:

(04)

where Xs is the stroke available, given by:

aWgS XXX (05)

with X SI as some initial length, P SI , thecharge pressure at X maxS , and , the polytropicgas constant. X S max is the maximum value towhich the gear can be extended. This form ofrepresentation of the pressure change is assumedto happen as a quasi-equilibrium process. Thesignificance of the polytropic gas constant is that itdescribes the type of process that occurs. Anaverage value is usually sufficient in application

Equation (04) was defined in such a mannerthat Pu will become very large when Xs is nearX S max, i.e. the gear is nearly completelycollapsed. This is a suitable representation of theprocess, with only the polytropic gas constant yas an unknown.

The pressures (PL and Ps) of the fluid in thelower cylinder and in the snubber are related to theflow rates of the fluid into and out of thoseregions. The volumetric flow rates through theorifice plate hole, Qc, and the snubber orifices, Qs,can be determined by combining the continuity

)(max SS

SISIU XX

XPP


176

equation and Bernoulli's equation for fluids. Flowis always from the higher pressure to the lowerpressure. Bernoulli's equation for anincompressible fluid states that along astreamline,

(06)

where P is the pressure at some point, g is thegravitational acceleration, V is the velocity of theflow, v is the specific weight of the fluid which isequal to the fluid density (p) multiplied by thegravitational acceleration (g), and Z is the heightdifference from some zero reference. Thisequation assumes that the viscous effects withinthe fluid are negligible, the flow to be steady andincompressible, and that the equation is applicablealong a streamline.Equating Bernoulli's equation(Eq. (06)) at two points in the flow along the samestreamline yields:

(07)

In the case of a landing gear, the potentialdistance between Z 1 and Z 2 can be neglected asthe distances involved are very small compared tothe other terms. Equation (07) with the continuityequation for incompressible fluids which states Q =A 1 V1 = A2V2 allows for the solution of thisequation in terms of one of the velocities.Assuming that P 1 > P2, i.e. the flow is from P 1 toP2, then solve for V 1 from the continuity equationas:

(08)

When the flow reverses, i.e. P 1 < P2, then thevelocity at point 2 is described by the aboveequation with the pressure terms switched and anegative sign on the square root. The idealvolumetric flowrate (Qideal) for an incompressiblefluid can be expressed as Qideal = A*V.Now we have:

VACCdQQ didealreal (09)

Substituting Eq. (08) into Eq. (09) for velocity:

21214

2

1 ))(1(

2 PPPP

DD

ACQ dreal (10)

For our landing gear, there are two flows thatare of concern, the flow through the orifice plateand the flow into and out of the snubber chamber.Define QsC as the flow rate into the snubberchamber in the compression mode, where thesnubber orifice area (As) becomes As

c, whichallows larger flow. The flow rate through thesnubber orifice during the extension mode isdefined as QS

E, and the area As becomes ASE, which

only allows small, restricted flow. In both cases,the flow through the main orifice plate is Q O .

Figure 5 - Control volume between piston and orifice

plate

Figure 5 shows the direction of fluid flow into andout of a control volume in the lower chamber as afunction of stroke mode (extension orcompression). In relating the flow rates to thepressures, defining a control volume as shown bythe dashed line in Figure 5 is necessary. Thestroke rate is defined as

aWgS XXX...

(11)where the compression mode is given by Xs >

0.0, and the extension mode by Xs < 0.0. The flowis assumed to be negative leaving the controlvolume, and is positive entering it. For anincompressible fluid, the volumetric flow rates forcompression and extension can be written as:

)1(

)(2

41

42

21222

1

22

1

DDPPVV

DDV

.)21( 2 constZVg

P

22

22

12

11 )

21()

21( ZV

gPZV

gP


177

0.0.

SLCSa XAQQ (12)

during the compression mode and

0.0.

SLESa XAQQ (13)

during the extension mode. Equation (10) definedthe general form of the equation for a flow rateandcan be written as:

ULUL

L

odoo PPPP

Dd

CAQ))(1(

24

(14)

where d0 is the effective diameter of the mainorifice, DL is the diameter of the lower chamber,and Cd is the discharge coefficient of the mainorifice. The flow through the snubber orificesduring this mode is described by:

with dsc as the diameter of a snubber orifice, DL as

described above, C dSC is the discharge coefficient

of the snubber orifice and Asc is the effective area

of the snubber orifice. Similarly,for the extension mode , where flow is into the

control volume (P L PU and P S ) .

LULU

L

odoo PPPP

Dd

CAQ))(1(

24

(16)

where the difference between this equation and Eq.

(14) is that the pressure terms have exchangedpositions and the whole term is now positive. Theflow rate through the snubber orifices during theextension mode is given by

where DR is the effective diameter of the annulussnubber chamber, ds

E is the diameter of a snubberorifice, AS

E is the effective area of the snubberorifices and C dS

C is the discharge coefficient ofthe snubber orifices in the extension mode. Tosimplify Eqs. (14), (15), (16), and (17), let thenon-pressure terms be redefined as:

))(1(

24

1

L

odO

DdCAE , 13 EE

))(1(

24

1

L

odO

DdCAE ,

))(1(

2

44

R

EE

dSES

Dd

CAEE

, respectively.

Substituting Eqs. (14) and (15) into Eq. (12) andEqs. (14) and (15) into Eq. (13) using this newnotation, rewrite Eqs. (12) and (13) as

0.0.

21 SLSLUL XAPPEPPE for

SX.

> 0 (12.a)

0.0.

43 SLLSLu XAPPEPPE for

SX.

< 0 (13.a)

Figure 6 - Control volume for the snubber chamber

)15())(1(

2

4SLSL

L

Cs

CdS

CS

CS PPPP

Dd

CAQ

)17())(1(

2

4LSLS

R

Es

CdS

CS

ES PPPP

Dd

CAQ


178

Additional information about the flow rate-pressure relationship can be gained by studying acontrol volume in the snubber chamber as shownby the dashed line in Fig.6 .

The variables AR and DR in Fig.6 are therebound chamber annulus area and effectivediameter respectively. Ps is the pressure in therebound chamber and ds

c and dsE are the diameters

of the snubber orifices in the compression modeand extension mode respectively. In the case ofcompression, where Xs > 0.0 and PL > Ps,

0.0.

SRCS XAQ (18)

Substituting the flow rate CSQ of Eq. (15) into Eq.

(18) yields: (19)From previous notation of E i this expression

becomes: 0.0.

2 SRSL XAPPE (20)Rearrange Eq. (20) to get an expression for thepressures in terms of the stroke rate as:

SR

SL XEAPP

.

2

(21)

2.2

1

)( SRL

UL XE

AAPP (22)

where Pu is given in Eq. (04). Square both sides ofEq. (21) and solve for Ps as:

P S = P L - (2E

AR ) 22.

SX (23)

Similarly, for the extension case with Xs < 0.0:

2.2

3

)( SRL

UL XE

AAPP (24)

P S = P L - (4E

AR ) 22.

SX (25)

These known pressures [Eqs. (04), (22), (23),

(24), (25)] can now be substituted into Eqs. (23)

and (24). Algebraic simplification of these

equations leads to the compression and extension

cases in terms of readily measurable quantities

as:(01a)(02a)(01b)

(02b)

Introduce a new notation using subscripts tosimplify the above equations: "1" and "2" will beassociated with compression (equation set (a)),and "3" and "4" with extension (set (b)). With thischange, the equations can be writtenin the form:

(01c)(02c)

where the coefficients of the stroke rate squaredterm are assigned the C /

i s, and the coefficients ofthe stroke position term are the K i 's.

fXAAE

AAA

EA

EAA

XX

PAALgMXM

SLRL

RRRL

S

SISILRUwgU

2..

02

1

2

2

2

1

..

)()(])()[(

)()(

fFX

AAE

AAAA

EAA

EA

XX

PAAgMXM

tS

CSL

RLCSR

RLR

S

SISIRLLaL

2.

2

1

2

1

2

2

..

)()()]()()[(

)()(

fXAEA

EAA

AAE

AA

XX

PAALgMXM

SRRRL

LRL

S

SISILRUwgU

2.2

4

2

3

.

02

3

..

])()[()()(

)()(

fXKXCLgMXM S3/1

2

S.

3/1Uwg..

U

fFXKXCgMXM tSSLaL 4/2

2.

4/2

..

0.0))(1(

2 .

4SRSL

L

Cs

CdS

CS XAPP

Dd

CA

fFX

AAE

AAAA

EA

EAA

XX

PAAgMXM

tS

ESL

RLESR

RRL

S

SISIRLLaL

2.

2

3

2

4

2

3

..

)()()]()()[(

)()(


179

The only unknown term left in these equationsis friction. As mentioned previously, friction inthis gear comes mainly from two sources, frictiondue to tightness of the seal and friction due to theoffset wheel (moment). The seal friction isassumed to be a maximum value statically andsome function of velocity in the dynamic state.The functional relationship between frictional forcelevel and velocity could be determined throughtesting. The friction due to the offset wheel is theresult of the moment produced by the nonaxiallyloaded piston within the cylinder.

It can be seen from Fig.7 that the force betweenthe piston head and the cylinder, N, is a result ofthe tire force, Ft, applied at moment arm, ma, fromthe centerline of the piston. The frictional forcedue to the offset wheel (F ow is assumed to be ofthe form (refer to Fig.7):

Figure 7 - Schematic of gear for friction model

development

NFOW (26)

Where N is the normal force of the cylinderwall resisting the side of the piston head, and isthe coefficient of friction between the two parts.To find the unknown force N, sum the momentsabout point O to zero to get:

0)(:0 stpXNmaFM St (27)

Where stp is the minimum distance betweenthe piston head and the lower seal when the gear isfully extended. Rearrange Eq. (27) by isolating N,and then substitute N into Eq. (26) to get anexplicit form of Fow:

(28)

(29)

The total friction in the landing gear, f, inequations (01c) and (02c) is now assumed to be:

This paper assumes that a proportionate part ofthe fuselage (half of the 80% of the total weightthat rests upon the main gear) is treated as a lumpmass centered at the centerline of the main uppercylinder. Also, this model takes into account onlyvertical loads on the strut. The tire is modeled asa nonlinear spring and damper. This tire modeldoes not take into account spinning stiffness(because the test tire does not spin) or spin-up drag.The fluid is assumed to be incompressible and allstructural members are assumed to be rigid, witheach having only a vertical degree of freedom.These assumptions are good only for straight-linetaxiing over runway profiles and landing impact(spin-up drag on the tire does not significantlyeffect the vertical loads on the strut). Any brakingor turning maneuvers are not covered in thedevelopment. The equations developed here arethe basis for a "rollout" simulation.

4. Conclusions

In this, the nonlinear equations of motion weredeveloped for a general, telescoping main landinggear.

These equations contain a pneumatic springthat is determined based on the polytropic gascompression law, a hydraulic damping that isproportional to the stroke rate squared,gravitational forces, lift, inputs from a runway, andfinally friction, which is composed of both aconstant seal friction and a variable bearingfriction. These equations explicitly contain theempirical parameters of polytropic gas constant,discharge coefficients for both the main orifice andthe snubber orifices, and the friction levels in thegear. These parameters are the only variables thatappear in equations (01) and (02) that cannot bedirectly measured.

Equations (01) and (02) are highly nonlinearand are discontinuous due to the differing valuesof friction and discharge coefficient as a functionof extension and compression. Future work willdiscuss more about the nature of these equationsand present a method of solving these equationsfor gear displacements and velocities.

stpXXFmaN

awg

t

)(stpXX

FmaFawg

tt


180

References

[1]Greenbank,S.J., Landing Gear-The AircraftRequirement, Proceedings of Institute ofMechanical Engineers, Vol. 205, 1991;

[2] Norman, S. - Aircraft Landing Gear Design:Principles and Practices, LockheedAeronautical Systems Company AAIAEducation series;

[3] O'Massey, R. C., Introduction to LandingGear Design, ASM Paper No. W70-18.1,March 1970;

[4]Rashant, K. Simulation of Landing GearDynamics and Brake-Gear Interaction,Germany, 2008

[5] Sonny,T.,Mason,H., Landing Gear Integrationin Aircraft Conceptual Design,Multidisciplinary Analysis and Design Centerfor Advanced Vehicles, Virginia PolytechnicInstitute and State University Blacksburg,Virginia,September 1996;


181

RESEARCH ON HOLLOW CATHODE EFFECT AND EDGE EFFECTAVOIDANCE IN PLASMA NITRIDING TREATMENT

AXINTE Mihai1, NEJNERU Carmen1, PERJU Manuela Cristina1, CIMPOE U Nicanor1,HOPULELE Ion1

1Technical University “Gheorghe Asachi” of Ia i-România, E-mail address:[email protected]

Abstract: This paper contains a description of the nitriding installation with unpolarized grid in order toavoid hollow cathode effect and edge effect. For the experimental study we have made wedge samples cut, andtriangular prism-shaped samples. We also made a theoretical study on these unwanted phenomena.

Keywords: hollow cathode effect, edge effect, grid

1. Introduction

Plasma technologies have been longstudied in surface engineering because theyprovide are important regarding technical andenvironmental bene ts over salt bath or gastreatments. Among them, the active screentechnology for plasma surface engineering offersmultiple advantages over conventional directcurrent plasma treatments, like: improved surfacequality and more uniform material properties [1].

In general, for the plasma nitridingprocess, the components to be treated are subjectedto a high cathode potential and the grounded wallwhile the furnace wall forms the anode. The partsare directly involved in the discharge process [2].Conventional direct discharge plasma treatmentsapply considerable cathodic potentials to thecomponents in order to reach the treatmenttemperature by means of bombardment withenergetic ions [1].The electric voltage appliedbetween anode and cathode during plasmanitriding is 400–700 V [2]. This makes thedischarge processes sensitive to non-conductivesubstances, and all parts must be thoroughlycleaned to avoid severe arcing during thetreatment.

The positive ions generated by glowdischarge are accelerated in the cathode fall regionnear the cathode surface and bombard the surfaceof the specimen. The ion bombardment causessputtering, transfers kinetic energy to thecomponent and raises its temperature. Manystudies have indicated that the geometry, size andratio of the surface to mass of the component have

substantial effects on the temperature distributionwithin the component and lead to inhomogeneousresponse to nitriding.[2]

Plasma characteristic, correlated with thespecific geometry and shape are responsible for anumber of problems:

1) Inhomogeneous batch heats notuniformly. Problems arise when the treated batchhas a significantly varying surface mass ratio. Ifthis ratio differs much from the same characteristicof a standard specimen with thermocouple, someof the parts in the batch can be overheated orunheated [3]

2) The hollow cathode effect is a specialsituation for the glow discharge between twoclosely separated cathode surfaces. This effectoccurs when the dimension of the cathode fallregion becomes as large as the separation distance.The loss of electrons is low due to the specialgeometry and because they are repelled by thenegative walls of the cathode, and in fact, theyoscillate between the sample and the wall (figure1). The plasma density (that is the electronconcentration) increases and reaches values as highas 1012 cm-3[29]. As a consequence, theproduction of ions rises too, and the ion fluxdensity on the substrate surface increases. Due tothese factors, such surfaces can be heated toextremely high temperatures even at relatively lowbulk substrate temperatures of 200–500 8C and thelocal heating can produce melting surface, whichmeans destroying the part [4]. The effect manifestsitself at a specific composition of the gas, which is

mailto:mihai.axinte:@gmail.com


182

characterized by pressure p[Torr], and specificdistance D [mm] between the face to face surfaces.

Figure. 1 Highlighting the hollow cathode effect. Crosssection through a narrow channel V shape

3. Edge effect is the growth rate of thelayer different on the edges and flat regions of thesurface of a part. The elementary volume on theedges receives more energy than the elementaryvolume on the flat surface in the same time period[3] (Fig. 2). Due to distortions of the electric fieldaround the corners and edges, the shape of plasmasheath, which is connected to the shape of samples,determines the ion flux distribution, which, in turn,affects the uniformity, hardness and surface phasesof coating, erosion rings occur, characterized bythe reduction of hardness [5].

Figure 2. Quantitative energy difference between flat,edge and corner areas, at edge effect.

To avoid this disadvantage anexperimental installation was build. Thisinstallation is equipped with an polarized screen. Insuch a case a glow discharge is largely“transferred” from the surface of the parts to thescreen surface, and the ion bombardment of the

parts becomes less intense [3]. This grid is made ofstainless steel, and has the role to modify theelectrical discharge field between anode andcathode, on the ionic triode principle. According to[6], the use of active screen virtually does notreduce the growth rate of the layer. Such a systemis a triode.

The triode ionic screen interposed betweencathode and anode is polarized, so it can changethe electrical field configuration. Modifying theelectrical field between anode and cathode,changes the discharge conditions and also thehollow cathode effect occurrence, by adjusting thegrid electric potential [3].

Constructive principle of the installation ispresented in figure 3.

Figure 3. Constructive principle for installation,Electrical block diagram

In the figure 4 is presented the detailedelectrical diagram. In ionic triode case theinterposed grid between anode and cathode iselectrically polarized, so it changes theconfiguration of the anode and cathode electricfield.

Changing the electric field between anodeand cathode changes the degeneration conditions


183

for arc discharge, and double-cathode formation byadjusting the electrical potential of the grid.

Figure 4 Detailed electrical diagram

The shape for the grid, or the active screenvaries depending the dimensions and the partconfiguration. It can have the following shapes:planar grid, circular grid, cage grid. Materials forthe screen: stainless steel, copper, steel sheet withdifferent diameter holes in it [7].

The installation function with active screenrepresents the grid activation in a certain stage forplasma nitriding, treatment technology

After cleaning the vacuum chamber andcreating the 10-1 vacuum value, the anodic currentis connected to the installation. Then the dischargecurrent is adjusted and proceeds to reducingvacuum level at the specific value by introducing

gas. Anode current is adjusted to the desired value.Due to the discharge changing conditions the gridpolarization is modified, conform to the necessarycircumstances.

When the temperature is near theprescribed one automatically the anodic dischargeelectrical current is reduced.

a) b)Figure 5. Samples different shapes: a) holes differentdiameter, b) prismatic shape with sharp edges.

In figure 6 is presented the plasmanitriding process and the of hollow cathode effectoccurrence due to the configuration of the sample.The sample has a V shape. This shape allows us tohighlight the hollow effect appearance conditions.

Figure 6. V shape cut sample dimensions

The sample has a cylindrical shape. Theheight is 50mm, exterior diameter is 10 mm,interior hole diameter is 8 mm, and the V shape cuthas the height 25 mm, figure 6. As it is graphically


184

explained in fig. 1 at the cathode fall intersectionthe electrons are caught between the two surfacesand the ionisation increases very much, also thetemperature. In figure 7 is presented the hollowcathode effect during functioning.

Figure 7 Hollow cathode effect on a V shape sampleduring plasma nitriding process, with no active screenbetween anode and cathode.

2. Conclusions1. The installation will be used to avoid classicalplasma nitriding defects by triode effect from thepolarized grid.2. The installation setings permits us to highlightthe cathode effect apearance conditions.3. Following researches on avoiding thedisadvantages attached to general plasma nitriding:edge effect, hollow cathode effect, arc discharge,not uniformly deposition layers.4. The ionic triode reduce the sputtering effect ,positively influencing the surface quality,decreasing the roghness value.

References:[1]S. Corujeira Gallo, H. Dong, Study of active

screen plasma processing conditions forcarburising and nitriding austenitic stainless steel,Surface & Coatings Technology 203 (2009) 3669–3675

[2]Y. Li et al., Plasma nitriding of 42CrMo lowalloy steels at anodic or cathodic potentials,Surface & Coatings Technology 204 (2010) 2337–2342

[3] S. Janosi et al., Controlled hollow cathodeeffect: new possibilities for heating low-pressurefurnaces, Metal Science and Heat Treatment, Vol.46, Nos.7–8 (2004) 310-316

[4] C. Alves Jr. et al., Nitriding of titaniumdisks and industrial dental implants using hollowcathode discharge, Surface & Coatings Technology194 (2005) 196–202

[5] C. Alves Jr. et al., Use of cathodic cage inplasma nitriding, Surface & Coatings Technology201 (2006) 2450–2454

[6]. C. X. Li et al , Active screen plasmanitriding of austenitic stainless steel, Proc. 4thEuropean Stainless Steel Science and MarketCongress, Paris, France, Vol. 2 (2002), pp. 297 –303.

[7]. M. Axinte et. al., Facility for study heatingand diffusion process, using a ionic triode in aplasma nitriding installation, Tehnomus XV,Suceava (2009), ISSN 1224-029X

[8] G. Vermesan, V. Deac, Bazele tehnologiceale nitrur rii ionice, Editura Universit ii din Sibiu,Sibiu, 1992


185

AN ALGORITHM FOR SOLVING QUALITY PROBLEMS,USING QUALITY TOOLS

Marius B U1, Mircea CIOBANU2

1 Stefan cel Mare University of Suceava, Romania, [email protected] Stefan cel Mare University of Suceava, Romania, [email protected]

Abstract: In this work is has been developed an algorithm, to address shortcomings incommunicating with customers and lack of coordination. As a support for this algorithm, there wascreated an Excel program which combines two of the tools proposed to solve the problem, respectivelyIshikawa diagram and Pareto chart.

Keywords: Quality tools, Ishikawa, Pareto

1. Introduction.

The essence of solving quality problems isestablishing, elaborating, identifying,implementing and monitoring the corrective /preventive actions for quality improvement. Inthe current standards (e.g. EN ISO 9004: 2010and other specialty papers [1], [11], [13], [14],[15]) there are presented various models andmethodologies to solve quality problems, toimprove quality, respectively to establish,develop, identify, implement and monitorcorrective / preventive actions. They vary innumber, form, sequence, level of detail andcontent of stages, phases, tasks, steps, specificsequences, and by grouping them in stages. Allof them are using a mixture of statistic andnon-statistical quality tools, known also as TheSeven Basic Tools of Quality and respectivelyas The Seven Management and PlanningTools.

In table 1 there is a briefing of these tools.There are presented the objectives of each oneof these tolls, and some hints when to usethem.

Table 1: The quality tools

2. The methodology of elaborating algorithmsin using the tools to solve quality matters.

Based on the theme or problem to be solved (inplanning, assurance, and quality control), thedevelopment of algorithms (strategies) for the useof methodological instruments in solving thequality problems (planning, assurance, control andquality improvement) suppose the browsing of aproposed synthetic methodology, depicted inFigure 1.

mailto:marius.baesu:@fim.usv.ro

mailto:mircea.ciobanu:@fim.usv.ro


186

Figure 1: The methodology of elaborating algorithms inusing the tools to solve quality matters.

The methodologies specific to the theme orproblem to be solved, could be graphicallyrepresented in the following forms:

-ordered list of stages, phases, tasks,sequences and methodological tools;

-organization chart, flow chart, blockdiagram with rectangles and words;

-organization chart, flow chart, blockdiagram, with words and symbols or drawingsof methodological tools;

-organization chart, flow chart, blockdiagram, with symbols or drawings ofmethodological tools;

-combinations of the previous forms.To indicate logical relationships between

the stages, phases, activities, sequencingmethodology, respectively the methodologicaltools used to address each stage, arrows shouldbe used.

The organizational charts, logic diagramsand block diagrams, as those presented below(figure 2), could be considered as a reference,as an example of a specific methodology tosolve a quality problem [4], [17].

Figure 2: Example of a specific methodology tosolve quality problems

3. Case studyFor the case study considered in this paper, data

regarding customers’ complaints about a servicedepartment were collected. The data were collectedduring September 2008 - March 2009.

The proposed methodology for identifying,detailing and improving the causes of thosecomplaints is as follows:

- collecting data (tables);- sorting data in descending order, according to

Pareto chart;- identifying the causes that led to an increased

number of complaints, with negative effects on thecompany image (using Ishikawa diagram)

- implementing the corrective and preventiveactions, resulting from brainstorming activities;

- collecting a new set of data, in the sameperiod of time as that when the initial data werecollected, the difference between being a calendaryear;

- building a new Pareto chart.- comparing data and drawing the conclusions.The algorithm is shown in Figure 3.


187

Figure 3: The algorithm proposed

As said before, for the case study considered inthis paper, data regarding customers’ complaintsabout a service department were collected. Thedata were collected during September 2008 -March 2009.

It is worth mentioning that during the datacollection period, the register of complaints waskept by "the workers on the shelf" and the vendors.

As noted, most complaints have occurredduring December-January, when the workload forcommercial workers is significantly higher(inventories, closing the year, etc..,)

The main problems, as is about to see, arecaused by the fact that there was not, at that time, acentral dispatch for service problems, lack ofcompetent people to solve problems or even toprovide some relevant solutions. By competentpeople means specialists in this field, because, asalready indicated, both the role of "dispatcher" andthe "primary dealer" was performed by commercialworkers, most of them mainly with basicknowledge of sales and technical matters. Inaddition, some of the claims submitted by serviceaimed at a new batch of heating products, just puton the market (it is late summer, early autumn2008), which, as noted later, had not undergoneenough testing.

3.1 Data collection

Data were taken from the book of suggestionsand complaints and tabled.

Table 2: Data collected between September 2008-March 2009

According to the methodology, the next stepbeing the establishment of a Pareto chart, in orderto identify the most common problems, the dataare arranged in descending order.

In Table 3 these data are arranged downward.

Table 3: Arranged data

According to the typology of drawing a Paretochart, the cumulative frequency of the recordeddata must be calculated.

On the Internet there are a lot ofExcel/Spreadsheet applications, freeware orshareware, that help in preparation of suchdiagrams. The only problem is that there is not anapplication that includes the same excelspreadsheet and a chart of data descendingarranged, and the cumulative frequency table andthe resulting Pareto diagram.

Figures 4 and Table 4 summarize the datacollected and the cumulative frequencies and thePareto diagram that shows further processing.


188

Table 4: The data frequencies calculation

Figure 4: The Pareto Chart

Figure 5 shows the excel application developedby the authors of this paper to create Pareto charts,application used by students in the QualityManagement applications.

Figure 5: Excel application

When finished, the chart reveals that themost frequent complaints are of the type: time

problem solving, information provided,treatment, working program.

So, looking at the graph, it appears that wefirst must operate on the four cases presentedin the previous paragraph.

Therefore, after a brainstorming session, itwas compiled a Ishikawa diagram, where wetried to find the factors (the causes)influencing the increasing number ofcomplaints over the service department.

Figure 6: The resulted Ishikawa diagram

Therefore, following the chart analysis anddiscussions it had been established a set ofmeasures:

• the founding of a central dispatch service,which:

- to receive complaints directly to a toll freetelephone number,

- to allocate a time resolution of each complaintand at the end of this period to contact thecomplainant and find out the steps taken and thestate of play in the case of larger problems;

• providing in each subsidiary stocks of spareparts, and in case of unique or the imported ones,ensuring the lowest possible delivery dates;

• allocation of the servicing to the collaboratingfirms;

• giving up some manufacturing companieswhose products brings the number of complaintsand guidance to other international companies;

• staff training in marketing;• overtime (extra hours) payment.So, being set all that, a new series of data was

collected, over the same period of time, (ieSeptember 2009 - March 2010), and after drawinga new Pareto chart, it was noted that thesemeasures had the desired results, ie fewercomplaints and complaints.


189

Figure 7:Data collected between September 2009 -March 2010

After the data collecting, we proceeded toarrange them in descending order, as we did in thefirst step, according to Table 5 and, then wecalculated the relative and cumulative frequency(Table 6).

Table 5:Data rearranged

Table 6: The data frequencies calculation

The application of the measures outlined abovehas led to a drastic decrease in complaints to theservice department, as shown in Figure 8.

Figure 8: Pareto chart for the data collected in thesecond phase

The Excel application adds a plus, meaning thaton a separate sheet there can be seen the twoPareto charts, the initial one and the final one, afterapplying the measures that have resulted from thefirst interpretation of Pareto charts.

As shown, the amount of complaints hasdropped by almost 70%.

The application brings a benefit, meaning that,after the drawing of the first chart and afterapplying the necessary measures (as a result ofbrainstorming sessions and use of Ishikawadiagram), starting with the recording of a new dataseries for the same amount of time, introducing anew data series (e.g. at the end of each month), itcan be observed, in real time, the effect of theimplemented measures by comparing the twocharts, so it could be applied in due time theinherent corrective or preventive actions.

Figure 9: Comparison of the two Pareto charts (theinitial and the final ones)

4. Conclusions

The elaboration of an algorithm for the use ofmethodological tools is helpful in solving qualityproblems, especially if supported by a program orsoftware.

In this work, it has been developed such analgorithm, meant to solve the deficiencies incommunications with clients, communicationbetween subsidiaries and the lack of coordination.


190

In support of this algorithm an Excel applicationwas created, which combines some of the toolsproposed to solve these problems, in this casestudy as these two statistical tools, respectively theIshikawa diagram and Pareto diagram.

The application brings a benefit, meaning that,after the drawing of the first chart and afterapplying the necessary measures (as a result ofbrainstorming sessions and use of Ishikawadiagram), starting with the recording of a new dataseries for the same amount of time, introducing anew data series (e.g. at the end of each month), itcan be observed, in real time, the effect of theimplemented measures by comparing the twocharts, so it could be applied in due time theinherent corrective or preventive actions.

References[1] u, N., Dima, O., Gur u, Gh., Gonzales

Barajas A., - Sisteme de asigurare a calit ii,Editura Junimea, Ia i, 1998.

[2] Ciobanu, M., Iacob, D., Mironeasa, C., -Ingineria calit ii, Editura Printech, Bucure ti,1999

[3] En tescu, A.-M., En tescu, M.-Al., - Calitate.Terminologie comentat , Editura Tehnic ,Bucure ti, 2000.

[4] Gherghel, N., - Ingineria calit ii. Aplica ii desintez i teste, Editura Cermi, Ia i, 2006

[5] Gherghel, N., - Instrumentele controluluicalit ii, Evaluarea i controlul calit ii, p. 117-254, Editura Junimea, Ia i,1998.

[6] Gherghel N., Seghedin, N., - Considera iiasupra diagramelor cauz — efect utilizate încontrolul calit ii, Tehnomus X, Tehnologiii produse noi în construc ia de ma ini, vol. V,

sec iunea Calitate i fiabilitate, p. 122 – 129,Suceava, 28-29 mai 1999.

[7] Gherghel N., - Noi tipuri i variante dediagrame cauz - efect pentru controlulcalit ii, Tehnomus X, Tehnologii i produsenoi în construc ia de ma ini, vol. V, sec iuneaCalitate i fiabilitate, p. 130- 137, Suceava, 28-29 mai 1999.

[8] Gherghel N., - Managementul calit ii însisteme de fabrica ie, Note de curs masterat,Univ. Tehn. „Gh. Asachi” Ia i, Facult. Constr.de Ma ., 2003 - 2005.

[9] Juran, J. M., - Planificarea calit ii, EdituraTeora, Bucure ti, 2000.

[10]Mittoneau, H., - O nou orientare inmanagementul calit ii: apte instrumente noi,Editura Tehnic , Bucure ti, 1998.

[11]Noye, D., - Ghid practic pentru controlulcalit ii. Principii, metode, mijloace, EdituraTehnic , Bucure ti, 2000.

[12]Perigord, M., - Etapele calit ii. Demersuri iinstrumente, Editura Tehnic , Bucure ti, 1997.

[13]Potié, Ch., - Diagnosticul calit ii. Metode deexpertiz i investiga ii, Editura Tehnic ,Bucure ti, 2001.

[14]Trandafir, M., Antonescu, V., - Calitatea.Metode i tehnici de lucru. Analiz , evaluare,control, O.I.D. I.C.M., Bucure ti, 1994.

[15]Zetu, D., Carata, E., ura, L., - Ingineriacalit ii în sisteme de fabrica ie, EdituraJunimea, 2000.

[16]Bass, I. - Six Sigma Statistics with EXCEL andMINITAB, McGraw-Hill, NY, 2007

[17]Durkee, J., - Tools for the management ofquality—part II, Metal Finishing, Volume105, Issue 3, pp70-71, March 2007.

[18]Gryna, F., Chua, R., Defeo , J. - Juran'sQuality Planning and Analysis for EnterpriseQuality, McGraw-Hill Series in IndustrialEngineering and Management, NY, 2005

[19]Luo, X. G., Kwon, C. K., Tang, J. F. -Determining optimal levels of engineeringcharacteristics in quality function deploymentunder multi-segment market, Computers &Industrial Engineering, Volume 59, Issue 1,pp. 126-135, August 2010.

[20]Pyzdek, T., Keller, P. A. - QualityEngineering book, Second edition, M. Dekker,N.Y 2005

[21]*** - Tools for Developing a QualityManagement Program: Proactive Tools(Process Mapping, Value Stream Mapping,Fault Tree Analysis, and Failure Mode andEffects Analysis), International Journal of R. O.B. P., Volume 71, Issue 1, Supplement 1, pp.S187-S190, 2008

[22]***www.bain.com/management_tools, 2010[23]***www.borland.com/us/rc/lifecycle-quality-

management, 2010[24]*** www.epa.gov/quality/qmps, 2010[25]***www.mastercontrol.com/quality-

management-software/total-tqm/tools, 2010[26]***www.sixsigmaiq.com, 2010[27]***www.skymark.com/resources/tools, 2010

http://www.epa.gov/quality/qmps


191

ABOUT NATIONAL STRATEGY OF ROAD SAFETYIN PRESENT AND FUTURE

CIUBOTARIU DANUT1, NECULAIASA VASILE2

1Ministry of Domestic Affairs and Administration, SPCRPCIV, Botosani, Romania, [email protected] Asachi Technical University, Iasi, Romania

Abstract: After the year 1990, the car park and also the number of drivers have almost tripled,being in a complete contradiction with the slow development of the infrastructure, leading to anincrease in the percent of material and human damages as a result of traffic accidents. As far asRomania is concerned at the level of the European Union, during 2001-2010 recorded a 25%increase of the number of deceased persons, which places our country on the last but one place inthe European top.

Keywords: deceased, seriously injured, traffic accidents.

1. INTRODUCTION

The continuous increase of the motorization indicestogether with the increase of the number of drivers aswell as the various infrastructure rehabilitation sites fornational and European roads led during the last years tothe blocking traffic, therefore influencing the behaviourof certain traffic participants and implicitly theaccidents dynamics evolution in Romania.

The present project tries to draw attention over theincreasing number of accidents with several victimsand also their severity.

2. THE TRAFFIC ACCIDENTS REPORTRECORDED AT THE WORLD LEVEL,EUROPEAN UNION AND ROMANIA IN 2009

a) At the world level, the traffic accident holds the firstposition among the causes of violent deaths [3,4]:- over 1,26 million deceased people – 3200 every day;- over 35 million injured people – 50.000 every day;- over 10 million handicapped people as a result of theaccident;

The most traffic events are recorded in developingcountries and in transition countries, although theyhave only 32% of the total number of vehicles.

b) At the European Union level, the situation of thetraffic events is as follows:- over 50.000 deceased people;

- 1,5 million injured people;- over 160 billion Euro for social costs;

Most of the experts consider that “roads can bedescribed as being far the most dangerous and exposedto accidents of all means of transport”; [5].The absence of a strong and well-defined strategy ofsupervision and control encumbers the development ofa safety culture of the traffic leading to an idle violentattitude among the traffic participants.

c) In Romania the number of the traffic accidentsreported to the number of inhabitants, vehicles, holdersof driver licenses is well above the EU average asfollows:- over 2400 deceased people annually;- over 6000 people seriously injured every year;- road traffic acquired a special significance in the last20 years as follows:- the car park developed from 2,2 million vehicles in1990 to 5.323.960 in 2009- the number of drivers increased from 3,3 million in1990 to 6.063.410 in 2009- the road infrastructure has remained at the sameparameters as in the 90’s.

The contradiction between the dynamics of theabove mentioned components and the slowdevelopment of road infrastructure, unable to ensurethe fluency and safety conditions of the crowdedtraffic, increased the threat of traffic events.


192

Figure 1: The evolution of serious traffic accidents during 1990 – 2008

Table 1: The evolution of serious traffic accidents during 1990 – 2008Year No. of accidents No. of deceased Severely injured1990 9708 3782 61371991 8948 3078 77891992 8781 2816 69601993 8791 2826 83021994 9381 2877 81981995 9119 2863 76981996 8931 2845 75041997 8801 2863 74511998 8457 2778 72211999 7846 2505 65942000 7555 2499 63152001 7244 2461 59632002 7236 2411 59762003 6689 2229 55852004 7068 2449 57752005 7211 2629 58852006 7164 2587 57802007 8503 2800 70892008 10592 3053 9350

Total:Accidents: - 158.025Deaths: - 52.351Severely injured: - 131.572

The outburst of road traffic after 1990 led to thedoubling of the number and consequences of theaccidents with victims compared to those before 1989,offering another dimension to this phenomenon.

3. THE CATEGORIES OF COSTS RELATED TOTRAFFIC ACCIDENTS [1]

- medical costs, material damages and losses for thesociety;- administrative costs (insurance policies, police, etc);- the evaluation of personal sufferings;- damages as a result of some slight accidents, withsmall material losses, which are not recorded in thepolice

statistical reports;- the life value reported to the value of the average lifetime;

In most countries the first two categories are taken intoaccount.There are two evaluation modalities:Approximate evaluation and estimated evaluation.

3.1. The approximate evaluationIn the formula used at the European level, thefollowing definitions have been taken into account:

“dead” = when the decease occurred during the first30 days after the accident;“seriously injured” = the injure which requires animmediate hospitalization;“slightly injured” = the other categories that are notincluded in the above-mentioned categories;

According to these conditions, we have the formula:

C = PNB (25,1D + 1,64R + 0,25r) (1)

Trend of severe accidents 1990-2008

0

2000

4000

6000

8000

10000

12000

1990 1992 1994 1996 1998 2000 2002 2004 2006 2008

AccidentsKilledSeverly injured


193

PNB – the gross domestic product for eachinhabitant;D – the number of people deceased in accident;R – the number of people seriously injured inaccident;r = the number of people slightly injured in accident;

3.2. The estimated evaluationAlthough it is not so accurate, it shows a larger degreeof precision than the approximate evaluation due to theadditional information introduced in the formula.

Epd = PNB (Vv - X) + 12 Cu (Vm - Va) + CMS (2)

Where:PNB = the gross national product;Vv – years length of life expectancy;Cu – the monthly pension rate due to the heirs;Vm – age at which the pension is paid to thoseentitled.Va – the age of the entitled people when theyreceived the first pension;CMS – medical costs as a result of the accident.

Romania wastes every year a billion Euro for the trafficaccidents, “direct social costs” calculated by the WorldBank.

4. THE NATIONAL STRATEGY FOR ROADSAFETY 2011-2020 [2]

The road traffic national strategy for 2011-2020represents a public policies document on a long termwhich proposes the creation of a new public policyrelated to road safety, taking into account the poorresults of the actual policies. During 2001-2008Romania recorded a 25% increase of the number ofdeceased people, which ranks our country on the lastbut one place in the European chart, according to figure2.

5. CONCLUSIONS

After examining the statistics we concluded thefollowing:-Most of the drivers do not acknowledge therelationship between the traffic rules and the safety ofthe road traffic, being willing to break the rules if theyconsider that this does not endanger their safety at thatmoment.-The slow development of road infrastructure incontradiction with the number of vehicles and holdersof driver licenses which almost tripled after 1990.-In this situation and in the context of adopting a newPlan of road safety actions for 2011 – 2020 by theEuropean Commission, plan which has as objective the50% number decrease of injured people and victimsfrom traffic accidents at the level of the EuropeanUnion, Romania has to adopt a long term strategywhich should aim to stop the increase of the number ofvictims and attaining the European objective.

REFERENCES

[1] Asandei Cezar – Doctor’s Thesis: Researches overthe dynamics of traffic events – pedestrian – vehicle– “Transilvania University”, Brasov 2001

[2] The European Council of Road Safety – ETSC,June 2010: PIN – the index of performance in thefield of road safety.

[3] Avramescu N, etc – The Dynamics of serioustraffic accidents 2002

[4] Statistics: the Traffic Police Department ofBucharest and the Community Public Departmentof Driving Licences and Vehicles Registration ofBotosani.

[5] Curt Rich, Drive to survive, USA, 1999.

Figure 2: Increase the number of decease

-55-50-45-40-35-30-25-20-15-10-505

1015

LV ES* PT EE FR

* LT IT* IL IE

*DE* SI SK CH BE

*FI* SE A

THU N

L CZ LU*

UK**

DK CY N

OEL* PL BG RO M

T

-36% (EU27)


194


195

RESEARCHES CONCERNING THE CORRELATION STRUCTURE OFPROPERTIES OF STEELS FOR METAL CONSTRUCTIONS

STRONGLY REQUESTED

Stela Constantinescu1, Maria Vlad1

1University “Dunarea de Jos “ of Galati, Faculty of Metallurgy , Materials Sciences and Environtment,[email protected]

Abstract: The RV52 steel plates are used to manufacture liquefied and compressed gas reservoirs,recipients and pressure vessels which operate at low temperatures, big pressures and highly loadedstructures . The making technology and thermique treatment determine high properties such as : strength,wear resistance and tenacity . The normalized plate sample were studied to determine the effect of thetension release heat treatment on the properties of steel plates used for metallic welded construction .From the relations expressing the link between the strength and tenacity relation o none hand and thestructural characteristics, on the other it is obvious that the only factor which leads both to increasingstrength properties and decreasing transition temperature is the finishing of the ferrite grains .

Key words: thermique treatment , strength, tenacity, ferrite grains

1.Introduction

The plates of RV52 steel are used tomanufacture liquefied and compressed gasreservoires , pressure vessels and devicesoperating at low temperatures and within heavilyloaded metal structures.

During the processing and operation in thebasic steel mass a number of phenomena canocurr such as:lamelar spreading, fissures in thewelded areas and easy breaking. Occurrence anddevelopment of these phenomena depend on thesteel chemical composition and the semiblankselaboration and processing conditions.Thenormalised plates of RV52 steel feature a ferite-perlite structure, the obvious tendency being thatof reaching an as low as possible content ofperlite with the corresponding decrease in thecarbon content.The heat normalising treatment ofthe thick plates took place in the roller continuingfurnaces from the rolling mill of SIDEX Galatiobserving the follwoing parameters : heating up totemperatures AC3 + 20o ÷ 40o C, air cooling.

The analysis of the relation ship betweenstrength and tenacity, on the one hand , and

structure characteristics, on the other hand, showsthat the heat tension-relieving treatment andespecially, the normalising treatment considerablyincrease the strength /resistance and decrease thetransition temperature due to the ferrite grainshrinking process [1].Manufacturing plates of high chemical andstructure homogenity leads to an isotrophycorresponding to these properties, which attractedthe researchers’ interest .

2. Experimental researches and resultsIn order to carry out the researches on thecorrelation of microstructure with the propertiesof steel RV52 , normalized and detensioned, thefollowing working variants have been established:

elaboration in a 50t electrical furnace of 15 tthick plates by ingot casting, normalisingtreatment , and sampling for heat treatmentdetension in laboratory;

elaboration in a 150t converter , by 25 t ingotcasting and continuing casting in slab , of thickrolled plates , normalizing treatment , andsampling for heat treatment detension inlaboratory [2,3]. The chemical composition of thesteel investigated is given in table 1.

mailto:stela.constantinescu:@email.ro


196

Table 1. The chemical composition of material.Chemical composition, [%]

C Mn Si P S V Al Ni

0,14 1,4 0,26 0,016 0,010 0,050 0,024 0,28

Sampling the steel RV52 plates before and afternormalising has been carried out acc to thefollowing scheme :- the samples of the plates made by ingot rolinghave been taken from the edges and the axiscorresponding to the upper, middle and bottomparts;

- the samples of the plates made by cast bramerolling have been taken from the edge and axis.

The metallographic structures of the sampleswithout heat treatment, from the 15t steel ingotsmade in the electrical converter , for upper,middle and bottom sides are illustrated in figures1a,1b,1c .

Figure 1. Metallographic structures of the samples without heat treatment from the 15t steel ingots made in theelectrical converter : a) upper; b) middle; c)bottom

The samples have been taken from the centralaxis of the plate corresponding to : a) upperside/head, b) middle part ; and c) bottom side/footof the ingot [4].

The microscopic structures of the plates madefrom steel elaborated in the LD converter andcontinuously cast, positions edge and axis, areillustrated in figures 2a, 2b.


197

Figure 2. Metallographic structures of the plates made from steel elaborated in LD converter : a) edge; b) axis

Metallographic slifs have been sampled from thenormalized plate samples. The slifs have beenprepared in longirudinal cross –section acc to the

classical procedure [5,6]. Determinations havebeen performed acc to STAS 5949-95, and theresults are given in table 2 .

Table 2 . Non- metalic inclusions score

ScoreTypes of non- metalic inclusions

Total score overthe same fieldSulphides

S)OxidesOL+OP

Silica iSF+SP(SN)

NitridesNT+NA

Non- metalic inclusions score1 2 3 4 5 6

a1

0,51

1,51,51

11

1,5

0,50,50,5

43,54

Max. scorefrom types ofinclusions

1 1,5 1,5 0,5

b0,50,51

1,51,50,5

1,51,51,5

0,500

43,53


1 1,5 1,5 0,5

c 111

0,51,51

11

1,5

0,500

33,53,5


1 1,5 1,5 0,5


198

The elaboration technique has also beenassessed by determining the non- metalicinclusions score as shown in table 2.The impurest zones, irrespective of the sampleposition, are those corresponding to the upper andmiddle side of the ingot . In the axis purity islower as compared with the sample edge [7,8].When elaborated in the electrical furnace, a higherpurity is reached than that in a converter. In thecase of continuing casting, the differences reportedbetween edge and axis are considered higher thatwith ingot casting. The purity results show that theC, Mn and S segregation zones are also zoneswhich include the non-metallic inclusions [9,10].A typical image in terms of purity is given infigure 3 .

Figure 3. Typical aspect of purity of RV52 steel plates,without attack.

Magnified x100

Micro-structural measurements have confirmedthe presence , in the max C and Mn segregationzones , of higher perlite proportions of thesesubstances as shown in figures 1a,1b,1c and 2a,2b.

The metallographic structure of the heat treatedplates is given in figure 4 which highlights a fine-granulation ferrite-perlite structure [11,12].

Figure 4. Micro-structure of normalized RV52 steelplates Magnified x 100

Variations and distributions of the mechanicalproperties : ultimate strength (Rm), yielding point(Rp0,2), breaking elongation (A5%), resilience on-v-grooved longitudinal samples, tested to -20 oCand –50oC, from upper, middles and bottom sidesare graphically illustrated in figures 5,6,7,8.

Sample position in comparison with the ingot Sample position in comparison with the ingotFigure 5. Tensil strength (Rm) values distribution Figure 6. Yield strength (Rp 0.2) values distribution

of samples in various zone of the sheet of samples in various zone of the sheet


199

Figure 7. Elongation A5 values distribution of Figure 8. Break energy at shock (KVL) values of samples in various zone of the sheet distribution on the longitudinal piece test at --200C.

Figure 9. Break energy at shock (KVT) values distribution on the transverse piece test at -200C.

The resilience of the cold- deformed and heat-treated samples under the above mentionedconditions has been determined at temperatures of- 20oC i – 50oC, acc to figures 9.In order to establish the influence of the heatingtemperature subsequent to cold deformation of thesamples subject to 4%,8% and 12% degree ofdeformation , these have been treated acc to thetreatment cycle described below :- heating: 2500C; 5000C; 6500C;- exposure: 160 min. As a result of 2 min/ mmexposure;- air cooling. [%]

From the analysis of the tension-relievinggraphics, it has been found that :- as a result of the tension-relieving heattreatment at 250 oC, the shock ultimate strengthat -20oC and –50oC decreases with respect to thevalues of of the samples from plates of 4%,8%and 12% degree of deformations , and the valuesof the normalized samples figures 10, 11.- determining the shock behavior is more obviouswhen the degree of deformation is higher.- with degrees of deformation of 12%, the shockultimate strength at -20 oC and –50oC takes lowervalues than those min admissible as indicated bySTAS 11502-90.


200

- heating to 650oC results in recovery of thetenacity properties of RV52 steel , without havingreached the level of normalization.- heating at 250oC has disastrous consequences,because tenacity is completely damaged both withrespect to normalization and deformation states;

- by heating at 500oC, the values of the shockultimate strength increase as compared withthose rached by cold deformation, while keepinghowever the influence of the cold deformation, i.e. lower values for higher degrees of deformation[13,14].

Figure10.The influence degrees of deformation and Figure11.The influence degrees of deformation andheting for values break energy at shock, KVL (-20oC) heting for values break energy at shock, KVL (- 50oC)

3. Conclusions

The researches were focused on determining thefactors that cause properties variation in differentzones of the RV52 steel plates manufactured byArcelor Mittal Galati .In order to obtain highly improved properties agood correlation should be achieved between theconditions of elaboration, deformation and heattreatments highlighted by the chemicalcomposition and structure .The cold plastic deformation of low degrees ofdeformation: 4, 8, 12% results in poorer tenacityproperties as compared with the normalised statevalues , which is more sensitive with higherdegrees of deformation.The best values of shock ultimate strength both at-20oC and –50oC have been obtained after atension relieving treatment at 650 oC for 6 %degree of deformation

The measurements of the sulphide , carbon andmanganese segregation , purity, microscopic

structures and the mechanical properties in thesezones have all shown that : The C, Mn, and S segregation phenomena andtherefore the variation of the properties acc todifferent directions are much more obvious in aless careful elaboration in the 150t converter, ascompared with the elaboration in the vacuum-degasing electrical furnace.The chemical, structural and propertiesanisotropy is much more obvious in case of 25 tingot casting as compared with 15t ingotcasting.In case of ingot plates , the segregation andanisotropy phenomena are more pregnant in theplate axis than in its edges, i. e. In the upper andmiddle sides than in the bottom side.Taking into account the results obtained , in orderto diminish the chemical segregation whichcauses the structure modification and propertiesanisotropy , it is necesary to take technologicalsteps able to eliminate this phenomenon. Suchsteps should include both the elaboration


201

technological parameters ( casting temperature,vaccum technology) and the casting parameters (ingot size) along with the use of some raw

materials ( liquid cast iron, refractory staff) whichshould cause a min. impurification of the steel.

4 References

[1] Constantinescu, S., Metals properties andphysical control methods, Didact i Pedag.Publishing House , Bucure ti , 2004.

[2] Constantinescu S. , Dr gan V., Destructiveand nondestructive testing of metals , Ed.Evrika Br ila , 2000.

[3] Walanbe T., Mechanical proprieties of Cr-Mo steels after eleveted temperature service,Partea I: Document II S-IX-116-79; Partea II:Document II S-IX-1167, 1990.

[4] Steven N., Jise, Steel , 1989, Vol.193, p.141[5] Sugiyama T., Kobe Seel Engineering

Reports,1995, Vol.25, nr.4, p.4.[6] Constantinescu S., Drugescu E., Studies on

Mechanism and Kinetics of PhaseTransformation in Superficial Layers UsingUnconventional Procedures. ResearchContract no. 5005, Galati 1995, p. 53- 88.

[7] Constantinescu S., Drugescu E.,Radu T.,Practical application of AE of the differentgrades of steel, in Proceedings of the25thEuropean Conference on AcoustionEmission Testing “ EWGAE 2002 “ PragueCzech Republic , 11 –13 september, 2002 , p.135-143.

[8]Constantinescu S., Influence ofmanufacturing process on chemical andstructural homogeneity of welded pipe sheetsfor tanks and vessels working under pressure,in Proceeding of the International Conferenceon Advances in Materials and ProcessingTechnologies, september 18 – 21 , 2001,Leganes, Madrid, Spain, p.57 .

[9] Constantinescu S., Researches on the optimumheat treatment for the low alloyed steels usedat high temperatures, Metalurgia Interna ional, nr. 6 /2005 , ISSN 1582- 2214 , p.3-7 .

[10] S. Constantinescu, L. Orac ,Characterizationof steels in terms of their physical andmechanical properties, BuletinulInstitutului Politehnic din Iasi, Tomul LI II

( LV II ) , fasc. 3, ISSN 1453 – 1690, p.51-58, 2007

[11]Vlad M., Constantinescu S, Stratulat, E.Improvent of steel sheet corrosionresistance by thermic coverage method withzinc-alluminium alloys, Annals of „Dunareade Jos” University of Galati, Fascicle IXMetallurgy and Material Science, ISSN1453 – 083X, 2007,Vol. 1, p. 32-38.

[12] Constantinescu S., Non-distructive control ofhot – rolled products, BRAMAT 2007,International Conference on MaterialsScience and Engineering, February 22-24,2007, Bra ov , ISSN 1223 – 9631, p.241-246.

