InFocus Issue 01 | 2007 Brakes Analysis of operational deflection shapes at Continental Automotive Systems leads to silent brake designs Page 4 Experimental Modal Analysis Basic tutorial on this important CAD process, essential for rapid automotive development Pull-out Interior Noise Improving damping material performance while reducing material usage Page 10 Automotive Vibration Testing: Finding Solutions for Development and Production Page 1 Optical Measurement Solutions
Transcript
InFocusIssue 01 | 2007
BrakesAnalysis of operational deflection shapes
at Continental Automotive Systems leads to silent brake designs Page 4
Experimental Modal Analysis
Basic tutorial on this important CAD process,essential for rapid automotive development
Pull-out
Interior NoiseImproving damping material
performance while reducing material usage
Page 10
Automotive Vibration Testing: Finding Solutions for Development and Production Page 1
Optical Measurement Solutions
Editorial
Page II
Polytec News
Page III
Vibration testing in automotivedevelopment and production
Page 1
Designing silent brakes
Page 4
Structure-borne sound intensityderived from laser scanningmeasurements
Page 6
Development of electro-hydraulicsteering systems
Page 8
Polytec Tutorial: ExperimentalModal Analysis
Pull-Out
MEMS sensor quality control andsurface topography in automotiveproduction applications
Page 9
Improving damping materialperformance
Page 10
Dynamic valve train analysis
Page 12
Transmission behavior of a dualmass flywheel
Page 14
Modal analysis of turbocharger compressor wheels
Page 16
Product News
Page VI
Events
Page VIII
II
Editorial
Eric WinklerVice President Optical Measurement Systems
Dr. Helmut SelbachManaging DirectorPolytec GmbH
Dear Reader,Did you notice that our popular magazine has both a new face and a newname: InFocus - Optical Measurement Solutions? The new name was chosento represent the whole range of our high-grade optical measurement systemsdesigned to meet a growing number of non-contact, customer-drivenmeasurement applications. To further emphasize this shift towards a broader application of optical sensing, we have renamed our “LaserMeasurement Systems” business unit to “Optical Measurement Systems”.
Our considerable experience in developing and manufacturing laser-based vibration and velocity measurement technology has been expanded toinclude optical sensor technologies that enable our customers to measuresurface topography using white light interferometry. In addition, we aredesigning and building spectrometers for production testing of materials and goods. These spectrometers are currently distributed in Europe only; but, will soon be available worldwide as the market applications grow.
As you can see, there is a lot going on at Polytec. Continue reading to findmany interesting contributions from our customers, including a specialfeature about automotive development and production. Pay special attentionto the pull-out tutorial about experimental modal analysis and to the manyinnovative products from Optical Measurement Systems.
Keep up-to-date and have fun reading!
Dr. Helmut Selbach Eric Winkler
Content
III
Polytec Vibrometer Users Awarded
2006 DGAQS Prize In recognition of their successful work on automatic testbenches for production control, two Polytec customers,Mr. Pfichner (then PARI GmbH) and Mr. Fuchs (Dr. FritzFaulhaber GmbH & Co. KG) received the 2006 DGAQSPrize. This prize is awarded biennially by the GermanSociety for Acoustic Quality Control (DGAQS).
Both of the projects are based on the use of PolytecVibrometers and were presented in the 2006/1 issue of our magazine (see www.polytec.com/LM-INFO). The test facilities were systematically designed anddeveloped starting with the selection of appropriatesensors and conditions. After a reliable correlationbetween measurement results, the parameters andcriteria of production faults could be established and the test bench automation completed.
Faulhaber uses vibrometers for testing of micro drivesystems, and PARI for testing of aerosol generators forliquid drugs.
Polytec’s PSV-400-3D Scanning Vibro-meter has become an invaluable toolespecially for automotive and aerospacedevelopment. Despite its high level ofsophistication and performance, it is
based on a simple physical principle:the Doppler effect. In our brand newdemo video, we show how the PSV-400-3D Scanning Vibrometer worksand how it is operated.
View the video trailer on ourweb page and order your individual DVD at
DGAQS director Prof. Dr. Kotterba congratulating Mr. Fuchs(Faulhaber) and Mr. Pfichner (formerly PARI)
New Instructional Video
Learn About 3-D Vibration Measurement On Your Desktop
Improved Quality – Lower Costs:
New Laser Velocimeters at Aluminum MillALUNORF in Neuss, Germany, the world's largest aluminum rolling and remelt plant, uses Polytec’s non-contact, laser-Doppler instrumentation to simultaneously measure length and speed of aluminum plates in themanufacturing process. Several LSV-6000 Laser Surface Velocimeters are integrated into the production line and provide length and speed data for process control.
By using Polytec’s innovative measurement technique, ALUNORF has implemented a fully automatic positioning system to position plates at the shears for high precision cut to length.
The speed measurement is used to synchronize roller conveyor speed with plate speed. This prevents abrasionand damage of the aluminum surface, as well as, reducethe maintenance frequency of cleaning the roller tables. The length measurement is used to position the plates atthe crop shears, resulting in a reliable and repeatable cutlength tolerance.
Laser Surface Velocimeters can measure surface speed and length of all types of materials and have been used in a variety of industries, providing reliable and accuratemeasurements from standstill to speeds of more than ±23,000 ft min–1 in either direction.
IV
Polytec NewsPolytec on the Scene:
The Secrets of Stradivarius and Guarneri Violin Masterpieces
Polytec joined forces with ProfessorGeorge Bissinger of East Carolina Uni-versity and two highly respected violinmakers to study the 3-dimensionalvibration response of three old Italianmaster violins using the Polytec PSV-400-3D scanning system. This wasa once-in-a-lifetime opportunity to testa Guarneri del Gesu and two Stradi-varius masterpieces. Scans of variousportions of each violin were stitched to provide a true 3-dimensional visuali-zation. 3-D data also permits volumechanges to be computed related to airforced through the f-holes, previously
shown to be a new radiation mechan-ism providing a major contribution to the sound near 500 Hz. Anotherindirect radiation mechanism – A1cavity-mode-forced body motion – wasseen quite strongly in one Stradivariusand one Guarneri del Gesu. The hope is to combine the Polytec vibrationaldata with acoustical and density/shapedata also measured during the visit inorder to develop a vibro-acoustic solidmodel. This would perhaps enable us to finally reveal some of the secrets of these masterpieces and apply thisknowledge to modern violin making.
British Center of MEMS Excellence:
UK National Standards Laboratory Purchases Polytec MSA-400
NPL is the United Kingdom’s nationalstandards laboratory, an internationallyrespected and independent center ofexcellence for R&D, and knowledgetransfer in measurement and materialsscience. To support the metrology anddynamic characterization of MEMS andother microstructures, NPL has added a Polytec MSA-400 Micro System Analyzer to its collection of working
in the micro and nano domain. TheThe Polytec system was chosen for itssmall spot size and hence excellentlateral spatial resolution, for its ability to measure through a vacuum windowand for its differential measurementcapability that eliminates any relativemotion between the tested device andits supporting structure. NPL’s newPolytec MSA-400 covers frequencies to1 MHz and measures dynamic motionin both the out-of-plane (Z axis) and in-plane (X-Y axes) directions. It is capable of measuring the frequencyresponse of resonant devices (such as cantilevers, membranes, accelero-meters, etc), as well as the time domain response of switches, actua-tors and other structures, displayingthat response as still or animateddeflection shapes for out-of-plane vibration, or Bode plots with motionamplitude for lateral measurements.
Data is easily taken, giving fast indica-tion of the overall frequency responseof a structure, with deflection shapesquantifying the motion at the resonantand other frequencies. Such infor-mation is essential to verify if design,manufacture and operation are correctfor the device. The MSA-400 will alsoensure that existing characterizationwork on PZT and other active materialswill also be enhanced and extended.
The purchase of such a system showsNPL’s commitment to supporting theemerging field of Microsystems andNanotechnology and it is hoped thatfurther work will ensure the develop-ment of calibration and certificationtechniques for the measurement ofthese challenging devices.
Read NPL’s new MSA-400/vacuumchamber application note on
More info on the ECU homepage:www.ecu.edu/news/poe/1006/violins.cfm
View audio slide show on www.newsobserver.com/1181/story/489521.html
1
Automotive Vibration Testing
Laser-based Vibration Measurement Technology Helps to
Increase Performance, Improve Time-to-market and Lower
Costs in Automotive Development
Laser vibrometry is firmly estab-
lished as the automotive industry’s
gold-standard for non-contact vibra-
tion measurement. Its advantages
include zero-mass loading, long
standoff distance, high precision
and sensitivity, fast set-up, ease of
operation, high-throughput, and
low operating costs. With so
many advantages over contact
transducers, laser vibrometry is
quickly revolutionizing design
development and experimental
modal analysis in the automotive
industry. It can be extended to the
most difficult measurement tasks,
such as red-hot, complex or micro-
scopic structures. Polytec’s compre-
hensive line of products and services
provide an optimal solution for
almost every automotive vibration
measurement application.
