Download - Swept Frequency Eddy Current (SFEC) Measurements of Inconel 718 a(1)

8/14/2019 Swept Frequency Eddy Current (SFEC) Measurements of Inconel 718 a(1)

1/136

Iowa State University Digital Repository @ Iowa State University

Graduate Teses and Dissertations Graduate College

2013

Swept frequency eddy current (SFEC)measurements of Inconel 718 as a function of

microstructure and residual stressRamya Chandrasekar Iowa State University , [email protected]

Follow this and additional works at:h p://lib.dr.iastate.edu/etdPart of theMechanics of Materials Commons

Tis Dissertation is brought to you for free and open access by the Graduate College at Digital Repository @ Iowa State University. It has been acceptefor inclusion in Graduate Teses and Dissertations by an authorized administrator of Digital Repository @ Iowa State University. For moreinformation, please [email protected].

Recommended CitationChandrasekar, Ramya, "Swept frequency eddy current (SFEC) measurements of Inconel 718 as a function of microstructure andresidual stress" (2013).Graduate Teses and Dissertations.Paper 13348.
http://lib.dr.iastate.edu/?utm_source=lib.dr.iastate.edu%2Fetd%2F13348&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://lib.dr.iastate.edu/etd?utm_source=lib.dr.iastate.edu%2Fetd%2F13348&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://lib.dr.iastate.edu/grad?utm_source=lib.dr.iastate.edu%2Fetd%2F13348&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://lib.dr.iastate.edu/etd?utm_source=lib.dr.iastate.edu%2Fetd%2F13348&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://network.bepress.com/hgg/discipline/283?utm_source=lib.dr.iastate.edu%2Fetd%2F13348&utm_medium=PDF&utm_campaign=PDFCoverPagesmailto:[email protected]:[email protected]://network.bepress.com/hgg/discipline/283?utm_source=lib.dr.iastate.edu%2Fetd%2F13348&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://lib.dr.iastate.edu/etd?utm_source=lib.dr.iastate.edu%2Fetd%2F13348&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://lib.dr.iastate.edu/grad?utm_source=lib.dr.iastate.edu%2Fetd%2F13348&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://lib.dr.iastate.edu/etd?utm_source=lib.dr.iastate.edu%2Fetd%2F13348&utm_medium=PDF&utm_campaign=PDFCoverPageshttp://lib.dr.iastate.edu/?utm_source=lib.dr.iastate.edu%2Fetd%2F13348&utm_medium=PDF&utm_campaign=PDFCoverPages


2/136

Swept frequency eddy current (SFEC) measurements of Inconel 718 as a function

of microstructure and residual stress

by

Ramya Chandrasekar

A dissertation submitted to the graduate faculty

in partial fulfillment of the requirements for the degree of

DOCTOR OF PHILOSOPHY

Major: Materials Science and Engineering

Program of Study Committee:

L. Scott Chumbley, Major Professor

Alan RussellIver E. Anderson

Nicola Bowler

Stephen Holland

Iowa State University

Ames, Iowa

2013

Copyright Ramya Chandrasekar, 2013. All rights reserved.


3/136

ii

DEDICATION

To my

Dad Chandrasekar

Mom Manonmani

Professor R. Bruce Thompson


4/136

iii

TABLE OF CONTENTS

DEDICATION ...................................................................................................................ii

TABLE OF CONTENTS .................................................................................................. iii

ABSTRACT .....................................................................................................................vi

CHAPTER 1: INTRODUCTION ....................................................................................... 1

1.1 Background ........................................................................................................... 1

1.2 Nickel based super alloys ...................................................................................... 1

1.2.1 Solid Solution Hardened alloys ....................................................................... 2

1.2.2 Dispersion Strengthened alloys ....................................................................... 3

1.2.3 Precipitation Hardened alloys .......................................................................... 3

1.2.4 Microstructure of Precipitation Hardened Ni- based alloys .............................. 4

1.2.5 Mechanical Properties of Precipitation Hardened Ni-based Super Alloys ....... 7

1.2.6 Failure of Ni-based superalloy parts .............................................................. 10

1.3 Residual Stresses ................................................................................................ 11

1.3.1 Compressive Stresses .................................................................................. 11

1.3.2 NDE Measurement of Residual Stress .......................................................... 12

1.4 Nondestructive Evaluation ................................................................................... 13

1.4.1 Eddy Current Measurements ........................................................................ 14

1.4.2 Limitations of Eddy Current Signals .............................................................. 15

1.4.3 Factors Contributing to Eddy Current Signals ............................................... 16

1.5 Swept Frequency Eddy Current (SFEC) Measurements ..................................... 23

1.5.1 Theory ........................................................................................................... 23