[13] Constantinescu, S., Orac L., MechanicalProperties of thin films investigated usingmicromachining techniques, METAL 2010,Ostrava, Czech Republic ISBN: 978-80-87294-17-8, p.926-931,2010 TANGER,spol.s.r.o., www.METAL2010.com

[14] Radu, T. ,Constantinescu, S., Vlad, M., .Morphologies of Widmanstatten structuresand mechanism formation in steels, Journalof Materials, Science Forum Vol. 636-637,pp 550-555, Trans Tech Publications,Switzerland,2010, ISSN 1662-9752,www.scientific.net / M6F. 636- 637.550

http://www.METAL2010.com/

http://www.scientific.net/


202


203

DETERMINATION OF THE EFFORTS THAT OCCUR IN THE CASE OFPERFORMING INNER THREADS WITH FORMING ROLLERS

ION CRISTEA1 , CRINA RADU1

1Department of Engineering “Vasile Alecsandri” University of BacauCalea Marasesti 157, 600115, Bacau COUNTRY ROMANIA [email protected]

Abstract: - The aim of this paper is to present a simplified model for the theoretical calculation of thedeformation efforts that occur in the case of performing inner threads by forming rollers. Firstly,determination of the normal stress distribution along the roll-blank separation line is presented and thenthe calculus of the resulting force of deformation.

Key-Words: -Inner threads, forming rollers

1. Introduction

Although known since the last century, internalthreading by plastic deformation is not widely usedbecause of a poor number of theoretical andpractical researches made on this area;consequently, it would be necessary to continueand deepen the researches in order to setup thetheoretical and practical basis concerning the toolgeometry and its influence on the materialbehavior.

The execution of inner threads by rolling ismade with thread rolling heads and it is specific tothe threads with internal diameter bigger than 30mm. Their breeding range is restricted because oftheir very high complexity.

Achieving inner threads by rolling differs fromtapping; in the first case, rollers have a rollingmotion over the interior surface of the workpiece,combined with a permanent plastic deformationmade in the part material. The operation is similarto the exterior rolling but in the last case the rollershave a larger diameter. In the case of rolling withheads, the material flow will appear in aperpendicular plan to the diametric plan of theroller [1].

The methodology of examination thetechnological deformability of metals and alloys ishighly complex, mainly because it is very difficultto objectively classify the metals plasticity underdifferent strain conditions and in function ofdifferent technological parameters (temperature,degree of deformation, deformation speed, etc).

If in the case of tap threading the threadingmoments increase up to 100%, in the case of pipethreading with vortex roll-process (fig. 1), theloadings are 10 - 15 times smaller compared to theclassical processing by cutting.

The forming roller 1 rotates around the axis ofthe device inserted inside the workpiece, eccentricto its axis. Every rotation will bring the formingroller in contact with the workpiece, forming thethread profile. Meanwhile, the workpiece rotatesaround its axis and moves on axial direction with arotation step. From theoretical point of view this isthe theory of final plastic deformation with acomplex kinetics of the deformation. To avoid thegreat difficulties in solving these problems [2], asimplified model for the theoretical calculation ofthe deformation efforts, based on the use of sliplines properties, is proposed [3].

Figure 1 Roller threading scheme

mailto:icristea:@ub.ro


204

2 Problem Formulation

Assuming that the material has rigido-plasticproperties and taking into account the marginalefforts for stress and speeds, it is necessary tocalculate the efforts acting on the spindle of rollinghead.

Considering the state of plane deformation, wedetermine the normal stress distribution along theroller-blank separation line and then the resultingforce of deformation. In this case the contact arcmay be replaced by a broken line [1], each segmentrepresenting the edge of penetrating wedge. Thus,the simplified scheme reduces to solving theproblem of plasticity of plane deformation state.

The differential equilibrium equations will bereduced to a system of two equations. [2]

0y

xyx

x (1)

0y

y

x

xy

The plasticity criteria of Mises and Treska-Saint Venant are analogous and differ only by thesize of the constant k:

0,25 ( x - y )2 + 2xy = k2 (2)

Further it will be analyzed the penetration ofthe symmetrical sectioned wedge into the rigid-plastic semiplane, by assuming that the plasticmass are extruded on both sides. The boundaryconditions of the problem for the right side of thesemiplan are set up, by expressing the friction onthe straight lateral contour lines through the angle:

t = k sin 2 , 0 /4 (3)

It is made the assumption that the roller actswith the speed v. The sliding field in the xy flowplane is shown in fig. 2. In our case the contour ofroller consists of two sectors. The first sector, BP,is parallel to the x axis while the second one, PA,is tilted with the angle /3 towards the first sector.Along the contact sector BP, the followingrelationships exist:

y = - P, xy = 0, vn = v0 (4)

Along the sector PA:

n = - n , nt = - t , vn = - v0 cos( /3) (5)

Along the free surface AG:

n = nt = 0 (6)

According to fig. 2, the angles , , 0 andare linked by relationships:

0 = - = - (7)

Figure 2 The field of plane stresses

By pressing depth it can be expressed the lengthof the contact sector c, with the followingrelationship:

hac 00 sin.sincos (8)

The uncompressed condition of the materialleads to the equality between the surface under thex-axis, pressed by the roller, and the area oftriangle situated above the x axis, represented bythe extruded material. Hence, the equation (9) isobtained:

2ah + h2 tg = (c + a) (c cos – h) cos ( - 0 )+

+ (c cos - h) tg (9)

The equations (8) and (9) allow us to establish alink between the angle and the depth h, as well asbetween and c.Further, the equations (1) and (2)need to be solved such that to satisfy the boundaryconditions and then the inverse problem must besolved. Assuming that the material is in plasticcondition, the stresses distribution along the borderof roller must be determined. From the equilibrium


205

condition we can determine the pressing force ofthe roller in semiplane.

By analogy with the problem of stabilized flowwhen introducing a sharp wedge, the case of thesectioned wedge was examined. In this case thegeometric similarity was violated, thus difficultieswere encountered in the determination of theparticle trajectory in the speed field. But there wereno difficulties to establish the stresses field at acertain deformation step.

In this case, for each step the deformation isconsidered small. The stresses field is determinedbased on the integrals of the plasticity equationsfor the OXY coordinates. The necessaryconstructions in the OXY work plan are presentedin Fig 2. In the areas BMP, PDEA and AFG,joined by central fields of slip lines, uniformstresses states occur.

In the AFG area the stresses field is determinedfrom the condition at the boundary AG: = nt =0. Hence:

= - k, f = /2 - (10)

where: - medium stress; f - angle betweenthe x axis and the highest component of the mainstress, 1.

As simple stresses states are analyzed, along thecharacteristics are valid the integrals of theplasticity equations:

= 2 kf + c1 (11)

where c1 is a constant determined in the AGFfield:

(12)

In the area PAED, the stresses field has thefollowing expresion:

= k cos 2 – n; f = / 2 - – (13)

while in the area BMP:

= k - p; f = 0 (14)

Knowing the c1 constant for the normal stresscomponents, it results:

n = 2 k ( - + + cos2 ) (15)

p = k ( - 2 + 2)

The punch pressure force, P, must be balancedby the normal and tangential stresses on the right:

P = 2aP + 2c(nsin - tcos ) (16)

Finally we will obtain:

P = 2ka [ + 2 - 2( - )] + 4kc[ 0 + + + cos · sin-1 · sin ( + )]·sin (17)

Establishing the pressing depth, h, the angleand the size of the coil facet, a, from the equations(8) and (9) it can be obtained the length of thecontact sector, c, as a function of 0:

0

0

sincossin hac (18)

By replacing eq. (18) into eq. (17), thefollowing relation results:

tghha

hha

aha

tghah

.cos.sincos

sin

cos.cos.sincos

sin.

.sincos

sin2

0

0

00

0

0

02

(19)

which is a transcendent equation in 0. It can besolved by halving the interval [0; /3].

4 Conclusion

An easier way to solve the transcendentequation consists in expressing the three variablesa, h and c, as a function of the thread pitch. Thiscan be done for metrics, trapezoidal and Whitworththreads by using the following relationships:

h = b1p; a = b2p; c = b3p (20)

The standardized values of the threecoefficients are given in table 1 [4]

Table. 1


206

Coefficients Metric Whitworth Trapezoidal

b1 5/16 tg/2

0,64 0,5

b2 0,25 0,137 0,36b3 5/16 sin

/20,72 0,517

By replacing eq. (20) into eq. (3) results:

b3p·[cos - sin( - 0)] - b2psin( - 0) = b1p (21)

By dividing with p we obtain:

b3 cos - ( b3 + b2 ) sin ( - 0 ) = b1 (22)

and hence:

23

130

cossin

bbbb

(21)

From the eq. (21) we can determine the angle 0as:

23

1300

cosarcsin

bbbb

(22)

From the eq. (4) we can determine the threadpitch, p:

2b1b2p2 + b12p2tg = (b2+b3)p2(b3cos - b1)cos( -

0) + + b3p(b3cos - b1p)tg (23)

and hence:

13322121

3

cos2cos

bbbbtgbbbbp

.tgbb 130cos

1 (24)

where the value of 0 is obtained from eq. (8).

References:

[1] G.A. Toropov, Dispozitiv pentru filetat interior.Brevet S.U. 1551458-1990.

[2] V.V. Sokolovskii, Construction des champs decontraintes et de viteses dan les problems delecoulement plastique, Inj. Journal T., 1963

[3] R. Lee, S. Tupper, Analysis of PlasticDeformation in a Steel cylinder Striking a RigidTarget, Journal of Aplied Mechanics , 21 (1),63-71, 1954

[4] I. Cristea, Tehnologii de executie si control afiletelor interioare deformate plastic, Ed.Junimea 1999


207

CONSIDERATIONS REGARDING THE USE OF CHUCK COLLETS INMECHANICAL SYSTEMS

Traian Lucian SEVERIN1, Vasile RATA1

1 tefan cel Mare” University of Suceava, Faculty of Mechanical Engineering, Mechatronics andManagement, University Street, No. 13, 720225, Suceava, Romania,EU. [email protected],

[email protected]

Abstract: In this paper the authors aim to present the theoretical study of of constructions such as clenchingmechanisms with chuck collets. For this to has been modernized the test stand clenching mechanisms with chuckcollets existing in the laboratory. Also in the paper were aimed experimental tests of clenching characteristics:clenching stress, dependence of clenching stress upon the main influencing factors

Keywords: clench, chuck collets, acquisition board, force.

1. General characterization of mechanicalsystems with chuck collets

The main use of chuck collets:- in the construction of grabbing systems of

robots’, of fixing mechanisms from workingtechnical structures, self-feeding systemswithin working centres, etc.

- technical helping systems to fix workingtools: drills, cutters, reamers, etc.

- control systems, positioning systems and / ormeasurement systems, etc.

Chuck collets have one or two conical inside oroutside surfaces, a series of longitudinal nicksdelimiting a series of elastic areas, called jaws.Besides chuck collets, the mechanisms of typeMA comprise rigid collets which act upon chuckcollet which is deformed until the active surfacesof the latter one get in contact with the piececentring surface, making its centring andclenching.

Chuck collets-mechanisms have someadvantages, some of which are mentioned below:- provide high centring accuracies (radius range

of centred surfaces being under 02...0.05mm);- allow clenching of thin walls, easily deforming

and law gauge pieces;- due to the relatively uniform repartition of

clenching stress , they are easily to be made andso low cost.

Their main disadvantage consists in theworking field limited by values of 0.05 D ,where D is the diameter of centring surfaces. Tocover a wide range of D dimensions, sets ofdifferent dimension chucks are required. If thecentring dimension tolerance does not fit withinthe working region, then the mechanism does notensure the clenching stress required by minimumdimensions or they are overdemanded, up tobreakage in the case of high dimensions whenbeing centred on outside surfaces. Also, under suchconditions, they lead to high accuracy and lowrigidity as a result of some (imperfect) linecontacts between the chuck collet and piece,between chuck collet and rigid mandrel or colletrespectively. In some constructions there mayoccur errors of axial orientation as well, especiallywhen the chuck collet has axial mobility.

2. Mathematical modeling of calculation2.1 Mathematical model for stress

calculationThe drive force calculation Q of chuck

collet and mechanisms is made generally foreach type of stress, depending on constructivepeculiarities.

The physics model for stress calculation


mailto:rata.v:@fim.usv.ro


208

Figure 1. Physics model of calculation

The calculation relations of drive forces Qdepending on total clenching forces S are:

- for mechanisms in fig. 1a and 1b:

Q = T1 +F2 = S tg( + 1) + 2S (1)

- for mechanism in fig. 1c:

S = Qctg( + 1) - Q tg 3. (2)

Calculation relations are valid whether chuckcollets 1 or rigid mandrels 2 are mobile.

Frontal anterior 4 (fig. 1a), or posterior (fig. 1b)bearings lead to appearance of axial friction forces

F2 = 2S between piece 3 and chuck collet 1whereas the frontal bearing 5 of chuck collet (fig.1c) leads to the appearance of radius friction forceF3 = 3Q.

The significance of noted items of relations is:- 1and 3 are friction coefficients betweenthe chuck collet 1 and rigid collet 2, frontalbearing 5 respectively;- 2 – friction coefficient between chuckcollet and piece, having values between

0.2... 0.7, depending on the shape of centringsurfaces of chuck collets (fig.3).

Chuck collets- mechanisms have becomewidely spread, especially in the construction ofdevices such as self-centred mandrels and chucks.They are frequently used in works from pig andpipe on turret lathes, semi-automatic or automaticlathes for large-scale series production and formachine equipment as well – universal tools meantespecially to fine mechanics. In this case thedevices are fitted with sets of chuck collets tosatisfy the current centring demands of pieces evenunder conditions of small series production orunique. The fixing of straight-shank tools oncutting machines is also made by the help ofdevices containing such mechanisms. Under theseworking conditions, constructive and functionalrestrictions are called for regarding the fixingaccuracy and axial shares carrying out.

Between the axial run of a chuck collet and therun ce of a jaw, the following relation may bewritten:

ce = ci tg . (3)

3 Experimental stand

Figure 2. Structure of experiment stand, 1-computer – labview soft, 2-dynamometer, 3-chuck collet, 4-dynamometerkey, 5- tensiometer bridge, 6-Labjack acquisition board


209

Figure 4 –Login scheme of data acquisition soft.

On the above presented stand, the value ofeffective clenching forces Fef at the level of chuckcollet jaws was taken and compared to thecalculation force FSc approximated by model FSc(1), (2).

At the same time, the precison degree ofclenching forces, depending on the level ofworking/stress force could be determined.

The research stand of functional characteristicsof fixing chuck collets – mechanisms uses a dataacquisition soft of a highly appreciable originality.

The experimental results obtained are shown intable 1

Tabel.1 Experimental results of chuck collets with experimental data processing

Nr.crt. Nr.div.(V) Kf(N/V) Fi (N) FSc (N) Fef (N)IntroducedMomentMi(N/m)

MeasuredMomentMf(N/m)

Momentmeasured

by ScMf’(N/m)

i=Fef/FSc

1. 1.74

1385

2409.9 1025.05 388.88 35 42 46.127 0.322. 2.481 3436.18 1401.58 435,.8 40 47 63.071 0.313. 2.743 3799.05 1615.92 555.55 45 60 72.716 0.344. 3.014 4174.39 1775.58 657.4 50 71 79.901 0.375. 3.218 4456.93 1895.75 712.96 55 77 85.309 0.3766. 3.345 4632.82 1970.57 842.59 60 91 88.676 0.4277. 3.480 4819.8 2050.1 916.6 65 99 92.254 0.4478. 3.605 4992.92 2123.74 981.48 70 106 95.566 0.4649. 3.724 5157.74 2193.84 1018.51 75 110 98.723 0.465

10. 3.862 5348.87 2275.14 1250 80 135 102.381 0.5411. 3.951 5472.13 2327.57 1268.51 85 137 104.74 0.54412. 4.105 5685.42 2418.23 1314.81 90 142 108.82 0.546


210

Fig.5. Variation of calculation forces and experimental forces

4. Conclusions

From the analysis of experimentally obtainedgraphs, the following conclusions may be drawn:

- experimental forces are smaller than thoseobtained by calculation, as friction forcesoccur between collet and conical bearing,between collet and semiproduct;

- an efficiency field in using the mechanismsin technical systems may be defined;

- for mechanical systems of automatizedstructures, corrections are necessary to betaken to have in view determination errors ofthese characteristics.

References[1.] Rata V, Severin T. – Managementul

proiect rii dispozitivelor mecanice.Tipografia MATRIX Bucure ti, 2008.

[2.] Rata V, Maiorescu A – Teoria proiect riidispozitivelor. Universitatea tefan cel MareSuceava, 1992.

[3.] Ro cule S. s.a - Proiectarea Dispozitivelor.Editura Didactic i Pedagogic Bucure ti,1982;

[4.] Tache V, Br garu A. - Dispozitive pentruma ini – unelte. Proiectarea schemelor deorientare i fixare. Editura Tehnic Bucure ti,1977;


211

THE CORROSION OF TUBULAR FURNACES IN PLANTS.ATMOSPHERIC DISTILLATION OF CRUDE OIL AND IN VACUUM FOR

FUEL OIL

Moro anu Marius1, Ovidiu Georgescu2,

[email protected]

Abstract: The paper aims to investigate the corrosion of the furnace tube elements of intalatiileAtmospheric crude oil distillation and vacuum of oil. Such phenomena occur in the ceiling andaccentuates intalatie distillation as we approach the staging area. Although the corrosion phenomena isnot frequently met for the tubes from the furnace, it produces serious damage because when a tubereached limit corrosion in one spot it must be immediately replaced taking into accomrnt the fact that theremaining tube is in perfect condition.

Keywords: corrosion, erosion, Sulfur compounds

1. Introduction

For unsulphurous fuel oil, the pipes from theformance are made from steel P235GH, and for thesulphurous fuel oil, the pipes are made from steelwith 5-9%Cr. The pipes are affected not only bythe corrosion phenomena, but also by the erosionphenomenon at the end of the cycle in areas wherethe speed of the fluid increases. The samephenomena appears in the ceiling and increases asit get closer to the transfer area. Solid salt particles,sulphates and fuel, resulted from the waterevaporation, are strongly hit by the meandres ofintroduction and by the ends of the tubes where thefuid changes it’s course. At high speeds remarcableerosions are generated. At the first tubes enteringthe ceiling, where the speed of the fuid is 25m/s,the speed of corrosion is 0.2 mm/year and 1,5-2,5mm/year for the tubes before leaving the furnace,where the amount of steem is bigger with a flowrate of cc78 m/s.

Although the corrosion phenomena is notfrequently met for the tubes from the furnace, itproduces serious damage because when a tubereached limit corrosion in one spot it must beimmediately replaced taking into accomrnt thefact that the remaining tube is in perfect condition.

In the furnace, the fuel oil is heated up at 300-3400C,the temperature at which magnesiumchloride hydrolyzes in proportion of 60-90%, and

the calcium chloride in proportion of 10%,hydrochloric acid being released.

MgCl2+H2O Mg(OH)Cl+HCl

CaCl+H2O Ca(OH)Cl+HCl

Basic magnesium and calcium salts decomposeat temperatures exceeding 3200C.

Mg(OH)Cl + H2O Mg(OH)2+ HCl

Ca(OH)Cl+H2O Ca(OH)2+HCl

Hydrochloric acid in gas form is corrosive forcarbon steel, meanwhile the sulfide hydrogenreacts even in the absence of water.

When processing the naphthenic oil with highchloride content, the transfer lines are sublimatedto a combined effect of erosion-corrosion becauseof the naphthenic acids and crystallized salts.Naphthenic acids dissolve the protective layer ofFeS and with the sodium hydroxide, introducedafter salt removal, form sodium naphthenate, withdetergent role which cleans the surface of themetal.

So the clean metal is subjected to the erosiveaction of the crystals ( NaCl, CaCl, MgCl2,Na2SO4).

mailto:marius.morosanu2003:@gmail.com


212

2. Work tehniqueThe experimental determinations regarding the

corrosion behavior of ferrous metals as a result ofthe aggressive action of circulating oil in thefurnace tubes were made with test-tubes of steelwith chemical composition presented in Table 2.1.

The samples were mechanically prepared thenthey were chemically etched with 5% HClinhibited solution, washed with demineralizedweak alkaline water and finally dried with acetonein a desiccator. After testing, individualobservations were made on the condition of thesurface and on the quantitative determination ofcorrosion.

From the pipe affected by corrosion, sampleswere taken and were chemically andmicrostructurally analyzed.

2.1 Experimental data anaysis

Figure 1 shows a schematic diagram of the ADfurnace with air heaters at the top, gas output.

Figure 1: Schematic diagram of the ADfurnace with air heaters

Tubular heat exchange elements aresimultaneously sublimated to both externalcorrosion due to combustion gases and internalcorrosion due to technologic fluid and defenserequirements, which vary from one facility toanother.

Resulting corrosion products of the outersurfaces due to combustion gases and on surfaceswithin the same tube determined by the action ofthe technologic fluid, are acting simultaneously onthe metall material. This simultaneously action notonly determines modifications of the materialstructure, but also can have a negative influence onthe main mechanical properties. Accurately assessof the changes occurring in the morphology ofsteel constituents and the nature of the subsidedand differently dispersed phases allow more realappreciation of the possible use of such tubularelements.

Intemal corrosion of the tubular elements of thefurnaces from AD and VD plants is determined onone hand by the aggressiveness of the technologicfluid, and on the other hand by operatingconditions. Processed raw material is made fromparaffinic, naphthenic and aromatic hydrocarbons,sulphurous hydrogen, mercaptans, variouschlorides, etc. For example, sulfur is dissolved infuel oil at a rate of 0.20% to 0.24% and chloridesfrom 20 to 250 ppm. With technology operating attemperatures above 4000C and pressure of 20bar,fuel oil undergoes thermochemical decompositionreactions, resulting a surplus of unwantedcorrosive agents, in addition to those existing inthe crude oil, or those witch are added as neededfor processing it. The intensity of metal tubularelements degradation processes by corrosion isadditionally influenced by the speed of thetehnologic fluid circulation and by theirturbulence within the pipes in the oven and transferlines. The fluid flow is biphasic, corrosionprocesses, as it consisted, depend not so much onthe total acidity of crude oil, but on the total acidityof the non-vaporised residue from the system(Table 2). In order here to calculate the totalacidity coefficient of the non-vaporised wasdetermined residue efficiency in distilled fractionsfrom native (TAC) of crude oil remaining aftereach fraction was removed.

The data presented in Table 2.2 show that thetotal acidity coefficient of the non-vaporisedresidue reaches maximum of 2.56 mg KOH / gcorresponding to the fuel oil exiting the furnace.The flow rate gradually increasing with thevaporization oil of a new quantity of hydrocarbs

No Typesofsteel

The concentration of elements, %

C Mn Si Cr Ni Ti P S1 S235j

R0,19

0,63 - - - - 0,050

0,070

2 P235GH

0,16

0,52 0,22 - - - 0,030

0,029

3 X6CrAl13

0,07

0,085

0,82 13,2 - - 0,035

0,025

4 X6CrNiTi18-10

0,06

1,25 0,75 18,5 11,2 0,45 0,040

0,015

5 Section withbreakingarea

0,11

0,42 0,29 - - - 0,021

0,024


213

determines a maximum of the total aciditycoefficient, of flow rate and temperature. All thesevalues are reached simultaneously at about exitingof the fourmance. Given the severity of thiscondition the speed of the corrosion should be veryhigh compared to the distillated fractions from thesame crude oil. However, this fact was notdetermined in reality. This fact can be explained byassuming that in this case crude oil containssubstances that ensures a protective layer of coke,as not in the case of distillates. This layer can bepartially removed if the fluid’s turbulence is high,fact that allows the installation of a localizedcorrosion attack.

Table 2.2: TAC Total acidity coefficient of distillatedfractions and non-vaporised residueDistillatedfraction

Efficiency(%)

TAVfraction(mgKOH/g)

TAVcalculatedfor non-vaporisedresidue (mgKOH/G)

Gasoline70…1000C

12 0,03 2,21

Petroleum I170…230

7 1,21 2,29

Petroleum II230..2800C

6 1,25 2,38

Gas oil I270…3100C

16 1,82 2,53

Gas oil II 310-3500C

3 1,99 2,56

Fuel oil >3500C

56 3 0,55

Corrosion products resulting from the attack ofnaphthenic acids are soluble iron naphtens whichcan not provide protection as that of iron sulphideresulting from the interaction with hydrogensulfide.

Along with naphthenic acid crude oil containssulfur compounds, hydrogen sulphide,mercaptans, etc., whose attack on steel plantenhances the corrosion processes specially at hightemperatures

Following this attack a layer of pores reducingthe speed of corrosion of naphthenic acids is made.

Test-tubes from carbon steel S235JR andstainless steel X6CrAl13 and X5CrNi18-10 withand without sulphide layer were tested in solutionsof naphthenic acids with TAC 145 mg KOH / gduring 20 days at a temperature of 1000C (Table3).

It can be seen from the data presented in Table2.3 that the existence of a compact layer of ironsulfide reduces significantly the corrosion rate dueto naphthenic acids.Material Protected

with FeSlayer

Corrosion rateKg

- (g/m2/h) P(mm/year)

S235jR No 0,092 0,102Yes 0,071 0,079

X6CrAl13 No 0,074 0,082Da 0,0092 0,0010

X6CrNi18-10

No 0,014 0,016Yes 0,0053 0,006

But in industrial conditions, it is possible that inthe presence of large amounts of naphthenic acidsin crude oil, with a high-speed of technologic fluid,the FeS film can be damaged (broken) whichallows the initiation and localization of new firmsof localized corrosion. In these circumstances,inner and outer surfaces of the tubular element arehighly stressed by the corrosion products.

The tube element from the AD furnace, whichcarries crude oil out of the furmace at atemperature of 3400C and pressure of 80 bar ofP235 GH steel, after 60,000 hours of operation,had shown corroded areas with cracks.

The extermal diameter of the pipe, outside thecracked area is 10% higher than the minimumdiameter, and in thefractured zone the remainingelongation is 12.4% with corresponding reductionof wall thickness to 0.9 mm and a reduction of91% compared to nominal thickness of the tubularelement.

In the breaking area, the tubular element showsan excessive oxidation due to local thermaloverstressing. This is due to the deposit ofcorrosion products that reducing heat transfer,determined the appearance of cracks with a depthof 1.2-2.0 mm. Te cracking and further materialbreaking were favorites by the greater thickness ofthe decarbonized layer, caused by the mentionedthermal overstressing, respectively by exceedingits resistance capacity under the action of internalpressure (Figure 2.2)


214

Figure 2.2. Cracks in the tube of the AD furnace

Metallographicall research conducted onprocessed cuts from both corroded and unaffectedareas (2% solution of metal activation) revealedmicrostructural changes resulting from corrosion.In the unaffected area no changes were found inthe feritico-pearlitic layers, specific for P235GHsteel (Figure 2.3) and the corroded area presents amodified feritico-pearlitic structure with deformedgrains.

Figure 2.3. Thematerial’smicrostructure fromthe unaffected area(x500)

Figure 2.4. Thematerial’smicrostructure from thebreaking area (x 500)

Also from the analysis of metallographic cuts, itis revealed a spheroidizing of the carbides from thepearlite and a precipitation at their ferrite grainslimits.

These specific types of precipitation arespecific for the steel which is maintained at highertemperatures, copmared to the recommendedtemperature range.

3. Conclusions

In the AVD plant which processes non-sulphurous oils, corrosion processes aredetermined primarily by naphthenic acidity anddistilled from crude oil, expressed by total aciditycoefficient (TAC), by the sulfur and chlorinecompounds, and operating conditions.

Sulfur compounds present in crude oil havenaphthenic corrosion braking action, as ironsulphide skin, as a product of the interactionbetween steel and sulfur compounds, is morecompact and more adherent than iron naphthenateskin.The flow rate of technologic fluid and its biphasicnature, corrode with the vaporization phenomenonto in balance and the start of thevaporos, liquidstages distress determines the intensification of thetransfer lines corrosion and of the furnace pipes,

because in these parts of the installation therecombines cavity, corrosion and erosion phenomenaThe breaking of the AD furnace tube was due toexceeding the endurance limit of the steel as aresult of local overheating due to corrosionproducts deposited layers, up to 8.2 mm thick.Overheating caused a pronounced decarburizationand the emergence of cracks up to 2 mm grooveson the esternal surface of the pipe.

References

[1] Shalaby H.M., Refining of Kuwait’ s heavycrude oil: materials challenages, Workshop oncorrosion and protection of metals, Arab schoolfor science and tehnology, December 3-7Kuwait

[2] Antonescu N.N., Moro anu M, Petrescu M.G.,Niculae A.D., Georgescu L., Georgescu Ov.,Issues on corrosion equipment in instalation bydistribution DAV naphthenic acids from crude oil,Poster WP15- corrosion in Refinery Industry,EUROCORR 2010, The European CorrosionCongres, European Federation of Corrosion 13-17 September 2010.

[3]Ropital F. Mecanismes et mecanique desinteractions plastic envirovmements. Casd’industrie petroliere, Plast Ox 2007 (2009),p265-275., EDP Sciences 2009

[5] Lobley G.R., Stress corrosion cracking: caseSaudies in refinery equipment, The 6 th sandiEngineering Conference, KFUPM, Dhahran,December 2002, vol.5, 2002, p17-26.


215

APPLICATIONS OF FULL FACTORIAL DESIGN EXPERIMENTSFOR LASER WELDING

Remus Boboescu1

1 Politehnica University Timi oara, [email protected].

Abstract:. Laser welding using Nd: YAG laser with continuous emission is applied for a low alloyedsteel. The study pursued molten areas characteristics in the material. On the weld cross section wasmeasured weld width, weld depth and the weld molten zone. Their variation was analyzed with powerand welding speed. A full factorial experimental design was applied for two particular values of thedistance between focal plane and the workpiece surface (defocusing depth). It presents mathematicalmodels, the ranking effects by Pareto charts, response surface method and the multiple ANOVA analysisof variance. It showed the main effect of laser power in determining the weld characteristics.

Keywords: laser welding, Nd:YAG laser, full factorial design, laser beam defocusing, mathematical model.

1. Introduction

Laser welding is widely applied machining ofmetals and in particular for steels. The laser beamis a concentrated heat source that allows for theintensity of the laser beam 104-107W/cm2 and forinteraction time between laser radiation and thematerial 10-3-10-2s obtaining molten zones. Meltingmaterial is used for welding. Several papers havepresented the features of the steel melted zone forlaser welding [1].

Main information about the weld characteristicsare given of the weld cross section. On this aremeasured the weld width, weld depth and the weldmolten zone area. Sectioning the welds in a stablearea provides information about the ability toperform laser beam melting material. Applied themathematical modeling for the molten zone ofwelds is useful for estimating the parameter valuesused in making welded joints. Type the fullfactorial experiment with the statistical method ofANOVA analysis of the multiple underlyingseveral methods of statistical analysis inexperimental research [2] [3].

This paper proposes a study on the laser weldsmade on the low alloy steel plates.

Apply a complete factorial experimental designtype22 for two different experimental situationsgiven by the focal plane position relative to theworkpiece surface (Defocusing) for welds madeusing laser irradiation Nd: YAG under continuous

regime. In experiments were varied laser powerand welding speed.

2. ExperimentsThe experiment consisted in made lines of

fusion (welds) ,110mm long, on Dillimax 500 steelplates with thickness of 10 mm (carbon steel,carbon content 16.0 %), figure 1.

Figure 1: Plate with welds a) Surface plate b) cross-sections through the plate

mailto:remus_boboescu:@yahoo.com


216

Was used a Nd: YAG Trumph Haas 3006Dlaser source with 3kW maximum power on acontinuous wave regime CW. Laser beam wastransmitted through a optical fiber with corediameter of 0.6 mm

The focus system made a focal spot with 0.6mm diameter. Lens focal length was 200 mm. Asprotective gas argon was used with a flow rate of20 l / min. Were used sheets of material withdimensions of 10130100 mm for which weremade between 5 and 8 welds, with a distance ofover 10 mm between welds,.

In experiments was varied the laser power,welding speed and distance between focal planeand piece surface (defocusing or defocusing depth)figure 2. Welds were cut in the stable part of theweld near the place where welding process wasstopped. Weld section was processedmetallographic. Weld width, near the piecesurface, and weld depth were examined using amicroscope with precision of 0.01 mm. Meltedarea was measured directly by its footprint.

Parameters varied in the experiments arepresentations in Figure 2. To focus within the piecedefocusing values are considered negative.

Figure 2: Parameters varied in the experiments

In the experiments were varied power and weldingspeed. To statistically analyze the effects ofparameters was necessary to introduce adimensionless parameter values. Transformationsbetween the two systems are based on real-codedrelationships following:

2PA [-] (1)vB 22.233.2 [-] (2)

The experimental plan is presented in Table 1 withactual values that coded for laser power andwelding speed.

Analysis procedure consisted of presenting theresults of the mathematical model, ANOVA tableshowing the correlation coefficients associatedwith the mathematical model, Pareto chart showingthe hierarchy of effects and response surface is agraphic representation of mathematical model. Forthe mathematical model were presented two formsfor real values laser power and welding speed andfor coded system values. The first allows rapidapplication of formulas and the second allowsdirect analysis of the values of regressioncoefficients.

Table 1: Varied parameters in the experiment

Based on these values were achieved Pareto charts.On the weld cross section was measured near

the surface of the weld the width w[mm], at thecenter of the weld the depth h [mm] and meltedarea MA [mm2]. These measurements are shownin Figure 3.

Figure3: Weld cross-section with the considered sizes

3. Effects of welding process parameters

The varied parameters laser power, weldingspeeds and defocusing have the following effectson molten zone dimensions:- Laser power. Increasing the laser power produceincrease the intensity on piece surface thereforemelted material amount increases. From a certainvalue, intensity not too high melting materialamount, favours material vaporization.- Welding speed. Increasing welding speeddecrease the interaction (time) between laserradiation and material. If the interaction time isless then the molten zone dimensions are smaller.- Defocusing. Defocusing by lowering the focal


217

plane within piece produce lower intensity at piecesurface by increasing laser spot area at piecesurface and from same issue will increaseinteraction time between laser and material.Focus within piece associated with the presence ofkeyhole in welding bath will increase the spread ofradiation in keyhole and coupling of laser radiationand material. Defocusing may thus have differenteffects on melting material. You can notpredetermine a clear trend of increasing ordecreasing the molten zone. Defocusing effectswill be analyzed based on experimental results.

4. Weld width

For the weld width at defocusing 0polynomial mathematic model is given by relations(3) (4). Statistical analysis of variation is given inTable 2.

PvvPw 111.0999.04335.07348.0 (3)ABBAw 05.055.055.08833.2 (4)

Table 2: ANOVA table for weld width w at 0

Pareto diagram on Figure 4 shows that the weldwidth increases with power and decreases withwelding speed. The two parameters are equal butopposite sign effects on weld width. Theinteraction between power and welding speed hasan effect on the decrease weld width. Themathematical model presented is statisticallysignificant data for the effects of speed and power.

Response surface in Figure 5 shows thevariation of weld width with power and weldingspeed at defocusing 0 .It is noted that on theexperimental field weld width increases withpower and decreases with speed. Maximum valuesfor the weld width are obtained at the highestpower and lowest speed. Lower values of weldwidth are not recommended because they areassociated with low weld depth.

Figure 4: Pareto Chart for weld width at 0

Figure 5: Response surface for weld width at 0

For the weld width at defocusing mm2mathematic model is given by polynomialequations (5) (6).Statistical analysis of variation isgiven in Table 3. Pareto chart in Figure 6 showsthat the weld width increase with power anddecreases with welding speed. Both effects arestatistically significant. Power effect is greater thanthe welding speed. Interaction between the weldingspeed and power decreases the width of the weld.

PvvPw 666.0444.0999.15826.0 (5)ABBAw 3.08.03.17166.2 (6)

Table 3: ANOVA table for weld width w at mm2


218

Figure 6: Pareto Chart for weld width at mm2

Response surface in Figure 7 shows thevariation of weld width with power and speed atdefocusing mm2 . It is noted that on theexperimental field weld width increase with powerand decreases with speed. The variations of weldwidth are low at low power and high weldingspeeds. It looks like that under weak irradiationweld width may be associated with the laser beamspot on the workpiece surface.

Thus the focus within the piece can produceacceptable weld width even if using low power orhigh values of welding speed.

Figure 7:Response surface for weld width at mm2

The variation observed previously shown thatthe weld width becomes more dependent on powerfor the laser beam focus within the piece.

Interaction effect between laser power andwelding speed increases weld width. It isassociated with dynamic phenomena occurring in

the weld pool that influence laser radiationabsorption.

5. Weld depthFor the weld depth at defocusing 0

polynomial mathematical model is given byequations (7) (8).Statistical analysis of variation isgiven in Table 4. Pareto chart in Figure 8 indicatesthat the weld depth increases with power anddecreases with welding speed. Effect of power ismuch greater than the effect of welding speed. Theinteraction between power and speed has a smallrole and does not present statistical significance.

PvvPh

06105.060495.0903425.1566625.1 (7)

BABAh 0275.03275.09675.13 (8)

Table 4: ANOVA table for weld depth h at 0

Figure 8: Pareto Chart for weld depth at 0

Response surface in Figure 9 shows thevariation of weld depth with power and speed atdefocusing 0 . It is noted that on theexperimental field weld depth increases sharplywith the power and almost no variation withwelding speed. Welds with high depth presentskeyhole welding regime.


219

Figure 9:Response surface for weld depth at 0

For weld depth at defocusing mm2polynomial mathematical model is given byequations (9) (10). Statistical analysis of variationis given in Table 5. Pareto chart in Figure 10shows that the weld depth increases sharply withpower. Power effect is statistically significant.Weld depth and decreases the interaction betweenwelding speed and power. The two effectsmentioned above are close.They are not given by the mathematical modelwith statistical significance. It looks like that forthe focus inside piece the effect of power becomesimportant. Dynamic phenomena occurring in theweld pool becomes more intense, which isreflected in the increasing role of the interactionbetween power and welding speed.

vPvPh 8436.06216.07554.28541.1 (9)BABAh 38.048.087.15383.2 (10)

Table 5:ANOVA table for weld depth h at mm2

Response surface in Figure 11 shows thevariation of weld depth with power and speed atdefocusing mm2 . It is noted that on theexperimental field there is a weld depth increasewith power. At high power has been a slightdecrease in the weld depth with welding speed.Effect of welding speed is obvious by reducing theincrease with power at high welding speeds.

Figure 10:Pareto Chart for weld depth at mm2

Figure 11:Response surface for weld depth at mm2

Effects on weld depth analysis show that itdepends strongly on the power in both cases ofdefocusing depth. Focusing inside the pieceincreases the welding speed effects due toincreased contribution of dynamic phenomena inthe weld pool. They are very sensitive to variationsin time of interaction between laser radiation andmaterial controlled by the welding speed.

6. Melted area at the weld cross section

For melted area mathematic polynomial modelat defocusing 0 is given by equations (11)(12). Statistical analysis is presented in Table 6.Pareto chart in Figure 12 shows that the meltedarea increases with power and decreases withwelding speed. The interaction between power andspeed decreases melted area. Power is the onlyeffect that has statistical significance. Contributionof welding speed by its effect and by its interactionwith power is lower than power effect.. But theeffect of welding speed is close to power effect. Itlooks like that for the melted area on weld cross


220

section are important both laser beam intensity atworkpiece surface and interaction time betweenlaser radiation and material.

vPvPMA 776.1222.0836.07664.1 (11)BABAMA 8.07.17.2866.3 (12)

Table 6: ANOVA table for melted area MA at 0

Figure 12: Pareto Chart for melted area at 0

Response surface in Figure 13 shows thevariation of cross section area of the molten weldwelding power and speed at defocusing 0 .High values for melted area are recommended forjoint the two pieces in laser welding. It is notedthat the melted area increases with the power anddecreases with welding speed. The decrease withspeed is lower at the low power than the highpower. This shows the role of interaction betweenpower and speed. Highest values of the melted areaare obtained at maximum power and weldingspeed on the experimental field.

For melted area mathematic model polynomialat defocusing mm2 is given by equations(13) (14). Statistical analysis is presented in Table7.Pareto chart in Figure 14 show that the meltedarea on the weld cross section increases withpower. Power is the only effect that is statisticallysignificant.

Figure13:Response surface for weld melted area 0

Welding speed and the interaction betweenpower and welding speed lower melting area. Thetwo effects are equal. It is noted that the overallcumulative effect of decreased melted area withspeed and with interaction between speed andpower is greater than the effect of increasing thepower given. It looks like the melted area isheavily dependent on the time of interactionbetween laser radiation and material.

vPvPMA

1625.416250.439375.793545.6 (13)

BABAMA 875.1875.1025.34843.3 (14)

Table 7:ANOVA table for melted area MA atmm2

Response surface in Figure 16 shows thevariation in area melted on weld cross-section atdefocusing mm2 .It is noted that on theexperimental field melted area increases with thepower and decreases with welding speed. Thedecrease with welding speed is stronger at highpower. Increasing with power is greater at lowwelding speeds. It looks like the importance ofinteraction between power and speed. Maximumvalues for the melted area are obtained atmaximum power and minimum welding speed. Atlow power and high welding speeds there is adomain with small variations for melted area.


221

Figure15:Pareto Chart for melted area at mm2

Figure16:Response surface for weld melted areamm2

Focusing within piece shows as additionalelements for the weld cross section melted areaincreased the welding speed contribution. This isaccomplished through both its effect and itsinteraction with power. It is noted that the effect ofwelding speed on melted area is stronger than theeffects of welding speed on weld width and welddepth.

7. Discussion

Mathematical modeling using full factorialdesign showed high values of correlationcoefficients and statistical significance wasobtained for sizes analyzed with power effect.For weld width and weld depth the focus withinthe piece led to an increase in power effect.For melted area on weld cross-section variationwas contrary effect, namely increased speed effect.It looks like that lower intensity laser beam to the

workpiece surface by focusing the laser beaminside piece lowers laser beam intensity to theworkpiece surface. The intensity is mainlycontrolled by power. So considered sizes are moresensitive to changes in power. Fused area is relatedto the total amount of melt obtained afterirradiation. At defocusing mm2 the laserbeam heat source is inside the Keyhole at its frontwall. Keyhole front wall angle depends on thewelding speed.

Strong inclination of the keyhole front wall ofthe laser radiation encourage reflection outsidekeyhole. This reduces the coupling between laserradiation and the material thus obtained decreasesthe amount of melt. The mathematical modelspresented are useful in determining the size of areathat can be melted by laser beam. The deepwelds are in keyhole regime. Molten zone shouldbe sufficient to melt the pieces and filler materialused in welding.

Mathematical modeling presented in this papercan be used to determine the characteristics ofwelds in concrete terms the imposition ofrestrictions related to carrying out the welding.For deep weld the weld width must be high.Focusing the laser beam inside the piece from thestart ensures high values for weld width due to theincreased the size of laser beam spot on theworkpiece surface. Experiments have shown thatinside piece weld width decreases in comparisonwith the weld at the surface but a certain amount iskept relatively constant with depth in the material.The high level of power ensures an appropriateweld width both of the weld surface and inside thepiece.

Increasing the depth of the weld is given bypower level. The high level of power producewelds in keyhole regime. Getting the welds underkeyhole regime is strongly associated with theformation of pores in the material. Increasingpower and decreasing the welding speed leads to asituation in which many pores are produced inwelds. Low welding speed produce significant heataffected zone. Appears important to control thelaser beam intensity at the workpiece surface bychanging the defocus. Decrease laser beamintensity independent of power provides amoderate keyhole welding regime is useful forreducing porosity in the weld.

For melted area is recommended the maximalvalues in all cases. This can be achieved byincreasing power and decreasing welding speed.


222

Be sure that it is not excessive variations do notproduce enlarged pores and heat affected zones.

8. Conclusions

This paper has analyzed the characteristics of across section through the weld associated with aperiod in which the welding process showedmaximum stability. The effects of power andwelding speed are considered for two situations forthe laser beam focus relative to the workpiecesurface.To control variation in the welding processby power and welding speed mathematicalmodeling associated with experimental researchhas shown the following:-Power is the main effect on the characteristics ofthe weld. We recommend a high level of power.-In relation to central point of a factorialexperiment is recommended to increase power andlower welding speed without reaching the extremevalues for both parameters.- It is recommended to focus the laser beam insidethe piece to achieve a moderate keyhole weldingregime that is associated with reduced porosity.- Modeling of the melted zone of the weld crosssection will be useful for the design of weldedjoints.

The paper presents several ways of presentingthe measured sizes. The mathematical model,Pareto diagram and response surface shows thesame variation. Each of them has a different utility.

From the technological point of view thesetypes of representations can be used to determinethe objective function values that characterize thewelds for data values of welding parameterspower, welding speed and defocusing. It canaddress the inverse problem required to determinethe values of parameters power, welding speed ,defocusing to obtain employment objectivefunctions, melted area, weld width, weld depthwithin certain limits. For this problem is to useutility of response surfaces. Response surface is afunction of grade 2.Applications of responsesurface in optimization of this type, refers to thedetermine of maximum and minimum extremes onthe experimental field from response surfacefigure. Thus the presentation of multiplemathematical models for the same experimentalsituation is useful.

For laser machining processes defocusing is animportant parameter. Defocusing can be variedcontinuously but that it may be the range is narrowbecause much lower intensity of laser beam to theworkpiece surface compromise the welding

process. Small variations in defocus make that itseffect can not be seized. Experimental situationpresented in the paper is a case in whichdefocusing variations are obtained withoutexcessive lowering laser beam intensity to theworkpiece surface.

Using mathematical modeling and statisticalanalysis of the changes is a powerful tool forunderstanding the causes of variations for analyzedsizes and control the welding process.

References[1] Kazuhiko Ono, Karoru Adachi, Yasuichi

Matsumoto, Isamu Miyamoto, Takashi Inoue,Ryuichi Narita, Influence of oxide film on weldcharacteristic of mild steel in CO2 laserwelding, ICALEO 1999

[2] Benyounis, A.G. Olabi, M.S.J. Hashmi, Multi-response optimization of CO2 laser-weldingprocess of austenitic stainless steel, Optics &Laser Technology 40 (2008) p:76–87.

[3] Lung Kwang Pan, Che Chung Wang, ShienLong Wei , Hai Feng Sher, Optimizingmultiple quality characteristics via Taguchimethod-based Grey analysis, Journal ofMaterials Processing Technology 182 (2007)p:107–116.

.


223

U.S. QUENCHING AND DIMENSIONAL STABILITY IN TIMEOF 100Cr6 STEEL

Nicolai BANCESCU1, Constantin DULUCHEANU1, Traian Lucian SEVERIN1

1 tefan cel Mare” University of Suceava, Faculty of Mechanical Engineering, Mechatronics andManagement, University Street, No. 13, 720225, Suceava, Romania,EU. [email protected],

[email protected], [email protected]

Abstract: Due to the relatively high susceptibility to deformation and cracking, the bearings steelcreates major problems to quenching, especially in the case of complex geometric configurations. On theother hand, the dimensional stability of bearing elements influences significantly bearing durability. Inthis paper, the authors present a series of experimental results on the dimensional time behavior of100Cr6 steel for several martensitic volume quenching.

Keywords: quenching, bearing steel, dimension stability.

1. General consideration

Sustainability of precision bearings largelydepends on the dimensional time stability providedby the structure elements obtained from the finalheat treatment. In this case the notion of structuredefines the nature, form, size, distribution ofphases and structural constituents also the size andsign of residual stresses induced in productfollowing due to the heat treatment applied.

Martensitic quenching is a heat treatment of thetechnologies that induces stress and strain indimensional stability with implications for longperiods of time. The transformation of austeniteinto martensite in quenching induces large amountsof residual material stress, due to the difference inspecific volume of the constituents (martensite hashigher specific volume than austenite from whichit came). Finally, on the quenching product surfacetensile stress is formed which is summedalgebraically with the tensions caused by thermalshock. Parts of the residual stresses are"downloaded" during its return through specificmechanisms. On the other hand in the structure ofquenching steel the residual austenite andmartensite phases are out of balance and sounstable over time. These phases tend to evolveover time, even at ambient temperature, the phaseconstituents close to balance or causingdimensional changes in some operational situationsaffecting the product behavior in exploitation.

Collective concerns of the thermal treatmentslaboratory FIMMM Suceava on optimization ofheat treatment of bearing steel dates back to 1985.In the article are presented results of researchregarding the influence of quenching regime fordimensional stability in time of 100Cr6 steel(symbolizing the ancient RUL1).

2. Research methodology

Typically, dimensional stability in time isexpressed as relative linear strain variation overtime; the measurements are made until the fullstabilization of the dimensions. To study thedimensional stability in time was designed andimplemented a device like in figure 1. With fiveworkstations, the device allows embedding ofcylindrical specimens with Ø10 x 150 mm,between a high rigidity wall and an elastic wall onthat are bonded strain gauge.

Figure 1. Device for testing dimensional stability

mailto:bancescu:@fim.usv.ro

mailto:dulu:@fim.usv.ro



224

To limit thermal deformation caused bychanges of ambient temperature environment aftermounting the samples device are placed in athermostatic chamber in which temperature ismaintained at 20 °C with an accuracy of ± 0.2 °C.Strain gauges in full bridge are connected to fivebridges N2305, selected in advance as stabilitywhile balancing adjustment. In order to ensurestable mechanical properties over time afterimplementation, the device was subjected tothermal stress relief treatment of 100 hours at 200°C. Immediately after heat treatment, samples weremounted in the device, gathering performing with

torque indicator handle wrench to maintain thesame measurement conditions. After placing thedevice in the thermostat room and aftertemperature stabilization was done to balance eachstrain gauge bridge. Calibration of each station wasdone with a orthotest. We worked with lots of fivespecimens for each group thermal treatments arepresented in Table 1. Austenitizarea for quenchingwas performed in a CARBOLYTE electric furnaceheated to 850 ºC, keeping time was 40 minutes. Ascooling environment mineral oil Lubrifin MET 1 Rtype II was used, recommended for quenchingbearing elements.

Tabel 1. Quench cooling variants of the samples groupGroupcode Getting austenite Mineral oil cooling (40 ºC)

LUBRIFIN MET 1 R Tip II Draw back Runningmaintenance

Cl 850 ºC/40 minutes Mechanical agitation - -Cl + R 850 ºC/40 minutes Mechanical agitation 150 ºC 2 h

Mg 850 ºC/40 minutes Continuous magnetic field 300 G - -Mg + R 850 ºC/40 minutes Continuous magnetic field 300 G 150 ºC 2 h

US 850 ºC/40 minutes Ultrasonic field 40,4 kHz and 4W/dm2 - -US+R 850 ºC/40 minutes Ultrasonic field 40,4 kHz and 4W/dm2 150 ºC 2 h

For magnetic field quenching was used an oiltank with 40 dm3 capacity, fiberglass located in thecenter of a DC powered coils. In the center of thebasin is done a magnetic field strength of 300Gauss, [1].

Ultrasonic field quenching was made in a tankwith a capacity of 40 dm3, stainless steel sheet, inwhich at the bottom was placed a piezoelectrictransducer with U.S. power of 200 W and theresonant frequency of 40.4 kHz. Power transducerwas made from an electronic generator, developedin the laboratory, capable of ensuring a U.S. fieldstrength of 4 W/dm2, [1].

Quality assessment was done by measuring theheat treatment hardness.


Figures 2.a. and 2.b. are the results of hardnessmeasurements and their statistics. From theiranalysis results the superiority of U.S. quench fieldover other hardening technologies experienced.

Measurement of dimensional stability in timewas made in a period of 204 hours. After this timeinstallation was found to stabilize the specimensdimensional. Measurement results are presented inTable 2 and chart in Figure 3.

a

bFigure 2. Results of hardness measurements for heat-treated groups: a) hardness values chart; b) statisticalmeasurements.