Photo: www.pixelquelle.de
Sound & Vibration CharacterizationThe dynamic and acoustic propertiesof an automobile are one of the mostimportant qualities affecting customerperception and vehicle sales. Polytecvibration measurement equipment isused by automotive manufacturersworldwide to improve and optimizetheir vehicles.
A luxury car manufacturer provides aperfect example. During pre-productionprototype testing of a new engineconfiguration, there was evidence ofalternator whine. Using a ScanningVibrometer, the noise hotspots wereidentified. This data was combinedwith a FE model permitting an intelli-gent redesign that reduced the noise
and increased the component dura-bility. Problem solved!
Read more about another projectdealing with structure-borne soundintensity of car bodies derived from laserscanning measurements on page 6.
2
Automotive Applications
FE-Test Correlation for Damping MaterialOptimization Material layout optimization can beperformed effectively using finite elementanalysis. The optimization results needto be validated by measuring the vibra-tion performance of real prototypes.
RMIT University, Melbourne, has appliedthe PSV-400 Scanning Vibrometer tothe research of automotive panels suchas the car door and bonnet. The naturalfrequencies and mode shapes beingmeasured by the vibrometer are com-pared with the results of an FEA fre-quency extraction procedure. A goodcorrelation between the simulationand experimental results gives confi-dence in the FEA model to perform an
optimization study using genetic algorithms that will redistribute the liner material.
Read more on page 10 about a studyperformed at Dow Chemical Companyto demonstrate the ability to improveliquid applied damping material per-formance while reducing the materialusage.
Valve Train Testing
Combustion and the associated enginevalve train movement are highly dynamicprocesses where extremely high speedsand accelerations can occur. Measuringvalve motion presents some special
challenges including separating thevalve motion from the superimposedwhole body displacement of the cylinderhead and measuring large displacementswith high resolution. A differential,high speed vibrometer is an excellentmeasurement solution and can provideaccurate valve motion graphs essentialfor optimizing the combustion process,fuel consumption, engine performanceand service life. The measurement rangeof up to 30 m/s allows measurementseven on high performance Formula 1engine systems.
For more detailed information, pleaseread the article on page 12 and viewPolytec’s Application Note VIB-C-03 at
How to See Brake Sounds Under certain operating conditions,the complex dynamics between thebrake caliper, brake pads and brakedisk can cause undesired audiblesquealing. Measurements withPolytec’s 3-D scanning vibrometeracquire complete vibration vector data showing the spatial dynamics ofthe brake disk. Using this technology,researchers at Robert Bosch GmbHhave managed to track down the
causes of undesired noises when brak-ing. The vibrometer can also measurethe sound field set out from a squeal-ing brake. For more information andto view a live animation of the brakesound field, visit
To learn how Continental AutomotiveSystems designs silent brakes by usingoperational deflection shape analysis,read the article on page 4.
Photo: RMIT
Experimental ModalAnalysisModal data describe the dynamicproperties of a structure and can assistin the design of almost any structure,helping to identify areas where designchanges are most needed. Predictingthe vibration characteristics of auto-motive components and systems is a standard CAE process in today’sautomotive development environ-ment. DaimlerChrysler studied thesuitability of Polytec’s 3-D ScanningVibrometer for data acquisition in carbody modal testing and discoveredthat vibrometry can make the samemeasurements as accelerometers butquicker and more accurately, cuttingmodal testing costs substantially.DaimlerChrysler’s technique andmeasurement results are summarizedin the Application Note VIB-C-01which can be downloaded from thePolytec website. To learn more aboutthe theory behind this method, readour tutorial “Basics of ExperimentalModal Analysis”.
Modal Analysis can also be applied inengine development (see page 16).
Photo: DaimlerChrysler
3
Tracking of Rotational Vibrations Rotational movements of automotivecomponents are always the subject ofintensive optimization efforts by prod-uct development groups. The non-uni-
form rotationinduced by the firingof individual enginecylinders leads to tor-sional vibrations inthe drive chain thatcause undesiredvibration and noise.
Rotational Vibrometers use remote laserprobes to avoid contact and allow a quickand easy examination of the torsionalvibrations while the components ofinterest are in operation.
Read more on page 14 to learn abouthow Rotational Vibrometers help deter-mine the transmission behavior of adual mass flywheel, and download our application note VIB-C-04 on
Industrial Vibration Sensors for Production Testing Acoustic quality control is a non-destructive process to assure thequality and reliability of productsand manufacturing processes. In thecar industry, non-contact vibrometerscan be used to test engines, gearboxes, steering gears, cam rings, turbochargers and fuel pumps toname a few examples. At TRW Auto-motive in Gelsenkirchen, several fullyautomatic test stations use IVS-300Industrial Vibration Sensors to provide
a 100 % inspection of motor pumpassemblies before integration into thesteering gear systems. Proper inspectionassures that drivers will experience performance and reliability withoutdistracting or annoying noise. Readmore and download our magazineissue 2006/1 specially featuring an article entitled “100 % Quality Controlin Industrial Production”.
Development of Micro-ElectromechanicalSystems (MEMS)As MEMS components in modern carsare increasingly taking on safety-relevanttasks, high sensor precision combinedwith lifelong reliability is of criticalimportance. Typical automotive appli-cations of MEMS include engine controlwith pressure sensors, airbag actuation
by accelerometers, vehicle dynamic con-trol, position sensors, light and moisturesensing, and distance sensors to avoidcollisions. For example, engineers atBosch are developing radar antennasincluding RF MEMS switches whosevibration behavior was designed withthe aid of Polytec vibrometers. Formore details please download the fullarticle about “Automotive Sensors“ at
Please read also the article on page 9to learn about MEMS pressure sensorsfrom Melexis, a world-class automotiveelectronics producer, and how Polytec’sMSA-400 Micro System Analyzer isused to reduce production cost.
Photo: Bosch
Photo: IAV
Photo: TRW Automotive
Vibration Tests onElectronic Circuits Electronical interconnects are acommon source for automotivefailure, very often caused by brokenwire bonds. Consequently, vib-ration testing of wire bonds isessential to finding and avoidingthese problems. Non-contact laservibrometers are ideally suited forthis purpose, while contact trans-ducers (accelerometers) are im-possible to use due to size andmass loading. By making a 3-DScanning Vibrometer measurementon a printed circuit board, anengineer can determine both in-plane and out-of-plane vibrationsof bond pads. In the exampleabove, FRFs and deflection shapesof opposite bond pads reveal fre-quencies that could be detrimentalto the bonds.
Figure 1: Calculated ODS (operationaldeflection shape) of a squealing brake system.
Quiet Please!Eliminate Brake Squeal with Numerical
and Experimental Vibration Analysis
Computer-aided simulation ofbrake noise has made impressiveadvances in the past few years.Likewise, 3D scanning vibrometryhas substantially extended thepossibilities of experimental vibra-tion analysis. Scanning Vibrometryenables the measurement of bothin-plane and out-of-plane opera-tional deflection shapes of therelevant components in the sametest setup. At Continental Auto-motive Systems the measurementof brake systems with scanningvibrometry including PSV-400-3Dsystems is fully integrated in thedesign optimization process tospecifically avoid brake noises.
Finite Element SimulationsBrake noise is an important concern in developing brake systems. Brakesquealing is still the most frequentlycited NVH concern. It is caused by self-excitation and includes usually just onefrequency. The noise appears whencertain temperatures and pressures areattained in the brake system. A firstoptimization to prevent self-excitationconsists of a complex eigenvalueanalysis with subsequent structuralmodifications to avoid unstablemodes. The analysis is based on thevibration equation
ƒ(t) = M··q + D ·q + Kq = 0
where ƒ(t) is the force, M is the massmatrix, D is the damping matrix, K isthe stiffness matrix and q is the displace-ment vector. Continental Automotiveengineers simulate the complete brakesystem with all adjacent componentsfeaturing several hundred thousand
degrees of freedom. On the one hand,the complex Eigenvalue analysis canbe started during the design processwithout any prototypes necessary. On the other hand, the method is too sensitive and shows more squealfrequencies than actually exist.Therefore, it is necessary to comparethe results to simulation and dynamo-meter testing and verify whether thecalculated ODS actually occur in practice. In Figure 1, the calculatedoperational deflection shape of asquealing brake system is shown.