1.5.2 Instrument ..................................................................................................... 23

1.5.3 Instrumental Factors Affecting SFEC Measurements ................................... 26

1.6 Measurement of Stress Induced by Shot Peening ............................................... 27

1.7 Problem Statement .............................................................................................. 30

1.8 References .......................................................................................................... 31

CHAPTER 2: EXPERIMENTAL PROCEDURE ............................................................. 39

2.1 Sample Preparation ............................................................................................. 39


5/136

iv

2.2 Shot peening........................................................................................................ 41

2.3 Electrical Conductivity Measurements ................................................................. 41

2.4 Mechanical Property Measurements ................................................................... 42

2.5 Residual Stress Measurements ........................................................................... 43

2.6 Microstructure Characterization ........................................................................... 45

2.7 Quantification of Precipitates ............................................................................... 48

2.8 Surface Roughness and Grain size Measurements ............................................. 50

2.9 Swept Frequency Eddy Current (SFEC) measurements ..................................... 51

2.10 References ........................................................................................................ 55

CHAPTER 3: EXPERIMENTAL RESULTS ................................................................... 56

3.1 Bulk Conductivity and Hardness .......................................................................... 56

3.2 Microstructure ...................................................................................................... 59

3.2.1 Grain Size ..................................................................................................... 59

3.2.2 Phase Identification ....................................................................................... 60

3.2.3 Phase Quantification ..................................................................................... 71

3.3 Stress Measurements using X-ray Diffraction ...................................................... 72

3.4 Surface Roughness Measurements ..................................................................... 80

3.5 Swept Frequency EC Signals .............................................................................. 83

3.6 References .......................................................................................................... 87

CHAPTER 4: CONSIDERATION OF EXPERIMENTAL FACTORS .............................. 89

4.1 Microstructure ...................................................................................................... 89

4.1.1 Grain Size Measurements ............................................................................. 90

4.1.2 Dislocation density ........................................................................................ 90

4.1.3 Phase Assemblage, Size and Distribution .................................................... 91

4.2 Physical Parameters ............................................................................................ 93

4.2.1 Magnetic Permeability ................................................................................... 93

4.2.2 Surface Roughness Measurements .............................................................. 93

4.3 Instrumental Parameters ..................................................................................... 95

4.3.1 Lift off ............................................................................................................ 95

4.3.2 Instrument calibration .................................................................................... 96

4.3.3 Instrument hardware ..................................................................................... 96

4.4 References .......................................................................................................... 96


6/136

v

CHAPTER 5: ANALYSIS OF EDDY CURRENT SIGNALS ........................................... 98

5.1 Conductivity Deviation Profiles ............................................................................ 98

5.2 Determination of Near Surface Conductivity Variation ....................................... 107

5.3 References ........................................................................................................ 120

CHAPTER 6: DISCUSSION ........................................................................................ 121

6.1 Effect of heat treatment on Eddy Current .......................................................... 121

6.2 Effect of Shot peening on Eddy Current ............................................................ 123

6.3 Combined effect of heat treatment and shot peening on eddy current signals .. 124

6.4 References ........................................................................................................ 125

CHAPTER 7: CONCLUSION ...................................................................................... 126

CHAPTER 8. RECOMMENDATIONS FOR FUTURE WORK ..................................... 128

ACKNOWLEDGEMENTS ........................................................................................... 129


7/136

vi

ABSTRACTThe goal of this thesis was to determine the dependency of swept frequency eddycurrent (SFEC) measurements on the microstructure of the Ni-based alloy, Inconel 718

as a function of heat treatment and shot peening. This involved extensivecharacterization of the sample using SEM and TEM coupled with measurements andanalysis of the eddy current response of the various sample conditions using SFECdata. Specific objectives included determining the eddy current response at varyingdepths within the sample, and this was accomplished by taking SFEC measurements infrequencies ranging from 100 kHz to 50 MHz. Conductivity profile fitting of the resultingSFEC signals was obtained by considering influencing factors (such as surfacedamage). The problems associated with surface roughness and near surface damageproduced by shot peening were overcome by using an inversion model. Differences insignal were seen as a result of precipitation produced by heat treatment and by residualstresses induced due to the shot peening. Hardness of the material, which is relatedboth to precipitation and shot peening, was seen to correlate with the measured SFECsignal. Surface stress measurement was carried out using XRD giving stress in thenear surface regions, but not included in the calculations due to shallow depthinformation provided by the technique compared to SFEC. By comparing theoretical

SFEC signal computed using the microstructural values (precipitate fraction) andexperimental SFEC data, dependency of the SFEC signals on microstructure andresidual stress was obtained.