225

Tabel 2. Dimensional sability over time for different variants of heat treatment [ m]Time[h] Cl Cl.+R. US US+R. Mg Mg+R

0 1,33 1,33 0,66 0,66 1,33 1,3312 2,66 2,66 0,66 2,66 2,66 4,0024 4,00 2,66 0,66 2,66 2,66 4,0036 4,00 4,00 1,33 3,33 4,00 5,3348 5,33 5,33 1,33 4,00 4,00 6,6660 5,33 6,66 2,66 4,00 4,00 8,0072 6,66 6,66 4,00 5,33 5,33 8,0084 6,66 6,66 5,33 5,33 5,33 8,0096 6,66 6,66 5,33 6,66 6,66 8,00108 8,00 8,00 6,66 6,66 6,66 8,00120 9,33 9,33 6,66 6,66 6,66 9,33132 10,66 9,33 8,00 8,00 8,00 9,33144 10,66 10,66 9,33 8,00 9,33 9,33156 12,00 10,66 10,66 8,00 9,33 9,33168 13,33 12,00 11,33 8,00 10,66 10,66180 13,33 13,33 12,00 8,00 10,66 10,66192 13,33 13,33 12,00 8,00 10,66 10,66204 13,33 13,33 12,00 8,00 10,66 10,66

Figure 3. Variation of dimensional changes in time of the specimens heat treated [ m].

4. Conclusions

1. Quenchig technologies significantlyinfluence the structure and properties of bearingsteel. The amount of residual austenite in thestructure, characteristics and size of residual stressof martensite is directly influenced by the

parameters of the cooling operation.Metallographic structural analysis for variousquenching regimes tested (Figure 4) is clearlyobserved differences between the quantity anddistribution phases.


226

a

b

c

Figure 4. Metallographic structure of quenchedsamples: a) classic quenching b) magnetic fieldquenching, c) U.S. quenching field. MO x 400

2. Activating U.S. cooling environmentaldrastically reduce calefaction on cooling anduniform heat exchange intensity in the first part ofcooling, as confirmed by less dispersion ofhardness values, as apparent from Figure 2.b.Furthermore, the U.S. vibration frequency of thefluid cooling determines the cooling capacitygrowth, and hence the cooling rate directly affectsthe amount of residual austenite in the structure.Compared with other types of heat treatment, U.S.

quench ensure sensitive reduce of the residualaustenite amount. (Figure 4.c.).

3. Mineral oil used as a cooling environmentcan be considered incompressible and ensuretransfer of mechanical energy from the transducerto the product. Additional energy intakedetermines the transformation of additionalaustenite quantities into martensite at quench withdirect implications on the content of residualaustenite, [2, 3].

4. Following the additional energy intake, evenduring the quench, product stress relief is realized,phenomenon confirmed by experimental research,[1].

5. Dimensional stability over time is directlyinfluenced by the proportion of phases andconstituents in the metallographic structure. Inbearing steel case is mainly on the amount ofresidual austenite in the structure, [1]. Amongvariants examined by heat treatment, the U.S.quenching ensures the most dimensional stabilityin time by dramatically reducing the amount ofresidual austenite in the structure in the first place.

6. As an alternative technology at industrialscale, the U.S. quenching has the advantage that itnot requires major investment or significantchanges in heat treatment lines or excessive energyconsumption.

7. U.S. heat treatment oil activation reduces thenegative effect of waste water and increases itsuse.

References

[1] Bancescu, N., Studii si cercet ri privindcresterea durabilit tii contactului cu rostogolireprin tratamente termice neconventionaleaplicate tribostratului, Tez de doctorat,Universitatea „Dun rea de jos” Gala i, 1995.

[2] Saban, R., Dumitrescu, C., Petrescu, M., Tratatde Stiinta si ingineria materialelor, Academia deStiinte Tehnice din România, Editura AGIR,Bucuresti 2006.

[3] Amza, Gh., Ultrasunetele. Aplicatii active,Academia de Stiinte Tehnice din România,Editura AGIR, Bucuresti 2006.


227

INFLUENCE OF POLYMER CONCENTRATION ON THEPERMEATION PROPERTIES OF NANOFILTRATION MEMBRANES

Balta Stefan1,2, Bodor Marius1, Benea Lidia1,2

1Universitatea Dun rea de Jos din Gala i, Facultatea de Metalurgie i tiin a Materialelor,2Centrul de Cercetare Interfe e – Tribocoroziune i Sisteme Electrochimice, Romania.

E-mail: [email protected]

Abstract: Driving force membrane processes seem to be most useful for water treatment. Membranes arevery effective in removing a wide variety of water contaminants. Therefore, the use of these processes in waterpurification to replace or to improve conventional treatment has increased. An inherent problem of membranes isfouling, the accumulation of materials (foulants) near, on, or within the membrane that causes a reduction in theamount of product water over time. As a result of fouling, capital and operating costs of membrane systems arehigher, making them less attractive. A membrane is an interphase between two adjacent phases acting as aselective barrier, regulating the transport of substances between the two compartments. The main advantages ofmembrane technology as compared with other unit operations in (bio) chemical engineering are related to thisunique separation principle, i.e. the transport selectivity of the membrane. Separations with membranes do notrequire additives, and they can be performed isothermally at low temperatues and - compared to other thermalseparation processes - at low energy consumption. Nanofiltration (NF) separate or remove small molecules orions from a solvent (most often water) by means of pressure – driven filtration through a dense polymericmembrane. Combination of selectivity with a high permeability to water and mechanical strength sufficient towithstand high pressures is achieved by using thin film composite membranes comprising a dense film of 10-200nm (active layer) supported by a thick asymmetric porous film. In this paper is describing the manufacturingprocesses of Polyethersulfone membranes (PES). A Polyethersulfone membranes was made with differentconcentration of polymer in N-Methyl-pyrrolidone (NMP) solvents. The influences of the polymer concentrationon the membranes permeation properties were studied. After the preparations all membranes were studied for acomparison with cross flow and dead-end equipments to see the flux and permeability of pure water. Thepermeation results and the SEM photography show the influence of the polymer concentration, increasingconcentration permeation properties are decreasing.

Keywords: membrane, nanofiltration, PES

1. IntroductionNanofiltration (NF) is widely applied in the

treatment of waste water and in the production ofdrinking water [1,2] because provide a feasibleprocess allowing a high retention of multivalentions as well as organic molecules. However, one ofthe main drawbacks of the nanofiltrationperformance is the fouling phenomenon, usuallyattributed to adsorption of organic substances onthe membrane surface. Membrane fouling lead indiminished the membrane performance, seriousdeficient production, and excessive operating costs[6,7]. Because of the fouling, the dyes rejection [8-10] and the permeation properties [11-13] of themembranes decrease due to a higherhidrophobicity of the membrane surface.Membrane fouling depends by the membrane

characteristics [14-17] and by the filtration mode(cross-flow or dead-end filtration) [18].In order to increase the effectiveness ofnanofiltration membranes, some properties such ashydrophilicity and fouling resistant should beimproved. The membranes were synthesized atfour different polymer concentrations 25, 27, 30and 32 wt.% by the phase inversion method. Thismethod means that after dissolution of the polymerin a solvent, the polymer solution has to be cast toa thin film with different thickness on a supportlayer. In our case the thickness of the polymerlayer was 250 µm. The support layer, with the thinpolymer film on it, is then immersed in a non-solvent bath, deionized water. Due to the diffusionof the non-solvent in the polymer film, the polymersolution becomes thermodynamically unstable,resulting in two phases: a polymer-poor phase (the

mailto:Stefan.Balta:@ugal.ro


228

pores of the membrane) and a polymer-rich phase(the matrix).To make a good comparison ofmembrane characteristics some properties likehydrophobicity, permeability and morphologywere studied.

2. Experimental2.1. MaterialsThe solvent used was 1-Methyl-2-pyrrolidone

and he support layer (type Viledon FO2471) usedfor the membrane manufacturing was obtainedfrom Freudenberg (Weinheim, Germany). Thepolymer, Polyethersulfone type Radel, wassupplied by Solvay (Belgium) and was used as thebase polymer. To determine the membranes fluxand the permeability was used distillated water.

2.2. Membrane preparationNET polymeric membranes were manufactured atfour different concentration of polymer (25, 27, 30and 32 wt.%) in NMP, using the phase inversioninduced by immersion precipitation method.Preliminary experiments made by othersresearchers showed that the membrane with 30 and

32wt.% of PES are the most suitableconcentrations to obtain NF membranes.

The casting solution was obtaining addingpolymer in the solvent solution and mixed at 400Con the mechanical stirring at 200 rpm for 24 hours.On a polyester support a thin film of the polymersolution with a thickness of 250 m was cast witha filmograph (K4340 Automatic Film Applicator,Elcometer). Membrane was immersed in distillatewater for precipitation and after 15 minutes waswashed to remove the excess of solvent. For everytype of membrane, four different solutions weremade and from every solution three membraneswas manufactured and tested to obtain the truevalues of the membranes properties.

2.3. Filtration experimentTo study the performance of the membranes,

permeability and flux, were used a dead-end (Fig.1a.) and cross flow filtration (Fig. 1b.)installations.

Figure 1. Filtration equipment: a) dead-end and b) cross flow

The pure water flux experiments were carriedout with a commercial cross flow unit onlaboratory scale. The permeability of the preparedmembranes were studied using two dead endmodules Sterlitech HP4750 at the roomtemperature and desirable pressure. The pressurewas realized with a nitrogen cylinder and apressure regulator, connected to the dead-end cell.The solution volume used for every experimentwas 250 ml and the permeate was collected in agraduate cylinder. The pure water flux wasdetermined at 10 bar pressure and the time wasmeasured at every 5 ml of permeate.

To determine the pure water permeability(PWP) was measured the water flux (Jw) at sixdifferent pressure ( P) from 5 to 20 bar. The PWPwas calculated by the following equation:

PJ

PWP w 1)

To determine the pure water flux andmembranes behavior for a long time all themembranes was tested in a cross flow installation.All experiments were realized at 24 oC and theapplied pressure was 8 bar. The membranes


229

surface area was 0.0059m2 and the time for everyexperiment was 24 hours.

2.4. Characterization of the membranesurface and morphology

To study the hydrophilicity/hydrophobicity ofthe membranes was used a Drop Shape AnalysisSystem DSA 10 Mk2 (fig. 2b.). On the cleaned anddry membrane surface was placed a distillate waterdroplet of 2 µl and the contact angle between themembrane surface and the droplet was calculated(fig. 2a). The final value o the contact angle forevery type of membranes was the average of 21measurements, seven determinations for threedifferent membranes.

Figure 2. Contact angle measurements: (a) theprinciple and (b) the setup

Scanning electron microscopy (SEM)measurements were performed for characterizationof the surface and cross-section of the membranes.For the cross-section analysis the samples wasprepared by fracturing the membranes in liquidnitrogen and sputtered with gold. The images weremade with a Philips XL30 FEG and Phylips FEI,QUANTA 200 instruments. . Surface SEM imageswere made with a Phylips FEI, QUANTA 200instrument with an accelerating voltage of 20 KeV

3. Results3.1. Pure water flux and permeability

For the determination of pure water flux thefiltration experiments were carried out with acommercial nanofiltration unit on laboratory scale.In all experiments the applied pressure was 8 barand the temperature was 240C. To minimizeconcentration polarization a feed velocity of 4.0m/s was used. The membrane surface area was0.0059m2. The evolution of flux was followed intime during 24 h. For analysis and comparison thevalues after 24 h of filtration were used.

Figure 3 shows that the permeation flux of PESmembranes. The polymer concentration has animportant influence on the pure water flux. Whenthe polymer concentration decrease the pure water

flux increase. The best membranes in term of waterflux are at 25 wt% of PES.

0

200

400

600

25 27 30 32

Flux

[J(L

/m2h

)]

PES concentration (wt.%)

Figure 3. Pure water flux at different concentration ofPES

Membranes with 25% of PES have a goodpermeability but because of the weaker mechanicalresistance have an important instability of flux intime (figure 4). Hence, the increase of polymericmaterial enhances the membrane mechanicalresistance.

Figura 4. Pure water flux for different PESconcentration

0

5

10

15

20

25

30

35

25 27 30 32

Perm

eabi

lity(

Lm-2

H-1

bar1

)

PES concentration (wt.%)

Figure 5. Permeability at different PES concentration.

Increasing the concentration of PES themembranes pore size decrease and in consequencepermeability decrease. Figure 5 confirm the sameinfluence of the polymer concentration on thepermeability like in the case of pure water flux.

0.00

100.00

200.00

300.00

400.00

500.00

600.00

700.00

5 10 15 20 25 30 35 40 45 50 70 75 80 85 90 95 100

Volume [ml]

Flux

[J(L

m2h

)]

32 NMP

30 NMP

27 NMP

25 NMP


230

Membrane at 25 wt.% of PES have thehigher permeability but because of them instabilityin time are not the ideal membranes to be selectedfor future experiments. Membranes a 27wt.% ofPES appear to be the best membranes fornanofiltration experiment.

3.2. Contact angleContact angle determination is a well-known

method to study the membrane surfacehydrophobicity, since a hydrophilic membranesurface gives rise to a low contact angle [16].

Figure 6 shows the measured contact angles forneat PES membranes for four different polymerconcentrations, indicating that membranehydrophilicity increases as decreases the polymerconcentration. The effect of polymer concentrationon membrane hydrophilicity should be explainedin terms of pore size and porosity considering.

Results are in concordance with permeationproperties. Membrane with 25 wt.% of PES havethe most hydrophilic surface.

Figure 6. Contact angle at different PES concentration

3.3. Membrane morphology characterizationThe permeation properties of neat membranes

can be better explained by the Scanning ElectronicMicroscopy (SEM) analyses. Figure 7 presentsSEM images of the cross-sections of 25, 27, 30 and32 wt.% PES. From the SEM images is observedthat polymer concentration has a clear effect on themembrane structure, which can be described interms of membrane pore size and porosityvariations

Figure 7. Cross section SEM photography of membranes at different PES concentration: a) 25%PES, b)27%PES,c)30%PES, d)32PES

Membranes with 25% of PES (figure 7 a) havemacrovoids in the structure who lead to a betterpermeability but, how was observed at the purewater flux, with an instability of flux in time. Thepore size and geometry change as the polymerconcentration increase, figure 7 a,b,c,d for 25, 27,30 and respectively 32 wt.% of PES, in the sametime suppress the macrovoid and increase the

thickness of the top layer [17]. For 30 and 32 wt.%of PES the permeability and pure water flux issmaller because the bottom with the top layer arenot connected by macrovoids and a spongestructure is formed. For membrane with 27 wt.% ofPES the bottom is connected with the top layer,porosity is uniform distributed and the poresstructure are like fingers.


231

Figure 8. Surface SEM photography at different concentration of PES: a) 25, b) 27, c) 30 and d) 32wt.%.

Figure 8 show surface SEM photography ofmembranes at different concentration of PES. Byincreasing the concentration of PES it wasobserved that the porosity of the membranesdecreases. Membranes with 25 wt.% of PES arethe most porous in comparison with membranes athigher concentration.

4. ConclusionsA systematic study of influence of polymer

concentration was carried out, testing a largenumber of membrane samples.

The polymer concentrations have a highinfluence on the membranes properties and have anegative effect on the water permeation andhydrophobicity. The results from permeationexperiment and from analysis of membranemorphology show the negative influence ofpolymer concentration. At 25 wt.% of PES,membranes have a good permeability but fluxshow instability in time. Because of this behaviourfor industrial application membrane with higherconcentration of PES need to be selected.

AcknowledgementsStefan Balta would like to acknowledge the

support provided by the European Union,Romanian Government and Dunarea de JosUniversity of Galati, through the project POSDRU– 6/1.5/S/15.

Bodor Marius would like to acknowledgethe support provided by the European Union,Romanian Government and “Dun rea de Jos”University of Gala i, through the projectPOSDRU – 107/1.5/S/76822.

References[1]. IC. ESCOBAR, AA. RANDALL, SK.

HONG, Removal by Nanofiltration: Full andBench-Scale Evaluation, J Water Sup ResTechnol Aqua 51 (2002) 67–76.

[2]. A. MATILAINEN, R. LIIKANEN, M.NYSTROM, Enhancement of the naturalorganic matter removal from drinking waterby nanofiltration, Environ Technol 25 (2004)283–91.

[3]. E. Matthiasson, B. Sivik, Concentrationpolarization and fouling, Desalination 35(1980) 59.


232

[4]. D.E. Potts, R.C. Ahlert, S.S.Wang, A criticalreview of fouling of reverse osmosismembranes, Desalination 36 (1981) 235.

[5]. B. Van der Bruggen, M. Manttari, M.Nystrom, Drawbacks of applyingnanofiltration and how to avoid them: Areview, Sep Pur Tech 63 (2008) 251–263

[6]. B. Van der Bruggen, L. Braeken, C.Vandecasteele, Flux decline in nanofiltrationdue to adsorption of organic compounds,Sep Pur Tech 29 (2002) 23–31

[7]. L D. Nghiem, PJ. Coleman, C. Espendiller,Mechanisms underlying the effects ofmembrane fouling on the nanofiltration oftrace organic contaminants, Desalination 250(2010) 682-687.

[8]. L. Braeken, R. Ramaekers, Y. Zhang, G.Maes, B. Van der Bruggen, C.Vandecasteele Influence of hydrophobicityon retention in nanofiltration of aqueoussolutions containing organic compounds . J.Membr. Sci. 252 (2005) 195-203.

[9]. JM. Arsuaga, MJ. López-Muñoz, A. Sotto,Correlation between retention and adsorptionof phenolic compounds in nanofiltrationmembranes, Desalination 250 (2010) 829-832

[10]. L. Braeken L, K. Boussu K, B. Van derBruggen B, C. Vandecasteele, Modeling ofthe adsorption of organic compounds onpolymeric nanofiltration membranes insolutions containing two compounds,Chemphyschem 6 (2005) 1606-1612.

[11]. J.M. Laine, J.P. Hagstrom, M. Clark, M.Mallevialle, Effects of ultrafiltrationmembrane composition, J. Am. WaterWorks Assoc. 81 (1989) 61–67.

[12]. M. Elimelech, X. Zhu, A.E. Childress, S.Hong, Role of membrane surfacemorphology in colloidal fouling of celluloseacetate and composite aromatic polyamidereverse osmosis membranes, J. Membr. Sci.127 (1997) 101–109.

[13]. C. Jucker, M.M. Clark, Adsorption ofaquatic humic substances on hydrophobicultrafiltration membranes, J. Membr. Sci. 97(1994) 37–52.

[14]. J. Cho, G. Amy, Interactions between naturalorganic matter and membranes: rejection andfouling, Water Sci. Technol. 40 (9) (1999)131–139.

[15]. D.B. Mosqueda-Jimenez, R.M. Narbaitz, T.Matsuura, Membrane fouling test: apparatusevaluation, J. Environ. Eng. ASCE 130 (1)(2004) 90–98.

[16]. C.R. DIAS AND M.N. DE PINHO, Waterstructure and selective permeation ofcellulose-based membranes, J. Mol. Liquids,80 (1999) 117–132.

[17]. H. Susanto and M. Ulbricht. PolymericMembranes for Molecular Separations.Membrane Operations. InnovativeSeparations and Transformations. WILEY-VCH, Weinheim. 2009.

TEHNOMUS - New Technologies and Products in Machine Manufacturing Technologies"

233

STATIC DEFORMATION OF A WORKPIECE FIXED INUNIVERSAL CHUCK AND LIFE CENTRE

Lauren iu Sl tineanu1, Lucian T caru2, Margareta Cotea 3, Teodor Tilic 4, MihaiBoca5, and Irina Grigora (Be liu) 6

1, 2, 3, 4, 5”Gheorghe Asachi” Technical University of Ia i - Romania, [email protected],[email protected], [email protected], [email protected], [email protected],

6University “ tefan cel Mare” of Suceava- Romania, [email protected].

Abstract: The rigidity could be defined as the capacity of a system or subsystem to oppose to theelastic deformation generated by the action of external forces. In this paper, the authors highlight somepossibilities to evaluate the static rigidity of the technological system corresponding to a turning process.Some remarks were elaborated by considering the workpiece as a bar fixed at one end and simplysupported at the other end. Experimental tests were developed to measure the static elastic deformation ofthe workpiece under the action of a variable loading force.

Keywords: elastic deformation, rigidity, turning process, deflection, applied force

1. IntroductionThe rigidity of a solid body is the capacity of

this body to oppose to the deformation generatedby the action of external forces [2, 3, 4]. Of course,there is the possibility to evaluate the rigidity of anindividual part, but there is also the possibility totake into consideration the rigidity of a system orsubsystem. In the case of the so-called classicaltechnological system constituted of the machinetool – tool – device – workpiece, the rigidity willrefers to its capacity to oppose to the deformationgenerated by the cutting process.

There are various methods of evaluating therigidity of the technological systems; thus, thereare static and dynamic methods. In the case of thestatic methods, a way to generate a force similar tothat appearing during the cutting process is used.As known, in the case of turning process, the maincutting force can be decomposed in componentsalong the main axes belonging to a coordinate axissystem attached to the technological system. Thedeformations appear for each axis along which acomponent of the cutting force acts. Thesedeformations can be measured and by determiningthe ratio between the force F and the deformation xgenerated along the direction of the force, animage about the static system rigidity along theaxis Ox is obtained. In some countries, there arestandardized methods for the static evaluation ofthe technological system; some recommendations

concerning the static evaluation of the lathe areincluded in an old Romanian standard.

On the other hand, there are methods in whichthe deformation generated by a real cutting processare determined; because there are some difficultiesin measuring the deformation during the cuttingprocess, the experimental researches were directedto the measuring the dimensions of the machinedsurfaces. By taking into consideration the results ofthe measuring the dimensions of the machinedsurfaces, a certain mathematical processing ofthese results allows the evaluation of thetechnological system rigidity in dynamicconditions.

Information concerning the rigidity of thetechnological system are important when someprecise machining processes must to be applied;this means that during the design of the cuttingprocess, the establishing of the depth of cut ap hasto take into consideration the deformation of theworkpiece, so that finally the dimensions of themachined surfaces be inscribed in a narrowtolerances field.

Some decades ago, Korsakov developed athorough theoretical study concerning the elasticdeflection of the technological system in the caseof the turning process; various possibilities offixing the workpiece were taken into consideration[2]. Jianliang and Rongdi [1] proposed a unitedmodel for predicting diametral error of slender bar,by applying a finite analysis to highlight theworkpiece deflection, when a follower rest is used

mailto:slati:@tcm.tuiasi.ro

mailto:ltabacaru:@tcm.tuiasi.ro

mailto:mcoteata:@tcm.tuiasi.ro

mailto:teodortilica:@gmail.com

mailto:mihaitzaboca:@yahoo.com

mailto:grigorasirina_sv:@yahoo.com


234

during a turning process. They concluded that thediametral error mainly depends on the location offollower rest, depth of cut and feed rate.

Having the intention to more detailed study theinfluence exerted by the technological systemrigidity on the machining accuracy, this paperpresents some considerations concerning the staticevaluation of the technological system in the caseof a turning process.

2. Initial hypothesizesOne of the machine tools existing in the

majority of the mechanical workshops is the lathe;essentially, the turning supposes a main rotationmotion of the workpiece, while the toolmaterializes a feed motion in the axial plane of theworkpiece. In such a situation, the turningaccuracy depends on the accuracy of the motionsperformed by the workpiece and the tool in relationwith a fix coordinates system; in our case, such acoordinates system could be defined byconsidering the rotation axis of the workpiece asOz axis (fig. 1, a).

There are three ways to clamp the workpiece.Thus, although it does not ensure a high alignment

a

b

c

dFigure 1: Workpiece deformation as a bar rigid fixed at one end and simply supported at the other end


235

of the machined surfaces when the workpiece must

be twice clamped, the positioning of the workpiecein the universal chuck and the life centre is themost frequently used in the case of the longworkpieces. A better accuracy from the point ofview of alignment when two positions of theworkpiece are necessary and the workpiece is longenough corresponds to the positioning between lifecentres; for such reasons, the workpiece clampingbetween the life centres is preferred within thefinishing cutting operations (turning, grinding).The both above mentioned ways of the workpiecepositioning suppose the existence of the centreholes on the frontal surfaces of the workpieces.The third way to clamp the workpiece, applicablein the case of short workpieces is by means only ofthe universal chuck; obviously, if two clampingsare necessary, there is a high probability that thealignment of the machined surfaces at the twoclampings not be high enough.

Within this paper, only the case of theworkpiece clamping in the universal chuck and lifecentre will be considered, this being widely usedway in industrial practice, although, as abovementioned, this way does not ensure a highaccuracy from the point of view of the alignmentof the surfaces obtained by two clampings.

During the turning process, the workpiece andthe other components of the technological system(machine tool, devices, tools) are affected bydeformation phenomena; the presence of thecutting forces determines the deformation of theabove mentioned mechanical components. Ofcourse, only elastic deformation could be accepted;the operation conditions must be established sothat the internal stresses to not exceed the limit ofelastic deformation of the material from which thecomponents of the technological system are made.

Due to this elastic deformation of thetechnological system components, the size of theactual depth of cut is not the same with the desiredone; under the action of the cutting forces, thecomponents are affected by elastic deformationsand the current depth of cut apc is lower that thedesigned depth of cut apd (fig. 2) :

pdpc aa (1)

This relation could suggest a way ofincreasing the machining accuracy; if thedifferences between the current depth of cut apcand the desired depth of cut ap are known, thedesigned depth of cut could be so established thateven by taking into consideration the elasticdeformation of the technological systemcomponents, an increased machining accuracy beobtained. A way used to obtain informationconcerning the elastic deformation of thetechnological system components is based on thestudy of the static rigidity which characterize thetechnological system. It is expected that the biggerthe technological system rigidity is, the higher themachining accuracy is.

The situation of the workpiece fixed in theuniversal chuck and the life centre (fig. 1, a) couldbe considered as similar to a bar rigid fixed at oneend and simply supported at the other end (fig. 1,b). A coordinate system xOyz could be taken intoconsideration; the origin of this coordinate systemcould be placed at the intersection of the rotationaxis with a frontal plane corresponding to the freesurface of the jaws of the universal chuck; this isthe case of the turning an external cylindricalsurface on a lathe, by using the mechanicallongitudinal feed.

The cutting force generated by the turningprocess can be decomposed in componentsdirected along the axes corresponding to thecoordinate system xOyz (fig. 3).

Figure 3: Components of the cutting force which actson the lathe tool during the turning process.

Figure 2: Decreasing of the depth of cut during theelastic deformation during the turning process.


236

The position of the point where the cuttingforce is applied changes during the turning processsimultaneously with the position of the lathe tooltip; this means that just the deformation of thetechnological system changes, by taking intoconsideration the position of the lathe tool tip.

As consequence of the turning process, thecross section of the workpiece is not the samealong the workpiece (the machined surface has alow diameter in comparison with the not machinedsurface), but in order to simplify the theoreticalmodeling, one can consider a workpiece having thesame diameter along all its length. In fact, in thecase of the workpieces having big diameters, thedecrease of the diameter as consequence of theturning process is not significant and theconsidering the same diameter along the entireworkpiece does not generate significant errors.

Also in order to simplify the mathematicalmodel, only the action of the cutting forcecomponent Fx will be considered; in real turningconditions, elastic deformations are generated yetalong the axis Oy and Oz, but one can appreciatethat from the point of view of the diametralaccuracy, they exert a less significant influence.

For a bar rigid fixed at one end and simplysupported at the other end, the relation valid for theelastic radial deflection is given [4] by the relation:

3

23

124

lEIzlzlzFf

z

x(2)

where Fx is the size of the cutting force componentdeveloped along a direction parallel to the Ox axis,z is the distance along the Oz axis from thecoordinate system origin to the tool lathe tip, l - thetotal length of the bar, E - the bar materialelasticity modulus and Iz – the second moment area

of the cross section corresponding to the barworkpiece.

The elastic deformation is not the same alongthe workpiece; just the different fixing of theworkpiece at the two end (considered as a bar rigidfixed at one end and simply supported at the otherend) determine the changing of the elasticdeflection along the workpiece. On the other hand,just the universal together with the main shaft ofthe lathe are affected by an elastic deformation;this means that the points where the workpiece isfixed at the two ends have not a fix position. Thismeans that if one tries to measure the elasticdeformation of the workpiece near the universalchuck, this deformation will be different incomparison with the elastic deformation generatednear the life centre (fig. 1, c and d).

3 Experimental testingIn order to verify the validity of some of the

previous mentioned considerations, anexperimental testing was developed [5]. Auniversal lathe type SNA 500x1000 was used(workpieces having diameter up to 500 mm and alength up to 1000 mm can be machined on such alathe); a workpiece having the diameter of 45 mmand a length of 817 mm was clamped in theuniversal chuck and the life centre (fig. 4).

Several dial gauges were placed in order toobtain an image concerning the elastic deformationof the workpiece (fig. 5), of the lathe main shaft inthe position corresponding to the universal chuckand of the life centre. We tried to place the holderof the dial gauges on the lathe bed. An elasticdeflection could also affect the lathe bed, when aforce is exerted on the technological system, butdue to the high rigidity of the lathe bed, this elasticdeformation is very low and one can consider thatthis deformation can be neglected.

Figure 4: Measuring the elastic deformation of the technological system components


237

The experiments were developed by applyingan increasing force Fx in various points along theworkpiece fixed in the universal chuck and livecentre; some of the experimental results wereincluded in the table 1. As one can see, in thecolumns no. 3, 4, 5, 6, 7, 8 there are the sizes of thedeformations corresponding to the universal chuck,to the life centre and to the tool holder; thedeformation registered by the dial gauges used tomeasure the deformations of the test piece in fourestablished places were inscribed in the columnsno. 9, 10, 11, 12, 13, 14, 15 and 16. For each case,there were measured the deformations both duringthe system loading (when the force increases) andunloading (during the decrease of the force size).

In order to obtain an image about the rigidity ofthe technological system at the level of theuniversal chuck and the live centre, the forceexerted on the test piece can be devised in twocomponents which act on the universal chuck andon the live centre, by taking into consideration the

distances among the point of applying the forceand the points where the test piece is supported.The moment equations allows the determining thecomponent Fx1=102 daN which acts on theuniversal chuck and Fx2=138 daN (this componentacts on the live centre) when the force applied tothe workpiece has the size Fx=240 daN.

Knowing the maximum sizes of the force Fxcomponents and of the elastic displacements x(from table 1) affecting the universal chuck(x1max=0.0242 mm) and the live centre (x2mx=0.065mm), the system rigidity at the level of thesesubassemblies could be estimated:

xFR x

x

(3)

By means of the relation (3), one obtainRx1=4214 daN/mm for the universal chuck andRx2=2123 daN/mm for the life centre. Taking intoconsideration the maximum elastic displacement of

Table 1: Elastic deformations during the increasing and decreasing of the radial forceapplied at the distance z=470 mm

Exp.No.

ForceFx,daN

Elastic deformationsUniversal

chuckLife centre Tool holder z=70 mm z =240 mm z =470 mm z =770 mm

Loa-ding

Un-loa-ding

Loa-ding

Un-loa-ding

Loa-ding

Unloa-ding

Loa-ding

Unloa-ding

Loa-ding

Unloa-ding

Loa-ding

Unloa-ding

Loa-ding Unloa-ding

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 161 0 0 0 0 0.0195 0.00 0.000 0 0.003 0 0.00 0.000 0.000 0.0000 0.00762 20 0.0005 0.003 0.002 0.026 0.00 0.033 0.010 0.0135 0.025 0.030 0.025 0.040 0.0203 0.02673 40 0.002 0.0045 0.006 0.033 0.025 0.050 0.020 0.0240 0.050 0.059 0.060 0.075 0.0356 0.04574 60 0.004 0.0065 0.010 0.037 0.035 0.058 0.029 0.033 0.075 0.075 0.093 0.110 0.0533 0.06355 80 0.006 0.009 0.018 0.040 0.040 0.065 0.039 0.042 0.100 0.110 0.130 0.150 0.0711 0.08266 100 0.0082 0.011 0.020 0.045 0.050 0.070 0.039 0.044 0.125 0.140 0.170 0.190 0.0876 0.09917 120 0.0105 0.013 0.028 0.050 0.055 0.076 0.040 0.044 0.155 0.165 0.210 0.220 0.1087 0.11689 160 0.0155 0.0175 0.036 0.060 0.065 0.083 0.040 0.044 0.210 0.220 0.290 0.300 0.1448 0.156210 180 0.0175 0.0195 0.044 0.064 0.070 0.087 0.040 0.044 0.235 0.245 0.325 0.340 0.1638 0.175311 200 0.0195 0.022 0.051 0.065 0.075 0.089 0.044 0.044 0.250 0.270 0.365 0.380 0.1829 0.193012 220 0.022 0.024 0.056 0.065 0.080 0.090 0.044 0.044 0.250 0.295 0.405 0.415 0.2007 0.208313 240 0.0242 0.0242 0.065 0.065 0.090 0.090 0.044 0.044 0.295 0.295 0.440 0.440 0.2172 0.2172

00.10.20.30.40.5

0 100 200 300 400 500 600 700 800 900

Coordinate z, mm

Def

rom

atio

n, m

m

F=100 daNF=200 daNF=240 daN

Figure 5: Workpiece deformation under the action of the force


238

the tool holder (x3max=0.090 mm) and the

maximum size of the Fx force (Fx=240 daN), therigidity of this subsystem could be considered asbeing Rx3=2666 daN/mm. These values highlightsa relatively low rigidity of the consideredsubassemblies (it is known that for a machining ofhigh accuracy, the rigidity must have values closeto 20000 daN/mm) and this fact could be generatedboth by the clearances between variouscomponents of the subassemblies and by the wearof the same subassemblies. An image concerningthe elastic displacements of the universal chuck,life centre and tool holder is presented in figure 6,where the mechanical hysteresis could be alsoremarked.

4 Conclusions

The study of the technological system rigidity isimportant due to its influence exerted on themachining accuracy; it is expected that a biggerrigidity of the technological system allowsobtaining a higher machining accuracy. Aworkpiece fixed in the universal chuck and the lifecentre could be considered as a bar fixed at oneend and simply supported at the other end; for sucha case, the elastic deflection of the workpiececould be estimated by means of mathematicalrelations used in the strength of materials. In such acase (workpiece fixed in the universal chuck andthe life centre), the rigidity is not the same alongthe workpiece. Theoretical considerations andexperimental results showed that the rigidity ishigher near the universal chuck and lower near thelife centre. In order to measure the elasticdisplacements of the universal chuck and life

centre and the elastic deformation of the test piece,

a set of dial gauges were used. The experimentalresearch proved also the presence of themechanical hysteresis phenomenon.

References [1] Jianliang, G. and Rongdi, H., A united model

of diametral error in slender bar turning withfollower rest, International Journal ofMachine Tools & Manufacture, 46, 2006,1002-1012.

[2] Korsakov, V. S., Tochinosti mekhanicheskogoobrabotka, Mashinostroenie, Moskva, 1961.

[3] Militaru, C., Fiabilitatea i precizia întehnologia construc iilor de ma ini, EdituraTehnic , Bucure i, 1987.

[4] Pico , C., Coman, G.. and Pruteanu, O.,Tehnologia construc iei ma inilor-unelte.Volumul I, Institutul Politehnic Ia i, 1970.

[5] Sl tineanu, L., Cotea , M., Be liu, I., Boca,M. and Pop, N., Consideration concerning theelastic deformation of the workpiece atturning, Proceedings of the 14th InternationalConference Modern Technologies, Quality andInnovation, Ia i - Chi in u - Belgrad,ModTech 2010, New Face of T.M.C.R., 20-22May 2010, Sl nic Moldova, România, 539-542.

-0,02

0

0,02

0,04

0,06

0,08

0,1

0 100 200 300

Defo

rmat

ion,

mm

Fx

universal chuck loadinguniversal chuck unloadinglife centre loadinglife centre unloadingtool holder loadingtool holder unloading

Figure 6: Elastic displacement of some subassemblies of the technological system


239

RESEARCH REGARDING THE USE OF QUALITY TECHNIQUES ANDINSTRUMENTS IN VIEW OF MAINTAINING AND IMPROVING

QUALITY MANAGEMENT SYSTEMS

Liliana Georgeta Popescu1, Mihai Victor Zerbes 1, Marilena Blaj1, Radu Vasile Pascu1,Roland Blaj2

1”Lucian Blaga” University of Sibiu, [email protected] of Agricultural Science and Veterinary Medicine of Cluj-Napoca,[email protected]

Abstract: In Romanian industry there is still a lot of think to be done in the field of maintaining andimproving quality management systems (QMS). It is imperative to design a strategy for relating with theemployees as well as the managers of the organizations in this field, aimed to make them better and easierunderstand the concept of QMS. The present paper sets forth as a starting point for this strategy the use ofquality techniques and instruments (QTI), especially the 5S technique which is a good premise for themaintenance and improvement of QMS. The 5S technique is a structured method for the maintenance andimprovement of working organization and standardization.

Keywords: inspection, cleaning, progress, 5s technique

1. 5s TechniqueThe ISO 9000 set of standards proves helpful in

the creation and implementation of a reliablequality management system, as it sets forth thefundamental principles requirements, glossary thatmake up the system as well as the guidelines forquality improvement. Mention should be made thatthese standards tell only “what to do” not “how todo” which shall be decided by each organization.

Customer requirements concerning the qualityof products and services entail the use of high-performing numerical command machine-tools etc.as well as highly qualified personnel. Thus aquality management system is essential forindustry organizations to understand, coordinateand correlate various processes necessary foraccomplishing high quality products and servicesthat gain a competition edge on the market.Furthermore, an adequately implemented,maintained and improved system shall unfailinglylead to efficiency and efficacy, thus toorganizational progress.

The use of the 5S technique represents a stepforward to the improvement of such a system. 5Sstands for the initial letters of five Japanese wordsthat define an effective organization of the workingplace and standardized working procedures. 5S ismore than mere administration, it differentiatesbetween a common and an outstanding

organization. The 5S technique represents astructured method for the improvement of workplace organization and standardization. A well-organized work place can motivate all employees.5S highlights security, output, efficiency and italso creates a feeling of propriety belonging.

The golden rule of the 5S technique is that theemployee should be able to identify anythingwithin his work place in less than 30 seconds and,respectively, in less than 5 minutes for anythingthat transgresses the boundaries of his work place,without any prior talk to a colleague, opening abook or computer searching.

5S is a method whose results can be noticed ina well organized and ordered work place, anenvironment providing “a place for anything is inits place, whenever you need something”. Thework place shall be clean, ordered and safe, thusthe employees become motivated and committed.This is a simple concept. However it requires long-standing commitment and hard work to attain it.

The “5S” technique requires the coordinatedunfolding of the following five categories ofactivities:Sorting: removing everything that is not necessaryat the work place: waste material, semi-products,unnecessary tools and equipment, scraps orobsolete documents;

mailto:lilianaescu:@ulbsibiu.ro


240

Figure 1: Activity 1 - SortingSetting in order: arranging the useful objectssubsequent to the previous operation and preparingthem to be used at any further moment;

Figure 2: Activity 2 – Setting in orderCleanliness: cleaning the entire work placeincluding the objects within the working area

Figure 3: Activity 3 – Cleanliness

Standardizing: setting clear rules for maintaininga perfect hygiene and pleasant atmosphere at thework place;

Figure 4: Activity 4 – StandardizingSelf-discipline: accurate observance of theestablished working procedures and assimilationby all employees of the proper method ofperforming previous operations of sorting,arranging, cleaning and standardizing and thus beable to apply them from their first working day;

Figure 5: Activity 5 – Self-discipline

2. Steps to Successful Implementation of the5S Techniquea. Everyone has to be involved.

It is essential to understand that theimplementation of the 5S technique can only beachieved by a team. Everyone is accountable forthe 5S implementation not only a few individuals,it is a process involving all individuals. Al


241

managers should take part in the decision-makingprocess in view of a better functioning of the 5S.b. Managerial approval

The 5S technique shall not be performedsecretly or surreptitiously as overtime activity.Instead, the 5S technique has to be approved by theorganization`s administration. Monthly meetingsmay be organized where the companyrepresentatives and managers may suggest moretopics.

c. The general manager has the final responsibilityThe 5S technique shall be taken seriously and

completely assumed only when the managers andgeneral manager himself are responsible and fullycommitted. There is nothing worse for the 5Stechnique than the managers transferringresponsibility of project implementation to theemployees.

d. The 5S technique should be properly understoodand assimilated

People should no longer ask questions such as:“Why do we have to stick red labels to objects?” or“Is the 5S implementation really necessary beforeany improvement?” Therefore 5S meeting shall beorganized in order to explain the concept andanswer the participant’s questions.

e. The manager should inspect the work placeThe manager should personally inspect work

places and desks in order to clarify positive andnegative aspects. The manager should point outelements subject to improvement and bring themup at 5S meetings.

f. Do not leave things half-done!Once initiated, the 5S process has to be

continued until completion. All participants to thisprocess have to be involved. As soon as theprinciples have been established and atmosphere ofdiscipline and order has to be maintained.

The implementation of 5S shall probablyincrease productivity, efficiency and thus anatmosphere of professionalism shall be maintainedand encouraged. Developing employee attitudesand discipline is more important rather than thephysical reorganization of the work place.

3. Stages in promoting the “5S”

3.1 Mandatory activities:

- setting up the work team;- training the work team;

-identifying work areas subject to “5S”implementation;

- organizing the four stages of “5S”;- training all employees to observe the “5S”;- 5S auditing;

Internal audits are necessary for theimprovement and regular reporting on the status of5S implementation and maintenance

2.2 Stages in the 5S audit accomplishment

- selecting the work area;- setting up the work team - in view ofimplementing the “5S” method, the pilot projectteam members shall coopt individuals who performtheir professional activity within the work area (i.e.foreman, adjuster, worker, etc.);- training the work team members;- assessment of the work area;i). accurate study of the area and collecting data inkeeping with the “Quotation” column (table 1);

Table 1: Sample of audit form

ii). each member of the work team shall assign aqualification (grade) for each assessment criterionin the audit form;iii). Summary of data collected from the quotationforms and establishing the average quotation level;iv) Identifying the objective (quotation level to bereached for the ongoing year), filling in the “5Sindicator” graph during the brainstorming sessionand posting the results in the respective area(figure. 6);

Figure 6: Sample form for filling in graph


242

v) Drawing up the improvement plan andthoroughly observing those criteria with poorgrading (table 2);

Table 2: Sample for the action plan

vi) Tracking the status of improvement plan by thesteering team;

vii) Resuming application of the method untilreaching the established level;

viii) Establishing a higher level;ix) Permanent application of the method.

3 ConclusionsThorough and continuous implementation andapplication of the “5S” technique may enable youto design the following objectives:

- zero preparation time;- zero loss;- zero defects;- zero delays;- zero labour accidents;

Here are the advantages of “5S” application:a. Diminishing costs by means of:- store decreasing- preventing object misplacement;- Preventing oil/water/air/energy/material wasteb. Efficiency accomplished by means of:- effective use of the area- doing away with useless search and waste

cleaning and inspecting critical points of theequipment

c. Maintenance of auxiliary equipment andcapacityd. Qualitye. Occupational health and safety accomplished bymeans of:- removal of dangerous areas;- improvement of work conditions- doing away with the causes of labour accidentsand occupational harm;- Decreasing errors entailed by carelessnessf. Personnel motivation

However, let us not forget that it is imperativeto:- Standardize habits in view of obtaining betterresults;- Communicate and train in order to attain quality;

- Communication is a two-way process- Proceed in such a way that each individual shouldfeel responsible and allow everyone to expressopinions on personal responsibility, as well as tobe able to respond whenever there is any anomalyrelated to one’s responsibilities. Implementation ofthe 5S technique shall lead to processimprovement, creating a positive image tocustomers, and increasing organizationalefficiency. It will also enable the employees to feelbetter and the organization thrive and becomemore competitive on the market.

The method can be used in different domains.For example we apply successfully 5S technique inthe factory (figure 7 a, b)

Figure 7 a: Before to implement 5Sinto a productive sector

Figure 7 b: After implementation 5S into aproductive sector


243

Than we apply 5S technique in the office (seefigure 8 a, b).

Figure 8 a: Before to implement 5S in office

Figure 8 b: After implementation 5S in the office

5S technique could be implemented even on thedrawer of the desk (figure 9 a and 9b),

Figure 9 a: Before to implement 5S in drawer

Figure 9 b: After implementation 5S in the drawer

and on the computer's desktop (figure 10 a, b)

Figure 10 a: Before to implement 5S on desktop

Figure 10 b: After implementation 5S on desktop


244

To conclude, implementation and maintenanceof the 5S technique may lead to progress.

Figure 11: The 5 steps

Success may be attained not only by means ofcertain applicable standards, instead it is advisableto study a number of guidelines, brochures, codes,sciences, standards, and especially methods andpractices successfully employed by others. Thenext step is to identify and select the optimum anduseful elements for the particular businesses andorganizations, which should be further promotedand implemented in an integrated manner as amanagement system which, in time, shall lead tooutstanding quality and successful managementand business.

References[1] Bosanceanu, M., Interdependenta calitatea

productiei-principiul 5S, mmq.ase.ro,Universitatea „Stefan cel Mare” Suceava.

[2] Budau, G., Zerbes, M.V., Ispas, M., Cercetariprivind necesitatea i oportunitateaintroducerii sistemului de management alcalit ii în organiza iile economice dindomeniul industriei lemnului, ProceedingsICWSE, Brasov, noiembrie 2003.

[3] Narayanan, L., The “5S” Philosophy. A betterWork Environment for Everyone, Availablefrom, http://www. tpmonline.com.

[4] Kifor, C.V., Oprean, C., Ingineria calitatii,Editura Universitatii „Lucian Blaga” Sibiu,2002.

[5] Imbunatatirea competitivitatii princresterea productivitatii in intreprinderiledin Romania, Bucuresti, 31 mai 2005;Available from, http://www.lean.ro/LEAN_imbunatatirea_competitivitatii.pdf.

http://www.lean.ro/LEAN


245

PROPERTIES OF HARD ALLOYS SINTERED FROMMETALLIC CARBIDES

Constantinescu Stela1

1University “Dunarea de Jos “ of Galati, Faculty of Metallurgy , Materials Sciences and Environtment,[email protected]

Abstract: The TiC or TaC alloys, the simple WC-C0 alloys with the same cobalt content are much more resistantto bending and breaking and feature improved electrical and heat conductivity. The oxidation resistance of the simpleWC-C0 alloys is considerably lower which results in a pregnant tendency for chip welding and low resistance.Hardness is affected by micro- porosity, granulation of WC phase, purity and composition, extent of homogenization ofthe liquid and carbons. The magnetic saturation increases with higher Co contents; it is worth mentioning that biggergrain alloys have a force considerably lower than that of fine grain ones.

Key words: heat conductivity, hardness, density, porosity, hard alloys

1. Intoduction The excessively high sintering temperature

results in a lower density which further negativelyaffects the mechanical strength. Due to over-heating and granulation growth by re-crystallization an acute decrease in the bendingultimate strength occurs.

Under or de- carburated materials alsocontaining fragile phases , feature poor ultimatestrength. The max values taken by the bendingultimate strength are reached at Co 20 - 25 % andsuddenly decrease with high Co contents. Attheses compositions there are no contact bondsamong the carbons, the carbon crystals beingindividual and surrounded by the Co metallic mass.

The alloys mainly used for short chip materialcutting (cast iron, porcelain, etc) are alloys of WC-Co type, in some cases with small additions ofother carbons. The same compositions but ofdifferent granulometric classes are also used forpieces exposed to wear ( wire drawing, moulds,etc).

The higher performance of the WC-Co alloysin short chip machining applications is accountedfor by its very good heat conductivity which is 2-3times higher than that of fast steels

The alloys WC-Co, WC-TiC-Co , are mainlyused for short chip material cutting and wear -resistant piece applications [1]. The hard- alloysproducers can resort to a wide range of possibilitiesto achieve variation of the properties of a WC-Co

composition and thus they can adapt their productsto the particular types of tools they may choose tomanufacture [2] . The present metallography technique allows for acorrect identification and evaluation of hardsintered carbons structures. The metallographyapproach is an indispensable method ofinvestigation and control in industry.

2. Researches and experimental results The experimental tests have shown certainfeatures previously identified as requirements for agood quality deposit layer, which are: highcompression strength, good impact resistance,good resistance to high temperatures and thermalshock, fine surface roughness, pure layer structurecombined with equal grain size, deposition ofmaterials with continuous change of composition,regarding the increase of deposit layer (sandwichlayers), deposition of materials impossible toobtain through other procedures, almost unlimitedopportunity to choose deposition materials, goodadherence between basis material and depositlayers. The reaction will be possiblethermodynamically if the calculated concentrations(partial pressures) of the reactants , underequilibrium conditions, are less than their originalconcentrations

The calculation of the equilibriumconcentrations from the equilibrium constantinvolves a good choice of the number of gas spaces

mailto:stela.constantinescu:@email.ro


246

which can be higher than two and the number ofindependent relations A relation implies theequilibrium expression depending on the freestandard reaction energy and temperature. Theother relation consists in that the system pressureis the sum of the partial pressures . If somereactants possess more than one valence state, thereaction should contain the reactant under itsmost stable valence state.

After making all the calculations the entirerange of deposition parameters (temperature,pressure, gas initial composition) is obtained.

The phases coming from the vapor whichcontain the diffusion element and the carryinggas pass through three main stages; vaporformation, transportation and deposition. Thesestages differ in terms of chronology and make up awhole process. According to its essence, such aprocess is a chemical transport reaction expressedby the solid substance interaction (A) with the gasor vapors (B).

Upon conveying the gas, two processes arepossible: isotherm , without forced flow;anisotherm , with forced flow

With the latter process, saturation with gasdiffusion by contact-free procedure is reported.

Figure 1: Metallographic appearance of freecarbon in fine-grained phase

When evaluate the structure of the hardsintered carbons structures by metallographymethod, the distinction is made between themetallographic aspect of the WC-Co, WC-TiC-Coby WC-TiC-TaC-(Nb)C-Co alloys. Particularattention is paid to the distribution of the sinteringbinder, namely the cobalt, to see the correlationbetween the alloy properties and cobalt content

In the case of two alloys having the samecobalt content but different WC granulations.it is found that a harder alloys involved finergranulations while milder alloys implies roughgranulations.

The cobalt, which takes the form of very fineinclusions with alloys 5% Co and 95% WC, ismuch more agglomerated between the WC crystalsalong with a slight porosity .

In figure 1 and 2, is illustrated structurally freecarbon metallographic appearance in two situationscompared with the standard scale.

Thus one can see that it always appears at grainboundaries of carbide WC - Co and that dependingon the carbon surplus in the balance of the alloymaterial, the amount of free carbon as a separateindependent phase may be higher or lower.

Figure 2: Metallographic appearance of freecarbon phase at the grain average

Macro aspect of the fracture allows accurateobservation of the sintered alloy grain size, itscolor and luster, the presence or absence of freecarbon and creating an impression that the exactproperties to be expected. Currently simplemacrostructure in fracture assessment allows afaster control routine in the production of sinteredcarbide parts.

An accurate assessment of the microstructure ofthe sintered alloy requires an appropriatemetallographic attack. Preparation of samples formetallographic study is relatively difficult due tothe hardness of the material.


247

Figure 3 illustrates the metallographic structureof the alloy containing 6% Co and 94 % WC andaverage granulation. The basic constituent, WC, isunder recrystallization form, called 2 whichstructurally stable crystals under triangle prismaticshapes of rectangular bases.

Figure 3: The metalographic aspect of the alloycontaining 6% Co and 94 % WC,sol.KOH-K3 (Fe(CN)6)

As to hardness, there is a tight correlationbetween the Co content and the WC-Co alloysproduced under identical conditions. As shown infigure 4, hardness decreases with the increase inthe cobalt content. The sintering temperature andthe exposure time along with the type of grindingand mixture homogenizing can decisively affectthe sintered metallic carbons alloys[3,4].

Figure 4: Variation of hardness, breaking andbending resistance and compression resistance of thealloys WC-Co , influenced by the Co content.

Hardness reaches max values with the optimumsintering temperature and then it decreases as aresult of carbon recrystallization and alloy super-sintering. An excessive sintering time, even if thetemperature is optimum, has the same effect ielower hardness. The WC-Co alloy density dependson the Co content and the extent of sintering.

Figure 5 shows the density variation dependingon Co content in the alloy; it implies that the realmeasured density takes values within 0, 5 - 3%,under the theoretical calculated values. This maybe accounted for by the residual porosity which isthe result of a normal sintering process because ofinsufficient mixture homogenizing or slight low-or high carburing which may occur with sintering[5,6] .

Figure 5: Alloy density vs Co content

As to the density variation, the specificpressure for pressing the compressed pieces has alower influence while the sintering conditions,such as the sintering temperature and its being keptat lower values, play a decisive role in determiningdensity (figure 6). Modern methods such as hotpressing or hot iso-static pressing, allow forreaching densities identical to the calculated ones[7].