Experimental Data
In Figure 2, the test setup for measuring the operational deflectionsshapes with a 3D Scanning Vibro-meter on the brake test stand isshown. The scan measurements are triggered when the brake brakesqueal is detected by a microphone.
4
Brake Development
Figure 2: The setup of the 3-D Scanning Vibrometer to measure the operational deflectionshapes of the brake caliper and rotor.
PSV-400-3D Scanning VibrometerThe PSV-400-3D Scanning Vibro-meter is the perfect measurementinstrument for gathering 3-dimensional vibration datafrom both simple and com-plex structures. It features anintuitive 3-D animation of themeasurement results with separa-tion of out-of-plane and in-planevector components as well as a powerfuldata interface toModal Analysis andFEM Software.
Polytec Product Information
3080
3100
3120
3140
3160
3180
3200
0 0,1 0,2 0,3 0,4 0,5 0,6
Friction Coefficient
Freq
uen
cy [
Hz]
Frequency variationdue to friction
Mode coupling Flutter instability
This guarantees that acquired data iscorrelated to specific sound behaviors.
The vibration behavior of the brakesystem is measured simultaneouslyfrom three different directions.
Instead of moving the sensor headsand stitching together the measure-ments from the different positions, surfaces not directly accessible fromthe sensor position are measured viamirrors. When measuring via mirrors, acoordinate transform of the measure-ments is necessary to make sure thatall measurements both with and withouta mirror are displayed in the same co-ordinate system. This is enabled auto-matically by predefining the mirrorpositions during the setup of theScanning Vibrometer. In Figure 3, the operational deflection shape of a brake caliper measured with theScanning Vibrometer is shown.
Extending the measurement to three dimensions, the 3D ScanningVibrometer has become an essentialtool to collect experimental data forthe optimized design of brake systems.It enables the measurement of both in-plane and out-of-plane operationaldeflection shapes of the relevant com-ponents in the same test setup. In contrast to conventional measure-ments with tri-axial accelerometers, itenables fast and efficient non-contactand thus reactionless measurements atall optically accessible surfaces, whilesimultaneously increasing the numberof measured points. An integrated dis-tance sensor known as the GeometryScan Module enables acquisition of the 3D spatial coordinates on a pre-defined measurement grid. If available,geometry data can also be importedfrom a FEM program.
Supression of Unstable ModesAs soon as the operational deflectionshapes and mode shapes are known,the causes of self-excitation can be elu-cidated on the basis of mode coupling.In Figure 4, the interaction of twomodes of the brake system (caliperand brake disk) are shown to dependon the friction coefficient. The dashedred line shows the frequency of onemode, the solid blue line the fre-quency of a close second mode.
By increasing the friction coefficient,the frequency of both modes arechanged. At first, the frequency of the first mode is decreasing while the frequency of the second mode isincreasing until the frequencies of both modes are the same at a frictioncoefficient of 0.1.
In this case both modes are coupledand result in a flutter instability as aresult of interaction between caliperand brake disk. When increasing thefriction further, the two modes remaincoupled. In order to eliminate thebrake noise, the structure is modifiedby shifting the resonance frequenciesso that the mode coupling disappears.
SummaryThe analysis of mode shapes based onexperimental data acquired with a 3DScanning Vibrometer enables selectivebrake design modifications to avoidbrake noises.
Figure 4: Flutter instability due to mode coupling as a function of the friction coefficient.
Figure 3: Operational deflection shape of abrake caliper measured by the vibrometer.
From Source to SinkBase Equation for IntensitySound intensity represents, in vectorform, the acoustic energy flow in amedium. Using continuum mechanics,a mechanical model for the intensitycan be derived from the strain and stress.In a rigid body, these are tensors withnormally six independent values, threein the normal direction and three in thetangential or shear direction (Fig. 1).
In basic mechanics, power is the prod-uct of force times velocity. Similarly, incontinuum mechanics, multiplication ofthe respective tensors results in theintensity vector in three directions. Eachcomponent Ji of the intensity vector isthe sum of three terms representingthe different stresses (normal andshear) averaged over time (Eq. 1).
Mechanics of Structure-BorneSound Intensity in PlatesIn general, it is not possible to accuratelymeasure stress and strain in a rigid bodyto satisfy Equation (1). The general in-tensity formula is applied to calculatedvalues (CAE). Any experimental appli-cation is limited to plates, where it ispossible to extrapolate from surfacedeflections (measured by scanningvibrometers) to interior responses.
Automotive engineers want to relate the interior vehicle noise level to
the forces applied to the body at the engine and chassis mounts. Although
the acoustics of solids are much more complicated than in fluids, where
measurements of intensity have been established for a long time, it is an
ideal measurement application for scanning laser vibrometry when limited
to thin plates. To further explain this point, a sound intensity analysis of
structure-borne effects is discussed for a car roof panel.
Using Scanning Laser Vibrometer Measurements to Derive
Structure-Borne Sound Intensity from Car Bodies
6
Figure 1: Incremental element of a rigid body with components of the stress tensor (see Equation 1).
Eq. (1)
Eq. (2)
Eq. (3)
Eq. (4)
Sound Analysis
After integration over the plate thick-ness and with restriction to the out-of-plane component, the power in the x-direction is given by Equation (2).
The next issue to consider is the timeaveraging of the product of two dynamicvariables y(t) = x1(t) · x2(t). In the timedomain, both variables oscillate withthe same frequency. Applying this, theintensity averaged over time can becalculated (Equation 3) from the meas-ured spectra in the frequency domain.
The real part denotes the active intensity.The imaginary part is known as reactiveintensity and describes the energy thatis oscillating over the surface of thestructure but does not travel on average.
The measurement systems used inmodal analysis are not able to measure
Experimental Results When exciting a structure, a physicalenergy flow occurs if an energy sinklike a damper is provided. This was firstimplemented using a simulated data setfor a plate with dimensions 700 mm x1100 mm x 1 mm and a single pointdamper. The simulated sample showsthe expected behavior in the activeintensity vector field and its divergence(Fig. 2).
Next, a steel plate supported by bungeecords was measured on a 21 x 10 pointgrid using the PSV Scanning Vibro-meter. A rectangular bitumen foil hadbeen added to the lower left corner asa damper so the energy flow wouldbecome defined. The exciter (Source
ConclusionWhen optimizing damping layers onautomotive panels to improve thenoise transfer function, a very effectivesound intensity analysis can be madeby using scanning laser vibrometry.The advantage of intensity over soundpressure measurements is that intensityrepresents the energy flow and revealsmore significant information that canbe used for NVH problem solving.
derivatives with respect to space, how-ever Equation (3) has derivatives up tothe third order and mixed derivativesthat are necessary to get the termsneeded for the intensity. In practice,this is the critical part of the method.The measurements can be done withina grid of measurement points duringthe scan and the interpolation is doneseparately in both surface directions x and z. Finally, a numeric algorithm is used to estimate the high orderderivatives, using either cubic splinesor harmonic functions.
MATLAB ImplementationA MATLAB program with a convenientuser interface has been developed to
7
Fig. 2: Intensity (energy flow) of a simulated plate at 420 Hz with an exciter (Source X) and a damper (Sink D).
Fig. 3: Intensity of a locally damped steel plate at 420 Hz with an exciter located at X.
Fig. 4: Scanning Vibrometer measurement setup (left) and operational deflection shape (right) for the car roof.
Fig. 5: Structure-borne intensity at 498 Hz of an isolated roof withsingle damping pad.
This work was done at the Ford Acoustic Centre Cologne with the support of Martin Flick.
read the Scanning Vibrometer data set, providing the mobility functionsand the coordinates of the measure-ment grid. After passing some integ-rity checks, an averaged frequencyresponse function (FRF) is plottedfrom the data. A single frequency isselected for the analysis and its FRFcan be plotted as a 3-D shape overthe measurement grid.
After checking the measurement data,an analysis can be performed at thefrequency of interest. Active andreactive intensity can be evaluated.The result is plotted as a vector plot or a divergence plot. In particular, the divergence of the intensity vectorfield (see Equation 4) is an effectivetool to find energy sinks and sources.
X) is visible in Fig. 3. The dissipationtakes place in a more distributed wayas compared to the analytical exampleshown in Fig. 2 with a point damper.
The real world sample is a roof takenfrom a passenger car by cutting at thetop of the A, B and C pillars (Fig. 4),then fixed with a bungee support, anda damping pad was attached to theupper right quarter of the roof.
In Fig. 5, the energy input of the shakerlocated at the lower left corner can beidentified clearly. The additional dampingpad does not show a significant effect.This is due to a reinforcing roof bow inthe middle between the two B-pillarswhich acts as a barrier to the soundwaves, so that only a part of the energyreaches the upper part of the roof.