8/136

1

CHAPTER 1: INTRODUCTION

1.1 Background

Superalloys are among the most compositionally complex alloys ever developed. Theycontain many alloying elements, producing different phases which are responsible forthe mechanical properties possessed by the superalloys. Superalloys are used inextreme conditions, such as corrosive environments or high temperature structuralapplications where oxidation is problematic. Ni superalloys in particular possess themechanical properties necessary for use at high temperature operating conditions. Dueto the number of alloying elements, a wide range of alloys can be developed by slightlyvarying the composition. This allows enhancement of the properties of the alloysdepending on the intended applications 1,2.

The ability to operate gas turbine engines at high temperatures and stress was themajor focus of this industry as companies sought to develop an alloy that can withstandthese needs. Development of gas turbines began in 1872 when Dr. Frank Stolzedescribed a device similar to a gas turbine engine, while in the early 1900s Charles

Curtis produced the first working models of gas turbine engines 3. In the 1910s austeniticstainless steels became the primary choice for an alloy that could handle the high temperature applications required 3. The next major step occurred in the 1950s, whenEiselstein introduced Inconel 7184, a Nickel based alloy with an operating temperatureof 1300 F5, and is arguably the first superalloy that was introduced. Continuous

improvement has occurred since that time.

1.2 Nickel based super alloys

The classification of superalloys is based on the alloying elements present in them. Themain alloying elements are iron, cobalt and nickel. The general operating temperaturerange of superalloys is above 1000F 2. The characteristics of Nickel based superalloys


9/136

2

apart from high operating temperature are their improved mechanical properties such ashigh toughness, ductility, and low cost. The low cost of nickel iron based superalloysis obtained by adding substantial amount of iron to the alloy without compromising theproperties necessary for the desired application 1. They are widely used in applicationssuch as turbine discs or forged rotors.

Ni-based superalloys are developed by adding various alloying elements to the basemetal to achieve the desired properties. The chemical composition of Ni- basedsuperalloy consists of 10-20% Cr, 5-10% Co, 25% Mo max, with 8% Ti and Al max. Inthe case of nickel-based superalloys, the base Ni alloy is commonly alloyed with Al, Ti,

and Nb. These additions help the superalloy obtain high strength. They also formintermetallics which can melt at high operating temperatures. For aircraft engineapplications the strength-to-weight ratio needs to be considered when considering themechanical property of any alloy. In the case of Ni-base superalloys, additions such as

Al and Ti, which are low density elements, reduce the weight of the superalloys whilestill providing good mechanical properties1,2. Based on the elemental addition used andthe strengthening mechanisms provided, superalloys can be developed to provide a)solid-solution hardening, b) precipitation hardening and/or c) dispersionstrengthening 2,6. These basic mechanisms of strengthening are briefly described below.

1.2.1 Solid Solution Hardened alloys

Solid solution hardening is achieved by the addition of a different soluble element to thematrix to increase its strength. Misfit of atomic radius between the solute and the matrixresults in distortion of the atomic lattice. The subsequent strain caused by this misfit

may be either tensile or compressive in nature, depending on the size of the atomintroduced7. In either case the strain inhibits dislocation movement 7. The typicalelements used in solid-solution hardening of Ni-based superalloys are aluminum, iron,titanium, chromium, tungsten, and molybdenum. This type of hardening also decreasesthe stacking fault energy present in the crystal lattice of the alloy. Due to low stacking


10/136

3

fault energy the movements of dislocation cross slip are arrested. This prevents thedeformation occurring at high temperatures 8.

1.2.2 Dispersion Strengthened alloys

Dispersion strengthening is a mechanism where a strengthening agent is added to thealloy. The strengthening agent may be particles of an entirely different phase or materialwith properties vastly different from the matrix. For example, oxide particles added to amelt can strengthen the matrix by blocking dislocation motion present in the matrix alloy.The advantage of dispersion strengthening is that usually a suitable particle, which isinert with respect to the matrix, can be added to almost any melt. The drawback of such

adding strengthening agents is that their crystal structure is incoherent with the matrixphase. Therefore, this mechanism provides strengthening through Orowan bypassingmechanism, where dislocations move either by cutting through or bypassing theprecipitates9. This will be discussed in section 1.2.5.