The breaking/bending strength of the WC-Coalloys increases with the cobalt content but therelation is not lineal [10]. The increase in thebreaking/bending strength takes place also in thealloys having more than 20% Co provided theseare sintered under special conditions (protectedagainst carburing) With alloys of 10% Co nopermanent deformations of the material can be


248

noticed during the breaking strength test, whilethey are quite obvious with Co contents higherthan 20%.

Figure 6: The effect of the sintering temperature andtime on the WC-Co alloy containing 6% Co density andhardness.

The compression resistance of the WC-Coalloys, as shown in figure 4 increases with the Cocontent and then significantly decreases after 4%limit is exceeded .[8,9].

The resilience or impact resistance of themetallic sintered carbon alloy is a measure of themechanical chock resistance. Chock resistancevariation vs. temperature is illustrated in figure 7.

The values of this curve have been determinedon groove-free samples, 6 x 6 mm cross section,leaning against seats 40 mm distance apart. Thecurve shape reaches a peak at the optimum alloy –sintering temperature [11]. The elongationstrength is quite difficult to determine for such afragile material like the hard sintered carbon alloyAll the composition WC-Co containing up to 10%Co show no permanent deformation of thematerial.

It can however be noticed elastic deformations,without plastic deformations, immediately afterbreaking alloys containing more than 25 % canfeature a measurable elongation [12,13].

Figure 7: Chock resistance variation vs temperatureon the WC-Co alloy containing 6% Co

The WC -Co alloy elasticity module isinteresting for applications involving elastic strain.As shown in figure 8, the elasticity moduledecreases with the Co content.

Figure 8: Elasticity module variation with Cocontent for WC-Co alloy.

As regards the magnetic properties of the WC-Co alloys, magnetization upon saturation is relatedto the Co content (the phase - double Co and Wcarbon) and the force acting upon it stronglydepends on sintering extent and the grain size.Thus the magnetic measurements can be of use inquality control [14].

The scope of application of WC-Co alloys is aconsequence of their properties. The applicationsof the metallic alloys depends on their compositionaccording to which there are 4 groups as shownbelow: Group I- 97% WC- 3% Co and 95,5% WC –4,5% Co has the following scope of application:cutting of graphite, ceramics and other metallic


249

materials, grinding, accuracy drilling of cast iron,non ferrous material machining, drawingmachines .

Group II- 94,5% WC – 5,5% Co and 93,5%WC – 6,5% Co is used for:a) global granulation sorts ( cast iron processing,

non- ferrous materials and alloys, synthetic andplastic materials, sensors for wear resistancetools and pieces requiring no high tenacity ,drawing machines .

b) fine granulation (machining of gray cast iron,mild cast iron, steels of ultimate strengthhigher than 1750 MPa , bronze, Si alloys,drawing machines .

Group III – 91% WC – 9% Co , 89% WC –11% Co , 87% WC – 13% Co used for machiningwood , synthetic resin , easy machining of steels,brass and bronze grinding ,plates for agriculturetools, wear pieces calling for high tenacity.

Group IV – 85% WC – 15% Co, 80% WC –20% Co , 75% WC – 25% Co , 70% WC – 30%Co used for: wear resistant parts calling for hightenacity ( cutting tools etc.) .

3. ConclusionsHardness is affected by micro- porosity,

granulation of WC phase , purity and composition ,extent of homogenization of the liquid and carbonsThe excessively high sintering temperature resultsin a lower density which further negatively affectsthe mechanical strength.

The max values taken by the bending ultimatestrength are reached at Co 20 - 25 % and suddenlydecrease with high Co contents. At thesescompositions there are no contact bonds among thecarbons, the carbon crystals being individual andsurrounded by the Co metallic mass.

Due to overheating and granulation growth byre-crystallization an acute decrease in the bendingultimate strength occurs. Under or de-carburatedmaterials also containing fragile phases , featurepoor ultimate strength .

The compression resistance first increases withthe Co content then considerably decreases afterthe Co 4% is exceeded.

The resilience of a Co 6% alloy increases withtemperature and reaches a max value at 1600oC.

The elongation vs compression resistance ratiois estimated as 1: 3 while the elongationresistance vs ultimate strength ratio is three timeshigher that that of steel

The magnetic saturation increases with higherCo contents; it is worth mentioning that bigger

grain alloys have a force considerably lower thanthat of fine grain ones.

4. REFERENCES

[1] Mito eriu, O., Constantinescu, S., Radu, T., .a.,Modern methods to perform the properties ofmetal materials, University „ Dun rea de Jos „of Gala i , 1998.

[2] Constantinescu S., Nitride coatings on widiasubstrate for mechanical applications ,Journal of Surface Engineering vol.52,nr1,p.77-81, ISSN 0267-0844, 2009, USA.

[3] Constantinescu S., Radu T., Modern methods toperform thin layers, Romanian MetallurgicalFoundation Scientific, Publishing House ,Bucure ti , 2003.

[4] Constantinescu, S., Metals properties andphysical control methods, Didact i Pedag.Publishing House , Bucure ti , 2004.

[5] Radu, T., Constantinescu, S., Vlad, M.,Morphologies of Widmanstatten structuresand mechanism formation in steels , Journalof Materials, Science Forum Vol. 636-637, pp550-555, Trans Tech Publications,Switzerland, ISSN 1662-97522010,www.scientific.net

[6] Constantinescu, S., Practical experience ofusing chemical vapour deposition coatings toresist wear, TEHNOMUS “New technologiesand products in machine manufacturingtechnologies”, Journal no.17, 2010,Ed.Universitatii Stefan cel Mare din Suceava, p.84-88, ISSN:1224- 029X, 2010.www.tehnomusjournal.fim.usv.ro.

[7] Constantinescu, S., Orac, L., Properties andapplication of B4C coats within thechemical vapour deposition, METAL 2009, ,p.96-101, ISBN 978- 80 87294- 03- 1, 19. –21. 5. Hradec nad Moravici , Cehia ,2009,www.metal2009.com

[8] Constantinescu, S., Orac, L., Characterisationof niobium carbide coating obtained bychemical vapour deposition, The Annals of“Dunarea de Jos” of Galati, Fascicle IXMetallurgy and Materials Science,No. 1, 2010,ISSN 1453-083X .

[9] Constantinescu, S.,Orac L.,Wear resistance ofchemical vapour deposition coatings,4thInternational Conference Welding inMaritime Engineering , BRAC - SPLIT-Croatia 13 -16. 05. 2009,[email protected]

mailto:branko.bauer:@fsb.hr


250

[10] Constantinescu, S., Nitride coatings on widiasubstrate for mechanical applications, Journalof SurfaceEngineeringvol.52,nr1,p.77-81,2009,ISSN0267-0844,www.ingentaconnect.com

[11] Constantinescu, S., Orac, L., MechanicalProperties of thin films investigated usingmicromachining techniques, Proceedings ofthe 18-th International Metallurgical &Materials,METAL 2010, p.926-931, ISBN:978-80-87294-17-8, 2010, Ostrava, CzechRepublic, www.metal2010.com

[12] Constantinescu, S., Researches on the effectof the tempering temperature on bothstructure and properties of thick plates madefrom steel of 700 N/mm2 yielding point,Revista METALURGIA INTERNA IONAL, nr.6 / 2004 , ISSN 1582- 2214 , p.12-16,www.metalurgia.ro.

[13] Constantinescu S., Non-distructive control ofhot – rolled products, IntenationalConference on Materials Science andEngineering, BRAMAT 2007. Organized by„Transilvania „ University of Bra ov andRomanian Academy of Technical Sciences,ISSN 1223 – 9631, 2007,

[14] Constantinescu ,S., Vlad, M., Mitoseriu, O.,Orac , L., Making thin coatings of bn bychemical deposition from vapour phase atnormal pressure, al-7-lea CongresInternational ISSIM - Iasi, BuletinulInstitutului Politehnic Iasi, Tomul LV(LIX),fasc.2, 2009, ISSN 1453 – 1690, p. 151-156.

http://www.ingentaconnect.com/

http://www.metal2010.com/


251

RESEARCH ON THE CORROSION BEHAVIOR OF ZINC COATING BYMEASUREMENT OF POLARIZATION RESISTANCE

Radu Tamara, Istrate Gina Genoveva,

University “Dunarea de Jos” of Galati Romania

Abstract: Measurement of polarization resistance of micro-alloyed zinc layers followed corrosionproducts characterization and corrosion behavior of the default layers analyzed. We have studied micro-alloyed zinc coatings with different percentages of nickel, tin, bismuth, lead, cadmium, exposed to acorrosive environment with 3% NaCl at room temperature. For experimental determinations was appliedelectrochemical method. A calomel saturated electrode was used as reference electrode and a platinumwire electrode as an auxiliary one. Analyzing the variation of polarization resistance for each sample, wecan observe how initially the sample surface becomes active, which causes a decrease in polarizationresistance followed in time by the formation of a layer of corrosion products which will cause a variationof polarization resistance.

Keywords: polarization resistance, zinc alloy, coatings.

1. Working method and apparatus used

In most atmospheric environments, Zncorrodes much less than steel, due to theformation of a protective layer consisting of amixture of Zn oxide, Zn hydroxide and variousbasic Zn salts depending on the nature of theenvironment. Thus the protection of steel by a Zncoating is mainly through the barrier effect.However, at the places where the Zn coating isdamaged and the steel underneath is exposed,such as at cuts or at scratches, the galvanic actionbetween steel and Zn can protect the exposed steelfrom corrosion.

The measured values of polarization potentialinclude the voltage drop (ohmic loss) due to theelectrolyte and reaction products film on the metalsurface. The value of this voltage drop is thepolarization resistance [1; 2]. Ohmic resistancedrops sharply with loss of power, unlikeelectrochemical polarization which decreases witha slower rate [3]. By measuring the polarizationresistance, for alloyed zinc coatings, it is intendedto assess the corrosion products filmcharacteristics and corrosion behaviour of thedefault layers.

Measurements were performed on samples ofsteel sheet coated with zinc compositionscorresponding Table 1, using as a salineenvironment corrosion 3% Na Cl at roomtemperature [4]. Layers of protection were

obtained by hot dip galvanized, with the classicalpreparation of the sample surface.

Since nickel has a much higher meltingtemperature than zinc micro alloyed a zinc-nickelalloy, with 2% nickel, was made. Bismuth,cadmium, tin and lead are easily assimilated intomolten zinc and micro alloying was achieved byintroducing these elements, finely choppeddirectly into zinc bath. Bismuth is a noveltyelement for micro alloying zinc melts, used toreplace lead, having the same effect on meltfluidity and reduction of surface tension withoutbeing toxic.

In all cases it was applied to the melthomogenize the mechanical mixing.

For measuring the polarization resistance, itwas used a potentiostate PGP type 201. A calomelsaturated electrode was used as referenceelectrode and a platinum wire electrode as anauxiliary one. The samples were prepared foranalysis by being degreased with acetone, washedand dried [5]. After connecting the cableconductor, the delimitation of the working surface(equal for all samples) was made by insulatingwith resin. Measurement of polarization resistanceat micro-alloyed zinc layers was made forcorrosion products and corrosion behavior of thedefault layers analyzed. The variation curves of thepolarization resistance (Rp) were experimentallydetermined as a function of time for galvanizedsamples and samples coated with alloy.


252


Results of the electrochemical analysis consistin the curve of the polarization resistance functionof time. Analyzing the variation of polarizationresistance for each sample( table 1), we canobserve how initially the sample surface becomesactive, which causes a decrease in polarizationresistance followed in time by the formation of alayer of corrosion products which will cause avariation of polarization resistance.

Knowing the Rp, the value of Icor per surface unitcan be calculated using the Stern and Geary formula(1)

22. cmSkRpmVB

cmAI cor (1)

The value used for the B coefficient was of26mV.

Surface, S, for each sample was 1 cm2.

Table 1. Chemical composition of coatings tested

Code Type of coatingChemical composition, [wt%]

Ni Bi Sn Cd Pb Al Zn0 Zn 0 0 0,0005 0,0004 0,0014 0,0005 diff.

I Zn- Ni-Bi-Sn 0,16 0,71 2,95 0,26 0 0 diff.

II Zn- Ni-Pb-Sn 0,16 0 2,88 0 0,72 0 diff.

III Zn- Ni- Pb-Bi-Sn 0,16 0,41 3,49 0 0,43 0 diff.

IV Zn-Bi I 0 0,27 0 0 0 0 diff.

V Zn-Bi II 0 0,36 0 0 0 0 diff.

VI Zn-Bi III 0 0,52 0 0 0 0 diff.

In the case of hot dip galvanized samples,without addition of other metals corrosion productformation leads to an increased polarizationresistance to a 1 k cm2 and because it did not

reached zero polarization resistance during theanalysis result shows that the corrosion productslayer has a slow process of compaction (Fig. 1).

Figure 1: Polarization resistance function of time for the sample coated with zinc


253

Sample I, zinc-coated complex micro alloyedwith nickel, bismuth, tin, cadmium, forming afilm of corrosion products with a polarizationresistance of 90 k cm2 very stable and dense.Value of 0 /cm2 (circuit interruption) wasreached after about five minutes (Fig. 2)Sample II, coated with zinc and nickel, tin, lead,

forming a film of corrosion products with a 45cm2 polarization resistance and respectively

27.5 k cm2, has a similar evolution in time as thegalvanized sheet (figure 3)

The sample III, coated with zinc andnickel, tin, lead, bismuth, forms a film ofcorrosion products with a very lowpolarization resistance of 0.38 k cm2 whichdid not provide surface passivity during theanalysis (Fig.. 4.).

Figure 2: Polarization resistance function of time for the sample coated with Zn- Ni-Bi-Sn

Figure 3: Polarization resistance function of time for the sample coated with Zn- Ni-Pb-Sn


254

Figure 4: Polarization resistance function of time for the sample coated with Zn- Ni- Pb-Bi-Sn

Sample IV, coated with zinc and 0.27%bismuth, shows a variation of polarizationresistance, very similar to the pure zinc coatedsample , polarization resistance is 1.23 k cm2

(Fig. 5)

Sample V, coated with zinc and 0.36%bismuth shows low polarization resistance (0.4

cm2) for corrosion products formed (Fig. 6).

Figure 5: Polarization resistance function of time for the sample coated with Zn-Bi I


255

Figure 6: Polarization resistance function of time for the sample coated with Zn-Bi II

Sample VI, coated with zinc and 0.52%bismuth shows a variation of polarization

resistance similar to the previous zinc with 0, 36%bismuth coated samples (Fig.7).

Figure 7: Polarization resistance function of time for the sample coated with zinc Zn-Bi III

Comparing the polarization resistance ofsamples coated, with different layers of zincmicro alloyed, with the polarization resistance ofthe zinc coated sample, we can observe:

- RpI RpII RpIV RpZn RpVI RpV and IcorI IcorII IcorIV IcorZn, result a higher

corrosion resistance of samples coated with zinc-nickel-tin-bismuth-cadmium (code I);


256

Coatings with zinc-tin-bismuth have acorrosion resistance similar to zinc coated orlower:

IcorZn IcorIV IcorVI IcorV.

Conclusions

- Analyzing the variation of polarizationresistance for each sample, we can observe howinitially the sample surface becomes active, whichcauses a decrease in polarization resistancefollowed in time by the formation of a layer ofcorrosion products which will cause a variation ofpolarization resistance.

- The sample coated with zinc-nickel-bismuth-tin, has form a layer of corrosion products with apolarization resistance of 90 k cm2 very stableand dense, presenting a very good corrosionbehavior compared with unalloyed zinc coating.

-Samples with layers alloyed with nickelshows a bigger corrosion resistance comparedwith unalloyed zinc coating: IcorI IcorII IcorIVIcorZn.

- Coatings with zinc-tin-bismuth have acorrosion resistance similar to zinc coated orlower.

References

[1] Bradea,T.,Popa, M.V., Nicola, M., Stiinta siingineria coroziunii, Ed. AcademieiRomane, Bucuresti, 2004.

[2] Bard, A.J., Faulkner, in ElectrochemicalMeihods: Fundamentals and Applications,New York, 1980.

[3] Vermesan, H., Coroziune, Ed.Risoprint ClujNapoca, 2005, p.5-221.

[4] Constantinescu, A., Detectarea si masurareaa coroziunii, Ed.Tehnica, Bucuresti,1976,p.14-22.

[5] Balint, L., Ciocan, A., Balint, S., Poteca u,Fl., Interaction of the zinc protective layerwith an aggressive atmosphere, AnaleleUniversit ii “Dun rea de Jos” din Galati,2006, Fascicula IV, ISSN 1221-4558, p.345-350.

[6] S. Baicean, L., Palaghian, F. Potecasu,A.Ciocan, C.Gheorghies, Macro andmicrostructural changes in naval steel underfatigue corrosion, Metallurgy InternationalReview ISSN 1582-2214, No.10 specialissue, 2009, p.34-41.


257

MATHEMATIC MODEL FOR OPTIMIZATION OF ZINC-NICKELALLOY CO-DEPOSITION PROCESS

1Violeta VASILACHE, 2Marius BEN A

1Universitatea tefan cel Mare, Suceava, Str. Universit ii, Nr.13, 720229, Suceava, [email protected] Transilvania Brasov, Bulevardul Eroilor, No.29, 500036, Brasov,

Abstract: Any optimization method implies a mathematical model which should resolve thequantitative requires of the problems. This model is based on the substrate effect and has to calculate thepartial current densities and so has to give prediction as far as concern the quantities of metalelectrodeposited and the energy involved.

Keywords: mathematical model, anomalous co-deposition, zinc-nickel alloys, mass transfer,electrochemical kinetic

1. INTRODUCTION

Electrodeposition of a simple metal has also asimplier mathematical model, but situation isdifferent for alloys deposition. In our proposedmodel kinetic parameters are analyzed and it wasmade the supposition that current distribution andmass transport are homogeneous on workingelectrode. Also we tried to determinate the reactionmechanism.

Figure 1. The scheme which shows partial currentdensities for A and B components. (a) both componentsunder activation kinetic control, present identic Tafelslopes. (b) both components present limitation of thecurrent. (c) both components under activation control,but with different Tafel slopes. (d) component Apresents current limitation, component B underactivation control.

2. EXPERIMENTAL CONSIDERATIONS

2.1. Electrochemical processe at cathode

In our experiments it was deposited zinc-nickelalloy on gold substrate, previous deposited bysputtering on glass plates. This methode waschosen because it permits to analyze the layerswith XRD and SEM-EDX techniques.

The mechanism of electrochemicalreactions which occur on cathode surface has twosteps, as Matlosz described Matlosz [4], [5]. Zincions are deposits on their own substrate, on goldsubstrate and on nickel substrate. Also nickel ionsare deposited on their own substrate, on goldsubstrate and on zinc substrate. More, there aresecondary reactions, Zn2+ ions are combining withhydrogen to form ZnH+, and similar Ni2+ ions arecombining to hydrogen to form NiH+. Theseintermediate species, formed in adsorption process,finally will be decomposed to metallic zinc andnickel respectively.

The mechanism of electrochemicalreactions could be written as follow:

Ni2+ + e- Ni+ads (1)

Ni+ads + e- Ni (2)

Ni + H+ + NiH+ads (3)

NiH+ads + H+ + 2e- Ni + H2 (4)

Zn2+ + e- Zn+ads (5)

Zn+ + e- Zn (6)

mailto:violetav:@fia.usv.ro


258

Zn + H+ ZnH+ads (7)

ZnH+ads + H+ + 2e- Zn + H2 (8)

Ni2+ and Zn2+ are disolved as metallic ions,hydrolized or not. Ni+

ads and Zn+ads which could

contain or not the group hydroxyl are adsorbed inintermediate reactions. Ni and Zn are metallicdeposits of nickel and zinc respectively [6,7]. Thekinetic of mass transfer is supposed to respectButler-Volmer equation. In far from equilibriumstates anodic reactions could be neglected.

2.2. Determination of partial currentdensities

For a binary alloy AB and a thicknes of depositd, partial current density of B element is,

BB

BB m

tmFni (9)

Here mB is mass of element B deposited inalloy, MB is atomic mass of element B, t isdeposition time and nB is number of electronsimplied in reaction of element B.

2.3. Normal and anomalous co-deposition ofzinc-nickel alloys

Electrodeposition of zinc-nickel alloys isgenerally an anomalous co-deposition, afterBrenner’s deffinition, because the metal less noble,zinc, is deposited preferential and its percent indeposit is higher than in electrolyte. Anyway,normal co-deposition of zinc-nickel alloys ispossible only in particulary experimentalconditions. Co-deposition of zinc-nickel alloysfrom different electrolytic bath was studiedpotentiostatic and galvanostatic, function ofdifferent variable parameters during theelectrodeposition [1,2].

2.4. Mathematical modeling of zinc-nickelalloy co-deposition. The model of substrateeffect

The initial nucleation of adsorbed nickelon electrode surface acts as a catalyser for zincdeposition, resulting an inhibition of nickeldeposition. Also it was shown that pure zinccannot be deposited from aqueous electrolytsolutions at UPD (underpotential deposition), butit could be co-deposited with nickel. Thesephenomena can be explaine by the fact that nickel

nucleation catalize zinc deposition. At potentialmore negative than zinc equilibrium potential, zincdeposition rate is enough higher and inhibits nickeldeposition resulting an anomalous co-deposition.

The alloy deposition performs unersubstrate effect. Not only nickel affects zincdeposition, but zinc too affects nickel deposition.

Figure 2. The diagram of zinc-nickel alloyco-deposition

In figure 2 is represented the diagram of theeffects of different substrates during electroplatingwith zinc-nickel alloys. The initial electrodesurface is divided in two parts. The first iscorresponding to Ni which is the surface coveredby nickel and the second is the surface covered byzinc, Zn. Every surface is then divided in fourparts. So, for nickel deposition, Ni 1 correspondsto the area of Ni substrate surface covered withNi(I)ads. Ni 2 corresponds to the area of Nisubstrate surface covered with NiH+

ads, Ni 3corresponds to the area of Ni substrate surfacecovered with Zn(I)ads. Free surface Ni(1- 1- 2- 3)corresponds to the area of Ni substrate surface non-covered.

For zinc deposition, Zn 6 corresponds tothe area of Zn substrate surface covered withZn(I)ads. Zn 5 corresponds to the area of Znsubstrate surface covered with ZnH+

ads. Zn 4corresponds to the area of Zn substrate surfacecovered with Ni(I)ads. Free surface Zn(1- 4- 5- 6)corresponds to the area of Zn substrate surfacenon-covered.

2.5. Theoretic model. General mechanism ofelectrode reactions

A mechanism of reactions was developed aseffect of substrate. his model is based on thesupposition that every individual component isdeposited after a two steps-reaction, as Matlosz [4]described. The nickel ions are deposited on theirown substrate and on zinc substrate. Also zinc ionsare deposited on their own substrate and on nickelsubstrate. More, hydrogenated species ZnH+ and


259

NiH+ are strongly bonded at electrode surface.Ni(II) will react giving NiH+

ads and these adsorbedspecies will react after with deposited nickel. AlsoZn(II) will react giving ZnH+

ads and these adsorbedspecies will react after with deposited zinc.

2.6. The mass transfer effect

The material balance in equilibrium statethrough diffusion layer for species Ni(II), Zn(II)and H+, 0<x<d, can be written

0)(IINiN (10)

0)( IIZnN (11)

0HN (12)

HOHW CCK (13)Supposing a constant diffusion coefficient,

D, the flow of every species i, into diffusion layeris dxDdCN ii / . Accepted values fordifuusion coefficients are 4 10-10 m2s-1 for Ni(II),5,09 10-10 m2s-1 for Zn(II), and 9,3 10-9 m2s-1 forsolvated protons, and 5,5 10-9 m2s-1 for hydroxidions [3]. Intermediate species, Ni(I)ads, NiH+

ads,Zn(I)ads and ZnH+

ads existe only on the electrodesurface so one their concentration is equal zero insolution [8], [9].

2.7. Electrochemical kinetic

The charge transfer kinetic is supposed thatrespects Butler - Volmer equation. Far fromequilibrium, anodic reactions could be neglected.A Tafel modified expression describes theelectrochemical reactions rate on surface and isadapted to calculate the partial current. Forexample, on the first step of deposition reaction onnickel substrate, partial current density, i11, couldbe wrote as:

i11=-Fk011CNi

2+Ni(1- 1- 2- 3)exp(-b11 11)

2.8. Software for predictive calculus of zinc-nickel alloy composition

To simulate zinc-nickel alloy deposition, it waselaborated a soft-ware, which uses the calculusrelationship from described mathematic model. Thecalculus stops when difference between a calculatedvalue and a previos one is smaller than 10-5.

The program lines are written as follows:

using System;

using System.Collections.Generic;using System.ComponentModel;using System.Data;using System.Drawing;using System.Text;using System.Windows.Forms;

namespace CalculTETAi{ public partial class frmMain : Form { public frmMain() { InitializeComponent(); }

double F, Ki0, Bi, EtaI, C0, ITotal, Ai, t,TetaS, CH, CNi, CZn, TetaIInitial, TetaZn, TetaNi;

double TetaICalculat; double Ci, Ii; private void m_btnCalculeaza_Click(

object sender, EventArgs e ) { m_tbTetaICalculat.Text = "";

try { F = double.Parse( m_tbF.Text ); Ki0 = double.Parse( m_tbKi0.Text ); Bi = double.Parse( m_tbBi.Text ); EtaI = double.Parse( m_tbEtaI.Text ); C0 = double.Parse( m_tbC0.Text ); ITotal = double.Parse(

m_tbITotal.Text ); Ai = double.Parse( m_tbAi.Text ); t = double.Parse( m_tbT.Text ); TetaS = double.Parse(

m_tbTetaS.Text ); CH = double.Parse( m_tbCH.Text ); CNi = double.Parse( m_tbCNi.Text ); CZn = double.Parse( m_tbCZn.Text

); TetaIInitial = double.Parse(

m_tbTetaIInitial.Text ); TetaZn = double.Parse(

m_tbTetaZn.Text ); TetaNi = double.Parse(

m_tbTetaNi.Text ); } catch {MessageBox.Show( "Introduceti valori

corecte", "Atentie", MessageBoxButtons.OK,MessageBoxIcon.Warning );


260

return; } int Contor = 0; int NumarBucle = 1000; TetaICalculat = TetaIInitial; do { TetaIInitial = TetaICalculat; Ci = C0 * Math.Exp( -Ki0 * Ai * t ); Ii = -F * Ki0 * CNi * CZn * CH *

TetaZn * TetaNi * ( 1 - TetaIInitial - TetaS ) *Math.Exp( -Bi * EtaI );

TetaICalculat = Ii / ITotal; Contor++; if( Contor > NumarBucle ) break; } while( Math.Abs( TetaIInitial -

TetaICalculat ) > 1 * Math.Pow( 10, -5 ) ); if( Contor > NumarBucle ) MessageBox.Show( "S-a depasit

numarul de bucle!", "Atentie",MessageBoxButtons.OK,MessageBoxIcon.Warning );

else { m_tbTetaICalculat.Text =

TetaICalculat.ToString(); } } private void m_btnIesire_Click( object

sender, EventArgs e ) { this.Close();

Figure 3. User interface with the software forcalculus of alloy composition

3. CONCLUSIONSThis model establishes a mathematic apparatus

to describe zinc-nickel alloy co-depositionprocesses, using the substrate effect model fordifferent concentration of electrolyte and fordifferent applied potentials. There was a goodcorrelation between experimental data and theprediction of this model.

4. ACKNOWLEDGMENTSThis paper was supported by the project

"Progress and development through post-doctoralresearch and innovation in engineering and appliedsciences– PRiDE - Contract no.POSDRU/89/1.5/S/57083", project co-funded fromEuropean Social Fund through SectorialOperational Program Human Resources 2007-2013.

5. REFERENCES[1]. Bard, A.J., Electrochemical Methods.

Fundamentals and Applications, John Wiley andSons, New-York, 2001

[2]. Di Bari,G.A., Modern Electroplating, FourthEdition, Edited by Mordechay Schlesinger andMilan Paunovic, John Wiley & Song, Inc., 2000

[3]. Teeratananon, M., Saidi, K., Fenouillet, B.,Vergnes, H., Journnes d’electrochimic, Poiters,France, june 2008

[4]. Matlosz, M., Journal Electrochemistry Soc.,140(1993)2272

[5]. Soares, M.E., Souza, C.A.C., Kuri, S.E., Corrosionresitence of Zn-Ni electrodeposited alloy obtainedwith a controlled electrolyte flow and gelatinadditive, Science Direct, vol.201, Issue 6, dec.2006,p.2953-2959

[6]. Schlesinger, M., Electrodeposition of Alloys,Modern Electroplating, Fourth Edition, John Wileyand Sons, Inc. New-York, 2000

[7] Brenner A, Electrodeposition of Alloys, Vol.I,Academic Press, New York, 1963

[8]. Vasilache,V., Gutt,Gh., Vasilache, T., Rev. Chim.(Bucuresti), 59, nr.8, 2008, p. 915

[9]. Vasilache,V., Ph. D. Thesis, University Stefan celMare of Suceava , 2008

[10]. Vasilache V., Gutt Gh. Vasilache T., Studies aboutelectrochemical plating with zinc-nickel alloys. Theinfluence of potential through stoichiometriccomposition, Revista de Chimie, Bucure ti,vol.59,nr.9 (2008)

[11] Vasilache V, Gutt S, Gutt Gh, Vasilache Tr, SanduI, Sandu IG. Determination of the Dimension ofCrystalline Grains of Thin Layers of Zinc-NickelAlloys Electrochemically Deposited. MetalurgiaInternational, 2009, 14: 49-53;


261

MODULAR DESIGN FOR A FAMILY OF MECHANICALANTHROPOMORPHIC POLY-MOBILE GRIPPERS WITH 4

FINGERS FOR ROBOTS

Ionel Staretu

Transilvania University of Brasov, Product Design and Robotics Department, Brasov, Romania, e-mail:[email protected].

Abstract: In this paper one anthropomorphic modular gripper for robots are described. The stages of synthesis,analysis design and functional simulation are presented. The structural synthesis of the anthropomorphic grippers forrobots can be made regarding the following main criteria: the number of fingers, the number of phalanxes, the relativedimensions of the phalanxes, the relative position of the fingers, the degree of freedom of the gripping mechanismand the characteristic constructive elements used. We choice a version with four identical fingers with threephalanxes on finger. The kinematic synthesis is used to obtain a correct closing of the finger and of the grippingmechanism. The function of position, the function of speeds and the function of acceleration for characteristic pointsare obtained from the kinematic analysis. The static synthesis solves the problem to obtaining the necessary grippingforce on each finger and the total gripping force. The calculation of strength was made in function of the internalforces which act between elements. With the constructive dimensions a 3D model can be obtained using CATIA soft.Some aspects regarding functional CAD and virtual simulations are shown too. For one variant of this type of gripper,with four fingers, the technical documentation is completed and the technical project has all the conditions forpractical achievement. There are two main constructive modules: the support – the palm and the finger.

Keywords: Robotics, Anthropomorphic grippers, Mechanism, Design, Functional simulation.

1. Introduction

The mechanical anthropomorphic grippers forrobots have as main mechanical component asimilar mechanism with the biomechanism of thehuman hand. This mechanism has only pivot jointsand two or more fingers with two or threephalanxes.

These grippers for robots comparatively withothers mechanical grippers (mechanical gripperswith jaws, mechanical tentacular grippers) havemore advantages like: a bigger degree of dexterity(99% for four fingers, 90% for three fingers and40% for two fingers comparatively with humanhand), a larger domain of utility (many types ofobjects can be grasped) and that the grippers cando micro-movements with the grasped objectbetween the fingers (if the degree of freedom isequal or bigger like the number of fingers).

In the paper are shown the stages of synthesis,analysis, design and simulation for a modulatedfamily of anthropomorphic grippers. There areshown the anthropomorphic poly-mobile gripperswith four fingers from this family.

2. Structural Synthesis and Analysis

2.1 Structural synthesis

The structural synthesis can be made regardingthe following main criteria: the number of fingers,the number of phalanxes, the relative dimensionsof the phalanxes, the relative position of thefingers, the degree of freedom of the grippingmechanism and the characteristic constructiveelements used [1,2].

For our family of grippers these criteria wereadapted in order to obtain a good performance:four identical fingers, three phalanxes on eachfingers, relative position of the fingers like inFig.1( for 4 fingers ), the degree of freedom M=n

Figure 1: The relative position of the fingers


262

(n – the number of fingers), linkage mechanisms.The main structural module is accordingly with afinger and it is shown in Fig.2.

Figure 2: The structural scheme of the finger

2.2 Structural analysisThe mechanism of the finger (Fig.2) is a

polycontour mechanism with two outsideconnection L=2 (v1, F1; vP1, FP1 – Fig.3,a) anddegree of freedom M=1.

The degree of freedom is obtained withci fMM , where Mi is the degree of

freedom for monocontour i mechanism and cf

is the degree of freedom for common joints(Fig. 3,b).

For each monocontour mechanism the degree

of freedom is obtained with KifM

(where if is the degree of freedom of the joints

and K is kinematic degree of the monocontour kmechanism [1]).

Figure 3: The block scheme and the graph of themechanism

For the mechanism shown in Fig.2, inaccordingly with the graph of Fig.3,b, thefollowing relations are obtained:

MI=fA+fB+fC+fD I=1+1+1+1-3=1MII=fD+fE+fF+fG II=1+1+1+1-3=1 (1)MIII=fL+fM+fN+fE III=1+1+1+1-3=1,and fC=fD+fE=1+1=2.The degree of freedom will be:M=MI+MII+MIII fC=1+1+1-2=1 (2)M=1 has the following significance: one

independent movement (speed): 11 sv and onefunction of the external forces: F1=F1(FP1). L-M=1represents one function of movement: vP1= vP1(v1)and one independent force: FP1 – the contact forcebetween finger and grasped object.

3. Kinematic Synthesis and Analysis

3.1 Kinematic synthesis

The kinematic synthesis is used to obtain acorrect closing of the finger and of the grippingmechanism. This situation is obtained with a goodcorrelation between the dimensions of thephalanxes and a good relative position of thefingers. The first and one intermediary position ofthe finger are shown in Fig.4.

Figure 4: Two configuration of the finger

3.2 Kinematic analysis

The function of position, the function of speedsand the function of acceleration for characteristicPi points are obtained from the kinematic analysis.The vectorial close chain method is usesuccessively for each monocontour mechanism.The vectorial equations are:

0'' ADDDCDAC (Fig.5,a),0GDFGEFDE (Fig.5,b) and0LEMLNMEN (Fig.5,c) [3].

a


263

b

cFigure 5: The kinematic schemes

The implicit form for the equation of positionsis: 72i= 72i(s1), i- the member of the fingers:i=1,2,3,4.

The functions for speeds are the derivativefunction of time of the functions for positions andthe functions for accelerations are the derivative ofthe functions for speed:

PiPi

iPi

va

v 72 (3)

4. Static Synthesis and Analysis

4.1 Static synthesisThe static synthesis solves the problem to

obtaining the necessary gripping force on eachfinger and the total gripping force.

4.2 Static analysis

The function of the external forces is obtainedfrom the theorem of balance between the powers ofentrance and emergence of mechanism:

vi·Fi+vPi·FPi=0, and

i

PiPii v

FvF (4)

The internal forces are calculated using thetheorem of the joints and, after-words, with thebalance static equations of the mobile elements [1].

5. Constructive Design and 3D ModelThe calculation of strength was made in

function of the internal forces which act betweenelements. With the constructive dimensions a 3D

model can be obtained using CATIA soft[6,7].There are two main constructive modules: thesupport – the palm (Fig.6,a) and the finger(Fig.6,b)[4,5].

a

bFigure 6: The main constructive modules

The family of anthropomorphic grippers isobtained using fingers in 5 relative positions (seeFig.1).

For instance the possible variants with fourfingers are shown in Fig.7: one variant with fingerswith parallel axes (Fig.7,a); one variant withfingers with parallel axes but with a interval (Fig.7,b) ; crossing axis (Fig.7,c); three fingers with one

a


264

b

c

d

eFigure 7: The family of the anthropomorphic

grippers with four fingers

in opposite and central position (Fig. 7,d) andthree fingers with one in opposite and lateralposition (Fig. 7,e).

For these five variants the technicaldocumentation is completed and the technicalproject has all the conditions for practicalachievement.

6. Functional Simulation A functional simulation (Fig.8,a,b,c) was made

to check the correct work and to identify thesolutions to obtain the optimum variant for thisgrippers.

a

b

c

Figure 8: The functional simulation

Other functional simulation is made with apiece(Fig.9). These gripper has four degree offreedom and its can grasp objects with regular orirregular forms.

a b


265

c d

e

Figure 9: The functional simulation with piece

This gripper, with one specific intermediarypiece, can be mounted on any industrialcommercial robot(see Fig. 10) . One of itsconfiguration can be obtained, during the gripper ismounted on robot, with change the relativeposition of the fingers, regarding the form of thegrasped object.

Figure 10: Example with the gripper mounted onABB robot

For functional simulation of the graspedoperations, the robot with the gripper weretransferred in virtual reality –VRML soft (Fig.11).

Figure 11: Transfer robot with gripper in VRML soft

Here we can test different grasping operationfor different objects. Then, the results, for onecorrect grasp, can be used for programming thereal gripper

7. Action and Command Scheme

7.1 Action scheme

The gripper is acted with four pneumaticmotors. The dimension of the piston of thepneumatic motor will be : s = Fm/p( where p is thepressure). For concrete adopted values is selectedthe motor: DSNU 25-25-P-A-MA-S2. Thepneumatic scheme is shown in Fig. 12.

7.2 Command scheme For command are used the following devices:

drossels (LRMA-1/8-QS-8), adapter ( SGS-M10x1,5), end component (CPE 14 – PRS –EP),expanding bloc ( CPE14 – PRSEO-2), end element( CPE14 – PRSGO – 2), blocked element ( CPE14-PRSB).

Figure 12: Pneumatic scheme

The control subsystem is make of eight sensorsCZN-CP15 type with the following characteristics:- 40 C degrees until + 85 C degrees; 0,2 until 100N grasp force; intensity: 1 Ma; period of life at 35N:10 million of operations and a signalconvector(1 M 36-22 Ex-U).

The general scheme for motor, command andcontrol subsystem for one finger is shown in Fig.13.The grasp process has the main followingstages: start signal for closing the gripper (electro-mechanical , electrical or voice); sensing or notsensing of the object by the tactile sensors;obtaining the grasping force; transfer of theobject; open the gripper ( similar as the closingstage of the gripper).


266

Figure 13: Command and control scheme

8. CONCLUSIONS

The next conclusion can be formulated inaccording to the considerations presented:

a) The main stages for to design theanthropomorphic mechanical grippers are:structural synthesis and analysis, kinematicssynthesis and analysis, static synthesis andanalysis, constructive design and 3D model andfunctional simulation.

b) These grippers can be obtained using twomain modules: the support – the palm and thefinger.

c) The family of the mechanicalanthropomorphic grippers for robots with two,three, four and five identically fingers has morevariants, what can be obtained in accordance withthe number and the relative position of the fingers.

d) The aspects shown in this paper can be usedat families of the anthropomorphic grippers withmore than four fingers, with five or six identicalfingers.

e) Each finger can be acted with onepneumatic motors and for command andcontrol can be used one classical commandscheme.

References[1] Staretu, I., Daj, I. Mechanisms and Machine

Elements (in Romanian), Ed. Lux Libris,Brasov, Romania, 2000.

[2] Staretu, I., Gripping systems (in Romanian),Ed. Lux Libris, Brasov, Romania, 1996.

[3] Staretu, I., a.a. Mechanical Hand.Anthropomorphic Gripping Mechanism forprostheses and robots (in Romanian), Ed. LuxLibris, Brasov, Romania, 2001.

[4] STARETU, I., BOLBOE, M., Synthesis,Analysis, Design and Functional Simulationfor a Family of Anthropomorphic Grippers forRobots, Proceedings of the SISOM 2007, pp.131-134, Bucharest, 29-31 May, Romania.

[5] STARE U, I., Anthropomorphic GrippingSystems with Jointed Bars or Wheel and wiresfor Industrial Robots – Constructive Synthesis,Analysis and Design, New Trends inMechanisms, Editors S.M. CRE U andDUMITRU N. Academica-GreifswaldPublishing House, 2008, pp. 133-144.

[6] STARE U,I., Sisteme de prehensiune(Gripping Systems)-editia a II-a, Ed. LuxLibris, Bra ov, 2010.

[7] STARE U,I., Gripping Systems, DercPublishing House, Tewksbury,Massachusetts,U.S.A., 2011.

.


267

THE INFLUENCE OF ENVIRONMENT TEMPERATURE VARIATIONON THE STRENGTH CHARACTERISTICS OF COMPOSITE MATERIALS

TYPE “ALUCOBOND"

Constantin DULUCHEANU1, Traian Lucian SEVERIN1, Nicolai BANCESCU1,

1 tefan cel Mare” University of Suceava, Faculty of Mechanical Engineering, Mechatronics andManagement, University Street, No. 13, 720225, Suceava, Romania, EU. [email protected],


Abstract: In this paper the authors present an analysis of the influence of temperature variation onthe strength characteristics of the composite material called „ALUCOBOND”;, a material used inconstruction on plating exterior wall of buildings and the execution of ornamental ensembles foradvertising purposes.

Keywords: composite material, environment, temperature, resistance

1. General consideration

The composite material called ALUCOBONDis made of aluminum plate of 0.5 mm, joined by acore of permanent polyethylene; blend is achievedat constant chemical and mechanical (hot-rolledand pressure) and the grip is perfect on all thematerial. This structure provides a surface rigidityof the composite material, resistance to bending,perfect flatness and ease of processing.

Even if they have a lightweight compared toother materials, ALUCOBOND boards have a verygood resistance to breaking or bending, while theplates have exceptional qualities in terms ofresistance at high pressures and buckling.Polyethylene core which is non-toxic have a gooddensity and improves resistance to fire.

ALUCOBOND dampens vibrations of 9 - 10times more efficiently than aluminum sheet; hebehaved as an electromagnetic coating and for thisreason is a very good solution for plating hospitals,airports, military bases, buildings in city centersetc. Also the manufacturers of these compositematerials say that ALUCOBOND does not changethe mechanical properties between -50 ºC and +80ºC, [3]


Because this composite material is used forexternal plating of the buildings walls and theexecution of ornamental ensembles for advertising

purposes which is used as a support for variousobjects, the authors proposed to determine whetherthe environmental temperature changes affect itsresistance properties.

To do this, lots of 10 pieces of 4 mm thick byALUCOBOND were subjected to followinglaboratory thermal stresses:

I. without thermal stresses (samples in deliverystate of material);

II. long cycle of heating-cooling;III. two long cycles of heating-cooling;IV. three long cycles of heating-cooling; V. six short cycles of heating-cooling.

A "long cycle" consisted of heating samples inan oven at 37 °C for one hour followedimmediately by cooling to -30 °C for one hour,cooling done in a laboratory refrigerator existing inthermal treatments at "Stefan cel Mare" UniversitySuceava. The "short cycle", heating and coolingwere performed at the same temperature (37 °C and-30 °C), but times were held at that temperaturewere only 15 minutes. The samples were made inaccordance with applicable standards (SR EN10002-1:2002) [2].

After the planned thermal stresses, the sampleswere traction tested. These tests were conducted inTechnology of Cold Pressing laboratory on atraction test machine of steel specimens (Figure 1),it is equipped with a linear displacement transducerwhich is connected with a strain gauge deck plate

mailto:dulu:@fim.usv.ro

mailto:severin.traian:@fim.usv.robancescu


268

by acquisition board LabJack U12 and makespossible the transmission of values to a computer[4]. Using special software, the data transmittedcan be automatically calculated and displayed

Figure 1. Traction test machine of steel samples

Acquisition and evaluation size program ismade in LabView programming environment; ittakes the sensor information in a chartrepresentation as presented in Figure 2. Theapplication contains a button to save data to a filetype. ".dat" The charts of the two sizes

are represented in different colors: red - "force"and green - "elongation".

Figure 2. Chart representation of the sizes results in thetraction test

The values recorded in the file. ".dat" are openwith Microsoft Office Excel program (Figure 3)are taken only force and elongation recordedvalues and placed in another Excel program thatcan generate the following representations charts:conventionally ( ); conventional ( ), realexperimental ( ), real experimental ( ), realtheoretically ( ), real theoretically ( ), [1].

Figure 3. Values recorded in the file “.dat”


269

3. Results and discussionAfter thermal stresses according to schedule,

the samples were subjected to traction tests, resultsare presented in Table 1 and Figures 4 6.

Tabel 1. Experimental resultsSample

lotsmax r

[MPa] [%] [MPa] [%]I 27,74 100 26,76 100II 27,66 99,72 26,68 99,70III 26,78 96,54 25,97 97,05IV 26,14 94,23 25,17 94,06V 26,40 95,17 25,53 95,40

Figure 4. Graphical representation of experimental results for sample 3 in group IV

Figure 5. The influence of thermal stresses on max

Figure 6. The influence of thermal stresses on r.


270

Analyzing the results is found a reduction of theresistance characteristics from 27.74 MPa for max,on samples in delivery state (group I), to 26.14MPa for samples in group IV (samples subjected tothree long cycles of heating-cooling), thatcorresponds to a reduction of 5.77%, in the case of

r values, the decline was from 26.76 MPa tosamples from group I to 25.17 MPa for those ingroup IV, ie 5.95% reduction. Similar results wererecorded for specimens in group V were subjectedto six short cycles of heating and cooling: max =26.40 MPa (4.83% reduction) and r = 25.53 MPa(4.60% reduction).

4. ConclusionsComposite material type ALUCOBONT is used

in plating exterior walls of the buildings and theexecution of ornamental ensembles for advertisingpurposes which is used as a support for variousobjects.

To check reaction of environmental temperaturechanges the samples for traction test weresubjected to thermal cycles involved heating at 37°C and cooling to -30 °C, maintaining thesetemperatures for 60 minutes (cycle "long" ) and 15minutes (cycle "short").

The analysis of the results was a decrease inresistance characteristics ( max and r) with valuescontained between 0.3% and 5.77%, depending onthe number of heating and cooling cycles andapplied times to maintain use.

References[1] Severin, L., Iacob, D., Severin, T.L.,

Tehnologia pres rii la rece, EdituraUniversit ii Suceava, ISBN 973-666-149-0,Suceava, 2005.

[2] SR EN 10002-1:2002.[3] * * *, www.geplast.ro.[4] * * *, www.labjack.ro.

http://www.geplast.ro./

http://www.labjack.ro./


271

MATHEMATICAL MODELS FOR SMOOTHING PROCESS BYMECHANICAL SHOCKS OF CYLINDRICAL SURFACES

Raluca Tanasa1, Traian Gramescu1

Technical University “Gheorghe Asachi” of Iasi-RomaniaDepartment of Machine Manufacturing Technology


Abstract: Smoothing by mechanical shocks is a processing method based on the impact effect of theballs with the work piece material. Specific phenomena such processing methods result in changes ofmechanical properties of the surface layer of the piece; also change the appearance of the original surfacelayer. The surface obtained after processing the outer cylindrical surface by mechanical shock isconsidered a chain of small cavities generated by the impact of the balls with the workpiece material. Theouter cylindrical surfaces roughness obtained after the smoothing using centrifuged balls, is an importantparameters of this process. The aim of this paper is to present ways to develop mathematical models ofsurfaces roughness. Also are presented experimental results that confirm the validity of mathematicalmodels and their level of approximation of the real situation.

Keywords: Smoothing, Roughness, Plastic deformation, Hardening

1. IntroductionIndustrial practice highlights a number of

technological possibilities for surface finishing ofpieces belonging to metal parts. The specialistsconsider that the cold plastic deformation of themetallic workpieces could be included into thegroup of cold manufacturing processes, when thetemperature during the process does not exceed 30% of the melting temperature of the respectivematerial. There are various cold manufacturingprocesses; even the machining processes (cutting,drilling, turning, milling, grinding etc.) have asinitial basic phenomenon the plastic deformation ofthe workpiece material. Along with cutting ornonconventional processes, cold plasticdeformation is used extensively in surfacefinishing.

Extensive research conducted over the lastcategories of interest to show their developmentprocesses [6]. The advantages of these processes,compared to the other are obvious: can be appliedto a variety surfaces, high mechanicalcharacteristics of the surface layer, surfaceroughness comparable to that obtained fromcutting, much lower costs related to processing. Amanufacturing techniques based on the cold plasticdeformation is caused by the impact effect of thehard metallic balls with the workpiece material.

From finishing processes of surfaces by plasticcold deformation, cold centrifuged ball strikingprocess, so-called ball peening [4] appears moreeconomically due to its simplicity and because itcan be applied to pieces with lower stiffness.

The method involves the hitting of theworkpiece material by balls found in motion, sothat each impact contributes to the generation ofsmall cavity in the workpiece surface layer.

The actions of hard balls sent with specific forceand frequency over metallic surface of theworkpiece have significant effects on the structureof the superficial surface layer but also on the sizeof micro-irregularities surface.

The balls can be directed to the workpiecesurface in different ways; they could be transportedby a gas jet [4], but there are also equipmentswhich direct the balls just to the workpiece surfaceto be processed.

During the process there is a clash of distortingelement (hard ball) to piece surface. The ballhardness is much higher than the one of the pieceto be processed. The shape of piece after crash willresult in form of a spherical cap with sizedependent of working conditions.

Observing the condition of generatingcylindrical surface, strikes (collisions) will becarried throughout the circumference of the pieceby turning it with speed np. It can be concluded

mailto:raluca.tanasa:@gmail.com

mailto:grat:@tcm.tuiasi.ro


272

that collisions constantly producing at part rotationangle at the center p (Fig. 1) there will always bean oscillation of level with a certain height Rmaxsurface, which is micro-irregularities of the processwith the same size [3].

Figure 1: Schematic diagram of the smoothing processwith centrifuged balls.

Methodology presented in this paper is todetermine theoretical mathematical expressions(models) of the roughness Rmax. Also are presentedsome experimental results obtained after applyingthe smoothing process by mechanical shocks todetermine the reliability of the models obtained,compared with experimental results.

2. Developing mathematical models

From Fig. 1 can be observed that the surfacedeformation is in the shape of a spherical cap, witha radius equal with balls radius, rb. This highlightsthe possibility that a complete rotation of the partobtain a number of spherical caps Nc, that can becomputed with Eq.(1).

2'

pc

c

rN

l. (1)

The spherical cap length l’c ( ADB arc) can beapproximated with length rope ( AEB ), the valueof segment ED is insignificant in relation to thesize of rope. Using geometrical relations existingin the triangle bO AE can be expressed the Eq. (2).

2max max2 2c bl r R R . (2)

By entering into Eq. (1) the obtained value of lc,the resulted relationship it will have the form ofEq. (3).

2max max2

pc

b

RN

r R R. (3)

On the disk tool are mounted a number of balls,Nb. After processing will be performed a number ofspherical shape on the parts surface, expressed bythe Eq. (4):

dc b

p

nN Nn

. (4)

In Eq. (4), with the nd is expressed the disk toolrotation, and np represent part rotation. CombiningEqs. (3, 4) results equation (5):

22max max2 0p p

bd b

r nR r R

n N. (5)

The solutions of Eq. (5) are:

22

maxp p

b bd b

r nR r r

n N. (6)

It will be considered for the value of Rmax only

22

maxp p

bd b

r nR r

n N.

Because for the higher values of Rmax than the ballradius rb, the circle of maximum section has theradius / 2b cr l , and the approximation isinconclusive.

The analysis of Fig. 1 shows the fact that to eachspherical cap from the surfaces part correspondinga center angle:

2p

cN. (7)

The Nc parameter from Eq. (7) represent thenumber of spherical cap resulted on the partsurfaces in one pass of disk tool over the piece.

Accepting the idea that the geometrical figureADBE represent a spherical cap, can be written thefollowing mathematical relations:

22 sin ;4

2 sin2

pp

pc p

ED r

l AEB r. (8)

Replacing the Nc parameter from Eq. (4) in Eq.(7) we get next relation:


273

2 pp

d b

nl

n N. (9)

The AEB rope is common to both circles withradius rp (piece) and rb (deformation ball).Relations based Eqs. (8, 9) can be written:

22 sin 2 sin

2pb

b pd b

nr r

n N. (10)

Applying mathematical rules can be obtainedthe expression for calculating the anglecorresponding to a center ball of a spherical cap:

22arcsin sinp p

bb d b

r nr n N

. (11)

b angle allows finding the height of sphericalcap (CE ) in the relationship

2

2

1 cos 2 sin2 4

212 sin arcsin sin2

b bb b

p pb

b d b

CE r r

r nr

r n N

(12)

To determine the expression for calculating theroughness surface Rmax we will consider the Eqs.(8, 9, 12).

max

2

2

212 sin arcsin sin2

22 sin

p pp

b d b

pp

d b

R CE ED

r nr

r n N

nr

n N

.(13)

From the analysis we see that for the theoreticalcalculation of height, Rmax, of micro irregularitiescan use different expressions (see Eqs. (6, 13).