D
X
XX
By integrating the scanning vibrometerinto the cover design process, the FEmodel was correlated with real worlddata and prepared for structuraloptimization. The final validation ofthe optimized design fully met theexpectations of the engineers.
A hybrid approach for vibration-related optimization of steering systemcomponents, employing numericalmethods like CAE and simulation earlyin the development phase, can only be efficiently done if tools are availablewhich enable a real-time verification of the model. The scanning laservibrometer is an important tool in this respect, especially when data isacquired for stationary events and constant operating conditions. Ascreening approach with accelero-meters can complement the vibrome-ter in some special cases, especiallywhen scanning certain transientvibrations during start-up.
TRW Automotive, with over 60,000 employees worldwide, is one of the
largest global automotive suppliers for safety systems such as wheel suspen-
sion components, braking systems, steering systems, airbags and restraint
systems. Acoustical and vibration tuning has always held a special spot with-
in the mainstream activities of TRW's product development. To accomplish
this task, the steering division uses two different types of non-contact Laser-
Doppler Vibrometers to analyze vibrations on automotive components and
systems.
In the production of steering systems,single-point vibrometers are used tocapture the surface velocities at speci-fied reference points on the steeringsystems in order to predict the noiseperformance in the vehicle beforedelivery (see page 3).
In the development centers in Düssel-dorf and in Birmingham, PSV-400 and PSV-200 Scanning Vibrometers are used in addition to single-point vibrometers. Stationary deflectionshapes of vibrating structures can beeasily determined with this equipmenteven when the surface is uneven. Inthe Figure 1, there are two deflectionshapes shown (1200 Hz and 3150 Hz)for a cover plate from an electro-hydraulic steering system. The fre-quencies needed to be shifted with thehelp of FE-based structural modifica-tions to improve the cover performance.Specific challenges of this task werethe material used, its complex materialproperties and some tooling-relevantaspects like lead time and costs.
Steering for QualityDetermination of Vibration Modes
in the Development of Steering Systems
Figure1: Vibration modes of a cover plate,determined by the FE model (above) andby scanning vibrometry (below).
8
Production and R&D
Dr.-Ing. Heinrich KostyraTRW Automotive GmbH Düsseldorf, [email protected]
Polytec Product Information
PSV-400 Scanning VibrometerThe PSV-400 is a quick and easy-to-use system for the analysis of full-field structural vibrations, helping to resolve noise and vib-ration issues in commercial, manufacturing and R&D markets.
Photo: TRW Automotive
Advancing Measurements by Light
Basics of Experimental Modal Analysis
What Good are Modal Data?
Modal data are extremely useful informationthat can assist in the design of almost anystructure. The understanding and visualiza-tion of mode shapes is invaluable in thedesign process. It helps identify weakness in the design and areas where improvementis needed.
The development of a modal model, fromeither frequency response measurements orfrom a finite element model, is useful forsimulation and design studies. One of thesestudies is structural dynamics modification.
This is a mathematical process which usesmodal data (frequency, damping and modeshapes) to determine the effects of changes in the system characteristics due to physicalstructural changes.
These calculations can be performed with-out actually having to physically modify the actual structure until a suitable set of designchanges is achieved.
Modal analysis is a method to describe a structure in terms of its natural
characteristics which are the frequency, damping and mode shapes – its
dynamic properties. Without using a rigorous mathematical treatment, this
article will introduce some concepts about how structures vibrate and some
of the mathematical tools used to solve structural dynamic problems.
E26
the input force is changed (Figure 3).There will be increases as well asdecreases in amplitude at differentpoints when sweeping up in time.
Even with a constant input force tothe system, the response amplitudevaries depending on the frequency of oscillation of the input force. Theresponse is larger when a force isapplied with a rate of oscillation thatgets close to a natural frequency (orresonant frequency) of the system and reaches a maximum when the frequency of oscillation is exactlymatched to the resonant frequency.Transforming the time data to thefrequency domain using the Fast
Figure 1: Dynamic modal model development and modal testing in the design process.
Figure 2. Simple plate excitation/response model.
Figure 3. Simple plate response upon a excitation frequency sweep.
Fourier Transform generates some-thing called the frequency responsefunction (Figure 4). There are peaks in this function which occur at theresonant frequencies of the system.
When overlaying the time trace withthe frequency trace, the frequency ofoscillation at the time at which the time trace reaches its maximum valuecorresponds to the frequency wherepeaks in the frequency responsefunction reach a maximum (Figure 5),provided that the frequency change is linear over time.
Regarding the deformation patterns at the natural frequencies, they take ona variety of different shapes dependingon which frequency is used for theexcitation force. The shapes can bemeasured by using either a non-contactscanning laser vibrometer or by placinga set of distributed accelerometers onthe plate and measure the amplitude ofthe response of the plate with differentexcitation frequencies. At each one ofthe natural frequencies a deformationpattern shows up that exists in thestructure (Figure 6).
The figure shows the deformationpatterns that will result when the exci-tation coincides with one of the naturalfrequencies of the system. At the firstnatural frequency, there is a first bend-ing deformation pattern in the plate(mode 1). At the second natural fre-quency, there is a first twisting defor-mation pattern (mode 2). At the thirdand fourth natural frequencies, thesecond bending and second twisting
In addition to structural dynamic modi-fication studies, other simulations canbe performed such as force responsesimulation to predict system responsedue to applied forces. Another very important aspect of modal testing isthe correlation and correction of ananalytical model such as a finite element model. These are a few of themore important aspects related to theuse of a modal model. A schematic isshown in Figure 1.
A Simple Model
Modal analysis is often explained interms of the modes of vibration of a simple, freely supported flat plate(Figure 2). A force that varies in a sinu-soidal fashion is applied to one cornerof the plate. While the rate of oscilla-tion of the frequency is changed, thepeak force will always be the samevalue. The response of the plate due to the excitation is measured with a laser vibrometer or with accelero-meters. In the following, the responsesignal at one corner of the plate is analyzed.
When measuring the response on a point of the plate, the amplitudechanges as the rate of oscillation of
deformation patterns are seen (mode 3and 4, respectively). Such natural fre-quencies and mode shapes occur in allstructures to some extent.
Basically, structure characteristics suchas mass and stiffness determine wherethese natural frequencies and modeshapes will exist. Design engineersneed to identify these frequencies andknow how they might affect the re-sponse of the structure when a forceexcites the structure.
Time Domain, Frequency Domainand Modal Space
In Figure 7, a simple cantilever beam is shown that is excited at the tip of the beam. The response at the tip ofthe beam will contain the response ofall the modes of the system. This timeresponse at the tip of the beam can be converted to the frequency domain by performing a Fourier Transform ofthe time signal. As described before,the frequency domain representation of this converted time signal is oftenreferred to as the frequency responsefunction, or FRF for short.
The cantilever beam will have manynatural frequencies of vibration: thereis a first, second and third bendingmode as shown in Figure 7, and thereare also other higher modes notshown.
Now the physical beam could also be approximated using an analyticallumped mass model or finite elementmodel (shown in black in the upper
As typically a large number ofequations must be solved, matricesare often used for organizing them.The matrix representation of all of theequations of motion describing howthe system behaves will look like
where M, C, K are the mass, dampingand stiffness matrices respectively; x
.., x
.
and x the corresponding acceleration,velocity and displacement and F is theforce applied to the system.
right part of the figure). This modelwill generally be evaluated usingsome set of equations where there is an interrelationship, or coupling,between the different masses, ordegrees of freedom (DOF), used tomodel the structure.
This means that if one pulls on one of the DOFs in the model, the otherDOFs are also affected and also move. This coupling means that theequations are more complicated inorder to determine how the systembehaves.
Figure 7: Composing time and frequency response from simple modal models.
Figure 5. Overlay of time andfrequency response functions.
Figure 6. Simple plate sine dwell responses. The deformation patternshave been generated using PSV Soft-ware.
E28
Usually the mass is a diagonal matrixand the damping and stiffness matricesare symmetric with off-diagonal termsindicating the degree of coupling bet-ween the different equations or DOFsdescribing the system. The size of thematrices is dependent on the numberof equations used to describe our sys-tem. Mathematically, something calledan eigensolution is performed and themodal transformation equation is usedto convert these coupled equations into a set of uncoupled single DOF systems described by diagonal matricesof modal mass, modal damping andmodal stiffness in a new coordinatesystem called modal space described as
where p is the vector of the modalcoordinates. So the transformationfrom physical space to modal spaceusing the modal transformationequation is a process whereby a com-plicated set of coupled physical equa-tions is converted into a set of simpleuncoupled single DOF systems. As isshown in Figure 7, the analytical modelcan be broken down into a set of singleDOF systems where the single DOFdescribing mode 1 is shown in blue,mode 2 is shown in red and mode 3 isshown in green. Modal space allowsthe system to be described easily usingsimple, single DOF systems.