3. Experimental research

To see the approximation of the theoreticalresults by applying the expressions referred tothose achieved in practice, were performedexperimental tests. To convert the surfaceroughness Rmax in Ra was used the approximaterelations [2]:

max

max

105

110

a a

a a

RR for R m

RR for R m. (14)

2.1 Equipment and work piece material

Research, [8] has been conducted with a specialdevice mounted on a universal lathe (Fig. 2). Thedevice features a disk tool with balls. It used aspecial tool to direct balls to the surface of theworkpiece. It consists of a cylindrical disk withholes in witch are placed in predetermineddiameter balls.

Figure 2: Device for smoothening by plasticdeformation with centrifuged balls.

At the periphery of this disc, there is a ringwhich is made a number of holes equal to thenumber of holes; diameter of these holes is smallerthan the diameter of balls inserted. In this way, theballs move short distances in cylindrical holes;when the disc is rotated, (Fig. 3) the balls are rolledon the ring under the action of centrifugal forceand parts thereof beyond the outer cylindricalsurfaces of the ring.

To be affected by the ball peening process,the hardness of the ball material has to exceedthe hardness workpiece material with at leastseveral HRC unities. The process of the ballpeening presented in this paper is applied inthe case of the workpiece materials able to beaffected by a certain plastic deformation andhardening; this means that usually, theworkpiece material must have a phasestructure corresponding to that obtained by anannealing heat treatment.

disk tool

devicework piece


274

In the considered case, the axis of theworkpiece and the axis of the disk tool are paralleland placed in the same horizontal plan.

When the disk tool does not perform therotation motion, the balls existing in the holes fromthe disk tool body has a position essentiallydetermined by the presence of their weight force.

Figure 3: Disks with deforming balls (differentdiameters of balls).

Striking force was achieved by centrifugal forceresulted at disk tool turning with nd speed.

The workpiece and the disk tool performrotation motions in the same direction; due to thisfact, the ball takes contact with the externalcylindrical surface of the workpiece in a positionsituated over the horizontal plan determined by theworkpiece and the disk tool axes.

As consequence of the impact, if the pressureexerted by the ball under the action of thecentrifugal force on the workpiece surface layer ishigher than the compression resistance of theworkpiece material, a small cavity is generated onthe workpiece surface. At the same time, the ball ispushed on a low distance in the conical holeexisting in the disc tool body, but it is obliged bythe ring to continue the motion along a trajectorywhich is partially circular. Because in the contactzone the workpiece surface and the ball performmotions in opposite directions, the shape of thecavity generated in the workpiece surface layer ismodified; of course, really the initial theoreticalspherical shape of this cavity is also affected by theelastic recovery of the workpiece material.

Experimental researches were complying withthe scheme from Fig. 1.

The experimental test samples for smoothing bymechanical shocks were selected from commonlyused materials in tree construction. Was consideredthat smoothing process by plastic deformation isapplied mainly journal shaft couplings that areused in friction [1]. The test pieces were made ofrolled steel containing 0.37% carbon, quality steelcontaining 0.45 % carbon and an alloyed steelcontaining 0.17 % carbon, 0.22 % molybdenum,0.95 % chromium, 1.3 % nickel.

Experiments were performed on cylindricalsamples of diameter = 45 mm and length L =300 mm. To ensure good crosses were madeduring experiments centering holes. Cylindricalsurface was divided into portions with a length of30 mm, Fig. 4; with "S” is noted the safety sectionleft to avoid possible influence of the chuck overthe finishing operation results.

5

F45

30

L=300

S

Figure 4: Dimensions and geometry of workpiece

Before being subjected to experimentalresearch, all samples were processed by finishingturning.

2.2 Smoothing conditions

Smoothing by mechanical shock occurs at anambient temperature, the temperature well belowthe recrystallization, which draws after itself thatall changes in the metal structure can be easilyrecognized, as a result of plastic deformationoccurred [5]. The considered parameters in order toexpress the theoretical point of view (Eqs. 6, 13),of a cylindrical outer surface roughness resultsapplying the finishing process by mechanicalshocks, were: parts radius (rp), part speed (np), disktool speed (nd), number of balls on disk tool (Nb)and ball radius (rb).

The experimental research carried out will takeinto account the initial roughness (Rai) of theworkpiece undergoing processing by mechanicalshock [7]. The parameter constant during theexperiment was piece diameter; the used diameter

disk tool

ring

ball


275

was = 45 mm. Smoothing process conditions aregiven in Table 1.

Table 1: Synthesis of smoothing conditions

Theo

retic

al p

aram

eter

sEx

peri

men

tal p

aram

eter

s Part radius(mm)

5 – 3017,5

Part speed(rot/min) 16

Disk tool speed(rot/min) 2890

Number of balls on disk tool(mm) 16

Ball radius (mm) 4Initial roughness

m) 2 – 4

3. Results and discussionThe experimental study of the ball meaning

process was made by using the deviceschematically presented in Fig. 2; the device wasmounted on a universal lathe, instead of the toolholder. The disk tool had a diameter Dd of 200mm; it presented a single balls row, the ball radiusbeing rb=8 mm.

In order to study and discuss the effect of theinput parameters and to determine the theoreticalmodels that express surface roughness near reality,were realized the construction from Figs. 5 – 8.These graphs show the effect of smoothingconditions in four different conditions in the casesof expression (6) and expression (13) consideredover the workpiece surface. In this condition, fourof the input parameters were constant, and variedonly one.

The effect of deforming ball diameter on thevariation of surface roughness, using theoreticalmodels, can be appreciated from figure 5. In thecase of first mathematical model (Eq. 6), as thedeforming ball diameter increases, the surfaceroughness decreases.

0

10

20

30

40

50

60

70

80

0.50 1.00 1.50 2.00 2.50rb [mm]

Rm

ax [

m]

Rmax th. (Eq.6)Rmax th. (Eq.13)

Figure 5: Theoretical results of roughness surfaceexpressed with theoretical models at different deformingball diameter

Figure 6 constitute the effect of the part speedon the theoretical surface roughness, in case of themathematical models. With increases of part speeda good value for roughness surface is indicated bythe Eq. 6. From the graph it is preferable to avoidsmoothing at high part speed, because surfaceroughness increases.

0

2

4

6

8

10

12

14

16

18

16 20 25 32 40

np[rot/min]

Rm

ax [

m]

Rmax th. (Eq. 6)Rmax th. (Eq. 13)

Figure 6: Theoretical results of roughness surfaceexpressed with theoretical models at different partspeed

Number of balls on disk tool represents animportant smoothing parameter that affects theresults of process. The increase of number of ballson disk tool will improve (decreases) the roughnesssurface. The shape of graph recommends using Eq.13 from the theoretical models.

0

2

4

6

8

10

12

14

16

18

1 4 8 12 16 Nb

Rm

ax [

m]

Rmax th. (Eq. 6)Rmax th. (Eq. 13)

Figure 7: Theoretical results of roughness surfaceexpressed with theoretical models at different number ofballs on disk tool

Effect of workpiece diameter on the theoreticalroughness is shown in Fig. 8. The graphicsrendering is seen as a small roughness is obtainedwhen is used the mathematical model 2 (Eq. 13).Also it is observed that lower values of surfaceroughness parameter, are obtained at differentdiameters of the part but maintaining aconstant optimal value of the other inputparameters in the process.

Mathematical models and experimental results(Fig. 8) were verified. Values obtained for a givensituation confirms the practical approximationmodel 2 (Eq. 13).


276

0

0.5

1

1.5

2

2.5

3

3.5

4

0 5 10 15 20 25rp[mm]

Rm

ax [

m]

Rmax th. (Eq.6)Rmax th. (Eq.13)OLC45 (Rai1)17MoCrNi14 (Rai1)OL37 (Rai1)OLC45 (Rai2)17MoCrNi14 (Rai2)OL37 (Rai2)

Figure 8: Theoretical results of roughness surfaceexpressed with theoretical models at different partradius; experimental points, resulted in same workingconditions, on different materials.

Deviations from theoretical values of theexperimental results may have several causes:- processed material behaves elastic plastic due to

the phenomenon of hardening;- as a result of the initial hypothesis, according to

which was considered the circular contourfootprint, occur deviations;

- due to superficial hardening phenomena, everyshot is applied near the previous. That why theprocessed material by mechanical shocks it ispartial hardened, fact that gives elastic plasticproperties, which change in practice the initialassumptions made.

- practice values deviation in comparison with thetheoretical values can be explained due to thefact that during the collision, the hard ball isdeforming because it elasticity resulting a printthat deviate from a spherical cap (as considered)Taking into account the elements of the local

strains theory for massive bodies collision isjustified the reduced roughness values increasingparts diameter.

Some deviations from theoretical values are dueto other factors of influence (initial surfaceroughness, longitudinal advance of the disk toolreporting to piece).

Also different values were obtained andvariation due to plasticity of tested materials. Notethat the processing of material with high plasticity(OL37), we obtain a higher roughness. Moretenacious material (elastic plastic) with a highercontent of perlite (OLC45, 17MoCrNi14) allowsobtaining lower roughness.

4. ConclusionsThe smoothing method presented in this paper

is a processing way of the workpiece cylindricalsurface which is based on the impact effects of theballs found under the action of the centrifugal forcewith the workpiece surface. Also as result of thesmoothing by mechanical shocks, the workpiecesurface layer is affected by a hardening process.Theoretical considerations were used to obtaindifferent mathematical relations to characterize thesuperficial surface roughness of the work pieceresulted after the impact phenomena. Someexperimental researches were finalized to offer animage concerning the influence exerted by certainwork conditions on the surface roughnessparameter Ra. Because each encounter is repeatedin the immediate vicinity of micro crater above ispossible that by increasing the superficial surfacelayer micro hardness the plastic deformationapproach to partially elastic deformation.

References[1] Dieter, G., Mechanical metallurgy (in

Romanian), Bucure ti: Editura Tehnic , 1970.[2] Dietrich, R., Garsaud, D., Gentillon, S.,

Nicolas, M., Precis methodes d’usinage.Methodologie, production et normalisation,Afnor, Nathan, Paris, 1981.

[3] Gr mescu T., Sl tineanu L., Upon sometheoretical aspects at metallic surfaces’smoothing with centrifuged balls, Ia i:Buletinul Institutului Politehnic, Tom LII(LVI), Fasc.5A, 2006.

[4] Korzynski, M., Dzierwa, A., Pacana, A., andCwanek, J., Surf. Coat. Technol., 204 (5), pp.615-620, 2009.

[5] Lin, Y.C., Wang, S.W. & Lai, H.-I., Therelationship between surface roughness andburnishing factor in the burnishing process. InlJ of Advanced Manufacture Technology,Vol.23, pp. 666-671, ISSN 1433-3015, 2004.

[6] Lupescu, O., N Surfaces smoothing by plasticdeformation (in Romanian), Chi in u: EdituraTehnica Info, ISBN 9975-910-67-X, 1999.

[7] Sl tineanu, L, Tanasa, R., Gramescu, T,Coteata, M., Academic Journal ofManufacturing Engineering, 6 (4), 109-113,2008.

[8] Tanasa, R., Researches on finishingtechnologies by mechanical shocks, PhDThesis, Iasi, 2010.


277

EXPERIMENTAL RESEARCH ABOUT INFLUENCE OFINCLINATION ANGLE DIRECTION UPON THE SURFACE

ROUGHNESS IN BALL END MILLING OF OLC45 (C45) MATERIAL

PA CA Ioan1, LOBON IU Mircea2

1 phd. student, North University of Baia Mare, Str. Dr. Victor Babes, nr.62 A , Baia Mare, Romania,e-mail: [email protected].

2 phd. prof., North University of Baia Mare, Str. Dr. Victor Babes, nr.62 A , Baia Mare, Romania,e-mail: [email protected] .

Abstract: Machining with ball end mills become a very usefull process to obtain pieces with complexsurfaces. Possibilities of machining with ball nose end mills represent a adequate process to achievesurfaces with a good quality. That fact lead to exclude some finishing operations that was necessary if thesurfaces was machined with a different process and tools. Elimination of this finishing operation have agreat economic advantages, because milling with ball nose end mills offer a adequate surface roughnessand precision. But, all of this are possible only with a good choise of different values of cuttingparameters and with an optimum inclination angle represented by value and direction of inclination. Thepresent paper includes an experimental investigation of surface roughness depending on the value anddirection of tool axis inclination in raport with surface machined.

Keywords: roughness, surface, inclination angle, direction.

1. Introduction

Obtaining a high quality of surfaces machinedwith ball end mills present a great interest forreserchers and persons involved in machining ofdie and mould surfaces. The surface quality in ballend milling is the results of influence of a numberof factors such as: the stability of machine tools,the correct choosing of cutting tools and cuttingparametters, properties of material machined andthe tool angle inclination upon the surface normal.To have good results all these factors should bechosen to have a minimal negative influence in theprocess.

Machining of dies and moulds surfaces inmilling process involved polishing operations thatrepresent sometimes approximately two thirds ofthe total manufacturing costs [4]. In this way themilling process with ball nose end mills represent asolution to reduce the costs of process and of finalproduct but only if the values of parametersinvolved in process are adequate. Chosen ofoptimum value of different parameters is dificultbecause it is necesary to reduce the time ofmachining and to obtain surfaces with goodroughness. In international context a great numberof researches is dedicated to surface quality

prediction. Some of them is dedicated toexperimental research, but a lot of them areproposing some mathematical models. Thatmathematical models proposed take into accountthe cutting speed, feed value variation and thefollowing parameters: cutting depth [8], cuttingdepth and radial depth [5], hardness of the materialand tool radius [7].

By using a cutting tool angle between the axisof tool and machined surface can contribute to ahigher quality of machined surface [1].

The paper [9] and [6] suggested that a toolinclination in the range of 10…20 degreesrepresents the optimum machining strategy forhigh speed milling in the die and mould makingindustry, but in other paper like [10] the optimalvalue of inclination angle in ball end milling ofblock materials was found at 15 degrees. Anotherreserch show that when piece was machined withinclination of tool axis of -17 degrees following Yaxis in one way the profile of the machined work-piece, show the improvement of the machinedsurface texture quality [2]. In paper [3] it is notedthat in practice the angle of 45 degrees forinclination of tool axis the quality of surface islow.

mailto:pascaing:@yahoo.com

mailto:mircea.lobontiu:@ubm.ro


278

Because surface roughness have an importanteffect upon product quality and economicity ofprocess, was necesary to study the influence ofinclination direction of tool axis upon surfacenormal and try to compare the surface roughnessobtained in some situation.

2. The analysis of tool axis inclinationMachine tools with five axis lead the possibility

to work with one or two angle of tool axisinclination in relation with normal to the surface,situation that allow to achieve better cuttingconditions and therefore a high quality machinedsurface. Working with this inclination angle thecontact zone between tool edge and material ischanging that involve another condition withmodification of chip thickness and area of contact.Also, the cutting forces are different and theenergy consumed are different. Using thisinclination angle we have possibility to reduce thecosts of process and to obtain a higher quality.

Inclination angle can be applied to tool aroundthe axis X,Y or Z corresponding to the structuralcharacteristics of machine tools and tocharacteristics of surfaced machined.

Inclination angles proposed for analysis in thispaper are around the X axis by angle A positive(Fig.1) or negative (Fig.2) and around to Y axis byangle B positive (Fig.3) or negative (Fig.4).

3. Effects of inclination angle upon effectivecutting speed

By using an inclination angle between tool axisand normal to surface we change somecharacteristics of process. One of them is effectivecutting speed thus value depending by effectivediameters. In order, the effective diameters dependby cutting depth, radial depth and inclinationangle. Increasing of effective diameters involve theincrease of effective cutting speed.

To establish the value of effective speedmilling Vc-eff coresponding to inclination of 15degress angle value, we used relation (1) [3]:

1000

RpaR

arccosnsinDn

efcV (1)

with the following condition:

Figure 1: Inclination of tool axis in B negative directionaround Y axis

Figure 2: Inclination of tool axis in B positive directionaround Y axis

Figure 3: Inclination of tool axis in A negative directionaround X axis


279

Figure 4: Inclination of tool axis in A positive directionaround X axis

RpaR

arccos90nR2ea

arcsin , (2)

where:-D – nominal diameter of tool [mm];-ap – cutting depth [mm];-n – spindle speed [rot/min];

n – inclination of tool axis [grade];-R – tool radius [mm].

Geometrical parameters values used in ourexperiments are presented in Table 1.

Table 1:Geometric parameters used

No. Geometricelements

Units ofmeasurement

Values

1 Diameter ofball end mill

mm 16

2 Cutting depth mm 0,23 Axial depth mm 0,24 Feed per

toothmm 0,1

5 Tool angleinclination

degrees 15

The results obtained for two values ofspeendle speed used in our experiments arepresented in Table 2.

4. Experimental work and conditions

Machining by using a cutting tool anglebetween the axis and machined surface cancontribute to a higher quality of machined surface

[1]. Analyzing the indication presented in paper[9],[6],[10], [11] and [2] about optimum value ofinclination angle we choose to apply in ourexperiment value of 15 degrees in raport withnormal to the surface.

Table 2:Effective cutting speed valuesNo. Speendle

speed n[rot/min]

Inclinationangle

[degrees]

Effectivecuttingspeed

[m/min]1 10000 15 234,62 15000 352

Once identified the optimum value ofinclination angle we found that is necesary toextend the studies by analysing the influenceof inclination angle direction upon surfaceroughness. We want to establish the bestdirection of inclination angle that lead to highquality of machined surface, but accordingwith some defaults machining conditions.

4.1 Experimental setupTo make the experimental analysis of

inclination angle direction upon surface roughnesswas necessary a number of experiment that wasmade by using a OKUMA MU400VA five-axisCNC milling machine equipped with a maximumspindle speed of 15,000 rpm (Fig.5.), that waspropriety of S.C. RAMIRA S.A. We used a toolholder SRM2160SNM and inserts SRG16C-VP15TF made by Mitsubishi Carbide that ispresented in Fig.6.

Figure 5: Five-axis milling machine


280

Figure 6: Tool holder SRM2160SNM

4.2 Workpiece configuration

The material used was OLC-45 (C45) withthe following caracteristics: 0,42…0,50% C,0,50…0,80% Mn, 0,17…0,37% Si, maximum0,040% P etc. The workpiece had a economicshape because we can machined six surfaceswith the same quatity of material (Fig.7.)

Figure 7: Workpiece configuration

4.3 Surface roughness measurements

In the technical literature and industry aremany parameters used to defined the surfaceroughness, but in generally is defined as theinherent irregularities of the workpiece affected bymachining process. The paper [3] show that themost popular of the 2D parameters is the averageroughness Ra. In order to obtain more informationabout surface machined we made themeasurements of surface roughness in feeddirection and perpendicular to it. We used thetester TR200 that we applied on new device thatprovides the conditions of perpendicularity andparallelism between roughness tester feeler andfeed direction. That measurements stand ispresented in Fig. 8.

4.4 Experimental conditions

In order to establish the conditions utilizedin our experiments we must say that allexperiments were conducted using in feed millingdirection and with parameters indicated in Table 1,2 and 3.

Figure 8: Roughness tester TR200 with adaptor device

Table 3: Experimental conditionsNr Inclination

angle[degrees]

Spindlespeed

[rot/min]

Inclinationdirection

1

15

10000

A0B+152 A0B-153 A+15B04 A-15B05

15000

A0B+156 A0B-157 A+15B08 A-15B0

5. Experimental results and discussion

Experimental data on surface quality of themeasurements obtained were analyzed in feeddirection and perpendicular to feed direction fortwo values of effective cutting speed and werecentralized in Table 4.

Table 4: Experimental dataNr.

Dire

ctio

n of

incl

inat

ion

Effe

ctiv

ecu

tting

sp

eed

[m/m

in]

Surfaceroughness [µm]

In

feed

dire

ctio

n

Cro

ss

tofe

eddi

rect

ion

1 A0B+15 234,6 0,176 0,3172 352 0,191 0,3413 A0B-15 234,6 0,189 0,4354 352 0,195 0,4985 A+15B0 234,6 0,335 0,3576 352 0,367 0,4917 A-15B0 234,6 0,317 0,3608 352 0,192 0,411


281

Direction of inclination angle of tool axisrelative to the surface normal is important becausesiginificant differences can arise in terms ofquality. Both, absolute value of the angle ofinclination and direction of this angle are veryimportant.

Variation of surface roughness when machinedOLC 45 at 15 degrees inclination angle valueapplied in different direction is presented in Fig. 9for measurements made in feed direction, and inFig.10. for measurements made perpendicular tofeed direction.

Figure 9: Variation of roughness for measurementsmade in feed direction

Figure 10: Variation of roughness for measurementsmade perpendicular to feed direction

The results releave the fact that it isrecomanded to avoid situation of inclination whenresults contact between material machined and tooledge were the effective cutting speed is reduce.This situation appear when we inclined tool axisaround X axis in direction A positive and aroundY axis in direction B negative. In this case wheninclination angle value is 15 degrees it is

recomanded to not increase the cutting speedbecause the surface roughness is worse, factobserved for situation when we mademeasurements in feed direction and perpendicularto feed direction. It is possible that for anothervalue of inclination angle the surface roughnessvariation to be different.

6. ConclusionsExperimental results show that direction of

inclination in ball end milling process have a greatinfluence upon surface roughness.

Using the direction of inclination angle aroundthe X axis with angle A positive or negative mustbe avoid because the surface roughness is weakercomparativ with direction in inclination madearound Y axis. Best results of surface roughness isobtained when we incline tool axis around Y axiswith angle B positive. Difference between surfaceroughness obtained with inclination angle in Bpositive direction, whose the best situation, andinclination in A positive direction is about 90% formeasurements made in feed direction and 40% formeasurements made perpendicular to feeddirection.

References

[1] Antoniadis, A., Bilalis, N., Savakis, C.,Maravelakis, E. and Petropoulos, G., Influenceof machining inclination angle on surfacequality in ball end milling, Proceedings ofAMPT2003, Dublin, Ireland, 8-11 July 2003.

[2] Boujelbene, M., Moisan, A., Bouzid, W. andTorbaty, S., Variation cutting speed on the fiveaxis milling, Journal of Achievements inMaterials and Manufacturing Engineering,Vol. 21, Issue 2, April 2007.

[3] Daymi, A., Boujelbene, M., Linares, M.,Bayraktar, E. and Ben Amara, A., Influence ofworkpiece inclination angle on the surfaceroughness in ball end milling of the titaniumalloy Ti-6Al-4V, Journal of Achiements inMaterials and Manufacturing Engineering,Vol.35, Issue 1, July 2009.

[4] Fallbohmer, P., Rodriguez, C., A., Ozel, T. andAltan, T., High-speed machining of cast ironand alloy steels for die and mouldmanufacturing, Journal of MaterialsProcessing Technology 98, pp. 104-115, 2000.

[5] Kadirgama, K., Noor, M., Rahman, M., Rejab,R., Haron, C. and Abou-El-Hossein, K.,Surface roughness prediction model of 6061-


282

T6 Aluminium alloy machining usingstatistical method, European Journal ofScientific Research, Vol.25, No. 2, ISSN 1450-216X, 2009.

[6] Ko, T., J., Kim, H., S. and Lee, S., S.,Selection of the machining inclination angle inhigh-speed ball end milling, InternationalJournal of Advanced ManufacturingTechnology 17, pp. 163-170, 2001.

[7] Krizbergs, J. and Kromanis, A., Methods forprediction of the surface roughness 3Dparameters according to technologicalparameters, 5th International DAAM BalticConference ,, Industrial engineering-addinginnovation capacity of labour force andentrepreneurs’’, Tallinn, Estonia, 20-22 April2006.

[8] Rashid, M., Gan, S. and Muhamad, N.,Mathematical modeling to predict surfaceroughness in CNC milling, World Academy ofScience, Engineering and Technology 53,2009.

[9] Schultz, H. and Hock, S., High-speed millingof dies and moulds-cutting conditions andtechnology, Annals of the CIRP 44, pp. 35-38,1995.

[10]Tonshoff, H., K., and Hernandez, J., C., Diemanufacturing by 5 and 3 axes milling, Journalof Materials Processing Technology 20, pp.105-119, 1989.

[11]Vidakis, N., Antoniadis, A., Savakis, C. andGotsis, P., Simulation of ball end tools milling,Proceedings of International Conference ICPR2001, Prague, Czech Republic, 29.07-03.08.2001.


277

EXPERIMENTAL RESEARCH ABOUT INFLUENCE OFINCLINATION ANGLE DIRECTION UPON THE SURFACE

ROUGHNESS IN BALL END MILLING OF OLC45 (C45) MATERIAL

PA CA Ioan1, LOBON IU Mircea2

1 phd. student, North University of Baia Mare, Str. Dr. Victor Babes, nr.62 A , Baia Mare, Romania,e-mail: [email protected].

2 phd. prof., North University of Baia Mare, Str. Dr. Victor Babes, nr.62 A , Baia Mare, Romania,e-mail: [email protected] .

Abstract: Machining with ball end mills become a very usefull process to obtain pieces with complexsurfaces. Possibilities of machining with ball nose end mills represent a adequate process to achievesurfaces with a good quality. That fact lead to exclude some finishing operations that was necessary if thesurfaces was machined with a different process and tools. Elimination of this finishing operation have agreat economic advantages, because milling with ball nose end mills offer a adequate surface roughnessand precision. But, all of this are possible only with a good choise of different values of cuttingparameters and with an optimum inclination angle represented by value and direction of inclination. Thepresent paper includes an experimental investigation of surface roughness depending on the value anddirection of tool axis inclination in raport with surface machined.

Keywords: roughness, surface, inclination angle, direction.

1. Introduction

Obtaining a high quality of surfaces machinedwith ball end mills present a great interest forreserchers and persons involved in machining ofdie and mould surfaces. The surface quality in ballend milling is the results of influence of a numberof factors such as: the stability of machine tools,the correct choosing of cutting tools and cuttingparametters, properties of material machined andthe tool angle inclination upon the surface normal.To have good results all these factors should bechosen to have a minimal negative influence in theprocess.

Machining of dies and moulds surfaces inmilling process involved polishing operations thatrepresent sometimes approximately two thirds ofthe total manufacturing costs [4]. In this way themilling process with ball nose end mills represent asolution to reduce the costs of process and of finalproduct but only if the values of parametersinvolved in process are adequate. Chosen ofoptimum value of different parameters is dificultbecause it is necesary to reduce the time ofmachining and to obtain surfaces with goodroughness. In international context a great numberof researches is dedicated to surface quality

prediction. Some of them is dedicated toexperimental research, but a lot of them areproposing some mathematical models. Thatmathematical models proposed take into accountthe cutting speed, feed value variation and thefollowing parameters: cutting depth [8], cuttingdepth and radial depth [5], hardness of the materialand tool radius [7].

By using a cutting tool angle between the axisof tool and machined surface can contribute to ahigher quality of machined surface [1].

The paper [9] and [6] suggested that a toolinclination in the range of 10…20 degreesrepresents the optimum machining strategy forhigh speed milling in the die and mould makingindustry, but in other paper like [10] the optimalvalue of inclination angle in ball end milling ofblock materials was found at 15 degrees. Anotherreserch show that when piece was machined withinclination of tool axis of -17 degrees following Yaxis in one way the profile of the machined work-piece, show the improvement of the machinedsurface texture quality [2]. In paper [3] it is notedthat in practice the angle of 45 degrees forinclination of tool axis the quality of surface islow.

mailto:pascaing:@yahoo.com

mailto:mircea.lobontiu:@ubm.ro


278

Because surface roughness have an importanteffect upon product quality and economicity ofprocess, was necesary to study the influence ofinclination direction of tool axis upon surfacenormal and try to compare the surface roughnessobtained in some situation.

2. The analysis of tool axis inclinationMachine tools with five axis lead the possibility

to work with one or two angle of tool axisinclination in relation with normal to the surface,situation that allow to achieve better cuttingconditions and therefore a high quality machinedsurface. Working with this inclination angle thecontact zone between tool edge and material ischanging that involve another condition withmodification of chip thickness and area of contact.Also, the cutting forces are different and theenergy consumed are different. Using thisinclination angle we have possibility to reduce thecosts of process and to obtain a higher quality.

Inclination angle can be applied to tool aroundthe axis X,Y or Z corresponding to the structuralcharacteristics of machine tools and tocharacteristics of surfaced machined.

Inclination angles proposed for analysis in thispaper are around the X axis by angle A positive(Fig.1) or negative (Fig.2) and around to Y axis byangle B positive (Fig.3) or negative (Fig.4).

3. Effects of inclination angle upon effectivecutting speed

By using an inclination angle between tool axisand normal to surface we change somecharacteristics of process. One of them is effectivecutting speed thus value depending by effectivediameters. In order, the effective diameters dependby cutting depth, radial depth and inclinationangle. Increasing of effective diameters involve theincrease of effective cutting speed.

To establish the value of effective speedmilling Vc-eff coresponding to inclination of 15degress angle value, we used relation (1) [3]:

1000

RpaR

arccosnsinDn

efcV (1)

with the following condition:

Figure 1: Inclination of tool axis in B negative directionaround Y axis

Figure 2: Inclination of tool axis in B positive directionaround Y axis

Figure 3: Inclination of tool axis in A negative directionaround X axis


279

Figure 4: Inclination of tool axis in A positive directionaround X axis

RpaR

arccos90nR2ea

arcsin , (2)

where:-D – nominal diameter of tool [mm];-ap – cutting depth [mm];-n – spindle speed [rot/min];

n – inclination of tool axis [grade];-R – tool radius [mm].

Geometrical parameters values used in ourexperiments are presented in Table 1.

Table 1:Geometric parameters used

No. Geometricelements

Units ofmeasurement

Values

1 Diameter ofball end mill

mm 16

2 Cutting depth mm 0,23 Axial depth mm 0,24 Feed per

toothmm 0,1

5 Tool angleinclination

degrees 15

The results obtained for two values ofspeendle speed used in our experiments arepresented in Table 2.

4. Experimental work and conditions

Machining by using a cutting tool anglebetween the axis and machined surface cancontribute to a higher quality of machined surface

[1]. Analyzing the indication presented in paper[9],[6],[10], [11] and [2] about optimum value ofinclination angle we choose to apply in ourexperiment value of 15 degrees in raport withnormal to the surface.

Table 2:Effective cutting speed valuesNo. Speendle

speed n[rot/min]

Inclinationangle

[degrees]

Effectivecuttingspeed

[m/min]1 10000 15 234,62 15000 352

Once identified the optimum value ofinclination angle we found that is necesary toextend the studies by analysing the influenceof inclination angle direction upon surfaceroughness. We want to establish the bestdirection of inclination angle that lead to highquality of machined surface, but accordingwith some defaults machining conditions.

4.1 Experimental setupTo make the experimental analysis of

inclination angle direction upon surface roughnesswas necessary a number of experiment that wasmade by using a OKUMA MU400VA five-axisCNC milling machine equipped with a maximumspindle speed of 15,000 rpm (Fig.5.), that waspropriety of S.C. RAMIRA S.A. We used a toolholder SRM2160SNM and inserts SRG16C-VP15TF made by Mitsubishi Carbide that ispresented in Fig.6.

Figure 5: Five-axis milling machine


280

Figure 6: Tool holder SRM2160SNM

4.2 Workpiece configuration

The material used was OLC-45 (C45) withthe following caracteristics: 0,42…0,50% C,0,50…0,80% Mn, 0,17…0,37% Si, maximum0,040% P etc. The workpiece had a economicshape because we can machined six surfaceswith the same quatity of material (Fig.7.)

Figure 7: Workpiece configuration

4.3 Surface roughness measurements

In the technical literature and industry aremany parameters used to defined the surfaceroughness, but in generally is defined as theinherent irregularities of the workpiece affected bymachining process. The paper [3] show that themost popular of the 2D parameters is the averageroughness Ra. In order to obtain more informationabout surface machined we made themeasurements of surface roughness in feeddirection and perpendicular to it. We used thetester TR200 that we applied on new device thatprovides the conditions of perpendicularity andparallelism between roughness tester feeler andfeed direction. That measurements stand ispresented in Fig. 8.

4.4 Experimental conditions

In order to establish the conditions utilizedin our experiments we must say that allexperiments were conducted using in feed millingdirection and with parameters indicated in Table 1,2 and 3.

Figure 8: Roughness tester TR200 with adaptor device

Table 3: Experimental conditionsNr Inclination

angle[degrees]

Spindlespeed

[rot/min]

Inclinationdirection

1

15

10000

A0B+152 A0B-153 A+15B04 A-15B05

15000

A0B+156 A0B-157 A+15B08 A-15B0

5. Experimental results and discussion

Experimental data on surface quality of themeasurements obtained were analyzed in feeddirection and perpendicular to feed direction fortwo values of effective cutting speed and werecentralized in Table 4.

Table 4: Experimental dataNr.

Dire

ctio

n of

incl

inat

ion

Effe

ctiv

ecu

tting

sp

eed

[m/m

in]

Surfaceroughness [µm]

In

feed

dire

ctio

n

Cro

ss

tofe

eddi

rect

ion

1 A0B+15 234,6 0,176 0,3172 352 0,191 0,3413 A0B-15 234,6 0,189 0,4354 352 0,195 0,4985 A+15B0 234,6 0,335 0,3576 352 0,367 0,4917 A-15B0 234,6 0,317 0,3608 352 0,192 0,411


281

Direction of inclination angle of tool axisrelative to the surface normal is important becausesiginificant differences can arise in terms ofquality. Both, absolute value of the angle ofinclination and direction of this angle are veryimportant.

Variation of surface roughness when machinedOLC 45 at 15 degrees inclination angle valueapplied in different direction is presented in Fig. 9for measurements made in feed direction, and inFig.10. for measurements made perpendicular tofeed direction.

Figure 9: Variation of roughness for measurementsmade in feed direction

Figure 10: Variation of roughness for measurementsmade perpendicular to feed direction

The results releave the fact that it isrecomanded to avoid situation of inclination whenresults contact between material machined and tooledge were the effective cutting speed is reduce.This situation appear when we inclined tool axisaround X axis in direction A positive and aroundY axis in direction B negative. In this case wheninclination angle value is 15 degrees it is

recomanded to not increase the cutting speedbecause the surface roughness is worse, factobserved for situation when we mademeasurements in feed direction and perpendicularto feed direction. It is possible that for anothervalue of inclination angle the surface roughnessvariation to be different.

6. ConclusionsExperimental results show that direction of

inclination in ball end milling process have a greatinfluence upon surface roughness.

Using the direction of inclination angle aroundthe X axis with angle A positive or negative mustbe avoid because the surface roughness is weakercomparativ with direction in inclination madearound Y axis. Best results of surface roughness isobtained when we incline tool axis around Y axiswith angle B positive. Difference between surfaceroughness obtained with inclination angle in Bpositive direction, whose the best situation, andinclination in A positive direction is about 90% formeasurements made in feed direction and 40% formeasurements made perpendicular to feeddirection.

References

[1] Antoniadis, A., Bilalis, N., Savakis, C.,Maravelakis, E. and Petropoulos, G., Influenceof machining inclination angle on surfacequality in ball end milling, Proceedings ofAMPT2003, Dublin, Ireland, 8-11 July 2003.

[2] Boujelbene, M., Moisan, A., Bouzid, W. andTorbaty, S., Variation cutting speed on the fiveaxis milling, Journal of Achievements inMaterials and Manufacturing Engineering,Vol. 21, Issue 2, April 2007.

[3] Daymi, A., Boujelbene, M., Linares, M.,Bayraktar, E. and Ben Amara, A., Influence ofworkpiece inclination angle on the surfaceroughness in ball end milling of the titaniumalloy Ti-6Al-4V, Journal of Achiements inMaterials and Manufacturing Engineering,Vol.35, Issue 1, July 2009.

[4] Fallbohmer, P., Rodriguez, C., A., Ozel, T. andAltan, T., High-speed machining of cast ironand alloy steels for die and mouldmanufacturing, Journal of MaterialsProcessing Technology 98, pp. 104-115, 2000.

[5] Kadirgama, K., Noor, M., Rahman, M., Rejab,R., Haron, C. and Abou-El-Hossein, K.,Surface roughness prediction model of 6061-


282

T6 Aluminium alloy machining usingstatistical method, European Journal ofScientific Research, Vol.25, No. 2, ISSN 1450-216X, 2009.

[6] Ko, T., J., Kim, H., S. and Lee, S., S.,Selection of the machining inclination angle inhigh-speed ball end milling, InternationalJournal of Advanced ManufacturingTechnology 17, pp. 163-170, 2001.

[7] Krizbergs, J. and Kromanis, A., Methods forprediction of the surface roughness 3Dparameters according to technologicalparameters, 5th International DAAM BalticConference ,, Industrial engineering-addinginnovation capacity of labour force andentrepreneurs’’, Tallinn, Estonia, 20-22 April2006.

[8] Rashid, M., Gan, S. and Muhamad, N.,Mathematical modeling to predict surfaceroughness in CNC milling, World Academy ofScience, Engineering and Technology 53,2009.

[9] Schultz, H. and Hock, S., High-speed millingof dies and moulds-cutting conditions andtechnology, Annals of the CIRP 44, pp. 35-38,1995.

[10]Tonshoff, H., K., and Hernandez, J., C., Diemanufacturing by 5 and 3 axes milling, Journalof Materials Processing Technology 20, pp.105-119, 1989.

[11]Vidakis, N., Antoniadis, A., Savakis, C. andGotsis, P., Simulation of ball end tools milling,Proceedings of International Conference ICPR2001, Prague, Czech Republic, 29.07-03.08.2001.


283

PARTICLES DIMENSIONAL ANALYSIS AND MICROSCOPICCHARACTERIZATION OF HYDROXYAPATITE POWDER

Aurora Anca Poinescu1, Rodica Mariana ION2,3

1 Valahia University of Targoviste, 2 Regele Carol 1 Street, [email protected],2 Valahia University of Targoviste, 2 Regele Carol 1 Street, [email protected]

3 ICECHIM, Bucharest, Analytical Department, 202 Splaiul Independentei, Bucharest-060021,Romania.

Abstract: The studies conducted for this paper we chose the wet precipitation synthesis method atroom temperature to obtain hydroxyapatite powder. Size particle analysis of hydroxyapatite was made inisopropyl alcohol. The surface study on nm scale and the surface topography were evaluated by usingAFM with tapping mode. The SEM pictures of the grown spherulite crystals shows many sphericalagglomerations and few crystallites of 0.1 m in size with pores.

Keywords: SEM and AFM microscopy, hydroxyapatite powder, wet precipitation.

1. IntroductionA challenging problem in health care is the

regeneration of partially or fully lost organs.There are two types of regeneration:

physiological and reparative.Physiological regeneration is a natural

process of disintegration and restoration ofmolecules, cells, and tissue (e.g., bone).

Reparative regeneration in the broad senseis the restoration of parts of body lost as a result oftraumas, healing of defects of tissue and organs,etc.

The use of calcium-phosphate-basedbiocompatible materials close in composition tothe inorganic component of bone tissue contributesto the creation of favorable conditions for thereparative functions of bone tissue.

In the past several decades, a number of implantmaterials based on calcium hydroxyapatite(Ca10(PO4)6(OH2), calcium phosphate ( -Ca3(PO4)2, ceramics of these phosphates,bioglasses, and composites have been used inorthopedics, neurosurgery, and dentistry.

In the case of fractures, the restoration of bonetissue is a complicated biochemical process, which

includes the synthesis of proteins and nitrogen-containing polysaccharides and deposition ofcalcium salts and is accompanied byhypermetabolism.

A serious problem related to thebiocompatibility of materials is the creation ofimplants that would ensure the delivery and in siturelease of a therapeutic agent with osteoinductiveand anti-inflammatory effects [1].

Hydroxyapatite surface coating applicationsinclude metallic orthopedic and dental implantswhere osseointegration HA promotes bothprocesses and reduce the release of metal ionsacting as a physical barrier, preparation for thereplacement of bone fragments. [2].

2. Materials and methods

The synthesis methods of hydroxyapatite aremany, the literature abounds in this area. Thestudies conducted for this paper we chose the wetprecipitation synthesis method at roomtemperature.

Figure 1 shows the flow of technology of wetprecipitation of hydroxyapatite powder by Sung[3].

mailto:poinescua:@yahoo.com

mailto:rodica_ion2000:@yahoo.co.uk


284

Figure 1: Modified chemical precipitation route for HApowder preparation. (Adapted from Sung) [3]

Chemical reagents used in this paper: Ca(NO3)2.4H2O and (NH4) 2HPO4 were separatelydissolved in deionized water. He added Ca(NO3)2in aqueous solution (NH4) 2HPO4, agitatedvigorously for about 1 hour at room temperatureand a precipitate was obtained milky, somewhatgelatinous, which was then shaken for 1h.

The mixture was sintered at 10000C for 24hours. After having been washed and filtered.After filtering the sticky product was compacteddry oven at 8000C. Dried powder was crushed in amortar with pestle and then calcined in an aluminacrucible at three different temperatures 8000C,10000C and 12000C for 1h.

Table 1

Method Sample

Temp.(OC)

Concentr.(M)

Reflux aftermixing

HA1 800 0.1:0.06

Reflux aftermixing

HA2 1000 0.1:0.06

Reflux aftermixing

HA3 1200 0.1:0.06

3. Results

The objectives of this work were:

to synthesize biocompatible hydroxyapatite;

to analyze the size particle of hydroxyapatite;to characterize by SEM and AFMmicroscopy hydroxyapatite powder obtainedby wet precipitation.

3.1. Size particle analysis of hydroxyapatitein isopropyl alcohol

Determination of average particle size andparticle size distribution will be done using a laserdiffraction particle size analysis by MalvernMastersizer Type S-type. It offers a wide range ofmeasurement in the range 0.05 to 3500 microns,excellent data reproducibility and flexibility inhandling evidence.

In the case of HA-A powder disperses notenough to be mechanically made a DLSmeasurement type. Apparent dispersions arecomposed of a particulate phase and no detectableperfectly clear, that the precipitating phase.

Precipitation is too fast and does not allow themeasurement.

So they resorted to grinding (in agate) toprecipitate directly in the presence of alcohol.After the samples were ultrasonic in an ultrasonicbath for 5 min (35kHz). Dispersions havedeveloped an opalescent appearance of whiteprecipitate respectively.

After several minutes, large aggregates havesettled, and then were collected samples from themetastable phase (upper).

That is metastable for about an hour theprecipitate as well. The average size (math) isabout 1800nm (1,8 microns).

But the physical interpretation is quite different,namely: Record several populations of particles.

0

5

10

15

20

25

0.1 1 10 100 1000 10000

Inte

nsity

(%

)

Size (d.nm)

Size Distribution by Intensity

Record 1: HA-A 1 Record 2: HA-A 2 Record 3: HA-A 3

Figure 2: Size distribution of the three HAp samplesafter intensity


285

Considering the first instability manifested bya set of measurements to another (see the three setsof measurements per color), but also change thenumber of population (number of peaks) is verylikely that the system be reviewed format ofelementary particles in different forms ofaggregation result from the process ofsedimentation. The system contains over 6000 nmparticles (the ultimate maximum).

0

10

20

30

40

0.1 1 10 100 1000 10000

Volu

me

(%)

Size (d.nm)

Size Distribution by Volume


Figure 3: Size distribution of the three HAp samples byvolume

In the figures 2, 3 and 4 are illustrated thedistributions of the three Hap samples aftergrading intensity, volume and numberrespectively.

0

10

20

30

40

0.1 1 10 100 1000 10000

Num

ber (

%)

Size (d.nm)

Size Distribution by Number


Figure 4: Size distribution of the three HAp samples bynumber

After global analysis of the system (above) hasdeveloped software filtering of unit intervalanalyzed in view of highlighting the possibleelementary particles between 0.6 - 100nm which inthe global analysis cannot be registered because of

the large aggregates shield. Such particles haveemerged about 40 nm single mode. These in adifferent associations with a different kinds ofaggregates generates large majority in the globalanalysis above.

0

20

40

60

80

100

1 10 100 1000 10000

Inte

nsity

(%

)

Size (d.nm)

Size Distribution by Intensity

Record 6: HA-A (0.6-100nm) 3

Figure 5: Size distribution of the HA-A samples byintensity

0

10

20

30

40

50

1 10 100 1000 10000

Volu

me

(%)

Size (d.nm)

Size Distribution by Volume

Record 6: HA-A (0.6-100nm) 3

Figure 6: Size distribution of the HA-A samples byvolum

0

10

20

30

40

50

1 10 100 1000 10000

Num

ber (

%)

Size (d.nm)

Size Distribution by Number

Record 6: HA-A (0.6-100nm) 3

Figure 7: Size distribution of the HA-A samples bynumber

For HA-A has conducted a global analysis.Reproducibility is better between the sets of


286

measurements, but measured particles seem moreindependent particles and not aggregates.

3.2. SEM microstructural characterisation

Microstructural observations were performedwith a Quanta 200 scanning electron microscopeequipped with software analyzer for quantitativeelemental analysis. For analysis by scanningelectron microscopy, hydroxyapatite powders wereplaced in LR White resin.

Figure 8: Electron microscopy SEM for Hap powdercalcined at 10000C, 2000 X.

In Figure 8 shows electron microscopy ofhydroxyapatite sample calcined at 10000C with2000x magnification and in the next figure samesample but at a 1000x magnification. We candistinguish micro pores (<10 m), allowingdiffusion of ions and fluid macro pore (100-600 m) to allow cell colonization. The micropores give the ceramics osteoconductiveproperties.

Figure 9: Electron microscopy SEM for Hap powdercalcined at 10000C, 1000 X

SEM images of crystals grown shows severalspherical clusters and few crystals of 0.1 m.

Figure 10: Electron microscopy SEM for Hap powdercalcined at 10000C, 10000 X

Subgrains size it is about 70 nm whichcorresponds to the synthetic HA powder size [5].For a good correlation, SEM images show severalclusters of high spherical spherulites andcrystallites less than 0.1 m in size.

3.3. AFM microstructural characterisation

Investigations by Atomic Force Microscope(AFM) were performed with an Agilent 5500 SPMsystem, described by PicoSPM controlled by aMAC Mode module and interfaced with acontroller PicoScan from Agilent Technologies,Tempe, AZ, USA.

All AFM measurements (256 samples / line ×256 lines) were made by scanning the surface at arate of 0.8 - 1.2 lines per second and wereconducted at room temperature, the mode ofpalpation.

For investigation with the atomic forcemicroscope, HA solutions were prepared freshbefore each experiment by suspending anappropriate amount of each sample in ethanol.Scanning movement is led by a piezoelectricscanner which scans the tip in a raster pattern onthe sample (or scans the sample on top).

AFM revealed a rough surface architecture forHA, the predominant size of grains being in therange of 90 - 100nm.


287

Figure 11: AFM image of the HA powder calcined at10000C

AFM is conducted in an ambient gas or liquidmedium and ethanol in our case. AFM shows thearchitecture for the HA surface, the grain is largein the range 70-100nm. Crystal size and sizedistribution reached critical nucleus depends on thesize, in conditions of super saturation, rather than astandard crystal growth, because aggregation ofvery small particles was observed.

Atomic force microscopy (AFM) in comparisonwith the (SEM) can be measured in all threedimensions (x, y, and z) with a single scan. 3-Dsurface topography was recorded on an area of 0.5-0.5 mm2.

Figure 12: 3-D representation of the HA crystalsobtained by wet precipitation

Figure 13: AFM analysis of hydroxyapatite powder

4. Conclusions

Hydrothermal techniques give hydroxyapatitepowders with a high degree of crystallinity andbetter stoichiometry having a wide distribution ofcrystal sizes. Nanometer sized crystals can beobtained at temperatures lower than 100 °C withprecipitation techniques.

The chosen methods was the wet precipitationbecause the wet-chemical precipitation route is themost talented route owing to its ease inexperimental operations, low working temperature,high percentages of pure products and inexpensiveequipment requirement.

In the SEM images of HA, was depicted smallcrystals (<100 nm) in the agglomerated particlesand the uniform grain size with a narrow sizedistribution corresponding to an improvedcrystallinity of HA powders, especially for thatsample after calcination at 1000°C for 1h. For agood correlation, the SEM pictures of the grownspherulite crystals show many sphericalagglomerations and few crystallites of 0.1 m insize.

The determination of size distribution of thegrown spherulite and sintered HA materials werecarried out by atomic force microscopy.

At higher temperature the deagglomeration ofbulk phases and agglomeration of nano phasesleads to the nano crystalline HA in this presentstudy.

The crystal size distribution attained dependson the size of the critical nucleus under thesupersaturation condition, rather than on standard


288

crystal growth, since the aggregation of very smallparticles was observed.

In 3-D graphical representation ofhydroxyapatite crystals (Fig. 13) can be seen onthe grain growth direction z due to the sinteringprocess (wet precipitate) and particularly thecalcination temperature.

One of the main attractions at the AFM is theability to show high-resolution surfaces in theliquid insulation. Since AFM does not rely on theconductivity, and image scanning mechanism willnot be disturbed by the presence of liquid

References:

[1] Zakharov N. A., Polunina I. A., Polunin K.E., Rakitina N. M., Kochetkova E. I.,Sokolova N. P., and Kalinnikov V. T.,Calcium Hydroxyapatite for MedicalApplications, Inorganic Materials, Vol. 40,No. 6, 2004, pp. 641–648.

[2] Monmaturapoj N., Nano-sizeHydroxyapatite Powders Preparation byWet-Chemical Precipitation Route, Journalof Metals, Materials and Minerals. Vol.18No.1 pp.15-20, 2008

[3] Sung, Y., Lee, J. and Yang, J.,Crystallization and SinteringCharacteristics of Chemically PrecipitatedHydroxyapatite Nanopowder. J. Cryst.Growth. 262 : 467-472, 2004.

[4] Lazic, S., Zec, S., Miljevic, N., Milonjic, S.,“The Effect of Temperature on theProperties of Hydroxyapatite PrecipitatedFrom Calcium Hydroxide and PhosphoricAcid”, Thermochim.Acta, Volume 374,Issue 1, pp. 13-22; 2001;

[5] Poinescu, A.A., Ion, R.M., Trandafir, I.,Bacalum, E., Radovici, C., Obtaining andcharacterization of a calciumhydroxyapatite, The XV-th InternationalScientific Conference “Tehnomus”, May 8-9, (2009), Suceava Romania;

[6] Santos, M.H., Oliveira, M., Palhares deFreitas Souza, P., Sander M.H., Gerais,W.L., Synthesis Control andCharacterization of HydroxyapatitePrepared by Wet Precipitation Process,Materials Research, Vol. 7, No.4, 625-630,(2004).


289

OVERVIEW OF LATEST MINERAL CARBONATIONTECHNIQUES FOR CARBON DIOXIDE SEQUESTRATION

Bodor Marius1, Vlad Maria1, Balt tefan1

1 Universitatea “Dun rea de Jos” din Gala i,[email protected], [email protected], [email protected]

Abstract: This relatively new domain, of carbon dioxide sequestration for mitigation purposes, is oneof the carbon dioxide capture and storage (CCS) alternatives. CO2 sequestration could be used in thefuture where geological storage, for example, is not an option due to the lack of suitable locations orcould represent a hazard to the surrounding areas. Also could be used in those locations where the neededmaterial (raw minerals or industrial wastes) are abundant.

The interest in this technique of CO2 storage is growing fast and results of researches on this topic arecoming from a larger number of countries every year. What is more interesting is that the research of theindustrial wastes carbonation is developing, giving in the future new purposes for these wastes.

Keywords: mineral carbonation, carbon dioxide, calcium carbonate, storage

1. Introduction

The greenhouse effect is known to be createdby different gases in the atmosphere. One of thesegases is CO2 and because its concentration isincreasing dangerously for the past one hundredyears is now considered to be the main cause of thetemperature rise on Earth.

In order to avoid the potentially devastatingconsequences of global warming and climatechange, the CO2 emissions into the atmospherecaused by human activities should be reducedconsiderably [1]. For this goal to be reached somecarbon dioxide capture and storage (CCS)technologies were proposed but for the momentnone of them were efficient enough in reducing theCO2 emissions. Despite the negative results,researches are still ongoing and in some cases theresults are promising as for example the storage inunderground cavities. The same is hoped to behappening with the CO2 storage in minerals andwastes and hopefully all the CCS technologies willcontribute together to a more efficient CO2reduction.

2. Carbon dioxide sequestration techniques2.1. Geological storage

One of the technologies, that have already beenemployed on a significant scale, but not largeenough to have a global CO2 emissions mitigation

impact, is storage of CO2 in underground cavities.This includes the so-called Enhanced Oil Recovery(EOR) and also Enhanced Gas Recovery (EGR),which are concepts aimed at improving the oil/gasrecovery potential of an oil/gas field by flooding itwith CO2 [2].

Perhaps the greatest problem related tounderground storage is the permanency of thesolution, as there will always be a risk of leakage.Therefore, this solution would require continuousmonitoring of storage sites for thousands of years.

2.2. Ocean storage

Another widely studied option for CO2sequestration involves injecting CO2 into the oceanat great depths, where the gaseous CO2 reacts toform carbonic acid (H2CO3). The carbonic acidthen dissociates into a (bi)carbonate ion andhydrogen ion in accordance with the equationbelow: [2]

CO2(g) + H2O(l) H2CO3(aq) HCO-3 + H+

CO2-3 + 2H+ (1)

Although ocean storage could provide a fastand relatively easy alternative for CO2 emissionsreduction it has lost its appeal in recent years,largely due to the uncertainty when considering the

mailto:bodormarius:@yahoo.com

mailto:maria.vlad:@ugal.ro

mailto:stefanbalta:@yahoo.com


290

environmental consequences (e.g. decreasing pHof ocean water) and the lack of permanency [1].

2.3. Storage below sea bedAn alternative to both geological storage and

ocean storage described above is CO2 storagebelow the ocean floor at depths of at least 3000 mof ocean and several hundred meters of marinesediment. In contrast to the previously mentionedoptions this option does not suffer from lack ofpermanency (ocean storage) or the demand formonitoring the storage site (geological storage).The idea is based on the fact that CO2 becomesdenser than water at sufficient depths ( 3000 m),but that it still needs to be trapped in order toprevent it from being released by ocean currents ore.g. earthquakes. Therefore, it should be storedbelow the seabed [3].

This alternative is still new and further researchis ongoing in order to verify the theories.

2.4. Mineral carbonation

The reaction between a metal oxide bearingmaterial and CO2 is called carbonation and can beexpressed by the following reaction:

MO + CO2 MCO3 + heat (2)where in practice M describes a (metallic) elementsuch as calcium, magnesium or iron. The reactionin Equation (2) is exothermic and the heat releasedis dependent on the metallic element bearing

mineral at hand (for the magnesium- or calcium-based silicate minerals- olivine: 89 kJ/mol CO2,serpentine: 64 kJ/mol CO2 and wollastonite: 90kJ/mol CO2 at 298 K). [2]

One major benefit of CO2 sequestration bymineral carbonation consist of the environmentallybenign and virtually permanent trapping of CO2 inthe form of carbonated minerals by using abundantmineral resources such as Mg-silicates [2]. Unlikeother CO2 sequestration routes it provides aleakage-free long-term sequestration option,without a need for post-storage surveillance andmonitoring once the CO2 has been fixed.

In addition to the benefits of mineralcarbonation, this option is the only CO2sequestration option available where largeunderground reservoirs do not exist and oceanstorage of CO2 is not feasible.

Another benefit of mineral carbonation is that,at least theoretically, the carbonation process couldproceed without energy input, but this has not yetbeen accomplished.

Attempts to speed up the carbonation reactioninclude using both dry and wet methods, additives,heating and pressurizing the carbonation reactor,dividing the process into multiple steps,pretreatment of the mineral source and more(fig.1).

There are several different elements that can becarbonated, but alkaline earth metals, calcium andmagnesium, have proven to be the most suitabledue to their abundance and insolubility in nature.

Figure 1: Main carbonation processes and variants.


291

In addition to the abundant magnesium andcalcium containing minerals, there are alsoindustrial solid residues that contain large amountsof Mg, Ca and even Fe.

Currently the most investigated mineralresources are olivine, serpentine and wollastonite.From the solid residues, steel slag has recentlyreceived a lot of attention but other industrialresidues also have been studied: asbestos-miningtailings, electric arc furnace (EAF) dust, cement-kiln dust, waste concrete, coal fly ash etc.

Figure 1 displays the various carbonation routesthat are currently being investigated.

3. Directions of mineral carbonation

Trapping carbon dioxide in carbonates can beachieved through various process routes asdescribed further, ranging from the most basicaccelerated weathering of limestone to advancedmulti-step processes.

3.1. Direct carbonation

Direct carbonation is the simplest approach tomineral carbonation and the principal approach isthat a suitable feedstock, e.g. serpentine or aCa/Mg rich solid residue is carbonated in a singleprocess step. For an aqueous process this meansthat both the extraction of metals from thefeedstock and the subsequent reaction with thedissolved carbon dioxide to form carbonates takesplace in the same reactor.

3.1.1. Direct gas-solid carbonation

Gas-solid carbonation is a simple approachtowards mineral carbonation. Here particulatemetal oxides are brought into contact with gaseousCO2 at a particular temperature and pressure. Thedry process has the potential of producing hightemperature steam or electricity while convertingCO2 into carbonates.

Unfortunately, the reaction rates of such aprocess have been too slow and the process suffersfrom thermodynamic limitations and furtherstudies around this alternative have mostly beenabandoned.

One of the important benefits of using industrialsolid residues as feedstock for carbonationcompared to the carbonation of mineral ores is thepossibility of utilizing a waste stream. The

possibility of simultaneously binding CO2 andlowering the hazardous nature of e.g. municipalsolid waste incinerator (MSWI) ash makes thiscarbonation route interesting [4]. However, thepotential CO2 storage capacity for this option islimited, simply because the amounts of materialthat may be carbonated are too small [4].

Nevertheless, where these wastes are presentthe carbonation should be taken into considerationbecause as shown by R. Santos et al. [5] duringthe carbonation the strength of the stainless steelslag is increased. This will give an added value tothe carbonated stainless steel slag wherefore thematerial can be used in secondary applications.The sale price of this product might increase.

3.1.2. Direct aqueous carbonation

The direct aqueous mineral carbonation-routereferring to carbonation preformed in a single stepin an aqueous solution, appears to be the mostpromising CO2 mineralization alternative to date[6]. High carbonation degrees and acceptable rateshave been achieved but the process is (still) tooexpensive to be applied on a larger scale [7].Ranging from 40–80 €/t CO2 mineralized (includesenergy use) compared to 0.4–6 €/t CO2 [2] storedfor geological storage.

Direct aqueous mineral carbonation can befurther divided into two subcategories, dependingon the type of solution used. Studies focusing oncarbonation in pure aqueous solutions have quicklymade way for additive-enhanced carbonationexperiments and a common solution type usedtoday, originally presented by O’Connor et al.(2000) [8], consists of 0.64 M NaHCO3 + 1.00 MNaCl. However, it has been reported [9] that thereare still improvements to be made regarding theabove mentioned solution.

When it is necessary to use additives incarbonation processes it is extremely important torecycle these, due to the large scale of anyindustrial application [7].

Work on finding optimal aqueous carbonationconditions is ongoing and even though it has beenstudied extensively, some questions still remainunanswered. For example increasing theliquid/solid ratio (L/S) has been reported to haveboth a positive and a negative effect on CO2conversion.

The advances made to aqueous solutionchemistry by McKelvy et al. [9] were significant,


292

but unless the (expensive) additives used cannot berecycled the process route becomes unattractive.Nevertheless, the studies conducted on directaqueous carbonation have improved the overallknowledge of aqueous carbonation reactionsconsiderably.

3.2. Indirect carbonation

When the mineral carbonation process isdivided into several steps it is classified as indirectcarbonation. In other words, indirect carbonationmeans that the reactive component (usually Mg orCa) is first extracted from the feedstock (as oxideor hydroxide) in one step and then, in another step,it is reacted with CO2 to form the desiredcarbonates.

3.2.1. Multistage gas-solid carbonation route

Direct gas-solid carbonation of silicate mineralshas been shown to be too slow for any large scaleimplementations, but a staged gas-solidcarbonation process could overcome the slowreaction kinetics. The process involves extractionof magnesium (oxide or hydroxide) in anatmospheric pressure step followed by acarbonation step at elevated temperature (>500 °C)and pressure (>20 bar) [10].

It has been found that the carbonation of MgOis significantly slower than the carbonation ofMg(OH)2 [11]. Using this observation Zevenhovenet al. [10] suggested (noting that Mg(OH)2production from serpentine in one step cannot bedone because of thermodynamic limitations) thatthe direct gas-solid carbonation process should bedivided into three-steps: 1) MgO production(Equation 3) in an atmospheric reactor followed by2) MgO hydration (Equation 4) and 3) carbonation(Equation 5) at elevated pressures:

Mg3Si2O5(OH)4(s) 3MgO(s) + 2SiO2(s) + 2H2O (3)

MgO(s) + H2O Mg(OH)2(s) (4)

Mg(OH)2(s) + CO2MgCO3(s) + H2O (5)

In addition to the faster carbonation kinetics inthe three-step gas-solid carbonation routedescribed above, the process is also preferablefrom an energy efficiency point of view compared

to the two step carbonation of MgO [10]. However,the three-step process is still too slow for largescale implementation, as preliminary testsperformed in up to 45 bar pressures have shown.

Progress has been made in enhancing the directgas-solid reaction rates, primarily by increasingpressure. Experiments showing that all three-stepsin Equations 3–5 are fast enough for industrialimplementation are still required.

Dividing the gas-solid carbonation route intoseveral steps could be beneficial, but there is notenough evidence yet for industrial viability.

3.2.2. Acetic acid route

In order to speed up the aqueous carbonationprocess, the use of acetic acid for the extraction ofcalcium from a calcium-rich feedstock has beensuggested by Kakizawa et al. [12]. In principal itconsists of two-steps as given in Equations 6 and7:

CaSiO3 + 2CH3COOH Ca2+ + 2CH3COO + H2O + SiO2 (6)

Ca2+ + 2CH3COO + CO2 + H2O CaCO3 + 2CH3COOH (7)

Equation 6 describes the extraction step andEquation 7 the precipitation step. In principal theacetic acid used in the extraction step could berecovered in the following precipitation step.

Inspired by the concept of binding CO2 incalcium extracted from a calcium silicate such aswollastonite using acetic acid [12], Teir et al. [13]investigated the possibility of producing a highvalue PCC material from calcium silicates andlater from other calcium-containing materials [14].

Steelmaking slags then became the centre ofattention as they can contain significant amounts ofboth CaO and MgO. Eloneva S. [15] reported that80–90 % pure calcite was produced from blastfurnace slag using acetic acid. However,significant amounts of sodium hydroxide wererequired for promoting the precipitation ofcarbonates from the acidic solution.

Research in Finland has focused on steel slagcarbonation and especially the possibility ofproducing valuable precipitated calcium carbonate(PCC) [14, 15]. The annual CO2 binding potentialof steelmaking slags is in the order of 70 – 180 MtCO2. Globally the CO2 sequestration potential forthis option is small, but for individual steel plants,


293

however, the method could provide significanteconomical benefits.

3.2.3. Two step aqueous carbonation

Two-step aqueous carbonation has beeninvestigated because the overall carbonationreaction can easily be divided into two-steps:extraction and precipitation that may beinvestigated and optimized separately.

By upgrading a waste product into a product ofhigh commercial value, expensive CO2sequestration processes could becomeeconomically feasible. One such approach hasbeen investigated by Katsuyama et al. [16] whostudied the use of waste cement for thedevelopment of high-purity CaCO3 by CO2carbonization.

Katsuyama et al. [16] studied the feasibility ofproducing CaCO3 from waste cement by firstextracting calcium from pulverized waste cementin a water slurry at high CO2 pressure (severalMPa), followed by the precipitation of CaCO3from the extracted solution at lower CO2 pressures,producing high purity CaCO3 (up to 98%) fromwaste cement at relatively high reaction rates. Theestimation of the cost of producing high-purityCaCO3 could be as low as 105 €/m3 whencompared to the commercial price of 154–269€/m3. In addition, if the produced CaCO3 could bepurified to meet the requirements of ultra-highpurity CaCO3 (>99% CaCO3) the potential profitscould increase substantially. The current cost ofultrahigh purity CaCO3 is around 7,700 €/m3, whileKatsuyama et al. [16] estimated a production costof only 250 €/m3.

Gorset et al. [17], describing a way ofproducing pure MgCO3 from olivine, claims thatthe process consisting of one dissolution step andtwo precipitation steps is rapid enough for largescale implementation. The process does not requirethe use of strong mineral or organic acids eventhough the dissolution step requires an acidicenvironment. The required acidity (pH 3–5) is tobe achieved using pressurized CO2 (50–150 bar)and a temperature around 100–170 °C, while thefollowing step consisting of MgCO3 precipitation,takes place in another reactor with preferably alower CO2 pressure (50 – 80 bar) and a highertemperature (140 – 250 °C) favoring theprecipitation of carbonates. Experimental resultsshowed a high degree of purity, between 99.28 and99.44% MgCO3, of the precipitated carbonate.

The pH-swing process developed in Japan isanother two-step aqueous carbonation processwhere at first the pH of the solution is loweredthereby enhancing the extraction of divalent metalions. In the second step the pH is raised to enhancethe precipitation of carbonates.

The principal reactions taking place inside theextractor (Equation 8) and the precipitator(Equation 9) are:

4NH4Cl + 2CaO·SiO2 2CaCl2 + 4NH3 + 2H2O (8)

4NH3 + 2CO2 + 2H2O + 2CaCl2 2CaCO3 + 4NH4Cl (9)

Equation 9, taking place inside the precipitator,consists of both CO2 absorption and CaCO3precipitation. In their study, Kodama et al. [18]investigated a CO2 sequestration process thatutilizes pH swing using NH4Cl. The energy inputrequirement for the investigated process using steelmaking slag as the mineral source was estimated ataround 300 kWh/t CO2, but the loss of a chemicaladditive (NH3) was considerable.

Different approaches of a two-step aqueouscarbonation process have been presented. Alloptions are good in theory, but it remains uncertainwhether or not these processes could lead the wayto any significant scale long-term storage of CO2 inthe future. More experiments for large scaleviability are required.

3.3. Other processes of carbon dioxidesequestration

Beside the process routes mentioned above,there are other processes and applications thatresemble mineral carbonation and are described inthe following section.

3.3.1. The precipitated calcium carbonateproduction

The production of valuable products (e.g. PCC)by utilizing CO2 was studied extensively. Anexample of this concept is Two-step aqueouscarbonation of solid residues. Various methods toobtain a product of desired properties have beenused and one of the simplest methods is that ofdirect aqueous carbonation without the use ofadditives.


294

3.3.2. Brines used in carbonation

The saline-based solution formed as a wasteproduct during the oil or natural gas extraction iscalled brine and it is found in large quantities instorage tanks. Due to its large amounts availableand a high concentration of elements fit to formcarbonates (mostly calcium and magnesium) thisbrine represents an option for carbon dioxidestorage by a carbonation process. Although thisbrine is capable of forming carbonates, its use in alarge scale process is for the moment notappropriate due to the slow kinetics. If the pH ofthe brine is raised the carbonation process is faster,but unknown factors about the parameters(temperature, pressure, brine composition and pH)need to be further studied.

3.3.3. Accelerated weathering of limestone

Another option is carbon dioxide capture andstorage by accelerated weathering of limestone(AWL). This option imitates the natural carbonateweathering according to the following reaction:

CO2(g) + H2O(l) +CaCO3(s) Ca2+(aq) + 2HCO3 (aq) (10)

The product of an AWL plant would be acalcium bicarbonate solution that could readily bereleased and diluted into the ocean with a minimalenvironmental impact [19]. However, there are stillmany issues to deal with, such as the energydemand of transporting large amounts of calciumcontaining (waste or mineral) material to the AWLplant that preferably should be located near a CO2point source as well as a possible disposal site (e.g.the ocean). In an ideal case (with access to freelimestone, e.g. waste fines, and a “free” watersource, e.g. power plant cooling water) the CO2mitigation cost by means of AWL could be as lowas 2.3–3.1 €/ton CO2. Rau et al. [19] suggests thatsome 10-20% of the United States point-sourceCO2-emissions could be mitigated this way.

Despite the potential positive effect ofbicarbonate disposal Rau et al [19] concludes thatfurther research is needed to fully understand theimpacts of AWL effluent disposal in the ocean.

Some of the most important applicationsand techniques for mineral storage of CO2 aredescribed above, yet almost all of theseoptions need further research before starting a

large scale process. Despite the inconveniencegiven by results lacking, all of them shouldrepresent an option when CO2 emissionreduction is taken into account.

4. Conclusions

It can be concluded that the carbonationtechniques using additives for reactivity risinggave better results than those not using additivesand for that, studies were performed with thepurpose of a better understanding on the reactionscomplexity regarding the carbonation. Despitesome good results the performed studies have not,for the moment, came with significant discoveriesregarding the reactivity in mineral carbonationprocess.

For the techniques using additives incarbonation the biggest problem represents therecycling of these additives and for the moment nosuch breakthrough was reported and if this willnever be accomplished, maybe method’s that don’tneed them will be implemented.

Besides the cost of the process that is for themoment larger than of those other carbon dioxidecapture and storage methods, the direct aqueouscarbonation seems to be the most promising optionof mineral carbonation. Other studies, however,show that the dissolution and the precipitationsteps should be separated, even if the costs will behigher. This separation would take out the need tobalance these reactions that are opposite. So,taking that into consideration, the indirect aqueouscarbonation is the most attractive option.

Despite the good results on the high purityprecipitated calcium carbonate process, for themoment, no large scale production was realized. Ifthis will be accomplished it is expected for theprice of carbonation process to be acceptable evenif is at a higher rate, due to the great value of thehigh purity PCC.

Before one or more mineral carbonationtechniques could be implemented on an industrialscale, much more research should be done andimportant improvements of existing options areneeded. Another way for the large scalecarbonation to become reality would berepresented by completely new techniques.

5. References


295

[1] Huesemann, M.H., Can advances in scienceand technology prevent global warming?Mitigation and Adaption Strategies for GlobalChange. (11), pp. 539-577, 2006.

[2] [Metz, B., O. Davidson, H. C. de Coninck, M.Loos, and L. A. Meyer (eds.)], IPCC specialreport on carbon dioxide capture and storage,Prepared by Working Group III of theIntergovernmental Panel on Climate Change,Cambridge University Press, Cambridge,United Kingdom and New York, NY, USA,IPCC, 2005.

[3] House, K.Z., Schrag, D.P., Harvey, C.F.,Lackner, K.S., Permanent carbon dioxidestorage in deep-sea sediments, 8th InternationalConference on Greenhouse Gas ControlTechnologies, 19-22 June, 2007, Trondheim,Norway.

[4] Rendek, E., Ducom, G., Germain, P., Carbondioxide sequestration in municipal solid wasteincinerator (MSWI) bottom ash, Journal ofHazardous Materials. (128), pp. 73-79, 2006.

[5] R. Santos, D. François, E. Vandevelde, G.Mertens, J. Elsen, T. Van Gerven,Intensification routes for mineral carbonation,2010.

[6] Lackner, K.S., CLIMATE CHANGE: A guide toCO2 sequestration, Science. (300), pp. 1677-1678, 2003.

[7] Huijgen, W.J.J., Comans, R.N.J. Carbondioxide sequstration by mineral carbonation:Literature review update 2003-2004, EnergyResearch Centre of The Netherlands, Petten,The Netherlands, 2005.

[8] O'Connor, W.K., Dahlin, D.C., Nilsen, R.P.,Turner, P.C., Carbon dioxide sequestration bydirect mineral carbonation with carbonic acid,Proceedings of the 25th International TechnicalConf. On Coal Utilization & Fuel Systems,Coal Technology Assoc. March 6-9, ClearWater, FL, Albany Research Center, Albany,Oregon, 2000.

[9] McKelvy, M.J., Chizmeshya, A.V.G., Squires,K.D., Carpenter, R.W., Béarat, H., A novelapproach to mineral carbonation: Enhancingcarbonation while avoiding mineralpretreatment process cost, Arizona StateUniversity, Center for Solid State Science,Science and Engineering of Materials GraduateProgram, and Department of Mechanical andAerospace Engineering, June 21-22, 2005.

[10] Zevenhoven, R., Teir, S., Eloneva, S., Heatoptimisation of a staged gas-solid mineral

carbonation process for long-term CO2storage, Proceedings of ECOS 2006, 12-14July 2006, Crete, Greece, pp. 1661-1669.

[11] Butt, D.P., Lackner, K.S., Wendt, C.H.,Conzone, S.D., Kung, H., Lu, Y.-C., Bremser,J.K., Kinetics of thermal dehydroxylation andcarbonation of magnesium hydroxide, Journalof the American Ceramic Society. (7), pp.1892-1898, 1996.

[12] Kakizawa, M., Yamasaki, A., Yanagisawa,Y., A new CO2 disposal process via artificialweathering of calcium silicate accelerated byacetic acid, Energy. (26), pp. 341- 354, 2001.

[13] Teir, S., Revitzer, H., Eloneva, S., Fogelholm,C.-J., Zevenhoven, R., Dissolution of naturalserpentinite in mineral and organic acids,International Journal of Mineral Processing.(83), pp. 36-46, 2007.

[14] Teir, S., Eloneva, S., Fogelholm, C.-J.,Zevenhoven, R., Dissolution of steelmakingslags in acetic acid for precipitated calciumcarbonate production, Energy. (32), pp. 528-539, 2007.

[15] Eloneva, S.,Reduction of CO2 emissions bymineral carbonation: steelmaking slags as rawmaterial with a pure calcium carbonate endproduct, PhD thesis, 2010.

[16] Katsuyama, Y., Yamasaki, A., Iizuka, A.,Fujii, M., Kumagai, K., Yanagisawa, Y.,Development of a process for producing high-purity calcium carbonate (CaCO3) from wastecement using pressurized CO2, EnvironmentalProgress. (24), pp. 162-170, 2005.

[17] Gorset, O., Johansen, H., Kihle, J., Munz,I.A., Raaheim, A., Method for industrialmanufacture of pure MgCO3 from an olivinecontaining species of rock, Patent,WO/2007/069902, 21.6.2007.

[18] Kodama, S., Nishimoto, T., Yogo, K.,Yamada, K., Design and evaluation of a newCO2 fixation process using alkaline-earth metalwastes, 8th International Conference onGreenhouse Gas Control Technologies, 19-22June, Trondheim, Norway, 2006.

[19] Rau, G.H., Knauss, K.G., Langer, W.H.,Caldeira, K., Reducing energy-related CO2emissions using accelerated weathering oflimestone, Energy. (32), pp. 1471-1477, 2007.

Acknowledgements

Bodor Marius would like to acknowledgethe support provided by the European Union,


296

Romanian Government and “Dun rea de Jos”University of Gala i, through the projectPOSDRU – 107/1.5/S/76822.

tefan Balt would like to acknowledge thesupport provided by the European Union,Romanian Government and “Dun rea de Jos”University of Gala i, through the projectPOSDRU – 6/1.5/S/15.


297

DEVELOPMENT OF MEANS AND A PLATFORM FOR RESEARCH

Liliana Georgeta POPESCU1, Mihai Victor ZERBES1, Radu Vasile PASCU1,Lucian LOBON 1, Livia Dana BEJU1

1"Lucian Blaga" University of Sibiu, email: [email protected]

Abstract: The paper describes the operation of a web platform made for an e-research center. Theplatform includes a number of procedures performed and the specific tools of research process designedto facilitate online research management, to facilitate dialogue between clients and professionals, enhanceteamwork, encourage exchange of ideas and the construction and design of prototypes of new products.

Keywords: e-research centre, platform

1. Introduction

PCs can create virtual spaces that are used in aninnovative manner, spaces for training,information, communication, collaboration,exploration, documentation, multimedia, wordprocessing, illustration, simulation and virtualreality spaces.

In this work is presented the development andtesting of a general software infrastructure for avirtual research center. The infrastructure mustsupport all activities concerned with technicalvirtual research, especially in problem solving,correcting and assessing solutions. To give amodern interpretation of distance research, it wasprimarily the relationship between the structure ofa center for research and dialogue can bedeveloped through it.

2. Development and characteristics of e-research platform

2.1. Development of e-research platformThe main phases of the development platform are:- Developing Virtual Research Centre - creation,

consolidation and expansion (infrastructure,operational methodology);

- The initiation, conducting joint research anddevelopment;

- The creation, development and maintenance offacilities and research platform;

- The study of models of product developmentand manufacturing process;

- The knowledge management;

- The research activities - joint design;- Dissemination of results and knowledge

transfer to industry;The input data refers to the participating

universities and research centers, human resourcesand materials, the existing knowledge base. Theoutput which are resulting from research anddesign activities, are the products used in industry,the knowledge that will increase database andfeedback which improve continuously the businesscenter.

Efficiency and viability of such a platform isbased on systematic and systemic structure thatallows continuous improvement, which could beextended.

The platform also allows the dissemination ofknowledge through databases.

2.2 Characteristics of e-Research Centreplatform

The platform includes a number of proceduresperformed and the specific tools of researchprocess designed to facilitate online researchmanagement, to facilitate dialogue between clientsand professionals) enhance teamwork, encourageexchange of ideas and design prototypes newproducts.

The website was done using Microsoft VisualStudio 2010, free version of the code editor,editing applications are of the HTML and ASPprogramming language, the applications formanaging databases are build in Access, and


298

statistical evaluation is accomplished throughExcel applications.

3. Platform functionality

Depending on the issues raised and shapedtheme, specific steps will be taken. Since designresearch is carried out by research teams indifferent locations, it is necessary to communicatea precise methodology supported by specialsoftware and a phased deployment of the researchmethodology to a fully efficient.

3.1 AccessThe access can be made from the Internet or

intranet from any computer connected to theInternet. The access to the general information andthose related to projects within the center is free.

3.2 Menu SectionsThe main window of access to the site, which is

open on the activities menu and the center whereyou can access the pages related to: ProductDesign, Design Services, Product Design, ProcessDesign, Maintenance is shown in Figure 1. In thesepages have general information about products,processes and services designed engineered byResearch Centre.

Figure 1: Access page to the Center ActivitiesThe access to the members' personal

information: professionals, associates or customersare secure. Each specialist, customer or employeeshall enter into the database by completing anapplication that is generated by accessing the"Application". Each user receives from theplatform administrator an username and initialpassword, which can be customized after the firstlogin. On acceptance of entry in the database, it

generates a message of acceptance, the newmember receives the right connection space (log-in).

3.3 Data BasesDatabases are build in Microsoft Access and

contains tables that store information aboutcustomers, employees, professionals, projects andnewsletters.

Figure 3: Tables from Access Data Base

3.4 Secure DataThe platform management of the database, ( the

add / change user names and / or passwords, theaccess to pages listing and changing the lists ofcustomers, experts and partners, the adding pagesand newsletters and projects) is secure and can beaccessed and performed only by the administrator.

3.4.1 Specialists DataInformation about the identification of members

can be viewed only by the site administrator.

Figure 4: Specialist Detail PageThis page allows you to add a new member,

following receiving the of application by theplatform manager. The window that can be viewedbehind the application, by the administrator of theplatform is shown in Figure 4. On the basis of


299

completed applications received from newmembers, new information is added by pressingthe "Add New Member" button. The input data canbe modified or deleted by pressing the "SaveChanges" and "Delete Member" buttons.

3.4.2. The Specialists SectionThis section provides information on

research areas and centers that can provideadvice in the field. Platform database containsdata related to research centers or universitiesin Romania. Customer registration databaseplatform is build based on the activity field, sothey can be easily accessed by selecting theindustry. Specialist is a generic term used forindividuals (experts, researchers, teachers, etc..)and for institutions, research centers, etc.. as theyhave skills and can work in an e-research project.

3.4.3 Customers SectionIn this section you can access to view / changepersonal information stored in the database.Customer registration in the platform's database iscarried out by their field activity, so they can beeasily accessed by selecting the industry ("SelectCategory" - Figure 5.

Figure 5: Window add new customer database

3.5 Communication Module

Communication includes discussion forumtools needed, on-line sessions, internal messaging.

The forum tools were developed to support thelisting of asynchronous discussions, the projectmanager coordinating discussions based on theproject development. There are also possibilitiesfor chats video – on-line sessions, internalmessaging (custom work groups). Addition pageon the server for the name and URL address of the.pdf files, which are newsletters (informing letters)for storage in the database.

4 Using e-research PlatformA general process for conducting network

research involves the following steps:- Establishing customer contact;The responsible for this section is the platformadministrator. When the client is seeking or

making a product order, process, or improve itsnew project, the first stage of the dialogue can beachieved by sending a text message, in an e-mailform. On the home page is the "Home" menu,which contains all the contact details of the center.- Registration by the administrator of the

proposal;The page allows the administrator to add a newproject based on what the client requested. Theproject may be based on designing a new product,or it can be based on improving a product or anactivity which is part of a bigger process (forexample, determinating the internal tensions for aparticular item- Designation of project manager;Depending on the category in which the project(field) fits, the platform administrator, sends theproposal of the customer to the specialists that areadded in that field related database. Depending onthe capacity and availability of specialist, theproject manager is appointed. Note that theadministrator can be also the project manager, atthe same time.- Composition of the Working Group;The forming of the working group is a task of theproject manager and of the platform manager,which will invite potential specialists in the field tobe part of the team. The specialists database of theplatform enables fast selection and information ofthe specialists that may form the group.- Registering platform project team and client

project;Table 'Admin users list is completed by theadministrator, with the team selected on a specificproject - a project that will receive a code. On theplatform can be run in parallel several projects, andeach of which can run separately. Also in thissection can be completed the database withcustomer data and potential new specialists addedto the project.- Establishing the schedule and procedure for

teamwork;Useful methodology to be used is that of

Project Management. The main stages that gothrough when applying this methodology are:- The defining of the project theme in detail:

specify the problem, purpose or majorobjectives, deliverables elements withdeadlines (times required), technicalspecifications, costs, management support, etc.

- Planning the projecta. Defining the structure of work activities (activityname, duration);


300

b. Implementation of network-type project planwill specify the order and durations of eachactivity;c. Allocation of resources to each activity (human,material, equipment);d. Optimization of planning in terms of duration;e. Estimated cost calculation.The project may contain in addition to the actualresearch or the optimization design modules ofproducts or technologies that achieve the designprototype and prototype implementation.- Establishing, analysis and completion basis;

In order to establish research directions,grading criteria, the project manager analyzes theclient wishes, according to the questionnaire thathe completed. To create questionnaires or tests, theplatform program contains Excel files. On theplatform are already built predefined questionnairetypes to be completed by the clients andspecialists. Depending on the theme of the project,these questionnaires can be changed or new onescan be designed.- Establishing product requirements

(classification criteria);The platform administrator is responsible for

the registration of a new project or proposal. Thepage allows the administrator to add a new projectbased on what the client requested. The projectmay be to design a new product, improving aproduct or an activity from a process.

After outlining the problem that is intended tobe solved and the models to be developed, it isnecessary a complex task of planning and trackinga project that aims to solve the problem. Usefulmethodology to be used is that of projectmanagement. The main stages that go throughwhen applying this methodology are:- Defining the theme of the project in detail:specify the problem, purpose or major objectives,deliverables with deadlines sites (times required),technical specifications, costs, managementsupport;- Planning the project:a) Defining the structure of work activities

(activity name, duration);b) Implementation plan network type of project

that will specify the order and durations of eachactivity;

c) Allocation of resources to each activity (human,material, equipment);

d) Optimization of planning in terms of duration;e) The calculation of cost estimates.

The project may contain in addition to theactual research or the optimization of module

design and products and technologies that achievethe design prototype. Project activities will beadded to the calendar center (figure 5), so, theteam/research team to be informed of the date andtime of deployment and type of action.

Figure 5: Add new action projects in calendarThere are cases when certain actions are

repeated, they recorded as repetitive actions (forexample - "Repeat Every Monday from 01 to 25 ofOctober" - "the action is repeated every 01 monthsfrom October 25), (figure 6).

Figure 6: Model of schedule worksheetThe project manager sets the team's working

procedure, inform and instruct the team in thisregard. Also in this stage, it lay down theresponsibilities of members. The administratormanages the material received and sends theinformation and their agenda to the professionals.The module allows managing the application-levelplatform structures in the database. The platform iscomposed of four administrative parts: systemadministrator or project manager, specialists,employees and customers. Between thesestructures there are specific relations of theorganization of the program.

The project manager sets the workingprocedure in the system depending on the type ofproject in order to choose or to repeat certain steps,holding several meetings (synchronous orasynchronous), requests for opinions. Thefollowing presentation is a general procedure for


301

conducting an online research. The platform isgenerally available for the following steps:- The proposal to address theme ideas;At the request of project manager, each teammember will complete in the assigned field (figure7) three proposals for resolving the theme.

Figure 7: Work sheet (synchronous-asynchronous)project team

- Improve and complete the initial ideas;Since the working sessions can be run both

synchronously and asynchronously, the projectmanager sets a deadline date and time whenproposals are loaded by each participant. If thissession takes place synchronously, this stageoverlaps with the previous stage, so theparticipants in this session can make changesimmediately after the initial data were added. Inthe event that not all team members are loggedsynchronously, it is recommended that the projectmanager to provide a period of several days toanalyze and make possible amendments to theproposals.- Analysis of proposals and determining

potential improvements;Ideas are centralized by the project manager andthe site manager is posting at the "ProjectDescription" section that is allocated to the projectin progress, the solutions that match the theme.Solutions that do not correspond to the theme orare not feasible are eliminated with the agreementof members.

- Comparison of proposals for selecting theoptimal solution;

The project team is the one that notes from 5 (themaximal value) to 0 - zero (minimum) proposedsolutions. The proposal ordering application afterscoring may be accessed by the platform manager,from the platform. Scores are awarded dependingon the requirements to be met by product / service /project. Data is loaded by each party specialist inits field, and the centralization is made by theproject manager.- Determination of the optimal solution;The Solutions hierarchy is done by collecting thescores given by team members. The field that isassigned to the project description is named"Project Description" on the "Project Information"page (fig.6.18) and it allows insertion of multipledata, as the window is provided with the cursor.All final data will be posted from shares concludedin this field manager. The fields assigned to thespecialists will be emptied (blank) for the nextsteps.- Review the solution by each team member;The review of the solution through the model takesinto account the highlighting of possible errors oromissions have not been studied or have not beenconsidered in previous steps and that can becorrected at this stage. For this action the projectmanager may convene an online meeting to get afairly and quickly feedback before a decision isbeing taken. In this case on the center calendar willappear a note named as a "Project ID xzy -Meeting September 15 at 10:00 am"- Adopt constructive solution;At this stage it is decided if the resulting solutioncan be adopted and implemented. Again, the teammembers "vote" and each will sustain its point ofview. The meeting is Synchronous , under a jointmeeting.- Modelling constructive solution/concept

design;Modelling can be done by one team memberdesignated for this purpose. It is indicated to beused a software that allows viewing and makingchanges to the design. The image is transmitted ona modelled solution. jpg,. pdf or. tif to the teammembers, for analysis and evaluation.- Improvements/changes in the constructive

solution of the model;Team members are given a deadline to review andpropose improvements to the model. They maypost their opinions on the "Project Information", inthe field assigned to each of them and then, duringa synchronously meting, the person in charge with


302

the modelling will present the improved variant.The final model image is sent by mail to the wholeteam.- Acceptance of Product;Team members will submit their final agreement interms of modelling. Agreement must be given andthe client, before the team ceased the activity. Ifthe customer finds design deficiencies of theproduct, can make some suggestions to improvethe maintainability design. The client can activelyparticipate - on-line at the stage of conception.Consent given by the team is practically ending theactivity of the design team.- Transmission of client project;Project Manager presents to the customer, thesolution / project that is the subject of the researchcontract.- The audit project;The project audit is highlighting the strong parts ofthe project and of networking team work and alsothe weak parts, to be clearer.- Update data base;

Platform manager responds of this activity. Theaccepted solution will be registered in the data baseand presented under "Research Center" which isunder "Product Conception" section, or "ProductDesign" section, depending on the type of research.The completed Project will be presented to thecenter members and prospective clients (if any)through the newsletter sent and posted on theplatform. The virtual product design involves notonly the virtual, but also the maintenance ofprocess modelling, human interaction - modellingand application object model. Virtual productmaintenance, analysis and testing can take place onreal products. To maintain the virtualmaintainability analysis system it has to bedeveloped key technologies so that the platformcan perform maintenance of virtual prototyping,modelling and simulation of maintenance actionsbased on failure analysis.

E-Research Centre should be seen as a livingthing that needs to be improved and adapted withthe evolution of business processes andtechnologies.

5. Conclusions

To achieve fair and effective planning of aproject within a virtual research center is necessaryto have databases that contain:- Database that aims general Engineering

knowledge, basis and specialty and liaise withexperts in the field;

- Database with information on staff: qualifiedpeople, location, contact address, scientificskills, managerial skills, labour costs;

- Database which includes laboratories andopportunities available to investigate, research,experiment, and model in these laboratories;

- Database on equipment, licensing, precisionlevel, input parameters, output parameters,user training required, running costs,consumables: characteristics, quantity, unitprice;

- Database which contains the existing software.- This information will be used to complete each

activity needs of research projects to be carriedout based on the methodology of projectmanagement.

References[1] Brîndasu, P.D., Dezvoltarea unui sistem de e-

learning i e-creativ design în domeniulsculelor alchietoate, Grant CNCSIS, A. Cod105/2006, 2007, 2008.

[2] Brînda u P.D., Beju D.L. s.a., E-learning i e-design în domeniul sculelor a chietoare,Editura Universit ii „Lucian Blaga“, Sibiu,2008.

[3] Popescu, L.G., Contributions on the efficiencyresearch activity in industry, Phd Thesis,"Lucian Blaga" University, Sibiu, 2010.


303

IDENTIFICATION AND QUANTIFICATIONS OF CHEMICAL ELEMENTSAND MICROSCOPIC CHARACTERISATION OF THE COPPER BASE

ALLOY AT DIFFERENT TEMPERATURE OF MOULDS

Cristina CAMBER IORDACHE, Maria VLAD, Simion BALINT,Vasile BASLIU, Mihaela GHEORGHE

Dunarea de Jos University of Galati, e-mail: [email protected]

Abstract: In this research paper, the structure identified by microscopic analysis for one copper base alloy attwo different moulds are comparing between them because the solidified conditions was different. The sampleswere prepared by induction furnace using electrolytic copper and zirconium sponge and the ingots were cast intothe steel moulds where solidifies its. Copper alloys are an knew group of casting Cu alloys, well-known ascorrosion resistant materials, which are widely used as the devices for their antibacterial properties. Dependingon the combination of alloying elements and other factors, like solidification temperature or micro alloyingelements, the alloys with different mechanical and physic-chemical properties are obtained.

Keywords: copper alloys, structure identification, optical microscopy imaging, chemical elements.

1. Introduction

Devices copper alloys are a special group ofindustrial copper alloys which have antibacterialproperties at normal room temperatures(approximately up to 25 C).

These properties of the alloys are resistant tosudden temperature changes. Due to this, in thedesign of this type of alloys, their mechanical andthermal strains have to be critically consideredwithout ignoring their environmentaggressiveness during exploitation. In order toaccomplish such set of requirements, copperalloys assigned for devices production need tohave an appropriate microstructure. Together withchemical compositions and process parameters,microstructure is as important parameter that has asignificant impact on the mechanical properties ofcast parts. [1]

Copper alloys contain various contents ofmajor and minor alloying elements. Cr-Zr hasstable equilibrium phase diagram with amaximum solid solubility of ~0,12 at%Zr at972°C and it reacts with copper and aluminium toform a thermodynamically stable phase. [2] Oneof the major difficulties of copper alloys withzirconium is that the last can lead particularintermetallic effect with deleterious effects on thephysical and mechanical properties of the castparts. Compounds are very hard with the result

that machining cast parts with relatively highzirconium contents can be difficult, resulting inhigh casting finishing costs. Heat treatmentprocesses do not change the size and distributionof this phase.

Are known that zirconium is desirable elementin copper alloys which improves the temperatureproperties and thermal stability of such alloys.Now, all the efforts are being made in attempts tomodify the adverse effect of intermetalliczirconium phases, e.g., decreasing their size andmodifying them into a less harmful morphology.Besides the major alloying elements that have ahuge impact on the solidification path of thesealloys, there are also some minor elements thatsignificantly change their solidification path. It isknown that the addition of elements such as Al,Fe and Mo can modify the zirconium phasemorphology into less harmful shapes. This means,that there is a possibility of controlling the castcomponents by optimizing Zr with usually Aladditions. [3].

The typical copper devices are very complexregarding their chemical compositions andobtained structures. There are alloying elements,Al, Fe, Mo, which have a significant impact onthe solidification path of these alloy Interactionsamong them create different phases andintermetallics, the shape and distribution of whichin the as cast and heat treated alloys depend on the

mailto:Cristina.Camber:@ugal.ro


304

corresponding process parameters. At elevatedtemperature, thermally stable intermetallicsshould stop or reduce the movement ofdislocations and increase the mechanicalproperties of alloys at elevated temperatures. Thestrengthening effect of such intermetallicsdepends on their stability at elevatedtemperatures. The more stable intermetallicsachieved the better strengthening effect.[4]

2. Experimental procedure

2.1. Sample preparation

The structural matter of a material has a directimpact on the physical and mechanical propertiesof a cast. For this reason, the control of themanufacturing process and the attainment of thedesired material properties require that the macroand microstructures present must be defined anddescribed, both qualitatively and quantitatively(Fig.1). [1]

Casting was performed directly into the steelmoulds. Alloy cast into the steel moulds (two)

were prepared by induction furnace type into thegraphite mould at a high temperature max.1300°C.The mass of the cast sample was generally 750g.

The shape of the chill-mould was circular withcone form without corrugate sides. The teemingwas performed into different ways. The melt wasteemed from above into moulds, down-hillcasting, at the different temperature of moulds.

One problem with downhill casting isproducing a good ingot surface. The quality ofsurface depends on how well the moulds havebeen prepared before the casting. In order to studythe surface quality of the ingot it was proceed toheat one of mould before to be used. The secondmould was used at the room temperature.

The stream of melt has a large area exposed tothe air. The exposure greatly increases the risk ofoxygen and nitrogen absorption by the melt.Those gases react chemically with alloying metalsin the melt, in our case with Al and micro alloyingelements, Fe and Mo in our case, which mayoccur in steel, and lead to the formation of oxides.The samples to study were cut horizontal from

A1

A2

B1

B2Fig. 1: Op ti ca l micrograp h (x=400) of th e jo int prepa re d at 130 0 °C and cast in two si milar mo ulds of steel beeing at d if fere nttemperature of mould.A-hot moul d (100 °C). B-could mou ld (10 °C). A1: mi cro gra ph of th e do wn part of the sampl e of the hot mould ; A2: micrograph of the middle part of th esample of the h ot mould, B1: micrograph of the do wn part of the sample of th e co ld mo uld; B2: micrograph of the middle p art of the sample of th e coldmoul d.

x400

x400 x400

x400

1

23

1

2

3

3

2

1

3

2

1


305

different parts of shape and after that wereprepared for analyses at the metallographic laband vertical form different part of shape toanalyse the internal character of crystals aftersolidification process.

Optical analysis was used for microstructureand phase identification. The measurements wereperformed with OLYMPUS BX 51M instrument.The elemental analyse was investigated withspectrometry of X ray fluorescence (FXR).

Table 1 presents a comparative view of theelemental composition of the tested copper alloy:

Chemicalelements

Cold mould Hot mould

Zr 14.76 15.29Cu 78.46 77.91Fe 0.53 0.31Mo 0.070 0.081Al 6.15 6.38

All the investigated alloys of the givenelemental contents were prepared underlaboratory conditions at Faculty of Metallurgy,Material Science and Environment, UniversityDunarea de Jos of Galati. Romania by the internalprocedure, our own recipe for melting andpreparation of copper alloy.

2.2. Sample characterization

The optical microscopy imaging is the firstmethod used for investigating the solidificationpath of metals and alloys. The resolution of theoptical instrument is generally very high, whichmakes this instrument useful for many researchapplications. A search of the available literatureshowed that optical microscopy imaging is apowerful tool that has been successfully used todetermine characteristic of solid phase duringsolidification processes. [5]

To determinate quantitative and qualitativeelemental analyses by FXR of the chemicalcomposition of metallic samples which wareinvestigated for differenced temperature ofmoulds was used the portable spectrometer.

The main aim of the present work was tocharacterize the solidification path of copper alloyCuZrAlFeMo and especially the homogeneity andthe assimilation of the chemical elements during

the solidification process in two similar mouldsbut being at different temperature.

3. Results and discussion

Optical investigationsOptical micrograph imagines of samples casts

by different heating and cooling rate of steelmoulds, i.e., presented in Fig. 1, relived that thesolidification of the alloy is significant dependingby the temperature of moulds.

The microstructure of the samples is shown inFig. 1. (A1, A2-B1, B2). As shown in Fig. 1 atboth samples casts in cold and hot mould themicrostructure is clearly defined with three microstructural constituents: crystalline grains andmechanical mixture of phases. Two microcrystalline constituents: 1-black grains withpolygonal form, 2-blau grains uniform distributedwith round form. The third micro structuralconstituent is copper based matrix phases.

Solidification as encountered in commonprocesses does not occur at equilibrium, sinceduring solidification of most castings,

The major differences between the shapesand the sizes of micro constituents, and theirarrangement/morphology in multiphasesystem are because both temperature andcomposition gradients exist across the castingbetween the moulds walls and alloy.

The four right-hand terms are the changein free energy because of temperature,

composition, curvature, and pressurevariation, respectively.

Table 1 shows obvious differences between thechemical elements as a result of the of the alloycast at the different temperature of moulds.

Copper alloy of the hot mould A, containedhigh percentages of the alloying elements Zr, Aland less percentage of Fe as micro alloyingelement while copper alloy of the cold mould Bcontained low percentage of alloying elements Zrand Al but a high percentage of Fe.

The temperature interval of the solidificationprocess was also shorter and it commenced at thecold mould.

Changes on chemical composition of alloyduring the solidification process can be the resultof different internal reactions.


306

Is known that the solid shell varies with thecasting method and the shapes and sizes of thecastings.

In our case the samples casts into steel mouldsat of the different temperature if moulds have thecolumnar character of the internal crystals. Thecross section area of the columnar crystalincreases with the distances from the cooledsurface of mould.

The growth rate of the solidification frontdecreases with the distance from the ingot surface.

The growth conditions and consequently alsothe structure morphologies are not constant.[6]

The dendrite arms and the crystal crosssections seem to grow in two or three differentsteps,

Eq (1) and (2).The relationship:

growth2

den=const, (1)

is valid in this case, but with differentconstants for each steps. At the constant eutectictemperature precipitation of the eutectic alloyoccurs.

Ingot with long solidification times the numberof new floating crystal decrease during thesolidification process and the central zone ofbranched columnar crystals is extended over thewhole section in the upper part of the ingot.[5]

Is known that that the necessary condition forsolidification is undercooling:

dr/dt= (TL-Tcrystal)n (2)

The temperature of the melt and the solidphase of an ingot as a function of time.

4. Conclusions

This paper presents the results of research oncopper based alloy in the different temperature ofmoulds cast.

The structure can be characterised as a veryfine surface zone that consists of a fine network oftin dendrite arms and not, as in ingots, of a greatnumber of fine-grained crystals.

At the ingot from into the cooled mould wecan observe the less quantity of micro alloyingelements assimilate comparing with the chemicalelements from the second ingot where the microalloying elements are assimilated more.

In the same time, the dendrites from thecooling mould became tin comparing with thesecond ingot.

Due to the high temperature gradient the thindendrite crystals grow inwards and form columnarcrystals. In the cooled mould its are stopped bythe formation of the equiaxed crystals that havegrown freely in the melt.[1]

With the aim of improve the corrosionresistance of the antibacterial copper based alloys,a number of copper alloys can be prepared bychanging the concentrations of alloy elementsand micro alloying elements in order to increasethe each possible phase or change of its shape inwhich the alloy element is dominant, since thephenomenon of change of the mean concentrationin the actual case, due to redistribution of liquidenriched or depleted by the alloy elements, leadsto a change of mean concentrations to valuessignificantly higher or lower than the nominalvalue, and to the possibility of the occurrence ofother phases.[4]

When a larger portion of micro alloyingelements exists in the matrix phase thesolidification occurs at different temperatures,which is good. However, a further increase in thecontent of micro alloying elements may lead tothe formation of a new phase, the strengtheningeffect of which is poorer. This means, the microallowing elements has to be added in a properamount to the alloy, depending on the Coppercontent [2], [3] and [5].

The results presented in this paper are only anintroduction to further ongoing research aimed atobtaining the required combination of the furtherantibacterial copper alloy with the goal ofproviding optimal characteristics for mechanicaland corrosion resistance as well as defining amathematical model that can predict the differentcases on the micro and macro levels ofsegregation.

Performed thermal analysis and microscopiccharacterization of Cu-enriched phases can bedetermined using thermal analysis. Theimportance of the analysis of these phases is dueto their strengthening effect that is usuallyenhanced and controlled by applying heattreatments that promote the precipitation ofcoherent or incoherent alloying or micro alloyingelements that improve the properties of Copperbased alloy.


307

Copper alloy with better properties than thoseachieved using conventional casting method canbe produced with the rapid solidification methods.

Acknowledgement

We would like to thank to the Project SOPHRD – EFICIENT 61445/2009 for financialsupport and we also thank to Prof. AnisoaraCiocan and Prof. Florentina Potecasu of theFaculty of Metallurgy, Material Science andEnvironment, University “Dunarea de Jos” ofGalati for providing facilities.

References

[1] Fredriksson, H., Akerlind, U., MaterialsProcessing during Casting, John Wiley&soons, Ltd, England, 2006.

[2] D.Arias, J.P. Abriata, Buletin of alloys phasediagrams, Vol.11 No.5, 1990.

[3] Moise Ienciu, Nicolae Panait, PetruMoldovan, Mihai Buzatu, Elaborarea siturnarea aliajelor neferoase speciale, EdituraDidactica si Pedagogica Bucuresti.

[4] Kebbache I, Debili, M.Y., Separation ofAluminum and Copper by IntermetallicCompounds after HF Induction Fusion,, JOM,2010

[5] ASTM[6] Scandinavian Journal of Metallurgy,

Processing and Materials Engineering Ed.Wiley, Ltd, England, 2006.

[7] Effenberg, G., Ilyenko, S., Al-Cu-Zr(Aluminium - Copper - Zirconium),SpringerMaterials{http://www.springermaterials.com).

http://www.springermaterials.com)./


308


309

MANGEMENT OF TECHNICAL DATA IN THE STUDY OF THEBRAKING SYSTEM OF CARS

CIUBOTARIU DANUT1*, NECULAIASA VASILE2

1Ministry of Domestic Affairs and Administration, SPCRPCIV, Botosani, Romania,email danutciubotaru@yahoo.

2Gheorghe Asachi Technical University, Iasi, Romania

Abstract: Braking process of a vehicle is complex, giving possibility to the driver to reduce car’sspeed or to stop it at big speed in a short distance. Using the diagrams registered for the differentworking conditions, defined accordingly with the experimental plan, there have been analyzed thevalues of deceleration, variation of speed and the covered space for each specified moment ofbraking. As result of the experimental researches, we have reached the conclusion that a goodbehavior at cars braking is registered when the diagrams of deceleration variation in relation withbraking time have a symmetric shape.

Keywords: vehicle stability, experimental braking process, ABS.

1. INTRODUCTION

Braking process of a vehicle is complex, beingdefined by factors leading to different effects fromthe point of view of impact, intensity andduration.

The braking system gives possibility to thedriver to reduce car’s speed or to stop it at bigspeed in a short distance.

Last generation cars are endowed with:-Service brakes;-Backup brakes;-Handbrake;-Auxiliary brake.From the point of view of the vehicle’s control

and safety, the most important is the servicebraking system (the foot brake).

The longitudinal deceleration ax is one of themain parameters of the braking system, which canbe calculated with formula:

(1)

where x is the longitudinal coefficient ofadherence of the tires with the running path and g– the gravitational acceleration (g = 9,81 m/s2).

Because, in reality, a car is not perfectfrom the point of view of the technical

characteristics, the effectiveness of braking islower.

The Directive of EU 71/320 [1] shows that thiscriterion has to satisfy the following condition:

(2)

In the specialty literature, the given value forthe adherence of the tires with the running path ona dry asphalt is x = 0.8.