Multiple Reference Analysis (MIMO)
MIMO means Multiple Input/MultipleOutput and is a measuring techniquewhere multiple inputs (excitations, references) are used to excite themeasurement object at the same timeand multiple outputs (responses) of the vibration are measured to obtain amatrix of FRF’s that describe how thevibrating system reacts to an excitation.
Multiple reference tests are an import-ant tool when the structure under test exhibits a high modal density, closelycoupled modes or repeated roots. The dynamic MIMO model is a linearfrequency domain model where spectra(Fourier transforms) of multiple inputsare multiplied by elements of an FRFmatrix to yield spectra of multiple outputs. The FRF matrix model can be written as:
Here m is the number of inputs (references), n the number of outputs(responses), f the frequency, vn(ƒ) thevector of outputs, rm(ƒ) the vector ofinputs and Hnm(ƒ) the matrix of the FRF’s describing the system. If Hnm(ƒ)is known, e.g. as a result from scanningvibrometer measurements where n is the number of scan points and m thenumber of the reference channels, the response of the system to a set of inputs can be predicted.
Modal Data and Operating Data
Modal data requires that the force is measured in order to determine the frequency response function andresulting modal parameters. Onlymodal data will give the true princi-pal characteristics of the system. In addition, structural dynamic modi-fications and forced response can only be studied using modal data(operating data cannot be used forthese types of studies). Also, corre-lation with a finite element model is best performed using modal data.However, modal data alone does not identify whether a structure is adequate for an intended service or application since modal data areindependent of the forces on thesystem. Operating data on the otherhand is an actual depiction of how the structure behaves in service.
This is extremely useful information.However, many times the operatingshapes are confusing and do notnecessarily provide clear guidance asto how to solve or correct an operat-ing problem (and modification andresponse tools cannot be utilized onoperating data). The best situationexists when both operating data andmodal data are used in conjunction to solve structural dynamic problems.
Laser scanning vibrometry is ideallysuited for modal tests because it provides an unambiguous phase reference, highly precise measure-ment data without mass loading problems and a high spatial resol-ution for detailed FEM correlations. Both complete and partial data setscan be exported to commerciallyavailable software packages forexperimental modal analysis (LMS,ME’scope, and others).
Acknowledgement
We wish to thank Dr. Peter Avitabile,Director of the Modal Analysis & Controls Laboratory at University ofMassachusetts Lowell. The main part of this tutorial is taken from his articleentitled “Experimental Modal Analysis – A Simple Non-Mathematical Overview”published in Sound & Vibration magazine, January 2001(http://macl.caeds.eng.uml.edu/umlspace/s&v_Jan2001_Modal_Analysis.PDF)
Melexis has designed and developed electronics for automotive systems for
over a decade. Besides pure electronic products like microcontrollers and
communication bus ICs, MEMS (Micro-Electromechanical Systems) play an
important role in the Melexis product portfolio. MEMS pressure sensors are
an important Melexis product used to measure oil and manifold air pressure
in automotive systems.
Micromachining technology is used byMelexis to manufacture these pressuresensors. The MEMS device senses thepressure through a temporary andreversible deformation to a specificallydesigned mechanical structure.
A typical structure consists of a thinmembrane only a few tenths of a milli-meter wide and etched into the solidsilicon substrate containing the elec-tronic circuit. This approach producessensors that are low cost and acceptable
Pass or Fail?Quality Control of Automotive MEMS Pressure Sensors Using
Polytec’s Micro System Analyzer
for high-volume automotive applica-tions. In development, Melexis usedthe MSA-400 Micro System Analyzerto characterize specific mechanicalsensor parameters. The vibrations ofthe electrostatically excited membranes(Figure 1) are measured by the vibro-meter integrated in the MSA-400. Themeasured eigenfrequencies are theinput for an algorithm that identifiesthe parameters of the membrane likethickness, edge length and intrinsicstress. The MSA-400 can also be usedto find defective dies prior to packaging,thus eliminating subsequent unnecessaryassembly steps, increasing throughputand lowering manufacturing costs.
White Light Interferometry is a fastand precise method used to measuresurface topography, providing an effi-cient tool for monitoring the qualityof automotive component sufaces. In contrast to most competing methods, Polytec's TopMap familyof Interferometers feature telecentricoptics and precisely image difficultsurfaces such as along high, steepedges or in deep, drilled holes.
The image shows an automotivecomponent with several parallel cir-cular planes. Using the TopMap
Interferometer, the distance, angles andwaviness of the planes can be deter-mined easily. The data feature a veryhigh repeating accuracy and can bevisualized as a cross-section or as aprofile cut along a circle line. Theseautomotive components are producedin high volumes with a production timeof only a few seconds. The measure-ments enable automatic collection ofquantitative data for statistical processcontrol, especially for pass/fail decisions,but also for adjustment of CNCmachines.
Polytec hasdesigned dedicatedmodels of the TopMapTopographyMeasurementSystem to meet specific require-ments from fast in-line throughputfor production floor applications tohigh-resolution topography formetrology lab measurements.
Figure 1: Shape of the 2nd modal frequency of a pressure sensor.
Measuring Surface Topography in Automotive Production Applications Using Scanning White Light Interferometry
Panel Vibration MeasurementsA Polytec Scanning Vibrometer wasused to measure the vibration responseof the floorpan. To cover the entirevehicle floorpan surface, the laserbeam was bounced off a static mirrorplaced at 45 degrees under the vehicle(Figure 1). Some of the measurementpoints used during the laser scan areshown in Figure 2. Small magneticretro-reflectors were placed on thefloorpan to ensure good signal qualityat each measurement point.
Two shaker locations were used tostimulate vibration modes in the floor-pan. The laser scan was completed byexciting one shaker at a time. Each shakerapplied a swept sinusoidal force on thevehicle body from 20 to 300 Hz.
At each sampling point, the scanninglaser vibrometer measured a frequencyresponse function (FRF) between thevibration and the excitation force.Laser vibrometer mappings of the FRFs(not shown) revealed the highest
Put a Damper on the Noise
Saeed Siavoshani, Ph.D., Jay Tudor and DevBarpanda, The Dow Chemical Company,[email protected]
Using Laser Vibrometry to Optimize Damping Material Layout
to Reduce Vehicle Body Interior Noise & Structural Vibration
A comprehensive hybrid technique was developed to optimize the application
of damping materials on vehicle bodies. The technique used finite element
analysis (FEA) and experimental measurements to complement each other.
In this example, it was used to suppress floorpan vibration and the associated
radiated noise. By optimizing the application of damping material, a 3 – 5 dB
reduction in the floorpan vibration level was achieved while saving 10 % in
material volume and mass. The optimized layout was validated on a body-
in-white using a scanning laser vibrometer.
10
Material Optimization
IntroductionA process was developed to optimizethe application of liquid dampingmaterial on an automotive platform.The approach consisted of four steps(Figure 1). First, the high vibration
areas of the vehicle floorpan weredetermined, indicating the areas inwhich damping materials are needed.Second, the vehicle transfer functionswere measured to quantify theamount of noise resulting from floor-pan vibration. Third, the material lay-out was optimized using FEA. Finally,an optimized product was measuredto validate the predicted vibration performance.
undamped floorpan. Confidence in the finite element model came fromcomparison to the original vibrometermeasurements and an understandingof the acoustic attenuation propertiesof the sprayable damping material.Then, the model was used to designthe optimal application of dampingmaterial to suppress interior noise levels.
Validation TestingThe improvement in performance predicted by the FE modeling was verified by repeating the vibrometermeasurements. In Figure 3, the aver-age FRFs are shown for all measure-ment points on the floorpan based onthe laser vibrometer data. A compari-son of the vibration level with andwithout damping material indicates an improvement of up to 15 dB is realized with the current dampingtreatment beyond 100 Hz.
mass by 0.8 kg (10 %) and the wetvolume by 0.24 gal (10 %).
ConclusionsThis study demonstrated the ability toimprove the performance of liquid-applied damping material while simul-taneously reducing its usage. Throughoptimization, the average vibrationlevel was reduced by 1 to 5 dB whileproviding a 0.8 kg mass savings. Thisvolume reduction of 0.24 gallons pervehicle results in a savings of approxi-mately $ 215,000/year to the autobody manufacturer. Ideally this opti-mization should be part of the initialdesign and placement of the dampingmaterial prior to a vehicle launch.
vibration areas are in the rear half ofthe vehicle. Concentrating the damp-ing material in these areas was optimaland minimized structure-borne noise.Conversely, the laser vibrometer alsoidentified low vibration areas whereless damping material was needed tosuppress the structure-borne noise;thereby, saving material, weight andcost.