Introducing this value, which applies only tothe vehicles and tires produced until 1980, inrelations (1) and (2), there results the decelerationof a vehicle in good technical conditions, atbraking on a dry asphalt

ax = 6.0 – 7,85 [m/s2]In the situation of actual cars, the maximum

coefficient of adherence has the value x = 1– 1,2[2] , if braking takes place on a running path withdry asphalt, situation when the deceleration of oldcars with tires produced at the present momentcan reach values in the range 7,35÷ 9,3 [m/s2].

The modern cars are endowed with systemswhich prevent wheels blocking (ABS) and the realbraking distance is quite close to the calculatedvalue obtained with the maximum values of thecoefficient of adherence. In this situation, thedeceleration can be close to g = 9,8 [m/s2].


310

2. ANTI-BLOCKING SYSTEMS (ABS)

Anti-blocking System is used at cars foravoiding the blocking of wheels during braking,for providing the stability of the vehicle and anoptimal deceleration.

Anti-blocking of wheels is realized by adaptingthe braking force at each wheel, function of thewheel’s adherence to the running path.

The advantages of a system [3] assure keepingthe vehicle under control, although the brakingand the side forces are increasing, on a minimumbraking distance with a reduced wear of tires.

Some of the conditions to be accomplished byan ABS are following presented:

1. ABS must function on all the range ofspeeds;

2. ABS must correspond to the conditions ofadherence with the running path;

3. To assure the steering of vehicle atslippages appearance;

4. To recognize aquaplaning and to respond toit;

5. To rapidly and automately correlate thehysteresis of tires at specific conditions;

6. The option of engine brake to be possible;7. If an error at the ABS functions occurs,

during braking, the braking system mustfunction normally on classical principles.

t

S

KGDGKPDP1

2

3

Vehicle speed

Figure 1 Variation of speed in relation to the breakingtime

The functioning principle of the anti-blockingsystem is based on rapid braking without wheelsblocking, by applying short and rapid actionshaving as objective the delay of wheels blocking.This process is realized by the action upon thepressure of brake fluid from the brake installation,regulated to avoid the wheel blocking regardlessthe adherence to the running path.

In fig.1 [4] there is presented the diagram ofvariation for the speed of a vehicle with ABS inrelation with the braking time, for the four wheelsof the car.

When the brake pedal is acted, the value ofwheels rotational speed is decreasing due to a truevalue depending to the vehicle’s speed (pt.1 fromfig. 1.). When the force which the brake is pressedwith is bigger or the running path is slippery, therotational speed of wheels significantly reduces(pt.2). In pt.3, ABS is activated and releases thebrakes. When the rotational speed of wheelsincreases again, the brakes are once againactivated, the process being repeated until thevehicle is brought to the desired speed.

2.1. Braking of vehicles without wheelsblocking

If the car is not endowed with a regulator ofthe braking force of the ABS, the total brakingforce is differently distributed between thevehicle’s bridges, from which reason the wheelsof the front bridge and those of the rear bridge arenot simultaneously blocked.

In the situation that the wheels have a goodadherence to the surface of the running path, therear wheels are first blocked, context in which thevehicle may lose transversal stability and if thefront wheels are blocked, the control of theadvancing direction can be lost.

A good behavior during braking is met whenthe motion of the car is continued in straight lineat brake action, even for great speeds, conditionimposed accordingly to the Directive of EU71/320. So, in the moment when the emergencybrake is acted, the motion of the car must continuein straight line, it means a bigger relative brakingforce upon the front wheels must be provided, incomparison with the force acting on the rearwheels.

The coefficient of distribution of the brakingforce [5] represents the ratio of the braking forceacting on the front wheels, to the total brakingforce, acting on all the wheels.

T = PST1 / (PST1 + PST2) (3)

PST1 and PST2 – represent the braking forcesupon the front wheels and respectively upon therear ones and they depend on structure of thebrake system. This coefficient ( T) can be


311

calculated for any type of vehicle [5], accordinglyto the known parameters of the brake system.

l l

L

h

P

Pin

z

stP

GzP

stP

1

Figure2 Scheme of forces in the process of brakingwithout wheelss blocking

In the situation of neglecting the loss of energy,for overcoming the resistance forces, themaximum of the deceleration for a vehicle at thelimit of rear wheels blocking will be:

ax = (4)

Where: l1 = the distance in longitudinal direction,between the mass center of the vehicle and thefront wheel, L represents the distance betweenwheels (the wheelbase) and hc is the height of themass center.The maximum deceleration of a vehicle at thelimit of front wheels blocking is:

ax = (5)

Where: l2 = the distance between the rear wheeland the car’s mass center.

The optimum coefficient of adherence opt, will be given by relation:

opt = (L .T – l2 )/ hc (6)

In the conditions when the correct value of thecoefficient of adherence exceeds opt, the rearwheels will be firstly blocked by braking andrelation (4) allows calculation of deceleration. Inthe situation when the value of the coefficient ofadherence is smaller than the optimum value opt,the front wheels will be the first blocked when thevehicle braking and the expression (5) will beused for calculate the deceleration.

When opt has a value close to zero or isnegative, only the rear wheels will be blocked inany situation of motion.

In the situation of a vehicle with unblockedwheels, the braking distance is calculated withformula:

S0 = (t1 + t2 + 0,5 . t3) . v0 + v02 /2ax (7)

In the situation when the initial speed of thevehicle (v0) is not known, the braking distance ofa vehicle with unblocked wheels (unlike the caseof blocked wheels, when the slip length can bemeasured) can be determined with certain errorsand this circumstance can influence thecorrectness of the obtained results.

In the situation of emergency braking, thebraking forces acting on the vehicles’ wheelsshould not exceed the forces of adherence of thetires at the running path, the calculated value ofthe deceleration of a vehicle with blocked wheelswill be smaller and the braking time and brakingdistance will be bigger, in comparison with thecase when the coefficient of adherence is known.

2.2. Comparison of the braking processes forvehicles with ABS and without ABS

From the study of vehicles’ theory, thedeceleration of a vehicle in good technicalconditions without ABS reaches the top, fig.3, justat the beginning of the braking process and thensome decreasings are happening, due to the factthat the top of deceleration is reached beforeblocking the vehicle’s wheels.

Figure 3 Diagram of braking with ABS-phase ofstarting the braking process

After the wheels blocking, a decrease of thedeceleration takes place, because braking of ablocked wheel is less efficient [6].

Maximum of deceleration lasts more duringthe braking of a vehicle without ABS at a lowerspeed; at a bigger speed, when the braking time isgreater, the top of deceleration lasts less. This factexplains the decrease of speed for vehicles


312

without ABS when a certain displacement speedis reached.

In fig.4 the diagram of variation ofdeceleration in relation with time is presented,during the period of speed variation of vehiclewithout ABS.

0,8

10

Dec

eler

atio

n [m

/s ]2

24

68

0

0,6

0,4

0,2

0

10 1

a) b)Figure 4 Diagram of braking without ABS: a)- phase of starting the braking process; b)- phase of real braking

In specific road conditions, the value of thevehicle’s deceleration may have a biggercorrectness, which imposes an experimentalresearch.

The deceleration of a vehicle on a running pathcan be measured using a special device(decelerometer) or can be calculated with relation[6], if the initial speed v0 and the braking distanceare known.

axn= (8)

3. EXPERIMENTAL RESEARCH UPONVEHICLE’S BRAKE

From those above presented, the necessity ofdeveloping some experimental researches uponthe process of cars’ braking has resulted. Forexperiments, two Daewoo cars in good technicalshape and with the following characteristics havebeen used:1) for vehicle without ABS

a) Vehicle category M 1;b) Wheelbase of the vehicle, at dynamic test:2360 mm;c) Mass: nominal: 1050 kg, maximum: 1550kg;d) Maximum speed: 190 km/h;

e) Tires dimensions: 185/65 R 15;f) Service brake with two independent circuits;1. Brake type : front: disk and rear : drum ;g) Brake with double circuit;h) Without ABS;i) Mass of vehicle to be loaded:-Test I: 1103 kg (667 kg + 436 kg)

-Test II: 1558 kg (798 kg + 760 kg)2) for vehicle with ABSa)Vehicle category M 1;b) Wheelbase of the vehicle, at dynamic test:2360 mm;c) Mass: nominal: 1050 kg and maximum:1600 kg;d) Maximum speed : 190 km/h;e) Tires dimensions: 185/65 R 15;f) Service brake with two independent circuits,in X;g) Mark and type of brake lining: front: ferrodo182 ; rear : DOW 8273 ;h) Brake with double circuit;i) Brake type : front: disk and rear: drum ;j) Vehicle endowed with ABS;

The experiments have been done in the testingground of Daewoo Craiova cars factory, ondifferent running paths with surfaces coveredwith: asphalt dry - wet, gook, glaze, snow, car


313

empty – loaded, engine clutched – declutched, andalso in normal conditions of circulation onnational roads in Botosani County, with theTesting Laboratory owned by SCROMTURINGIA.

In the experiments, a device for measuringvehicles’ deceleration (decelerometer) typeMAHA VZN 100 was used. The device isappropriate for such measurements both forvehicles with hydraulic and pneumatic brakesystem and for those with ABS, as in fig.5.

The device is compound of: control board, fordata displaying and registering; transducer wheelwith its fixing support; flexible drive cable anddynamometric pedal (specific for braking tests).

The transducer wheel transmits rotationalmotion by the flexible cable to the control board,for displaying and registering, where thedisplacement speed can be read at any moment. Inthe registering apparatus, a time base isembedded, which electrical signals, at a period ofone second, are marked on the registering paper,moving by rolling on the internal drums, with aspeed proportional with the speed the wheel isrunning on the ground.

The used device measures the processparameters and draws the corresponding diagrams

for speed, deceleration and the covered space,function of the type of braking.

The conformity of the measuring device VZN100 with the standards stipulated for the device’schecking is certified by the German Associationfor Technical Inspection (TUV) [5]. This device isalso used by the Lithuanian centers for technicalinspection for establishing the efficiency of thebrake system, basing on the maximumdeceleration.

Using the diagrams registered for the differentworking conditions, defined accordingly with theexperimental plan, there have been analyzed thevalues of deceleration, variation of speed and thecovered space for each specified moment ofbraking.

As result of the experimental researches, wehave reached the conclusion that a good behaviorat cars braking is registered when the diagrams ofdeceleration variation in relation with brakingtime have a symmetric shape.

Based on the found parameters of vehicle’smotion, we can provide the variation of themaximum values of deceleration for specificsituations.

Figure 5. Principle scheme of a car endowed with the 5th whell


314

5. CONCLUSIONS

Based on the bibliographical researches in theconsidered technical field and on the experimentalresearches, certain conclusions of undoubtedimportance can be presented:

I. In the developed experiments, thedeceleration established for vehicles with ABShave varied from 8,05 m/s2 (at the speed of 40km/h) up to 8,45 m/s2 (at the speed of 80 km/h),and often the maximum values have been close to9 m/s2 .

For vehicles without ABS, the decelerationhave varied from 7,05 m/s2 (at speed of 40 km/h)up to 6,87 m/s2 (at speed of 60 km/h) and up to6,65 m/s2 at a speed of 80 km/h. When v0increases, the difference of deceleration orvehicles with and without ABS increases with11,1%, 18% and respectively 24%.

II. There was calculated that the average ofbraking distance for a vehicle with ABS is smallerwith 4,25% in comparison with those for a vehiclewithout ABS.

If the initial motion speed is 40 km/h, and if v0= 80 km/h, approximately the deceleration ismodifying up to 15,59%;

If v0 = 60 km/h, the braking distance forvehicles with ABS was of 20,80 m, and forvehicles without ABS it was of 22,42 m, thedifference being of approximately 7,35%.

III. The value of the braking distance can becalculated based on the braking processparameters, such as: braking time t3 , the value ofdeceleration axn and the initial speed v0.

There is proposed the use of the values of theparameters determined within the experimentalresearch for the examination of situations specificfor traffic accidents, for a more correct defining ofthe methods for calculating the brakingparameters for cars with or without ABS.

IV. As a result of research accomplishing,there has been noticed that the used cars arecorresponding and respect the stipulations of theRegulations of braking R-13ECE – ONU. That iswhy the regulations of EU impose for all the carsbuilt since 2011 the obligation of endowment withABS.

REFERENCES

[1]. ***, Council Directive 71/320/EEC of 26 July1971 on the approximation of the laws of theMember States relating to the braking devicesof certain categories of motor vehicles and oftheir trailers.

[2]. Mitunevicius, V.-1999-Application of brakingresponses of vecles for expert’s examinationsof vehicles.Im Transbal-tica-99:Collection ofscientific reports of the InternationalConference (Transbaltica-99;tarptautineskonferencijos moksliniu pranesimu rinkinys,Vilnius, 8-9 April 1999).Vilnius: Technika,1999, p.221-226(in Lithuanian).

[3]. Neculaiasa Vasile-1996 Vehicle Dynamics,Polirom Publishing House, Iasi, 1996;

[4]. Gillespie,T.D.-1992 –Fundamentals of vehicledynamics.Society of Automotive Engineering,Inc.400 Commonwealth Drive Warrendale, PA15096-001.1992.250 p.

[5]. Illarionov, V.A.-1997-Expert’s examinationof traffic accidents Moscow;Transport,1997.255 . (in Russian).

[6]. Sokolovsku, E.-2004- Investigation oninteraction of the wheel with road its elementsin the context of examination of trafficaccidents: Doctor;s thesis:tehnologicalsciences:Transport engineering (03T) Vilnius,2004.147 p.


315

COMPOSITE SHAPE MEMORY ALLOYS FOR METALLICSTRUCTURES

A. Enache, I. Cimpoe u, S. Stanciu, I. Hopulele, R. M. Florea, N. Cimpoe u*

The ”Gh. Asachi” Technical University of Iasi Faculty of Materials Science and Engineering,Prof. dr. docent Dimitrie Mangeron Rd., 61 A, 700050, Ia i, Romania

Abstract: Shape memory alloys exhibit nice and interesting properties based on them internal solid statetransformation from martensite to austenite under temperature modifications. Damping capacity represent one ofthe latest attraction of shape memory alloys based on them big internal friction values having applications inmany domains as dissipaters, dampers or structural elements for noise and vibrations attenuation. Compositewith metal matrix materials represent interesting materials that use both materials properties. In this study ashape memory alloy composite material is analyzed by his interface between SMA matrix material and a steelreinforcement element point of view. Using scanning electrons microscopy and EDAX analysis few results wereobtains and comment to establish the interface nature.

Keywords: shape memory alloy, composite, diffusion

1. Introduction

Among the various types of composites, thefamily of discontinuous metal matrix composites(MMCs) containing particulates, whiskers, wires,precipitates, fibers, nodules and platelets, arefavoured because they offer improvements to themechanical properties of the monolithic alloyswhile remaining relatively easily deformable.Metal matrix composites are compound materialswhose microstructure consists of a metallic alloyinto which a particular reinforcing component isintroduced. MMCs offer advantages in applicationswhere low density, high strength and high stiffnessare a primary concern. The availability of varioustypes of reinforcements at competitive costs, thefeasibility of mass production and high dampingcapacity [1,2] make this type of MMC, metallicmatrix and wires as reinforcement elements, moreattractive. However, these materials may sufferfrom inhomogeneous distribution, size or shape ofwires, low ductility, inadequate fracture toughnessand inferior fatigue crack growth performancecompared to that of the matrix [3,4].

To optimize the mechanical and physicalproperties, in particular the damping conditionsand damage tolerance of such materials one canutilize shape memory alloys as reinforcement.Shape memory alloys (SMAs) have received greatattention because of their shape memory effect

(SME) and many investigations are conducted ontheir basic performance and applications.

The transformation under compression canresult in stress in the matrix, which in turnenhances mechanical properties such as yield stress[5–10], fracture resistance [11,12], an capacity ofsuppression of crack growth [20] and thermo-mechanical fatigue [9,13]. Two SMAs whichgenerate large amounts of strain and are capable ofgenerating a large force upon transformation backto the austenitic phase are NiTi alloys and Cu-based alloys. Copper-based SMAs are particularlyinteresting because of their low-cost and relativeease of processing. On the other hand, NiTi alloystend to be more thermally stable and to have alower density, higher yield and ultimate tensilestrength; they are also more resistant to corrosionthan Cu-based alloys [14].

Traditional restoration techniques often do notgive structures sufficient resistance againstmaximum expected earthquakes and/or might betoo invasive. Therefore, there is an ongoing effortto find techniques that can guarantee structuralstability and at the same time respect the integrityof the structure. Special devices that exploit thesuperb damping properties of Shape

Memory Alloys (SMA) are under development.SMAs have found applications in many areas dueto their high power density, solid state actuation,high damping capacity, durability and fatigueresistance. When integrated with civil structures,


316

SMAs can be passive, semi-active, or activecomponents to reduce damage caused byenvironmental impacts or earthquakes.

Though most of the research activities ofSMAs’ applications in civil structures are still inlaboratory stage, a few have been implemented forfield applications and found effective [15-19].

Having a composite material with a metallicmatrix of a copper based shape memory alloy inthis study were analyzed some concerning aboutthe metallic interface nature. The results present ahighly well form interface between the matrix andreinforcement element, both metallic materials thatimprove the mechanical behavior of the material.


To obtain the shape memory alloy based oncopper for the composite matrix was used alaboratory furnace with graphite crucible usingcopper, zinc and aluminum high purity materialswith reduce percentage of iron [21].

After the matrix were melted, the liquid alloy,was poured in a special semi-form prepared forarch material reinforcement element stabilizationwith the geometrics and main elements presentedin figure 1.

Figure 1: Composite realization semiform having theelements: 1-metallic body form; 2-cylindric cavity for

reinforcement element fixation and matrix body form; 3-articulation; 4-rod suport for arch fixation at extension;

5- rod used for supporting the extension arches forwires; 6- hand support, 7-reinforcement fiber; 8-arch

used for metallic fiber extension

Each melting provide material for nine sampleswith different materials as reinforcement element ifis necessary or similar like in this case.

In figure 2 is presented the geometrical shape ofthe composite samples obtained and furtheranalyzed.

Figure 2: Catia software design of the shape memorybased composite

Chemical composition was determined throughspark spectrometry analysis using Foundry Masterequipment (for matrix and reinforcement elementschemical analysis) and EDAX analysis as well forinterface study. In this study different EDAXsoftware applications were used to determine thechemical variation of the elements on Line,Mapping or Point mode with automatic or elementlist considerations.

Microstructures of the composite in differentheat treated states of material were obtained with ascanning electron microscope (SEM) LMH II byVega Tescan brand using a secondary electrons(SE) detector at a 30 kV lamp power supply.


A composite material made by CuZnAl shapememory alloy as matrix and arch wire steel asreinforcement element was obtain by classicalmelting method.

Composite material surface was analyzed usinga scanning electrons microscope to observeespecially the interface between those two metallicmaterials that form the composite. Frommicrostructure images, figure 3 a) and b), can beobserve that a thick interface is formed incomposite after the material suffer a 30 minuteshomogeneous thermal treatment.

A proper chemical attack of CuZnAl and Fe-Calloys to evidence both microstructures is difficult,as can be seeing from figure 3 were only theinterface is marked better, interface based onaluminum compounds. On the matrix someformation of martensite variants can be seeing andon reinforcement element no microstructure isevidenced.

Microstructural analyses represented in figure 3evidence an interface formed between the shapememory matrix and the elastic element withdifferent dimensional areas and relatively straight.


317

a)

b)Figure 3: SEM microstructure of the composite

material a) 1000 x power and detail b) at 2500x

The interface dimension is around 1-2 µm withareas modify as inhomogeneous.

In figure 4, on a selected line presented in a), isrepresented the chemical elements copper, zinc,aluminum, iron, manganese and silicondistributions. The selected line represents theelements distributions on matrix, the reinforcementelement and interface as well in the same time. Theline selected is of 20 µm perpendicular on theborder area and evidencing the mass transfer typeat the interface.

In figure 4 b) is presented a complete chemicalanalysis of the elements variation between matrixand reinforcement element with accent on theinterface. Can be observed a diffusion typematerial transfer between those two metallicalloys. On the interface area formation of somenew chemical compounds are observed as FeAl3and a migration of elements characterize theinterface find a higher aluminum percentage on thesteel part than the shape memory alloy materialwere a loss of aluminum is observed.

In figure 4 b) the interface areas is analyzedthrough mapping elements technique as wellmarking the aluminum conglomeration on theinterface, a separation layer on the interface andthe loss of Al element from the matrix. Siliconelement is also presented on the border formingdifferent stabilization compounds like AlSi2.

a)


318

b)Figure 4: Composite material interface area analysis a) selected area for analysis and b) distribution of elements

on the interface area

In figure 3 b) two points were selected forchemical composition investigations the resultsbeing presented in table 1 and table 2 for Pct 1 andPct 2. First point is selected on the interface andthe second one near the border on thereinforcement element part. Even the first point issituated more to the matrix part the iron elementhas a very high percentage presence based on

diffusion process movement of the elements.Aluminum is also present having a FeAl basedinterface. The other elements participate at theinterface formation in smaller percentages andloosing in front of the iron aluminum compoundsas X-ray signal and quantitative percentages.

Table 1: Chemical composition of the interface in point 1 selected in figure 1 b)

Element AN series Net [wt.%] [norm. wt.%] [norm. at.%] Error in %Iron 26 K-series 99081 71,50894 81,84484 75,4816 1,844726

Aluminium 13 K-series 8076 6,7355 7,709049 14,71579 0,381417Copper 29 K-series 5054 5,971241 6,834324 5,539327 0,197696

Zinc 30 K-series 1144 1,479678 1,69355 1,333941 0,084248Silicon 14 K-series 1761 1,102398 1,261739 2,31386 0,086212

Manganese 25 K-series 921 0,573593 0,6565 0,615477 0,084937Sum: 87,37135 100 100

Table 2: Chemical composition of the interface in point 2 selected in figure 1 b)

Element AN series Net [wt.%] [norm. wt.%] [norm. at.%] Error in %Iron 26 K-series 120555 80,01884 91,8602 89,77184 2,05754

Copper 29 K-series 3090 3,256238 3,738104 3,210522 0,127146Aluminium 13 K-series 1858 1,924301 2,209063 4,46842 0,141725

Zinc 30 K-series 701 0,820734 0,942188 0,786391 0,063644Manganese 25 K-series 1126 0,611817 0,702355 0,697745 0,088394

Silicon 14 K-series 655 0,477439 0,548092 1,065084 0,057611Sum: 87,10937 100 100


319

On the other part of the interface, at thereinforcement element side, copper is diffusedin a reduce percentage of 3.738 % wt buthigher than the copper presence in the arch-wire steel material.

Silicon element decrease as well in percentagecomparing to the other side of the interface factthat represent a smaller formation of silicon basedcompounds. Mechanical tests of the compositematerial with present the material behavior,interface modifications and material properties asdamping element.

4. Conclusions

A composite based on a shape memory alloymatrix were obtain and analyze by the interfaceformation point of view. The reinforcementelement proposes was a steel archwire with 1 mmdiameter. The interface between those two metallicmaterials was between 1 and 2 µm thick especiallybased on FeAl compounds. In formation of theinterface diffusion process interfere with accent onaluminum and silicon elements decreasing thecopper and zinc percentages. Interesting variationsof chemical elements distribution can be observeon the border between materials anyway a nicebond between the metallic materials was created.

Acknowledgement

This paper was realised with the support ofPOSDRU CUANTUMDOC “DOCTORALSTUDIES FOR EUROPEAN PERFORMANCESIN RESEARCH AND INOVATION” ID79407project funded by the European Social Found andRomanian Government.

References

[1] Cimpoe u, N., Stanciu, S., Meyer, M.,Ioni , I., Cimpoe u Hanu, R., Effect ofstress on damping capacity of a shapememory alloy CuZnAl, Journal ofOptoelectronics and Advanced Materials,Vol. 12, Issue 2, pg. 386-391, IDS 570DE,ISSD 1454-4164, 2010.

[2] Paun, M. A., Cimpoesu Hanu, R.,Cimpoesu, N., Agop, M., Baciu, C.,Stratulat, S., Nejneru, C., Internal frictionphenomena at polymeric and metallic shapememory materials. Experimental and

theoretical results, Materiale Plastice, vol.47, nr. 2 , pg. 209-214, 2010.

[3] Davidson, D.L., Eng. Fract. Mech. 33,965–977, 1989.

[4] Davidson, D.L., Campbell, J.B., Page,R.A., Metall. Trans. A 22, 377–391, 1991.

[5] Furuya, Y., Sasaki, A., M. Taya, Mater.Trans. JIM 34, 224–227, 1993.

[6] Hamada, K., Lee, J.H., Mizuuchi, K.,Taya, M., K. Inoue, Metall. Mater. Trans. A29, 1127–1135, 1998.

[7] [7] G.A. Porter, P.K. Liaw, T.N. Tiegs,K.H.Wu,Mater. Sci. Eng. A 314 (2001)186–193.

[8] Porter, G.A., Liaw, P.K., Tiegs, T.N., Wu,K.H., JOM, 52–56, 2000.

[9] Shimamoto, A., Taya, M., Trans. JSME 63,26–31, 1997.

[10] Tsoi, K.A., Stalmans, R., Wevers, M.,Schrooten, J., Mai, Y.X., Proc. SPIE 4234,125–133,2001.

[11] Furuya, Y., Taya, M., Trans. JIM 60,1163–1172, 1996

[12] Shimamoto, A., Zhao, H.Y., Abe, H., Intl.J. Fatigue 26, 533–542, 2004.

[13] Shimamoto, A., Furuyama, Y., Abe, H.,Key Eng.Mater. 334–335, 1093–1096,2007.

[14] Loughran, G.M., Shield, T.W., Leo, P.H.,Int. J. Solid Struct. 40 271–294, 2003.

[15] Saadat, S., Salichs, J., Noori, M., Hou, Z.,Davoodi, H., Bar-on, I., Suzuki, Y. andMasuda, A. An Overview of Vibration andSeismic Application of NiTi Shape MemoryAlloy, Smart Mat. and Struc., , Vol. 11, p.218–229, 2002.

[16] Van der Eijk, C., Zhang, Z.L. andAkselsen, O.M., “Seismic Dampers Basedon Shape Memory Alloys: MetallurgicalBackground and Modeling”, Proc. ThirdEuropean Conference on Structural Control,ECSC, Vienna, Austria, 12-15 July 2004, p.M1-5.

[17] Song, G., Ma, N., and Li, H.-N.,“Applications of Shape Memory Alloys inCivil Structures”, Eng. Struc., 2006, Vol.28, no. 9, p. 1266-1274.

[18] Janke, L., Czaderski, C., Motavalli, M.,and Ruth, “J. Applications of Shape


320

Memory Alloys in Civil EngineeringStructures-Overview, Limits and newIdeas”, Mat. and Struc., 2005, Vol. 38, p.578-592.

[19] Dolce, M. and Cardone, D., MechanicalBehaviour of Shape Memory Alloys forSeismic Applications; Austenite NiTi WiresSubjected to Tension Int. J. of Mech. Sc.,Vol. 43, no. 11, p. 2657-2677,2001.

[20] W. Huang, Mater. Design 23 11–19, 2002.[21] Mares, M. A., Carcea, I., Chelariu ,R.,

Roman, C., Mechanical and tribologicalproperties of aluminium matrix compositesafter ageing treatments, Euromat 97 -proceedings of the 5th european conferenceon advanced materials and processes andapplications: materials, functionality &design, vol. 1 - metals and composites pg.423-426, 1997.


321

STUDY OF THERMOMECHANICAL FATIGUE FOR SHAPEMEMORY ALLOYS TYPE CuZnAl

Achitei Dragos Cristian1, Hopulele Ioan1, Perju Manuela Cristina1, Mihai Axinte1

1Gheorghe Asachi Technical University of Ia i, [email protected]

Abstract: The paper presents the design elements of a prototype installation for the study of specificthermal/mechanical fatigue of shape memory alloys, and fatigue curves obtained respectively.

Keywords: shape memory alloys, fatigue curve, prototype installation

1. Introduction

Memory shape alloys from some installationsare submitted to mechanical stresses and to thermalstresses so that their integration within a fatiguecalculus system supposes taking into considerationtheir function.

Figure 1. Actuators

Memory shape alloys have a series of propertiesalike the other ordinary metallic materials. Betweenthose, an important characteristic is their capacityof changing the geometric shape from a lowtemperature to a high one.

Under certain circumstances, shape changingcan be reversible so that the materials canmemorize two geometric shapes such as the hightemperature shape and low temperature shape.

These transformations accomplish because of aneffect named memory shape effect. By memoryshape effect, one can understand that the materialcan make labor work when passing from the coldshape to the warm one.

The pieces subjected to variable (cyclic) stressesdestroy to stresses, which are inferior to staticfracture resistance. The maximum cyclic stress

when a material does not brake calls fatigueresistance of the material.

In order to dimension the constructive shapeof an element with memory shape from amechanic device, a series of numeric values arenecessary, among these values being also fatigueresistance, labeled by the minimum value of thedeformation recovered after a certain number ofuse cycles.

In addition, besides the phenomena met inclassic crystalline materials, memory shapealloys present supplementary connectionmechanisms at phase change, which arecharacteristic only for memory shape alloys.

Taking into consideration a device wherememory element makes a double sense memoryshape effect, within work system (resortcoupling uses for recovery), fatigue resistancelimit defines by the number of cycles untilrecovery tension lowers to a minimum value(approximately 70% from the initial one).

According to the cycling type, a memoryshape alloy can present irreversible deteriorationphenomena of the microstructure defined byspecific categories of fatigue.

If cycling though double effort, fatigue isthermal and by cycling through simple effort ofmemory shape thermo-mechanical fatigueappears.

The fatigue of metals is the phenomenon thatproduces the breakage of different pieces, undertemperature variation conditions and other workparameters, too.

mailto:dragos_adc:@yahoo.com


322

Mechanical fatigue implies breakage productionin the following stages: defects accumulation,cracks formation and conduction, in stationaryregime at first and non stationary at final breakage.

Figure 2. Fatigue cracks appeared in the areas markedwith 1, they conducted in the interior of the piece and thefinal breakage produced quickly in the areas markedwith 2.

The breakages that appear when applying somevariable loads are called fatigue breakages probablybecause it noticed generally they appear only after aconsiderable period.

A fatigue breakage is very dangerous because itappears as a preliminary warning. The properfatigue breakage is fragile, without important globaldeformations. At macroscopic scale, breakagesurface is usually perpendicular on the direction ofthe main normal stress.

A fatigue breakage can be easily recognizedbased on the aspect of the fracture surface, whichpresents a smooth area and a rough area.

Many times, the way breakage advanced isindicated by a series of concentric circles, whichadvance from the initial point of the breakagetowards the interior of the section.

The stresses producing fatigue breakage at hightemperature cannot appear due to some mechaniccauses. Fatigue fracture can be produced byvariable stresses due to temperature, underconditions where stresses due to mechanic causesdo not produce.

Thermal stresses produce when the variation ofa piece dimension is stopped in a certain way.

If the breakage appears due to a single applianceof thermal stresses, the stress calls thermal shock.

If the breakage appears after many appliances ofsome thermal stresses of low values, the solicitationcalls thermal fatigue. In the equipments working at

high temperatures, there are usually created thepremises for breakages due to thermal fatigue.

Figure 3. Cracks appeared due to thermal fatigue.

2. Experimental installation

In order to study thermo-mechanicalphenomenon of memory shape alloys it designedand realized a prototype with a complexconfiguration.

Figure 4. Standard samples, from memory shapealloys type CuZnAl, for thermal fatigue test.

The number of cycles necessary for a loadingis 104 – 105 cycles and functioning time for theachievement of these cycles is of the order oftens hours.

Standard sample made of memory shapealloy will be subjected to traction, being caughtinto the dies, loading realizes by a system oflevers, and at their ends, some weights areattached.

Samples with memory shape alloys weretested with the chemical composition presentedin table 1.

Table 1Cu Zn Al Fe Si Mn75.4 18.6 5.85 0.021 0.026 0.007Conventional notation: Cu75.4Zn18.6Al5.85.


323

The scheme of the prototype installationpresents in figure 5.

Figure 5. Prototype installation for thermo-mechanicalfatigue testing: 1 – command panel; 2 – rigid metallicframe; 3, 4 – system of bearings and levers for loadmultiplication on the sample; 5 – counter weight; 6 –engine assembly, reducer and arm for weight lifting; 7– weights; 8 – comparator; 9 – shell of the heatingchamber; 10 – sample made of memory shape alloy; 11– jigs; 12 – fixed die.

The dies used for traction are positions in ametallic chamber, foreseen with a transparentvisiting cover. The sample will be cyclic heatingand cooled in the range of temperatures 40–100°C, by means of an installation that blowswarm and cold air.

Figure 6. Metallic chamber where the jigs arepositioned, measurement system of temperature ofsample.

According to the weights attached to leverssystem, the sample will be stressed with a loadwhose size is directly proportional with the usedweight. This proportionality achieved by means ofthe levers system. For a weight of 5 kg, tractionforce on the sample ends will be of 500 kgf.

During experiments, the sample made ofmemory shape alloy will suffer an elongation,which will be determined with a comparator.

The parameters followed on command paneland a computer using the software named XMEMcan control the entire experimental process. Aswell, the number of heating-cooling cycles will berecorded by XMEM software.

Figure 7. Detail of XMEM software in function.

A very important problem of this installation isthe synchronization of the thermal cycle withmechanical cycles. Importance that consists in thefact that memory shape alloys need to determinethermo-mechanical fatigue during materialeducation. Material education realizes in double-effect memory shape materials.

The samples were also analyzed by means of adilatometer to observe modifications on thecharacteristic points and implicitly if ‘amnesia’phenomenon appears – if the alloy losses itsmemory shape effect.


The samples are subjected to a variable,chosen arbitrarily, thermal /mechanical fatiguecycles. After a certain number of cycles, thespecimen was analyzed dilatometric, followingcontraction changes during heating (EMF) andhow the temperature transformation martensite -austenite varies.

The Cu75.4Zn18.6Al5.85 alloy in quenching and3% range elongated presents the martensite –austenite transformation domain at 65.8°C and102.2°C temperatures. After 100 cycles, thedomain have moved at the 40,7°C and 99,5°Ctemperatures.

The maximum of contraction have a growthfrom 25 m to 75 m. At 6000thermal/mechanical fatigue cycles we have acontraction from 180 m.

0

20

40

60

80

100

120

0 1000 2000 3000 4000 5000 6000 7000 8000number of cycles [N]

max

imum

con

tract

ion

[m

]

maximum contractiontransformation point Mstransformation point Af

Figure 8. The maximum variation of contractionfunction the number of cycles of thermal / mechanicalstress.


324

After 12865 thermal fatigue/mechanical cycleson the sample surface, subjected to experimentaltests on the prototype installation, have appearedsome cracks (figures 9 and 10).

Experimental tests on this sample have beenstopped. If it had continued, these cracks could bemoved inside the sample, and finally they wouldhave broken it.

Following shows the appearance ofmicrocracks micrographs at 500X magnification.

Figure 9. Microcracks on surface sample fromCu75.4Zn18.6Al5.85 alloy, magnification 500 X.

Figure 10. Microcracks on surface sample fromCu75.4Zn18.6Al5.85 alloy, magnification 1000 X.

4. Conclusions

The applications of memory shape alloysimpose the determination of functioning periodfor pieces that work under the conditions of underload cyclic heating.

The determination of the characteristics for amemory shape alloy allows components producersto assure a certain guarantee on these products.

The achievement of the prototype installations,as well the creation of a standard of thermo-mechanical fatigue for a series of analyzed alloyswould be a solution for the producers of memoryshape components.

The development of thermal/mechanicalfatigue process is characterized by a diminution ofcontraction, and finally the destruction of thesample.

4. References

[1] Stefan, M., Vizureanu, P., Bejinariu, C.,Badarau, Gh., Manole, V., Studiulpropriet ilor termice ale materialelor,Editura Tehnopress, Ia i, 2008, ISBN 978-973-702-566-1.

[2] Stanciu S., Materiale cu memoria formei –metode de investiga ie si aplica ii in tehnica,2009, editura Universitas XXI.


325

STUDY REGARDING THE AUDIT OF MANAGEMENT PRINCIPLES

Costel Mironeasa1, Silvia Mironeasa2, Georgiana Gabriela Codin 3

University tefan cel Mare, Suceava, [email protected], [email protected]

Abstract: In this study we developed a check list and we applied a scale to measure theimplementation quality of the management principle, as set in the standard ISO 9000. The auditwas for the organizations that are included in small and medium organizations category, fromNorth Moldavia and Banat.

Keywords: principles, audit, discriminate analysis, standards.

1. Introduction

Management must fulfill quality criteria.Applying the principles of quality management tomanagement practice is the base transition to thetotal quality of applying the management functions[1]. Modern management systems must answer thechallenges linked to quality assurance, employeesafety, information security, environmentalprotection or social responsibilities and for somespecific organizations, to the insurance of foodsafety.

The term Total Quality Management (TQM)seems to be an universal remedy that can beapplied to all organization regardless of the profileprocesses which take place in the organization. Inpractice, this concept is transposed in a new culture[10] with values [11], principles and procedures[12]. The management principles can be found inthe TQM and are especially adopted by theorganizations which implement the requestsspecified in the standard 9001 [7].

The standard 9000 sets forward the eightprinciples regarding:

1. Customer focus2. Leadership3. Involvement of people / stakeholders4. Process approach5. System approach to management6. Continual improvement7. Factual approach to decision making8. Mutually beneficial supplier relationships

The implementation of the standards has aspurpose the management of the managementsystems, regarding the compliance of theseprinciples in order to create a culture oriented

towards the quality of the processes from thesystem [3]. The requirements expressed by thestandards SR EN ISO 9001 [9]; SR EN ISO 14001[5]; SR OHSAS 18001 [8]; SR EN ISO 27001 [6]etc. are implemented in order to answer twoimportant directions: ensuring satisfaction ofclients/interested parties and ensuring compliancewith the requirements of the standard. Requirementsspecified in table 1 reflect the managementprinciples from the analyzed standards.

The management systems are implemented inorganizations following the quality assurancecycle: plan, do, check and act. Client orientationand the interested parties involve a strategicapproach of all the processes that are part of thesystem. The evaluation of the functionality of thesystem can be obtained through monitoring,control and audit [4].

The management of the organization uses theaudit as one of the instruments which can helpevaluate the status of the audited entities. Audit, asan instrument used by management determines: ifthe management principles are complied to, if thereare established objectives for the processdevelopment, if there are performance criteriawhich establish targets for the development ofactivities which account as a process, which are theresults of the processes and if the resultscorrespond to the pre-established criteria, the levelof compliance of the requests to the objectives, etc.The audit has the great advantage that it allowsobtaining much more complete information thanthe administration of questionnaires to determinethe status of a system or process.

mailto:costel:@fim.usv.ro

mailto:silviam:@fia.usv.ro


326

Table 1. Management principles reflected in the requirements from standards ISO 9001; ISO 14001; ISO18001and ISO 27001Crt.no.

ManagementPrinciples ISO 9001 ISO 14001 /

OHSAS 18001 ISO 27001

1. Customer-FocusedOrganization

5.2; 7.2 (7.2.1; 7.2.2; 7.2.3); 7.3.6;7.5.2; 7.5.4; 8.2.1

4.3 (4.3.1; 4.3.2);4.4.3; 4.4.6

5.1;

2. Leadership 5.1; 5.3; 5.4 (5.4.1; 5.4.2); 5.5(5.5.1; 5.5.2; 5.5.3); 5.6 (5.6.2;5.6.3); 6.1; 6.2.2; 6.3; 6.4; 7.3.1;7.3.4; 7.4.1; 7.5 (7.5.1; 7.5.2;7.5.3); 8.2 (8.2.3; 8.2.4); 8.3; 8.4;8.5 (8.5.1; 8.5.2; 8.5.3)

4.2; 4.3.3; 4.4(4.4.1; 4.4.2;4.4.3; 4.4.6;4.4.7); 4.5 (4.5.1;4.5.2; 4.5.3); 4.6

4.2 (4.2.1; 4.2.2;4.2.3); 5.1; 5.2.1; 8.1(8.2; 8.3)

3. Involvement ofPeople

5.5 (5.5.1; 5.5.3); 6.2.2 4.4 (4.4.1; 4.4.2;4.4.3)

5.1; 5.2.2

4. Process Approach 02; 7.1; 7.5.2; 8.2.3; 8.4 4.5.1 02;5. System Approach to

Management02; 4.1; 5.1; 5.4.2; 5.5.2; 5.6(5.6.1; 5.6.2; 5.6.3); 6.1; 7.1;7.4.2; 8.1; 8.2 (8.2.1; 8.2.2; 8.2.3);8.4

Introduction; 1;4.2; 4.3 (4.3.1;4.3.2); 4.4 (4.4.1;4.4.2; 4.4.3); 4.6

4.2 (4.2.1; 4.2.2)

6. ContinualImprovement

1.1; 4.1; 5.1; 5.3; 5.6 (5.6.1; 5.6.2;5.6.3); 6.1; 8.1; 8.4; 8.5.1

4.1; 4.2; 4.3.3; 4.4(4.4.1; 4.4.2); 4.6

4.2.2; 7.2; 8.1; 8.2;8.3

7. Factual Approach toDecision Making

01; 4.1; 5.6.3; 6.2.2; 5.1; 5.3; 5.5(5.5.1; 5.5.3); 6.3; 6.4; 7.2.3;7.3.1; 7.4.2

4.4; 4.4 (4.4.1;4.4.3; 4.4.3;4.4.6); 4.6;

4.1; 5.2.1; 7.2

8. Mutually BeneficialSupplierRelationships

02; 5.1; 5.3; 5.5 (5.5.1; 5.5.3); 6.3;7.2 (7.2.2; 7.2.3); 7.3 (7.3.1;7.3.3); 7.4.2; 7.5.1; 8.2.1; 8.4

4.2; 4.4; 4.3.1;(4.4.1; 4.4.3;4.4.6; 4.4.7); 4.5.1

5.1;

2. MethodologyWe focused on two regions from Romania:

Banat (Timi oara and Re a counties) and North-Moldavia (Suceava and Boto ani counties). In total34 organizations have been audited from which 19in the Moldova area and 15 in the Banat area. Ineach organization all top managers and 82intermediary managers were audited: 46 in theMoldova area, 36 in the Banat area. The number ofmanagement systems implemented in theorganizations was different, between one systemand up to four systems. Quality managementsystems (QMS) were implemented in all theorganizations. Besides this system the followingsystems were also implemented: EnvironmentalManagement System (EMS); Occupational Healthand Safety Management Systems (OHSMS);Information Security Management System (ISMS).The audited management systems were found inthe following combinations: QMS + EMS; QMS +EMS + OHSMS; QMS + EMS + OHSMA +ISMS.

Information gathering was achieved by directinterview with each process owner found in theorganizational process of the organization. Theywere addressed questions which had the purpose toobtain answers referring the way in which themanagement principles are implemented.

3. Results and discussions

The evaluation of the way in whichmanagement applies the management principleswas achieved by following two directions: theevaluation of the way in which managementperceives the level of importance of themanagement principles and the evaluation of theway in which the management principles havebeen implemented. The implementation stage ofthe principles was done by the followingconsiderations of the auditor: met, partially met,and not applied.

The evaluation was done through statisticalmodeling and it was applied to the results in table 2in order to see to which degree are theredifferences between the way the organizations,from the analyzed regions, approach the


327

management principles. The data from table 2 wasadjusted according to the number of participants.The number of managers which answered wastaken into consideration and also the number ofqueries for each principle. The obtained data

represent the average of each considered criterion.Each principle was assigned a number of questionswhich were addressed to the top and executivemanagers.

Table 2. The resulting dataRegion Management Criteria P1 P2 P3 P4 P5 P6 P7 P8

Moldova Top M 0,789 0,880 0,833 0,805 0,861 0,900 0,861 0,860PM 0,114 0,082 0,105 0,128 0,096 0,074 0,090 0,079NM 0,096 0,038 0,061 0,068 0,043 0,026 0,049 0,061

Executive M 0,746 0,786 0,594 0,742 0,769 0,815 0,629 0,645PM 0,132 0,145 0,264 0,099 0,146 0,061 0,182 0,225NM 0,121 0,069 0,141 0,158 0,085 0,124 0,189 0,134

Banat T M 0,822 0,926 0,933 0,838 0,873 0,927 0,919 1,056PM 0,122 0,041 0,089 0,143 0,079 0,053 0,038 0,122NM 0,078 0,033 0,022 0,019 0,048 0,020 0,043 0,078

E M 0,343 0,386 0,389 0,349 0,364 0,386 0,383 0,440PM 0,051 0,017 0,037 0,060 0,033 0,022 0,016 0,051NM 0,032 0,014 0,009 0,008 0,020 0,008 0,018 0,032

No. of questions 12 18 6 7 11 10 14 6

The obtained data were processed according tothe multiple discriminator analysis method (). Thebasic technique of the analysis is to estimate theposition of an element towards a line which betterseparates two different population classes. Withthe current method, estimations can be made uponthe belonging of certain elements to one ormultiple categories. Application of thediscriminator analysis sought to determine aminimum number of variables that can express theseparation of behavior which managers havetowards the approach of management principles.The determination of the associated model followsthe calculation of the sums, the square sums andthe sum of products for the eight characteristics,

without taking into consideration the differencesbetween the categories of managers from the tworegions. The next step was to determine deviationsfrom their mean values of the square and productsums with the relation (1).

12

XXXXyy

12

1

12

1ji12

1

12

1jiji (1)

The obtained data was entered in a system witheight equations with eight unknowns in order todetermine the value of the coefficients ai, i=1...8(2).

y1y1 y2y1 y3y1 y4y1 y5y1 y6y1 y7y1 y8y1 a1 z1

(2)



y1y4 y2y4 y3y4 y4y4 y5y4 y6y4 y7y4 y8y4 × a4 = z4






328

The determined coefficients ai, i=1...8 offerinformation about the relative importance of eachmanagement principle as assessed by the twostudied groups. In order to determine theimportance of one or another characteristic thecoefficients of the associated model are compared.The calculated values were adjusted to their scoredispersion values, without taking into account theregion and type of manager. The values of thecoefficients are given in table 3. The coefficient

with the largest algebraic value will have thehighest degree of importance, in this case a2 =1,397.

Since the difference between the average scoreobtained by the managers from Moldova and Banatfor a management principle Pi, i=1...8, has thesmallest value for P8, this attribute has a lessimportant discrimination degree for the analyzedcategory. The difference between the averagegiven points is 0,037.

Table 3.Principle no. P1 P2 P3 P4 P5 P6 P7 P8Coefficients ai 1,997 3,83 1,870 2,276 -3,538 -3,781 2,745 -4,540Standard Deviations 0,311 0,3647 0,3269 0,3205 0,3442 0,3739 0,3355 0,3534Coefficients ai 0,6205 1,397 0,6113 0,7295 -1,218 -1,414 0,921 -1,6042Importance level 4 1 5 3 6 7 2 8Average difference 0,092 0,097 0,087 0,097 0,097 0,097 0,097 0,037

4. ConclusionsDiscriminator analysis provides us with

information on the level of importance which themanagement principle has. The use of the audit hasallowed obtaining results through which it can beestablished that the leadership principle is the mostimportant. The second principle, as level ofimportance, is linked to a management feature,Factual Approach to Decision Making. Thesefindings show that managers believe that in orderto implement a management system, the mostimportant elements are linked to management. Asurprising element is the location on the secondsmallest spot of the principle referring to ContinualImprovement. This is a sign of weakness of themanagement system which can be explained by alack of understanding of the benefits arising fromreducing the risk of non-conformities.

5. References

[1] Fausto, G., The golden integral qualityapproach: From management of quality toquality of management, Total QualityManagement, Jan 1999, Vol. 10 Issue 1, p 17-35.

[2] Gunaselaran, A., Enablers total qualitymanagement implementation inmanufacturing: a case study, Total QualityManagement, vol. 10, no. 7, 1999, 987-996.

[3] Irani, Z., Beskese, A., Love, P.E.D., Totalquality management and corporate culture:constructs of organisational excellence,Technovation, Volume 24, Issue 8, August2004, Pages 643-650

[4] ISO, 2002, ISO 19011, International Standard:Guidelines for Quality and/or EnvironmentalManagement Systems Auditing. InternationalOrganization for Standardization, Geneva,Switzerland.

[5] ISO, 2004, ISO 14001, Environmentalmanagement systems. Requirements withguidance for use

[6] ISO, 2005, ISO/CEI 27001, Informationtechnology. Security techniques. Informationsecurity management systems. Requirements

[7] ISO, 2005. ISO 9000, Quality managementsystem - Fundamentals and vocabulary

[8] ISO, 2007, OHSAS 18001, Occupationalhealth and safety management systems.Requirements

[9] ISO, 2008, ISO 9001, Quality managementsystems - Requirements.

[10] Matta, K., Davis, J., Mayer, R., Conlon, E.Research questions on the implementation oftotal quality management, Total QualityManagement, no. 7, 1996, 39-49.

[11] Stephan M. Wagner, S.M, Eggert, A.,Lindemann, E., Creating and appropriatingvalue in collaborative relationships, Journal ofBusiness Research, Volume 63, Issue 8, 2010,840-8


329

INTEGRATION OF SIX SIGMA AND LEAN PRODUCTION SYSTEM FORSERVICE INDUSTRY

Militaru Emil1

Abstract: Six Sigma is one of the most popular quality initiatives recently. Lean Production System is theworld famous production system developed and practiced by Toyota mobile company for a long time. It isbased on two concepts: “Just-In-Time” and “Jidohka”. Both are based on varying thinking to improve thebusiness process, enhance quality, production and competitive position. Besides, the integration of them isviewed as a new trend in the next management wave.

Keywords: six sigma, lean production system, service industry

1. INTRODUCTION

Nowadays, the development of industrialcore intrinsic technologies is more and moreimportant due to the dynamic competition in theglobal market. Companies have to keep andcontinuously upgrade their intrinsic technologiesin the professional field to gain the sustainablecompetitive advantage. However, they also haveto continuously upgrade their managementtechnologies, and keep sensitive to the latestissues as well as their integration with company’scurrent system. Otherwise, they still cannotsurvive in the market even though their intrinsictechnologies are advanced.

No matter how the management technologiesbe developed, they must emphasize theircontribution to business performance, customersatisfaction and continuous improvement of theproducts or services. Moreover, the integration ofdifferent systems is an important issue today andtomorrow.

This research will focus on “Six Sigma” and“Lean Production System” to discuss theirintegration based on the background and thoughtsmentioned above. Six Sigma is one of the mostpopular quality initiatives recently. LeanProduction System is the world famous

production system developed and practiced byToyota mobile company for a long time. It based

on two concepts: “Just-In-Time” and “Jidohka”.Both are based on the variation in thinking inorder to improve business process, enhancequality, production and competitive position.Besides, the integration of them is viewed as anew trend in the next management wave.

Moreover, regarding the industrycharacteristics, service industry is quite differentfrom manufacturing industry. Even though thereare more wastes and improvement opportunities,the application of Six Sigma, Lean ProductionSystem or their integration in service industry isquite few neither in literatures nor practice.

This research proposes the Lean Six Sigmaintegration model based on the research gap andthe practical need, and then adapt it for serviceindustry. The model is named as “Lean SixSigma for Service (LS3)” in this research. Itbalances the viewpoints of internal and externalcustomers, and gives consideration to the Leanspeed as well as Six Sigma high quality. Also,this research tries to contribute to theenhancement of management technologies.

2. Introduction to Six Sigma

Six Sigma is the major focus of many companiesfor its powerful breakthrough performancedemonstrated in GE, Motorola etc. recently. SixSigma can help companies to reduce cost, increaseprofits, keep current customers and create newcustomers. In brief, Six Sigma is a methodology


330

to reduce the variation of every process and theirinterfaces to achieve a very high quality level.

In statistical theory, six sigma is an idealtarget value, and expressed as: 6 . It means whenthe process or product we observed under anormal distribution, the probability of a specificattribute value shifts from the mean about positiveor negative six standard deviation would be 0.002part per million (ppm). Motorola company founda phenomenon that the process mean would shiftaround the center point of specifications in a long-term processing, and the shifting range would beabout positive or negative 1.5 standard deviationsfrom the center point of specifications.

Hence, Motorola company modified thestatistical meaning of six sigma. The definitioncan allow the sample mean shifts from the centerof the population, and the observed process orproduct would out lie the six sigma limits only 3.4times per million operations under the originalspecifications. In addition, the sigma performancecan also be expressed by “Defect Per MillionOperations (DPMO)” shown as Table 1.