Floorpan Vibration to InteriorNoise LevelsCompared to the body-in-white forthe vibrometer measurements, a fullyassembled vehicle was used to deter-mine the amount of noise originatingfrom the vibrating floorpan that reacheda driver’s ears. First, the interior acousticmodes were measured and calculatedby FE modeling (title image) and com-pared to direct interior sound levelmeasurements. Similar to structuralvibration modes, these acoustic modes
highlight the most important acousticfrequencies for modeling. Then, thetransfer function between the floorpanvibration and the resulting interior cavitynoise was determined by aligning theshakers to the previously used attach-ment points. From these studies, inte-rior airborne noise at driver ear levelwas related to floorpan vibration.
Optimization Using FiniteElement Analysis
Once the transfer function was under-stood, a finite element technique wasused to study the vibration response of the vehicle floor. This techniqueidentifies the “noisy” region in the
The optimized layout determined by FEA was created by removing oradding material to a vehicle bodyapplied with the current layout. InFigure 4, the vehicle body is shownwith material removed in areas of lowvibration and material added in areasof high vibration. The vehicle was thentested using the laser vibrometer tomeasure the vibration level.
The optimized damping material layout results in a vibration reductionof about 1 dB for the front shaker loca-tion and a reduction of about 3 – 5 dBfor the rear shaker location (Figure 5).This performance increase was achievedwhile reducing the damping material
11
Figure 1: Laser scan test setup. Figure 2: Points measured in laser test.
mirror
Figure 5: Optimization results.
Figure 3: Current damping performance.
Figure 4: Vehicle body with optimizeddamping material layout.
12
Valvetrain Analysis
By combining Rotec’s Rotation Analysis System and Polytec’s High Speed
Vibrometer, development engineers can measure and analyze dynamic and
high-speed valvetrain motion, even on racing engines, ensuring that valve-
train components satisfy strength, durability and accuracy requirements.
IntroductionModern valvetrain systems must pro-vide both large cross sections for thegas exchange process and high-speedopening and closing of the valves. Thiscombination results in high structuralexcitation and component stresses fromthe fast changes in valve velocity andacceleration combined with large liftvalues. Development programs must
Faster,Higher, StrongerDynamic Valvetrain Analysis with Rotec’s Rotational
Analysis System and Polytec’s High Speed Vibrometer
Figure 1: Measurement setup for valvetrain testing using Rotec’s RAS.
TEST RIG TEST DATA ROTEC-RAS
assure that valvetrain components satisfystrength and durability requirementsand that they operate within tightspecifications and tolerances.
With the increasing complexity of val-vetrain systems, the requirement forcomprehensive valvetrain testing canbe addressed through application-spe-cific test and analysis protocols in cus-tomized valvetrain motion software. In
response to this need, RotecGmbH has developed a PC-
based Rotation AnalysisSystem (RAS) to performsignal acquisition and noiseand vibration analysis onengines and transmissions.A large number of thesesystems are used world-wide by automotive test-ing and developmentdepartments.
Measurement SetupIn Figure 1, a typical measurement setup for valvetrain testing is shown.Camshaft speed and angle are measured by either fitting an incre-mental encoder to the shaft or byscanning a toothed wheel with a mag-netic pickup. On both fired engines andnon-fired test benches the valve lift isgenerally measured with inductive orcapacitive displacement sensors.
Polytec’s High-Speed Vibrometer (HSV)system is an excellent sensor to measurevalve velocity on motored test benches.The advantages of the HSV include non-contact, high-resolution measurementup to 30 m/s and linear output signals.Valve velocity is measured at frequen-cies up to 50 kHz and valve lift can be
Sync. Pulse (optional)
Camshaft Speed and Angle
Valve Lift
Valve Velocity
Additional Analog Signals
13
measured up to 250 kHz. Differentialmeasurement compensates forunwanted vibration and movement(see information box).
By combining both the RAS and HSVsystems, engineers can make demand-ing dynamic measurement and analy-sis of valvetrain motion, even on highperformance racing engine test rigs.Synchronous to the camshaft speedand valve lift and velocity signals, addi-tional test data such as valve springloads can be acquired.
The RAS rotational speed channelsrequire square-wave TTL level signalsas input. The time interval between rising (or falling) edges for each pulseperiod is recorded using a 10 GHz/40-bit high-speed counter/timer. TheRAS analog channels sample at 400 kHzwith 16-bit resolution. In valvetraintesting, the speed signal is used fortransforming the time equidistant sam-pling of the lift and velocity signals intoangle equidistant data. Consequently,a toothed wheel and proximity probe(instead of rotary encoders) may be usedfor measuring camshaft speed and angle.Signals from gear wheels with missingteeth may also be processed, a signifi-cant advantage of the RAS software.
Exemplary ResultsThe RAS valvetrain software offers avariety of options for analyzing valvemotion versus speed and angle. In Fig. 2,a speed run-up measurement is plottedin 3-D. The valve lift signal which de-termines the valve lift versus cam angleand speed is measured by the PolytecHSV and shown in Fig. 2a. The valvevelocity (Fig. 2b) is also measured bythe HSV. However, since the camshaftspeed changes over the course of themeasurement, it is more meaningful torepresent valve speed in m/rad insteadof m/s. This option is integrated into theRAS software. The normalized valveacceleration in m/rad2 is shown in Fig. 2c.This is the 1st derivative of the meas-ured valve velocity (HSV) sampled bythe RAS. The software allows for low-pass filtering before differentiating.
There are several methods of calculat-ing valve closing velocities and closingangles. In general, a threshold value oflift during the closing sequence is spec-ified. Then, beginning at maximum liftand looking along the cam angle, theclosing velocity and angle are foundwhen the valve lift falls below thethreshold lift. Alternatively, havinglocated a specified lift threshold andlooking along the cam angle, the first
Figure 2: a) Valve lift versus cam angle and speed; b) Valve velocity versus cam angle andspeed; c) Valve acceleration versus cam angle and speed.
HSV High-Speed VibrometerThe HSV copes with high vibrationspeeds up to 30 m/s and providessingle or differential velocity anddisplacement measurement capa-bilities for high-speed applicationslike valve train testing on perform-ance engines, power tools andimpact testing.
local maximum of valve acceleration isfound. The valve closing velocity andcam angle are then determined at thisposition (Fig. 3).
The contour plot (Fig. 4) shows valvevelocity versus cam speed in the closingangular range where valve bouncing isapparent. The valve seats at approx.288 degrees cam angle. The alternatingred and green colors show the valveimpacting the seat before finally coming to rest.
Conclusion and OutlookThe RAS valvetrain software offers manyother capabilities such as comparingmeasured 3-D plots with theoreticalcurves or determining lift loss normal-ized to angle during the opening andclosing phases. Valve open and closeduration is also of interest. Valvetrainmaterial and geometrical parametersmay be used to investigate cam andtappet component strains (Hertzianstress). In conclusion, the use of high-resolution measuring equipment andapplication-specific analysis software helpsatisfy the demands for meaningfulresults and shorter development cycles.
14
Rotational Vibrations
Vehicle drive trains equipped with a combustion engine experience
torsional oscillations caused by the crankshaft. Considerable amplitudes
can occur at various positions of the crankshaft affecting the mechanical
stability and acoustic properties of the drive trains. To provide a design
that avoids or minimizes such phenomena, engineers need knowledge
of the dynamic properties of the drive train components. Using Polytec
Rotational Vibrometers, a dual mass flywheel can be characterized, demon-
strating how the dynamic transmission behavior can be determined on a
test rig for drive elements installed at the University of Kaiserslautern.
operated as a generator so that a tor-sional momentum is generated thatloads the drive element under test.The driving torque can be superim-posed on a well-defined oscillationmomentum at an excitation frequencyfexc > 450 Hz. With a high-resolutionmeasurement of both torsion angleand torque momentum, the dynamicresponse behavior can be determined.
Measurement of TorsionalVibration Using RotationalVibrometers For the measurement of the torsionalvibration, two rotational laser vibro-
Drive TrainsUnder TestRotational Vibrometers Help Determine the Transmission
Behavior of a Dual Mass Flywheel
Figure 1: Dynamic test rig for drive elements.
Prof. Dr.-Ing. Bernd Sauer and Dipl.-Ing.Andreas Nicola, University of Kaiserlautern,Germany, [email protected]
High-dynamic Test Rig for Drive ElementsA sophisticated dynamic test rig fordrive elements (Figure 1) is available at the Institute for Machine Elements,Gears, and Transmissions at the Uni-versity of Kaiserslautern. This system
allows the vibrational testing of drivetrains and their components under thespecial influence of torsional excitationsand the derivation of the dynamictransmission behavior at various loads,rotational speeds, excitation frequen-cies and amplitudes.
The test rig uses a twisting motionproduced by a high-dynamic electricmachine that drives a braking motorvia the test item. The braking motor is
15
RLV-5500 RotationalLaser VibrometerPolytec's rotational vibrometersare advanced non-contact angularvelocity and displacement sensors,perfect for measuring rotatingstructures such as crankshafts,axles and pulleys. As proof of itssuccess, automotive design andtest engineers have skillfully usedrotational vibrometer data in bothresearch and development toreduce engine noise and toincrease product durability. Thenew RLV-5500 Rotational LaserVibrometer features an expandedrpm range of up to 20,000 rpm,an excellent optical sensitivity andS/N ration due to digital decodingtechniques, and a very compactsensor head that can be flexiblymounted.
Figure 2: Rotational vibrometers and dualmass flywheel on the test rig.
Polytec Product Information
Figure 2: Dual mass flywheel on the test rig.
meters are used. The measurementprocedure is highly precise, robust andmobile. The setup allows the measure-ment of drive speed n, dynamic frac-tion of the rotational speed ��, anddynamic oscillation speed �� withoutcontact during operation. A detaileddescription of the operating principleof the Rotational Vibrometer can befound at www.polytec.com/rotvib.
Example: Dual Mass Flywheel In every combustion engine, a flywheel is used as energy storage tokeep the piston motion running evenwhen there is no work cycle. At thesame time, it smooths out the torsionalexcitation of the crankshaft and avoidsvibrations. In the majority of cases, thisis accomplished solely with flywheelmass. An alternative method is to use a dual mass flywheel (DMF). In theDMF, the flywheel mass is split intotwo masses that are torsional linked by elastic springs. By varying the ratiobetween inertias and spring stiffness, a desirable low Eigenfrequency can befound. The DMF acts as a mechanicallow-pass filter at the transition to thedrive train.
In the title image, the test setup fordetermining the dynamic transmissionbehavior under various conditions isshown. The DMF is driven from theleft side by a motor at stationary speedwhile superimposing a torsional oscilla-tion. On the right side, it is loaded bythe generator with a constant torquemomentum. Between specimen andelectric machines, the momentum ismeasured by the torque sensors
and the dynamic oscillation angle is measured by the two rotational laser vibrometers.
The experimentally acquired responsebehavior of the dual mass flywheel atvarious revolution speeds during a fre-quency sweep between 0 Hz and 40 Hzis shown in Fig. 2. The excitation wasdone with a constant angle amplitude.A speed of 500 rpm corresponds to anEigenfrequency of 13 Hz at a maximalamplitude ratio of �2/�1 = 3.5. The Eigenfrequency moves to higherfrequencies with higher speeds. The amplitude amplification alsogrows with higher speeds.
Assuming that the modal masses areconstant, the increase of the Eigen-frequency is due to a stiffening of theexisting springs. The reason for thechange in stiffness is supposed withinthe radial deformation of the spring.Apparently, this deformation pressesthe spring to an external contact sur-face so that friction is induced at thecontact points, decreasing the effectivenumber of springing turns and increas-ing the stiffness. The increasing ampli-tude at higher speeds shown in Figure4 is caused by a decrease in systemdamping, a fact that could be con-firmed by further investigations.
Conclusions and ProspectsThe potential to investigate torsionalvibrations with the institute’s drive element test rig in combination withrotational laser vibrometry is exciting.Because of the flexibility of the testfacility and data acquisition equipment,it is possible to gauge other drive traincomponents such as torsionally stiffand flexible couplings, cardan shafts,and vibration dampers and absorbers.It is also possible to perform acousticinvestigations of gearboxes (e.g. rattlebehavior) and to acquire knowledgeabout the dynamic stiffness and fre-quency attenuation of gears. The equip-ment is mobile so that measurementscan also be made on-site with cus-tomers’ test rigs and running engines.
16
Engine Development
IntroductionThe maximum power rating of largediesel engines currently used in shipsor power stations is about 21,000 kWin the 4-stroke range and 97,000 kWin the 2-stroke range. These enginepowers can only be achieved by usingturbochargers with the optimum util-ization of the compression process,enabling increases in performance of300 %. Manufacturers of large dieselengines realized this fact at an earlystage; MAN Diesel SE has developedand built turbochargers for more than70 years.
The compression process is repeatedlysubjected to conditions where an excitation frequency caused by inter-ferences in the air inlet and outlet (e.g.guide baffle, etc.), and the natural frequencies of the compressor wheellead to an increased vibrationresponse. The resulting dynamic alternating load must not exceed thefatigue strength in order to ensure a reliable operation of the com-pressor wheel.
The following steps are used in thedevelopment of compressor wheels toprevent fatigue:
FE Analysis
Finite Element Analysis (FEA) providesapproximate modal parameters describ-ing “natural frequency” and “naturalmode” values which enable a roughassessment of the loads occurring dur-ing operation. Additional experimentalinvestigations are necessary since thedamping and the mistuning caused by manufacturing deviations are notexactly known.
Experimental Modal Analysis
The experimental investigation of the real structure for the determinationof the modal parameters primarilyserves for the comparison with theresults of the Finite Element Method.Additionally, further influences on thevibration behavior can be determined,e.g. caused by the preparation withsensors for operational vibration analysis.
Turbo Power for Large Diesel Engines
Extremely high pressure ratios and volume flow rates are now achieved
in turbochargers for large diesel engines. The protection of rotating
components against high-cycle fatigue is extreme important to reduce early
failures. The combination of modern techniques such as laser vibrometry
and modal analysis allows an exact insight into the vibration behavior of
compressor wheels at ambient temperature. Combining modal analysis
with measurements carried out on the rotating component provides the
basis for the determination of the loads during operation.
Modal Analysis of Turbocharger Compressor Wheels
for Large Diesel Engines
17
vibration modes. The modes have verysimilar blade deflection shapes and canbe distinguished by the fact that theindividual blades oscillate opposite inphase.
The measured FRF’s show the strongcoupling of the individual modes. Inorder to separate the individual modes(Fig.4), the data is exported to themodal analysis software Visual ModalPro by ME’scope. The PSV softwaresupports simple methods to exportdata to external modal analysis soft-ware. For an intuitive graphical pres-entation the natural modes can be re-imported into the PSV software after the modal parameters have been successfully identified.
Summary and Outlook
The PSV-400 Scanning Vibrometer provides data allowing a fast and high-quality verification of FE models.The determination of modal par-ameters of delicate structures is possible without any mechanical in-fluence, a significant advantage over traditional contact transducers such as accelerometers that can have a substantial mechanical influence.
Future use of the PSV-400 with MISO (Multiple Input Single Output)should enable a better separation ofsuperimposed modes. This will be afurther step in the continuous processtowards understanding and develop-ing complex radial-flow compressorstructures.
Figure 1: Force Response Function of the vibration amplitude (above) andthe phase (below).
Figure 2: Experimental setup for the measurement of modal parameterswith superimposed scan grid and operational deflection shape.
Figure 3: Comparison of a mode usingselected measuring points and nodalpoints from FEA.
Figure 4: Separation of the modes bymeans of an ME’scope curve fit. Upperdiagrams: Bode diagram of the measuredcurve (black lines) and result of the curvefit (red lines). Lower diagram: separated modes.
Vibration Analysis DuringOperation
Adding to the results from the FE ormodal analysis, the strain is deter-mined at selected points of the com-pressor wheel at different operatingpoints by means of strain gauges inorder to measure the load of the component.
By unifying the results of the individ-ual steps, the vibration behavior andstress occurring at the component at different operating conditions arediscovered.
Tests for Experimental ModalAnalysisThe dynamic behavior of linear struc-tures can be described by three modalparameters: natural frequency, damp-ing and natural mode. They are specified from any number of ForceResponse Functions (in short FRF) bymeans of curve fitting (Fig. 1).
The PSV-400 Scanning Vibrometer hasproven to be an exceptional tool forthe measurement of these FRF’s. It’sadvantages over contact transducersinclude adjustable measuring rangesthat match the excitation intensity andsystem response, and elimination ofmass loading at the measuring points(Fig. 2).
In addition, the PSV-400 enables thescanning of a great number of measur-ing points within a very short period oftime and the import of FEM meshes asmeasuring points for a simple verifica-tion of natural modes (Fig. 3).
When measuring the modal parameters,a shaker is used for the excitation ofthe structure to measure the excitationforce necessary to determine the FRF’s.A comparison with the system responsewhen using a loudspeaker for non-contact excitation ensures that themechanical connection of the shakerto the compressor wheel does notinfluence the vibration mode.