Yield DPMO ShiftfromMean

PopularAge

6.68 % 933200 ± 030.9 % 690000 ± 169.2 % 308000 ± 2 1970s93.3 % 66800 ± 3 1980s99.4 % 6210 ± 4 Early

1990s99.98 % 320 ± 5 Mid 1990s99.9997

%3.4 ± 6 2000s

Table 1. DPMO and Sigma Performance

Six Sigma means the world leading qualitylevel. More and more companies understand touse Six Sigma to improve the process quality soas to achieve the business dramatic performance.This is because Six Sigma requires thequantitative measurements and analyses of thecore business processes as well as suppliers’involved processes.

Originally, Six Sigma methodology isapplied to manufacturing industries. However, theapplications of Six Sigma are no longer be limitedin manufacturing processes today. Keim (2001)

demonstrated Six Sigma is very suitable toimprove the service performance by two realcases. Paul (2001) pointed that the recent trends inSix Sigma are: emphasis on cycle time reduction,smaller business deployment, and integration withother initiatives.

As the Six Sigma market grows, so does theavailability of organizations to assist indeployment and integration. This availability oftechnical expertise allows smaller businessesrealistically consider Six Sigma deployment withminimal economic investment. Besides, due to thecentral concern of Six Sigma is to pursue thecustomer satisfaction and business performance,we can view Six Sigma a main structure whileintegrating with other initiatives. As for theintegrating initiatives such as Lean ProductionSystem, Total Quality Management or QualityCosts etc. depend on the different requirements ofeach company.

2.1 Introduction to Lean Production System

Lean Production System (also called ToyotaProduction System) is the world famousproduction system developed and practiced byToyota mobile company for a long time. It basedon two concepts: “Just-In-Time” and “Jidohka”.This kind of production system is very flexible tothe dynamic change of market demands, and LeanProduction System is established by many smallgroup improvement activities to eliminate allkinds of wastes in the business.

An important literature written by Spear andBowen (1999) published in Harvard BusinessReview pointed that, the Toyota ProductionSystem and the scientific method that underpins itwere not imposed on Toyota – they were not evenchosen consciously. The system grew naturallyout of the workings of the company over fivedecades. As a result, it has never been writtendown, and Toyota’s workers often are not able toarticulate it. That’s why it’s so hard for outsidersto grasp. In the article, Spear and Bowenattempted to lay out how Toyota’s system works.They tried to make explicit what is implicit.Finally, they described four principles – threerules of design, which show how Toyota sets upall its operations as experiments, and one rule ofimprovement, which describes how Toyotateaches the scientific method to workers at everylevel of the organization. It is these rules –and notthe specific practices and tools that people


331

observe during their plant visits – that in theiropinion form the essence of Toyota’s system.Hence the two authors called the rules as the DNAof the Toyota Production System.

These rules guided the design, operation, andimprovement of every activity, connection, andpathway for every product and service. The rulesare as follows:

Rule 1: All work shall be highly specified asto content, sequence, timing, and outcome.Rule 2: Every customer-supplier connectionmust be direct, and there must be anunambiguous yes-or-no way to send requestsand receive responses.Rule 3: The pathway for every product andservice must be simple and direct.Rule 4: Any improvement must be made inaccordance with the scientific method, underthe guidance of a teacher, at the lowestpossible level in the organization.

All the rules require that activities,connections, and flow paths have built-in tests tosignal problems automatically. It is the continualresponse to problems that makes this seeminglyrigid system so flexible and adaptable to changingcircumstances.

2.2 Four Characteristics of Service Industry

Recently, due to the economic and internationaltrading environmental change, the structures ofmany companies are also changed. The growth ofservice industries rapidly chases the growth ofmanufacturing industries. Especially for thecurrent situation in Taiwan, many factories aremoving to mainland China. Hence, the needs forservice industries to fill in the space of economicactivities become very huge. That’s why serviceindustries play an important role in the economicdevelopment recently.

This research concludes the fourcharacteristics of service industries based on theliteratures written by Kotler (1997), Regan (1963)and Zeithmal, Parasur& Berry (1985) as follows:

1. Intangibility: It means that servicescan be consumed and perceived, butthey cannot easy to be objectivemeasured like the manufacturedproducts. That’s why there is usually aperception gap between the service

provider and consumer.

2. Variability: It means that services aredelivered by people, so the servicequality may change depending ondifferent time, people and consumerperception. That is, the variability ofservices.

3. Perishability: Unlike the tangiblemanufactured products, services cannotbe inventoried. They are deliveredsimultaneously while the demands fromconsumers appear. Once the demandsdisappear, the services perish.

4. Inseparability: Since the delivery andconsumption of services almost bedone simultaneously. Hence theinteractions between servers andconsumers play an important role onthe evaluation of service quality.Consumers evaluate the service qualityon the moment of consuming theservice.

2.3 MODEL CONSTRUCTION

This research proposes an integration model ofSix Sigma and Lean Production System forservice industry called as “Lean Six Sigma forService (LS3)”. In practice, the first stage is to“Lead” the process improvement project byhearing the “Voice of Customer (VOC)”. Theproject identification and its scope must beclarified so as to serve the customers moreefficiently and effectively by the improvement.

The “Lead” stage provides the project team awell-defined scope of the problem they are faced.Hence the major mission of the “Study” stage is tomeasure the current status or level by quantitativedata, and then to analyze how the problem affectsthe process. By the collection of “Voice ofProcess (VOP)”, the project team can try toconverge the problem and begin to find out itsroot causes.

Moreover, no matter what the processindicators are, the project team has to well definethem first, and to explain the purpose as well asthe use of each indicator. Most important of all,the performance indicators’ definition and theirevaluation methods must be agreed and confirmedby the people involved. It’s very important todetail record the performance levels and action


332

results for the further enhancement of serviceprocesses.

After studying and analyzing the currentstatus of the service process targeted, the thirdstage is to draw up problem solvingcountermeasures. The countermeasures must betransformed to the “Voice of Server (VOS)” so asto “Smooth” the whole service process. It is hopedto reduce the defects and speed up the servicedelivery by the verification of performanceindicators. Therefore, the project team willpropose some education and training plans for thecoming countermeasures’ conduction.

The project team can measure if the projecttarget achieved or not by the proposedcountermeasures so as to continuously monitorand control the better results. At this time, theproject team has to “Sustain” the operatingstability of the service process. Therefore, thepurpose of this stage is to confirm the result, theeffectiveness of the countermeasures and if thereis any side effect. Once these things are confirmedfeasible, the project team can view the knowledgeand experiences as the base of knowledgemanagement and technology accumulation.Finally, the knowledge and experiences must bediffused and deployed throughout the organizationso as to be the “Voice of Business (VOB)”.

The LS3 operating model proposed by thisresearch shown as follows:

Figure 1. Structure of Implementing LS3

The key points and tools of implementingLS3 are concluded by this research and shown asTable 2. Moreover, the tools of LS3 are alsoshown as Figure 2.

Table 2. Key Points and Tools of ImplementingLS3

LS3 Activities Tools

Lead

Lead the processimprovementprojects byhearing the voiceof customer(VOC)-- Identify the

processimprovementproject

-- Define theprojectperformanceindicators

-- Select theproject teammembers

-- Accomplish theproject charterand jobassignments

-- Market survey-- Project charter-- Annual police

deployment-- Quality function

deployment-- Value stream

analysis

Study

Study andanalyze thecurrent status oftargeted processto get the voice ofprocess (VOP)-- Observe the

actual process,and measurethe baseline

-- Analyze thecollected datato understandthe presentsituation

-- Confirm theproblem andcritical-to-quality

-- Processmapping

-- Measurementsystem analysis

-- Motion andtime study

-- Multi-varianalysis

-- Cause andeffect matrix

-- Processcapabilityanalysis

-- Time valueanalysis


333

Smooth

Propose thecountermeasures,and transformthem into thevoice of server(VOS) to smooththe serviceprocess-- Draw up the

improvementcountermeasures

-- Implement thecountermeasures to acceleratethe servicedelivery

-- Confirm theresults byperformanceindicators

-- 5S-- Operation

balancing-- Rapid operation

switching-- Visual

management-- Eliminate,

combine,rearrangement,simplify

-- Processreengineering

-- Failure modeand effectanalysis

Sustain

Sustain andcontrol theproject results,and spread outthe organizationto be the Voice ofBusiness (VOB)-- Standardize the

effectivecountermeasures to sustain theresults

-- Continuouscontrol theimprovementlevel

-- Design the jobvalue ofemployees in

-- Control chart-- Check list-- Process

standardization-- Error proofing-- Education and

training

the serviceprocess

-- Knowledgediffusion andapplication

Figure 2. Tools of Implementing LS3

2.3. CONCLUSIONSDue to the limitation of practical resources, theLS3 model demonstration by a real case could notbe included in this research. Therefore, thisresearch used the questionnaire survey to verifythe theoretical logic and feasibility of LS3

structure. We interviewed several LeanProduction and Six Sigma experts andconsultants, and we expect to provide a base ofverification by their experiences and knowledge.Finally, we conclude the agreements andsuggestions of the experts as follows:

1. All the experts and consultants agreedthe theoretical structure of the proposedmodel by this research, and expressedthe “very much agree” level on thelogic, implementing steps and theircontents.

2. All the experts and consultantsexpressed the “very agree” level on theproposed model with PDCAmanagement cycle.

3. All the experts and consultantsexpressed at least the “agree” level onthe fitness for use of the proposed tools.


334

Some experts considered that thenames of some tools originally beapplied in manufacturing industryshould be changed to be wellunderstood in service industry. Inaddition, all the experts and consultantsconsidered that the tools mostly appliedin manufacturing industry should alsobe applied to service industry. This isbecause there are huge demands andexpectations of these tools in serviceindustry based on their experiences andobservations. Hence they deeply agreedthe proposed model by this researchcontains the tools for reference.

4. All the experts and consultantsexpressed the “very agree” level on theproposed model really integrated theprinciples of Six Sigma and LeanProduction System. Some expertssuggested we could emphasize theprinciples of Lean Production System alittle more.

5. All the experts and consultantsexpressed the “very agree” level on thefeasibility of applying this model to theservice industry.

6. All the experts and consultantsexpressed the “very agree” level on thepractical value of this proposed model,and they also expressed that they willrefer the structure to demonstrate it ifthere is any suitable opportunity in thenear future.

Therefore, i conclude the agreements andsuggestions by the experts as follows: all theexperts agreed the proposed model by thisresearch on the whole structure, implementingsteps and tools planned. They also expressed thepractical value and operational feasibility of thismodel is very high. Moreover, all the expertsagreed this model on the fitness for use in theservice industry, and they will refer this model todemonstrate it when there is any opportunity inthe near future.

References

[1] Breyfogle III, F.W., Cupello, J.M. andMeadows B. (2001), “Manging Six Sigma- APractice Guild to Understanding, Assessing,

and Implementing the Stagy that YieldsBottom-Line Success”, John Wiley & Sons,Inc., pp. 18-20;

[2] George, M.L. (2003), “Lean Six Sigma forService”, McGraw-Hill Inc., pp. 28-42;

[3] Jiang, Jui-Chin, Chen, K. H. and Wu, M. C.(2004), “Integration of Six Sigma and LeanProduction ”, 33rd International Conferenceon Computers and Industrial Engineering;

[4] Wu, M. C. (2003), “Integration of Six Sigmaand Lean Production ”, Master Thesis,Industrial Engineering Department, Chung-Yuan Christian University;

[5] Jiang, Jui-Chin, Fu, K. P. and Ku, S. F. (2002),“Construct the Integration Model of SixSigma and Lean Production ”, 8th AsiaPacific Quality Organization ConferenceProceedings, pp. 282-289;

[6] Jiang, Jui-Chin, Hsiao, F. Y. and Huang L. C.(2002), “Integration Model of Six Sigma forHospital” 8th Asia Pacific QualityOrganization Conference Proceedings;

[7] Keim, E. (2001), “Service Quality Six SigmaCase Studies”, Quality Congress.


335

OPTIMAL PERTURBATION FOR MIXED CONVECTION HEATTRANSFER IN RECTANGULAR CHANNEL

Rachid SEHAQUI1

Faculté des Sciences Ain Chock UFR de mécanique BP 5366 Maarif , Casablanca , Maroc Fax:(212)(22) 23-06-74 Tel : (212)(22) 23-06-80E-mail: [email protected]

Abstract: We studied the growth velocity perturbation at the entrance of a rectangular channel formixed convection heat transfer. The flow is bidimensionnel and the fluid is newtonian,incompressible obeying to Boussinesq approximations. Particular forms of entry velocity profileallowing a better heat transfer inside a channel for a given regime of flow is determined. Optimalcontrol method based on adjoint equations is used to obtain optimality condition.

Keywords: adjoint equations, convection mixte, optimal contro,l optimal perturbation

1. Abridged English versionThe study of the mixed convection in a horizontalchannel, uniformly or partially heated, plays asignificant role in various technological processes,such as, the chemical vapour deposition of solidlayers, cooling of microelectronic equipment's,enhancing thermal efficiency of compact heatexchangers or the calculation of the solar collectorenergy. Physical phenomena that appear in mixedconvection, in a horizontal channel are complexand depend on the Reynolds number Re, theRayleigh number Ra, the Prandtl number Pr, andthe geometry of the channel characterised by theaspect ratio B = l /h. Previous investigations haveessentially focused on the classification of thedifferent regimes of flow following the values ofdimensionless numbers [1] – [6].In the last decade, increased attention has beendevoted to the development of techniques capableof enhancing our ability to control the unsteadyflow in a wide variety of configurations such asengine inlets and nozzles, combustors,automobiles, aircraft, and marine vehicles. Gad–el-Hak [7] and Gad-el-Hak and Bushnell [8] providean excellent introduction and overview of variouscontrol methodologies. In practice, for a givencontrolling input, we choose an objectivefunctional to be maximised, and by a formalmaximisation procedure we obtain a system of

differential equations (state equations and adjointequations) and condition of optimality, whosesolution gives the optimal control. The results canthen be applied to develop mechanisms for control.This approach used in particular by Cathalifaudand Luchini [9] they attacked the problem ofcontrolling linear optimal perturbations in spatiallydeveloping boundary layer on flat and concaveplates by considering the wall receptivity. Theoptimal perturbations are the result of recentdevelopment in the field of hydrodynamic stability.For a given definition of disturbance energy, wecan calculate the input disturbance of the boundarylayer that maximises the output disturbanceenergy, i.e the optimal perturbation. For the spatialalgebraic growth of steady disturbances leading tobypass transition in the boundary layer over a flatsurface, this ‘optimal’ input disturbance wascalculated by Andersson [10].Many flows exhibit the following scenario: theflow stars out in laminar state which is instable;then a linear instability develops, and grows up tothe point where non-linear interactions finally leadto turbulence. In many applications, it might bevery interesting to delay this transition towardsturbulence and thus to maintain the flow laminar.The first objective, within this framework, of a

mailto:sehaqui:@hotmail.com


336

system of flow control will be to control heattransfer by entry velocity profile.This paper is organised as follows. In section 3, themethod of optimal control is presented. In section4 the method of adjoint equations and the optimalcontrol. In section 5, the results are presented anddiscussed

2. Introduction

Le transfert thermique par convection mixte àl’intérieur d’un canal rectangulaire présente unintérêt considérable en vue d’applicationstechnologiques telles que : dépôt chimique descouches minces, refroidissement des piècesélectroniques, augmentation de l’efficacité d’unéchangeur de chaleur, calcul de l’énergie d’uncapteur solaire et plusieurs autres domaines dessciences de l’ingénieur. Les phénomènes quiapparaissent dans le cas de la convection mixtesont complexes et dépendent de plusieursparamètres : le nombre de Reynolds Re, le nombrede Rayleigh Ra, le nombre de Prandtl Pr et durapport de forme B = l / h. Les conditions auxlimites tant celles fixant le profil de vitesse àl’entrée du tube que celles relatives à la géométriede la conduite ont une influence sur le transfertthermique. De nombreux travaux ont cherché àcorréler le transfert thermique aux conditions auxlimites thermiques imposées aux parois du tube ouà la forme géométrique de la conduite [1]-[6].Cette dernière décennie, une attention accrue a étéconsacrée au développement des techniquescapables de contrôler la stabilité des écoulementdans plusieurs configurations. Gad–el-Hak [7] etGad-el-Hak and Bushnell [8] fournissent uneexcellente introduction aux diverses méthodes decontrôle. En pratique pour contrôler lesécoulements d’entrée, on choisit une fonctionnelleobjective à maximiser, et par un procédé formel demaximisation nous obtenons un systèmed’équations différentielles (des équations d’état,des équations adjointes) et la conditiond’optimalité, dont la résolution permet d’obtenir lacommande optimal de contrôle. Cette approche aété en particulier employée par Cathalifaud etLuchini [9] ont traité le problème de contrôle desperturbations optimales linéaires dans le cas de lacouche limite pour une plaque plane et concave.Les développements récents dans le domaine de lastabilité hydrodynamique sont à l’origine del’intérêt accordé aux perturbations optimales. Pourune définition de l’énergie de perturbation donnée

on peut calculer la perturbation à l’entrée de lacouche limite qui permet de maximiser l’énergiede perturbation à la sortie, c’est à dire laperturbation optimale. Andersson [10] a étudié lacroissance des perturbations régulières permettantd’éviter la transition, dans le cas d’un couchelimite qui se développe au dessus d’une surfaceplane.On rappelle dans la section 3 la méthode ducontrôle optimum. Dans la section 4 la techniquedes équations adjointes. Dans la section 5 onexpose les résultats obtenus.

3. Méthode du contrôle optimum

L’idée du contrôle optimum est d’optimiser lafonctionnelle représentant la quantité physique quenous cherchons à contrôler. Nous allons décriredans ce paragraphe la méthode classiqued’optimisation basée sur les multiplicateurs deLagrange, Gunsburger [11]. Soit q la variabled’état et soit r la variable de contrôle ( que nousvoulons contrôler ). On suppose que q et r vérifientle problème suivant:

0,rqF (1)

Ou rq FFF , . L’équation (1) est appeléeéquation de contrainte du problème direct ousystème d’état. Fq est la contrainte relative à lavariable d’état q et Fr celle relative à lavariable de contrôle r. On définit lafonctionnelle gain rq, qui peut représenterpar exemple une distribution d’énergie, detempérature ou de frottements visqueux…etc.Le problème qui consiste en la recherche de qet r vérifiant les contraintes (1) et dont lafonctionnelle rq, est minimale est appelé« problème de minimisation ». L’approchegénérale de la minimisation est de former levecteur multiplicateurs de Lagrange. Lafonctionnelle Lagrangienne est ainsi définiepar :

,,,,, rqFrqrqL (2)

Ou ,.. représente le produit scalaire entre les

variables d’état (q,r) et les variables adjointes


337

rq , . Le problème d’optimisation consiste à

la détermination de ,, rq , pour ,r,qLstationnaire et vérifiant le système :

scontraintedeséquations0

adjointeséquations0

optimalité'dconditions0

rq

LLqLrL

(3)

4. Méthode des équations adjointes

4. 1. Problème direct

Dans le système de coordonnéescartésiennes(x*,y*), on considère l’écoulementplan d’un fluide newtonien incompressible deviscosité cinématique , dans un canalrectangulaire, de longueur l et de hauteur h. Latempérature du plan inférieur (y* = 0) est imposéele plan supérieur (y* = h) est adiabatique. Onintroduit les quantités adimensionnelles suivantes :

TTT

TUpp

UVV

UUU

hyy

lxx f*

,*,*,*,*,*2

000

où ,,, TpVU désignent respectivement lavitesse longitudinale, la vitesse transversale, lapression et le champ de températureadimensionnels et et,0U,T une différence detempérature de référence, la vitesse moyenne, lamasse volumique. Les équations de Navier-Stokesadimensionnelles s’écrivent sous la forme :

0VU yx (4)

UPUVUU yyxyx1Re (5)

TRaUPVVVU yyyyx 2RePrRe1

TPeTVTU yyyx1 (6) (7)

oùhURe 0 est le nombre Reynolds,

3ThgRa le nombre de Rayleigh, Pr

le nombre de Prandtl, la diffusivité thermique, gl’accélération de la pesanteur, le coefficientd’expansion thermique et Pe = Re Pr le nombre dePeclet. Cet écoulement principal fait l’objet d’uneperturbation (u,v,p, ) à l’entrée (e), et on étudiel’influence de cette perturbation à la sortie (s). Oncherche à maximiser un gain G défini à partir del’énergie de perturbation. On appelle perturbationoptimale la donnée des fonctions u(xe,y), v(xe,y)qui conduisent au gain maximum. Les équationslinéarisées pour la perturbation s’écrivent :

0vyxu (8)

upUuVUu yyxyyx1Rev

2RePrv

Re1vv RapVVuU yyyyxx

yyxyyx PeUTTuU 1v(9)

A ce système on associe les conditions aux limitessuivantes :u(x,0) = v(x,0) = (x,0) = 01yx ,

u(0,y) = ue(y) , v(0,y) = ve(y) , (0,y) = e(y).

4. 2. Problème adjointLa majoration du gain G(u,v,p, ) moyennant

les équations (8)-(11) conduit à l’étude de lafonctionnelle :

dydx

PeTTuyxd

RapVVuyxc

upUuVUuyxbuyxa

puGpudcbaL

f

e

x

x

yyyyxx

yyyyxx

yyxyyxyx1

01

121

1

VvU,

PrRe-vRevvU,

Re- v,v,

,,v,,,v,,,,,

Dans laquelle a, b, c et d sont les multiplicateurs deLagrange. La recherche de la solution optimale

revient à écrire 0LGrad pudcba ,,v,,,,, . Aprèsdes calculs basés sur des intégrations par parties,on peut écrire le problème adjoint en introduisant


338

le groupement A = a+bU. Les variables adjointesdu problème sont : A, b, c et d.

0cb yx (10)

0Re 1 TdVcbbVUbA xxyyyxx

0Re2 -1 TdccVcUbUUbA yyyyxyyy

0RePr

12 cRad

PedVdU yyyx

Aux équations (13)-(16) on associe les conditionsaux limites suivantes, issues du système (3) :

0,,0,0,,,0

01,0,1,0,1,0,

yAxdyxdyxcyAxcyAxbyxb

yxdyxdyxcyxcyxbyxb y

4. 3. Méthode de resolution

On superpose à l’écoulement de base uneperturbation, on associe à cette perturbation unefonctionnelle énergie appelée gain que l’oncherche à maximiser. Pour se faire on utilise uneformulation variationnelle permettant de mettre enévidence le problème adjoint assorti des conditionsaux limites adéquates. Le problème direct de laconvection mixte bidimensionnelle est de natureparabolique et on le résout de l’entrée à la sortiedans le sens de l’écoulement, c’est à dire des xcroissant. Le problème adjoint consiste en unsystème d’équations paraboliques dans le sensrétrograde résolu de la sortie à l’entrée ; lesvariables adjointes sont appelées variables desensibilité. La recherche de la perturbationoptimale, se fait par itérations successives eninitialisant le processus par une perturbation ue

0(y)donnée arbitrairement. A la énième itération , leproblème directe sera donné par :

0vny

nxu (11)

nyy

nxy

nny

nx upUuVUu 1Rev

nnyy

ny

nyx

nnx

RapVVuU 2RePrv

Re1vv

nyy

nxy

ny

nnx PeUTTuU 1v

Avec les conditions aux limites :u n(x,0) = v n(x,0)= n(x,0) = 01y

nx , dans le cas ou le gain

G est définit comme suit :

1

02

1

02

),(

),(

dyyxu

dyyxuG

e

s (12)

et la condition d’entrée est donnée par

1

0

21

21

0

21

1

2 dyy

dyy

yy

u

uAu

ns

ne

ne

ne

(13)

La résolution du problème direct permet à chaque

itération de déterminer yun

s , ce qui donne la

condition d’entrée pour le problème adjoint :

1

0

2

2

dyy

yy

uuAne

nsn

s

(14)

Le problème adjoint à l’itération (n) s’écrit :

0ny

nx cb (15)

0Re 1 TdVcbVbUbA xn

xnn

yyn

yxnn

x

0Re2 -1 TdccVUcbUUbA ynn

yyn

yn

xn

yynn

y

0RePr

12

nnyy

ny

nx cRad

PeVdUd (16)

Et la condition d’entrée écrite ci dessus. Une fois leproblème adjoint résolu à l’itération (n) laconnaissance de la fonction An calculée àl’entrée permet de passer à l’itérationsuivante :

1

0

2

21

0

2

1

2 dyy

dyy

yy

u

uAu

ns

ne

ne

ne


339

6. Résultats

-0,1

-0,05

0

0,05

0,1

0,15

0 0,167 0,333 0,5 0,667 0,833 1

YU(Y

)

Série2Série3Série4Série5Série6

iter = 1

iter = 7iter = 8

iter = 10iter = 11

Figure 1. Profil de vitesse optimum obtenu après plusieurs itérations

0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

600 1000 1500 2000 2500 3000 3500

Série2Courbe de stabilité

Re

G

Figure 2: Variation du gain en fonction du nombre de Reynolds


340

Figure 3: Variation du nombre de Nusselt local(a):profil de vitesse non perturb´e,(b):profil de vitesse optimum

Références bibliographiques

[1] J. N. N. Quaresma, R. M. Cotta, Exactsolutions for thermally developing tube flowwith variable wall heat flux, InternationalCommunications in Heat and Mass Transfer 41(1994) 729-742.

[2] Dae-Young Lee, Sang-Jin Park, SungTack Ro, Heat transfer by oscillating flow incircular pipe with a sinusoidal walltemperature distribution, Int. J. Heat MassTransfer 38 (1995) 2529-2537.

[3] P. Stefano, An analytical approach to fullydeveloped heating of laminar flows in circularpipes International Communications in Heatand Mass Transfer 22 (1995) 815-824.

[4] Tatsuo Nishimura, Naoya Kojima, Masstransfer enhancement in symmetric sinusoidalwavy-walled channel for pulsatile flow, Int. J.Heat Mass Transfer. 38 (1995) 1719-1731.

[5] G. Russ, H. Beer, Heat transfer and flowfield in a pipe with sinusoidal wavy surface - Inumerical investigation, Int. J. Heat MassTransfer 40 (1997) 1061 –1070.

[6] G. Russ, H. Beer, Heat transfer and flowfield in a pipe with sinusoidal wavy surface-IIExperimental investigation, int. J. Heat MassTransfer 40 (1997) 1071-1081.

[7] Gad-el-Hak, M., Flow control, AppliedMechanics Review, Vol 42, 10, (1989), 261-293.

[8] Gad-el-Hak, M., Bushnell, D. M., Separationcontrol: Review, Journal of Fluid Engineering,Vol 113, 1, (1991), 5-30.

[9] P. Cathalifaud, P. Luchini, Algebraic growth inboundary: optimal control by blowing andsuction at the wall. Eur. J. Mech. B- Fluids 19(2000) 469-490.

[10] P. Andersson, M. Berggren, D.Henningson, Optimal disturbances in boundarylayer,Physics of fluids 11 (1999) 134-150.

[11] Gunzburger M., Inverse design andoptimization methods : Lagrange multipliertechniques, Von Karman Institute for FluidDynamics, Lecture Series (1997).

TEHNOMUS - New Technologies and Products in Machine processing Technologies

341

STUDY ABOUT VIRTUAL AND ACTUAL MANUFACTURINGPROCESS WITH THE ROBOT

Traian Lucian SEVERIN1, Romeo IONESCU1,

1 tefan cel Mare” University of Suceava, Faculty of Mechanical Engineering, Mechatronics andManagement, University Street, No. 13, 720225, Suceava, Romania, [email protected]

[email protected]

Abstract: In this paper the authors aim to identify the possibility of broadening the use of industrialrobots for the processing of wood. In the robotics laboratory has attempted a stand where the robot willbe able to process the wood pieces with tool attached to the robot. We try to point out the steps forapplication development; our target was cutting process of wooden objects. Industrial applications couldbe very many and diverse. The robot programming was developed for working on multiple facets of apiece type cube it used the robot simulation processing. To realize a comparative study of robotmanufacturing with three axes CNC machines was used NX cam software which allowed us to simulatethe same processing.

Keywords: robot, milling, processing, manufacturing, simulation,

INTRODUCTION

In this paper the authors aim to identify thepossibility of broadening the use of industrial robots forthe processing of wood. The wood industry can bedivided into four main areas: the furniture industry, theconstruction-related sector, the wooden materialsindustry, and other sectors. The jobs that robots arecalled on to perform in the wood industry includepainting, handling, grading, and repairing of woodenparts and products. In the last years, the flat-packfurniture and construction materials industries areentering a new era of robotized, flexible manufacturing.The manufacture wood technology with flexibletechnology that can quickly and accurately adapt tovarying surfaces and consistency of materials.

The requirements on personnel and machines arejust as great in high-precision machining applications inthe manufacture of large work pieces. Now we can seehow an industrial robot in wood industry offersincreased productivity at a lower cost.

We can offer an example polishing and buffing ofwood products is an interesting application and manyrobots are integrated for wood products. In figure 1 wecan see how a major guitar manufacturer had integrateda work cell with robot to polish and buff guitar bodies.In that application, the robot picks up the guitar bodyand presents it to sanding and buffing wheels [11].

Material handling of wood products is anotherimportant job that robotics accomplish for end-users. Inmany workshops where working with wood, in thesecondary operations of wood product processing for

cabinets, windows and doors manufacturing, roboticsare taking pre-cut parts and assembling them orpresenting them to joining machines.

Figure 1. Details of robotics guitar buffing [11].

End-users ask why they should invest in robotics tohelp manufacture wood products when people havebeen performing these tasks. Most times the answercould be: tedium suffered by those in the woodworkingindustry is one reason that robots are used but is notonly that. The robots can increase productivity, qualityand efficiency of wood processing objects.

In the robotics laboratory has attempted a standwhere the robot will be able to process the wood pieceswith tool attached to the robot. We try to point out thesteps for application development. Our target wascutting process of wooden objects. Industrialapplications could be very many and diverse.


mailto:romtit:@fim.usv.ro


342

EXPERIMENTAL

The choice of the parts to be processing is veryimportant because choosing a piece with a quitecomplex geometry can lead to the demonstration ofall robot´s real skills. For the basic study and thework one can choose first a relatively simple part,in our case a cube, on whose sides variousprocessing can be done. This part allows us todemonstrate the fact that the robot is not limited byits freedom degrees, but by the clamping device.This means that if we could clamp the cube on onesole corner then we could make processing on allthe cube´s sides without any other movement ofthe device.

Next step consists of the study of thepositioning and clamping device of the part. Thisdevice has to fulfill the following conditions: to berobust in order to prevent vibrations; to beuniversal in order to allow the clamping of a widevariety of parts; to be positioned in an optimumposition (height, distance) from the robot to alloweasy processing on all surfaces; to allow a simplepositioning, a good screw on with a reducednumber of operations; to ensure the elimination ofthe waste resulted in the processing operation.

The choice of the milling device, itsmodification and clamping to the robot form thenext stage of the study within the modeling andsimulation of processing with the help of the robot.The mill must be mounted on the robot so toprevent any interference or collision with therobot´s parts or with the part´s clamping device inthe moment of processing and also must be rigidlyclamped to ensure a quality manufacturing.

After the study of all these problems, we canstart the actual simulation of the robot processingand we can make its program.

For the stand, parts as well as the splintingdevice the software CAD/CAM NX was used,software which offers a wide range of programs inone integrated solution, to allow the user to use thelasted technologies in terms of used tools andprocessing processes. NX CAM supports the latestgeneration of multiple function tools, includingmilling, drilling and turning on a 5 axis machine.NX allows a wide range of flexible processing on 5axis machines and with tool control on variousaxes.

NX CAM is completely related to the other NXsolutions so that the NC programmers cannotdirectly access technical design drawings,assemblies and tools in a processing environment.Through modeling by manufacturing´sassociatively, the changes of the model areautomatically conveyed into processing operations.Programmers and mechanical engineers can workwith modeling parts, can create and mount devices,can develop paths and can even mould wholemachines through a 3D simulation of theprocessing in this work environment.

For the part to be processed we will start fromthe simple cube type part on whose sides varioustypes of transformations can be done. As it can beseen in figure 2, on each side of the cube thedifferent forms from figure 2 can be milled, somesimpler, some more complex. For thetridimensional parts the 3D, NX CAM softwarewas used.

After a favorable result, more complex shapescan be tried, with complex material removal, likeconic wholes, uneven surfaces, with variableradius. It is to be mentioned that now days thereare specialized software which convert the courseof the tool necessary for the surface processingfrom the design software into the robot´s specificlanguage.

Figure2. Cube type part with various manufacturing


343

Figure 3 processing forms to be done on the cube´s sides.

In order to be able to compare the robotprocessing with classical methods of processingfrom today, we will use the same parallelepipedused for the presentation of the CAM software,cube which actually exists.

For the processing device we had to take intoconsideration the already existing stand from therobotics lab. Making a new independent standwhich might be fixed on the floor could lead to anoverload of the robot´s space of work and thereforeto a reduction in the operation space necessary forthe human factor.

Figure 4 – Securing device of the semi-finishedattachable to the already existing stand from therobotics lab. a) real b ) virtually designed in NX

As for the choice and adaptation of the millingdevice to perform the task, we started from abought hand mill, specially designed to mill

wholes, slots, edges as well as wooden and plasticjoints.

The specifications of this milling device make itgood for the work parameters necessary formanufacturing: power 1020 Watts, tension /frequency: 230 V~50 Hz, weight: 3 Kg,revolutions: 11500-34000 rpm, mill diameter: 8mm, adjustment: sliding rods, revolutionsadjustment.

Figure 5. Milling machine STERN ER 1020

For this mill its modification was considered inorder to be adapted and attached to the robot. Afterseveral meetings a certain shape and a clampingsystem were adopted, even if this meant thecancellation of certain mill elements.( figure 6).

Figure 6 – Milling device attachable to the robot Kuka KR125


344

For the simulation of the robot application thespecial software Kuka SimPro was used, fact thatallows the off line development and programmingof robot applications on Kuka robots also offeringthe possibility to overlap production times withtraining times for the new processing tasks, a bettersolution examination and the possibility ofimproving robots´functioning programmes withoutbeing pressed by time. [8].

For the virtual flexible cell the robot´s specificparameters were taken into consideration and alsothe clamping device position and the orientation ofthe semifinished, meaning an external coordinationsystem according to which the robot will beprogrammed. Figure 7

After programming the robot from this virtualflexible cell it was noticed that the milling devicecan perform any type of manufacturing, within thelimit imposed by the robot. In figure 7 one can seethe robot´s positioning during the processing ofeach proposed shape on each of the cube´s sides.

During simulation various aspects had beenfollowed: accuracy of manufacturing, lack ofinterference and / or collisions between variousmoving parts, avoidance of impossible to reachpositions from the robot, taking notes andobservations in order to modify and resize someelements from the final assembly in order toimprove the processing process, obtaining a

program which can be directly transferred onto therobot for real manufacturing.

Figure 7. Virtual flexible cell manufactured with theKukaSimPro software.

For the simulation of a classic manufacturing,NX CAM software was used. It is to beremembered that in the simulation we will use a 3axis numeric command machine, which means thateach side of the part to be milled will be processedseparately.

Figure 8. Simulation of form milling on cube´s sides

Figure 9 Simulation of classic processing on a CNC machine with the help of NX software.


345

In figure 10 one can see that the program is ableto give also the time for the processing of the entireassembly, or for each manufacturing. It is also tobe considered that the part must be repositioned fora new manufacturing, fact that introduces newtimes for each manufacturing

Figure 10. processing times obtained during classicsimulation with NX software.

Below we present the programming way of theKuka KR 125 robot. In order to program theprocessing of the intended shapes on the cube´ssides we followed the following steps: we used theNX software in which we traced with a dotted linethe course to be followed by the tool; to obtain thecoordinates of the points which give us the coursedescribed by the robot´s tool we used as startingpoint the point given by the intersection of thediagonals.

Therefore, an example of points to be obtainedis the following:

Palpated coordinates for the center of the sideare:

x =16; y = 69,5; z = 39,5 [1]

Manufacturing is done in the positive directionof the axis X, so the table below will not show thiscoordinate.

Table 1.Coordinates for U shape manufacturing

Coordinatescontrasted to thecenter of the side

Coordinatescontrasted to base

originNr.crt.

y z y z

1 45 30 114.5 69.52 45 -2 114.5 37.53 35 -12 104.5 27,5

4 25 -2 94,5 37,55 25 30 94,5 69,5

Therefore, for the processing of the „U” shape 5points are used, plus the points entering and exitingthe material. A linear movement will be usedbetween points 1 and 2, respectively 4 and 5 andbetween 3 and 5 a circular movement will be usedusing point 4 as intermediary.

Figure 11. Design of „USV” processing through points

Figure 12. Extract from „U” shape processingprogram.

As we can see from picture 11 the robot startsfrom the position „home”, in position P1, then goestraces a straight line and enters the material,performs its first linear movement, then thecircular movement, followed by the second part ofthe „U”, and at the point P7 it finishes theprocessing and comes out of the material.

This is the procedure which was followed forthe processing on all possible sides of this cubewithin the robotics lab.


346

a)

b)

c)Figura 13. processing of the other three sides

The figure below shows some aspects of thecube processing with the help of the KukaKR125 robot.

a)

b)

c)Fig, 14. processing examples on cube sides

The times obtained are similar to the timesobtained in the simulation with the CAM software:

- Loading a program: 32 seconds;- Star program: 2,53 minutes;


347

- Rectangular program: 6, 10 minutes;- Circle program: 5, 50 minutes;- USV program: 2,40 minutes.

It is to remember that having only one clampingdevice, the times for placing the cube in the workposition is eliminated fact which leads to a betterprocessing time.

CONCLUSIONS

After the simulation with the specializedsoftware the following have been noticed:Table 2. Comparison between the two processingmethods

Characteristic With KUKAKr 125

With 3 axisCNC

Work speedMaximum (of

the millingmachine)

Maximum

Number ofclampings 1 4

Cutting depth 3 mm 3 mmStrength Medium Maximum

Wotking plan Any plan Worksvertically

Analyzing table 2 we can draw the followingconclusions:

The CNC machine is sturdier than therobot, although at the level of the processingdone with the robot we are not talking aboutsturdiness as if the mill is properly anchored wewill not get vibrations.

In order to completely process theparallelepiped part the robot will do this in oneclamping, because the degrees of freedom allowsit. For the 3 axis processing device a newpositioning for each side is necessary, fact whichleads to a longer processing time.

According to the type of mill, on the robotwe can adopt a maximum speed of the millingdevice or we can mount a broach with very highmilling speeds on the robot while the 3 axismachine is limited to the maximum speedimposed to the machine.

The robot, because of the big number offreedom degrees can work on several plans,sidelong or on complex spatial courses, while the3 axis processing device is limited to the verticalplan, and therefore in processing the part.

If we took into consideration a 6 axisprocessing device then the cost price would be

much higher than of a robot which has the sameamount of work, so it would not be profitable.

The processing device has a protectedvolume, so that all waste from processing is notthrown away, and the robot must be added avacuum cleaner for the dust and for theremoved material. Otherwise, it can work inspecial arranged rooms, where the removedmaterial does not harm the environment.

References[1]. Ionescu R., Semenciuc, D., Robo i industriali,

Principii de baz si aplica ii, Editura OID-ICM,Bucure ti, 1996.

[2]. Ionescu, R., Amarandei, R., Robo i Industriali.Programe de simulare, Universitatea „ tefan celMare” Suceava, 2003.

[3]. SEVERIN Traian Lucian, IONESCU Romeo„Study about the opportunity in robotic drilling”,in volume of The 15th international conferenceTEHNOMUS NEW TECHNOLOGIES ANDPRODUCTS IN MACHINE processingTECHNOLOGIES”, Suceava May 8-9 2009,pag.195-200, ISSN 1224-029x.

[4]. Romeo IONESCU, Traian SEVERIN, ViorelPETRARIU, Dumitru AMARANDEI “Studyabout machining with industrial robots help”, The4th International Conference ROBOTICS’08,Brasov, 13-14 noiembrie 2008.

[5]. Ionescu, R., Severin, T., Petrariu, V., Amarandei,R., Study about CAD/CAM utilization in pieces’spatial processing with robots help. Iasi, 2008

[6]. IONESCU R., The initial training schemes of theengineers between Rumanian industrial reality andEuropean teaching, 3rd Balkan Region Conferenceon Engineering Education 2005 LBUS, Sibiu,Romania, 12 - 15 September, 2005

[7]. SEVERIN Traian Lucian, IONESCU Romeo„Aplica ii robotizate dezvoltate în laborator cuperspectiva de extindere în industrie”, în volumulcelei de-a XIV Conferin e tiin ifice cu participareinterna ional „Tehnologii i produse noi înconstruc ia de ma ini” Tehnomus XIV, Suceava 4-5 mai 2007,pag. 455-460, ISSN 1224-029x.

[8]. Proiect licenta, David Voiculescu, Prelucrareaunor piese din industria constructoare de ma ini cuajutorul robo ilor industriali,FIMMM, Suceava.

[9]. K u k a GmbH, Programarea robotului KUKAKR125, KUKA Augsburg, 2005

[10]. Derek K., Rapid Traverse Technology andTrends Spotted By The Editors of ModernMachine Sho http://www.mmsonline.com/.

[11]. Brumsson B., , Is the Wood Industry Ready forRobotics? Robotics online, Posted 11/14/2007,http://www.robotics.org/content-detail.cfm/Industrial-Robotics-News/.

http://www.mmsonline.com/.

http://www.robotics.org/content-


348


349

CONSIDERATION OF REFLECTION COEFFICIENT AS A FUNCTION OFFREQUENCY FOR WEDGES OF DIFFERENT MATERIALS

Liliana Petre-Cainiceanu, Constantin D.Stanescu, Tudor Burlan-Rotar

Politehnica University,. Bucuresti

Abstract: The wedges made of foamed poly-urethane-ether could withstand large de- formationswithout any damage. They would show a considerable extension before breaking, but a very low stiffnessmade additional support of the wedge tips necessary for horizontal mounting.

1.INTRODUCTION

The wedges made of foamed poly-urethane-ethercould withstand large deformations without anydamage. They would show a considerableextension before breaking, but a very low stiffnessmade additional support of the wedge tipsnecessary for horizontal mounting.Figures. 1 and 2 show the reflection coefficient asa function of frequency for the different materials.On the basis of these investigations four differentmanufacturers were invited to give in tenders forthe supply and mounting of wedges in the twoplanned anechoic chambers, as the opinion wasthat the price should also be taken into account inthe total considerations.

Figure1 The reflection coefficient as a function offrequency for wedges of different materials

Figure2. The reflection coefficient as a function offrequency for wedges of different materials.

After a thorough testing of wedges from the twolowest quotations, Sillan wedges of specificdensity 120 kg/m', manufactured by GrOnzweigund Hart-mann, Ludwigshafen (Germany) werechosen as the best alternative.

Figure 3 shows the reflection coefficient as afunction of frequency for the wedges chosen.

The measurements at frequencies above 250Hz are carried out in a smaller tube than thatshown in Fig. 3. Both whole wedges and parts ofwedges were investigated.

In order to obtain an estimate of the maximumdeparture from a free field to be expected in ananechoic room lined with these wedges, a digitalcomputer was used to calculate the sound field fora point source in rooms of different size and withdifferent values for the reflection coefficients ofthe walls. The calculations were made with theassumption that the sound waves were reflected inaccordance with Snell's law at the outer walls of


350

the room and that phase coincidence of thereflected sound existed at the measuring point.

Figure3. The reflection coefficient as a function offrequency for wedges used in the anechoic rooms

Calculations were carried out for differentpositions of the sound source. With theseassumptions and a reflection coefficient of 0.1 oneshould have a departure from free field ofmaximum ± 1 dB at 2 m and ± 3 dB at 11 m in aroom of free dimensions of 12 m X 10 m X 9 m.

3.Sound insulationAs the anechoic rooms were also intended for

a number of psycho-acoustical measurements, itwas desirable to have a background noise level ofabout 10 dB below the threshold of hearing atfrequencies above 200 Hz, whereas forfrequencies below 200 Hz a level lower than 10dB re 2 X 10-5 N/m' was required.

For extraordinary outdoor sound levels, suchas those caused by jet aircraft at low altitudes,slightly higher background levels would beacceptable.

It would, however, be an advantage if normalacoustical measurements were not appreciablydisturbed even in these cases.

These requirements necessitate a wallconstruction, which has a transmission loss ofabout 80 dB at 200 Hz, as a background level of60-70 dB re 2 X 10-5 N/m2 must be expectedoutdoors during working hours. In order toachieve this, a double wall construction must beused where the inner part, the anechoic roomitself, has no stiff mechanical connection to theouter building.

The outer walls of the building consist ofbrickwork and reinforced concrete with a layer ofheat insulation between, total thickness 30 cm.

The roof is made of 20 ern reinforced concretewith heat insulation. The inner boxes, with theanechoic rooms, have walls, floor and roof madeaf 40 ern reinforced concrete. The space betweenthe inner and outer walls is 1.1 m and acousticallydamped by covering both the inner side of theouter shell as well as the outer side of the innershell with 5 cm wood wool cement slabs. Also theinside of the outer roof is covered with thismaterial.

The large and the small room are placed on 24and 4 rubber vibration isolators respectively, seeFig. 4 .

Figure 4. Vibration isolators for the large anechoicroom.

Each vibration isolator is loaded by about 50tons, and a resonance frequency of about 7 Hzwas intended for the rooms placed on the rubberpads. The vibration isolators, made from amixture of natural and artificial rubber (hardness50° shore) are protected by a layer of neopreneand placed so that they can be inspected andreplaced if necessary.

4.DoorsThe doors are important parts of the constructionwith regard to lining and sound insulation. For thelarge anechoic room the entrance, which is 2.5 mX 2.7 m, is closed by two door sections travellingon rails and made up from three layers of steel (10+ 5 + 5 m) with concrete in between and coveredwith wedges. When the doors are closed thewedge lining of the room is thus continuous. Forpractical reasons there is also a hinged doorcovered with mineral wool, which can replace oneof the door sections for less criticalmeasurements. The opening in the outer wall isclosed by an air tight steel sliding door and the


351

corridor outside the door has no windows and isheavily damped.The 2.2 m X 1.6 m entrance to the small anechoicchamber is closed by two hinged doors made fromsteel plate and concrete. One of the doors is linedwith wedges. In front of the other, which is linedwith a 5 cm layer of Sillan, there is a light, wedgelined, sliding door which can be slided into thecorner of the anechoic room where there are nowedges, but only a 5 cm Sillan lining. The wedgelining of the room is thus continuous when thedoors are closed, and at the same time theconstruction is simple and saves space.

5.FloorsThe floors of the anechoic rooms should notinfluence their acoustical properties appreciably.Wire netting was decided upon, which has beenused for many anechoic rooms with satisfactoryresults. The mesh size was chosen to be 50 mmusing 2 mm diameter steel string for the largeroom and 3.5 mm for the small room. The smallerwire thickness in the large room makes itnecessary for people to use large overshoes withwide rubber sales when walking around in theroom, in order to protect the strings and todistribute the weight over a larger area.In the smaller anechoic room, which is used partlyby the students, the influence of a thicker wire hasbeen accepted, in order to avoid the use of extrashoes. The wires are fixed to adjustable tensiongrips situated on the outside of the inner concretebox. Each wire is pre-stressed by 200 kg, so thatthe deflection with a person (75 kg) situated at thecenter of the floor is about 1 cm.The maximum load for each netting is 1,000 kg,however it must be distributed in such a way thatthe load per meter of the periphery of the loaddoes not exceed 100 kg.The wire nettings are electrically isolated from thebuilding. This is done in order to reduce the riskof electric shocks from measuring instruments andto make it possible to fix the potential of the floorat any desired level independent of that of thesurroundings.The nettings are placed at the same level as thefloors of the outside rooms and the central planesof the anechoic rooms, where most of themeasurements will be taken, are situated about 1.5m above the nettings. Parts of the rooms whichare below the nettings can be reached throughremovable gratings in front of the doors.

A fine mesh perion netting is placed below thesteel nettings, in order to catch small itemsaccidentally dropped.

6.LightingAs the linings of the rooms are highly insulatingwith respect to heat, the lighting must give aminimum of heat radiation. At the same time it isnecessary to have quite a high light intensity inorder to be able to work with small microphones,hearing aids etc. As the wall lining is also verylight absorbing and as no reflectors can betolerated around the lamps, which must also beplaced as far away as possible from the measuringarea, the lighting installation must have thehighest possible conversion efficiency.The main illumination comes from 200 Wmercury vapour lamps, 12 in the large room and 2in the small room. As these lamps ignite slowlyand are probably not completely noiseless, thereare also a set of ordinary 200 W incandescentlamps which may be used separately, 8 in thelarge room and 2 in the small room. It is arelatively simple matter to remove the lamps ifthis should be necessary for acoustical reasons.

7.VentilationThe ventilation system is dimensioned with aview to obtaining a suitable rate of change of airin the rooms, and at the Same time to avoid anyappreciable increase of background noise levelwith the system operating. In both rooms theventilation system consists of a row of air inletsalong the bottom of the walls of the room andoutlets along the top of the opposite wall. Thefans are placed in the basement outside theanechoic rooms, and on its way into the rooms theair passes through two long, heavily dampedconcrete ducts, which are connected by flexibletubes. The exhaust air is similarlytaken through two concrete ducts connected byflexible tubes to the fans with outlets into openair.For the large anechoic room the ducts are 16 mand 10 m long with internal cross-section 70 cm X80 cm. The thickness of the concrete wall is 10cm. The acoustical attenuation is obtained bylining one of the inner walls with 50 cm mineralwool, which is divided into sections by thin metalplates to avoid sound transmission through thematerial along the ducts.The ducts for the small anechoic room are 7 mand 5 m long, and their concrete walls are 10 cm


352

thick. The internal dimensions are 45 cm X 60 cmand the sound absorbing lining consists of a 20 cmlayer of mineral wool along one wall, divided intosections as that for the large room.The inlet and outlet nozzles in the anechoic roomsare made of sheet metal and to avoid soundradiation from possible vibrations in the nozzles,these are covered externally with a 5 cm layer ofmineral wool. The inlets and outlets in the largeroom have the dimensions 6 cm X 100 cm and inthe small room 6 cm X 40 cm.Also the ducts leading to the open air inlets andoutlets are heavily damped so that noise from thefans will not disturb measurements in roomssituated on the same side of the buildings as theseopenings.The air can be changed 4 to 5 times an hour in thelarge room and 12 to 17 times per hour in thesmall room. The air speed at the inlets and outletsinto the rooms is about 2.9 m/sec. Starting and stopping of the ventilationmachinery and control of the air temperature isconducted by control knobs outside the doors foreach room. It is not, however, possible to obtain alower air temperature than that of the outside airas there is no refrigeration system. A refrigerationsystem was discussed during the planning stages,but due to the particular insulating properties ofthe wedges, which made calculations difficult itwas decided to leave space for a refrigeration unitbut to delay the procurement and installation untilthe room had been used for some time.Experience from 1966-67 does not indicate anyneed for refrigeration of the air.

8. Conclusion

The measurements carried out show that bothanechoic rooms can be used even for veryexacting measurements, the data for the largeroom probably representing something near thelimit of what one can achieve today.From the background noise measurements it isseen that it has been possible to attenuate thenoise from the ventilation system to such a degreethat the ventilators may be used even duringpsycho-acoustical tests.The insulation against noise in the vicinity of thebuildings is extremely good, in as much as evenwith sound pressure levels of the order of 110 dBre 2 X 10-5 N/m' the sound pressure inside theroom just reaches the threshold of hearing. It istherefore possible to carry out a large number of

normal acoustical measurements in these roomseven under extreme outside conditions.

Bibliography

[1]. Rivin, A. N.: An Anechoic Chamber forAcoustical Measurements.

[2]. Soviet Physics-Acoustics, 7, (2002), No.3.[3]. Wiuff, A E.: Eksamensprojekt 2005 (Master

Thesis), DTH, Copenhagen,[4]. 2006 (in Danish).[5]. Nissen, Preben: Bending Moments in

Members of Tension and Stressed WiresSubject

[6]. to Concentrated Transversal Loads.Bygningstekniske Med- delelser, 38 (1997), nr.1 (in Danish with an English Summary).

Date post:	13-Apr-2015
Category:	Documents
Upload:	botez-marta
View:	103 times
Download:	2 times

Cuprins Tehnomus XVI

Documents