Compressor wheels with n blades generate the same number, n, of veryclose and hence strongly coupled
Dipl.-Ing. Rüdiger Rehm and Dipl.-Ing. Joseph Woyke,MAN Diesel SE, Augsburg, [email protected]@de.manbw.com
Product News
VI
Measure Rotatonal Vibrations
New RLV-5500 Rotational Laser Vibrometer
Polytec's rotational vibrometers areadvanced non-contact angular velocityand displacement sensors, perfect formeasuring rotating structures such as crankshafts, axles and pulleys. Con-trolling drive train torsional vibrations is critical to designing reliable vehicles,electric power generators and aircraftpropulsion systems. Rotational vibrationdata used in the early stages of productdevelopment has enabled automotivedesign and test engineers to skillfullyreduce engine noise and to substan-
tially increase product durability. A largestandoff distance makes positioning the laser probe fast, safe and conveni-ent and enables the precision measure-ment of operating machinery at several locations without interruption. The RLV-5500 Rotational Laser Vibro-meter is a new class of instrument that benefits from digital decodingtechniques, an improved S/N ratio and an expanded rpm range of up to20,000 rpm. The sensor head design is new, more compact, and smallenough to get closer to the measure-ment object. An integrated air purgesystem protects the optics from produc-tion oil spray and dust. Even onboardmeasurement of an operating drivetrainin a moving vehicle is now possible.
The TopMap Metro.Lab is a high-precision white light interferometerwith a large z-dynamic range andnanometer resolution. This non-contact topography measurementsystem is designed to characterize flatand curved surfaces. The Metro.Labcan measure flatness and generaltopography on hard or soft, robust or delicate surfaces and determineparallelity of two or more surfacesseparated by as much as 70 mm. By combining increased throughput,simple operation, precision measure-ment and an affordable cost, Polytecis changing traditional contact metro-logy. The Metro.Lab is a completemeasurement station for character-izing the topography of large surfaces
without contact. Complex parts canhave many precisionflat and curvedsurfaces that must be measured andcompared to design specifications.The Metro.Lab’s wide field-of-view is a critical advantage when fastthroughput is desired. With a largevertical dynamic range of 70 mm andtelecentric optics, the measurement of flatness, ripple and parallelism is simple even under traditionallydifficult circumstances such as com-paring the top and bottom surfaces of a blind hole. Classic surface par-ameters currently measured with touchprobes can also easily be investigatedwith the Metro.Lab. Since the com-plete measurement area is captured in one measurement run instead of
Features and BenefitsCompact, fiber-coupled sensor head
Precision measurement of rpm,angular velocity and displacement
Expanded rpm range to 20,000 rpmat 10 kHz bandwidth
Data independent of shaft geometry
Improved S/N ratio through digitaldemodulation and filtering
Compatible with all data acquisitionsystems
Integrated air wipe to cool and protect the optics
composing it from individual lines, measurements are often completed in just a few seconds.
Key FeaturesNon-contact, non-destructive,optical interferometer Measurement on surfaces near steepedges (e.g. drilling holes) is possibledue to telecentric opticsIncreased flexibility due to large 70 mm z-dynamic rangeFast measurement over large field-of-viewLarge 80 mm x 80 mm field-of-viewavailable
Non-Contact Curvature and Flatness Measurement Made Simple and Efficient
New TopMap Metro.Lab Topography Measurement System
VII
Version 2.4 Now Available
New PMA In-Plane Vibration Measurement Software Released Stroboscopic Video Microscopy measurements provided by the MSA-400 Micro System Analyzer arenow even easier with the softwarerelease of PMASoft version 2.4. A widenumber of new functionalities andimprovements makes this release thebest software yet for in-plane measure-ments. Featuring a whole series of newbenefits, the PMA software can now
Display and save evaluation results withthe video window, without starting ameasurement, to give a first impression
Compact, Digital, Versatile
New Compact Vibrometer adds to the Family
For displacement-only measurement,the OFV-2510 is the best choice forlower bandwidth measurements and is based on proven fringe countingtechnology. All OFV series sensor heads(also legacy versions) can be connectedto OFV-2500.
More CLV-2534 Compact LaserVibrometersThe OFV-2500 series sensor heads arealso evolving. The CLV-2534-2 Com-pact Laser Vibrometer is a 350 kHzhigh resolution digital version and theCLV-2534-3 is a 3.2 MHz high band-width system with a velocity limit of 10 m/s, offering optional direct displacement output. The main appli-cations are end-of-line productiontesting, hard disk inspection andgeneral R&D.
New Optical Accessories
In addition to the camera option, whichoffers a video image of the object undertest, there are many new accessoriesavailable for both the OFV-534 SensorHead and the CLV-2534 Sensor Head,helping to increase the range of possible
Polytec’s CLV-2534 Compact LaserVibrometers and OFV-2500 Vibro-meter Controllers are designed tomeet the needs of sophisticatedindustrial vibration measurement.A series of new versions is nowavailable that benefit from digitaldecoding technology and feature a wide number of new functionali-ties and improvements.
Expanded Choices in OFV-2500Vibrometer ControllersThe digital OFV-2500-3 VibrometerController features 0.02 µm/s resol-ution limit at 350 kHz bandwidth. Ananalog, large-bandwidth OFV-2500-2version complements the digitalmodel by featuring high measure-ment velocities up to 10 m/s andbandwidths up to 3.2 MHz. Engin-eered for value, the OFV-2500-2model replaces the retiring OFV-2500-1 model with better specifica-tions but at same sales price. TheOFV-2500-2 can be ordered with anadditional integrator for direct dis-placement output.
applications. Combining the sensor head with the new VIB-A-510 LED IlluminationUnit and a microscope objective (e.g. VIB-A-20xLENS) turns the CLV-2534 or the OFV-534 into a measurementmicroscope for microstructures. With a spot size of only 1.5 µm andthe camera option, the laser is always positioned and focused correctly.These accessories are easily mountedonto the sensor head.
Where a large stand-off distance andsmall field-of-view are required, theVIB-A-520 telescopic objective replacesthe standard objective, making theOFV-534 and CLV-2534 Sensor Head aperfect tool for hard disk drive compo-nent measurements and quality tests ,featuring a 320 mm working distanceand a spot size of only 15 µm.
For measurement locations requiring aturn to get access, two 90° DeflectionUnits (for laser only and for both laserand video image) are available.
Display a Sine fit for periodic signalsand a curve fit to specified peaks
Provide extensively enhanced peakanalysis with a band cursor providingstatistical parameters and harmonicoscillator curve fitting and a harmoniccursor that plots additional cursor linesat higher orders of the base frequency
Import PSV scan grid points from out-of-plane measurements to evaluate a micro device in all three dimensions
Events
Advancing Measurements by Light
Tech
nica
l sp
ecifi
catio
ns a
re s
ubje
ct t
o ch
ange
with
out
notic
e.O
M_I
F_In
Focu
s_20
07_0
1_85
00_U
S
Reference our web site for the most up-to-date information and links on trade shows, technology seminars and training classes!
CEO/Publisher: Dr. Helmut Selbach Editorial Staff: Dr. Arno Maurer,
Dr. Phil Mitchell
Production: Regelmann Kommunikation
Image credits: Bosch, Continental AutomotiveSystems, DaimlerChrysler, IAV,J. Bienert, Melexis, MAN DieselSE, RMIT, Rotec Munich, TRWAutomotive, Univ. of Kaiserslautern, www.pixelquelle.de.
For a more in-depth view of our technology, register fora complimentary Technology Seminar. Current usersshould consider attending specialized Training Classesdesigned to improve theirunderstanding and effective-ness in using their Polytecequipment.
For more informationplease contact your localsales manager or email usat [email protected](North America) [email protected] (all otherregions).
May 02 – 03, 2007 2007 US Vibrometer Users Conference Ann Arbor, MI
May 07 – 10, 2007 AISTech 2007 Indianapolis, IN
May 08 – 10, 2007 Automotive Testing Expo Europe Stuttgart, Germany
May 08 – 11, 2007 Control 2007 Sinsheim, Germany
May 15 – 17, 2007 Noise and Vibration St. Charles, IL
May 22 – 24, 2007 Sensor & Test Nuremberg, Germany
May 12 – 16, 2007 METEC 2007 Düsseldorf, Germany
Jun 18 – 21, 2007 LASER 2007 Munich, Germany
Jul 09 – 12, 2007 14th International Congress on Sound and Vibration Cairns, Australia
Sept 12 – 14, 2007 Automotive Testing Expo China Shanghai, China
Sept 19 – 20, 2007 Diskcon USA 2007 Santa Clara, CA