1.2.3 Precipitation Hardened alloys

Precipitation hardening is achieved by generating finely distributed precipitates in thematrix from a supersaturated solid solution during heat treatment. In the case of nickel-based alloys, finely distributed precipitates are produced by adding elements such astitanium, aluminum, and niobium. As these elements have limited solubility in the alloymatrix, low temperature anneals of solid solutions enables the production of finelydistributed precipitates. Secondary phase precipitates such as - Ni3 (Ti, Al) or -Ni3Nb phase reduce the movement of dislocations and hence increase the strength. Asthe precipitates are produced within the matrix they are coherent with the matrix. In this

case as stated by Stoloff et. al, Dislocation movement is inhibited both by the strainfield surrounding coherent precipitates and the particle itself. Thus, precipitationhardening is inherently more effective than simple dispersion strengthening since twodifferent strengthening mechanisms are operative. Movement of a dislocation in thematrix containing precipitates can only take place either by cutting through or by


11/136

4

bypassing the particles 10. The drawback of precipitation hardening is that since theprecipitates develop from the matrix, heat treatment becomes critical in creating theright size and/or dispersion of precipitates for maximum strengthening. Since the alloysexamined in this study are strengthened by precipitation hardening a brief discussion ofmicrostructure is now in order.

1.2.4 Microstructure of Precipitation Hardened Ni- based alloys

Nickel has a face centered cubic (FCC) crystal structure and the major phase isdesignated as after the high temperature FCC Fe phase 2,11. In precipitationstrengthened Nickel-based super alloys containing titanium and/or aluminum, the

strengthening secondary phase precipitates are Ni 3 Al or Ni3Ti. Since these precipitatesare also FCC and ordered, both Al and Ti atoms occupy specific sites within the FCCframework and are designated as . The crystal structure of Ni3Nb precipitates asbody-centered-tetragonal (BCT) and are designated 1,2. In the early stages of heattreatment when the precipitates are still small, the precipitates are generally coherentwith the matrix, thus restricting the dislocation motions and providing maximumstrengthening effect. In such a case, movement of dislocation can only take place bycutting through or by bypassing the particles.

The alloys A-286, V-57, Nimonic 901 and Inconel 718 are the major precipitationstrengthened alloys. In these alloys is the major st rengthening precipitate 7 with ,also present in Inconel 718. Inconel 718 is a highly weldable superalloy and thepresence of is due to the addition of niobium. Thus, dual strengthening precipitatesoccur when combining the elements such as aluminum, titanium and niobium, resulting

in strengthening of the superalloy by the presence of both the secondary phases ( and) 8. The chemical composition of Inconel 718 is given by Table 1.1 5. Secondaryalloying elements are added to change the corrosion resistance, creep properties,strength, and grain refinement behavior of Ni-alloys11.


12/136

5

Inconel 718 is frequently used for cryogenic storage tanks, turbine engine components,and in corrosive environments such as oil well drill-shafting and well-head parts. Otheruses include jet rocket, nuclear fuel, and pump body components. Inconel 718 is alsoideal for hot extrusion tooling and is used for any number of parts including nuclear fuelelement spacers and high-strength bolts 2 due to its attractive combination of durabilityand weldability. Inconel 718 also can withstand a wide range of temperature extremes,which makes it a useful alloy for both cryogenic and high temperature applications 2.

Table 1.1 Chemical composition of Inconel 718.

Alloying Element Ni Cr Nb Mo Ti Al Co C Mn Si Cu Max (wt. %) - 21.00 5.50 3.30 1.15 0.80 1.00 0.08 0.35 0.35 0.30

Min (wt. %) 50.00 17.00 4.75 2.80 0.65 0.20 - - - - -

Figure 1.1 -1.4 shows several typical microstructures of Inconel 718 at various heattreatment conditions2,11. The (Ni3 Al and Ni3Ti) precipitates that form upon heattreatment to strengthen the matrix generally occur as spherically shaped precipitateswhile (Ni3Nb) tends to be disc shaped 1,2.

Figure 1.1 SEM micrograph of ring-rolled IN 718 after solution treatment at

1025 C/1h11.


13/136

6

Figure 1.2 TEM micrograph (bright field) of spray-formed IN 718, heat treated at 875

C/6 h showing early precipitation of delta plates. The larger ellipsoidal particles presentare 11.

Figure 1.3 SEM micrograph of spray-formed IN 718, heat treated at 850 C/24 h11.


14/136

7

Figure 1.4 SEM micrograph of spray-formed IN 718 after heat treatment at 950

C/50h11.

As for all precipitation hardened alloys, the actual strength reached depends on acombination of heat treatment temperature and time-at-temperature 12. The mechanicalproperties of the precipitation hardened alloys changes rapidly in the vicinity of thesolvus temperature of the strengthening precipitates 1, and so heat treatments must becrried out very carefully. A brief review of mechanical properties associated withprecipitation hardening in Ni-based alloys is provided in the next section.

1.2.5 Mechanical Properties of Precipitation Hardened Ni-based Super Alloys

High creep and creep-rupture strengths are required for applications such as turbineairfoils and disk engines2 due to the high operation temperatures and high stresses

encountered in service. Good ductility is also required to avoid fracture, which will leadto engine failure.2. Such mechanical properties and creep strength for high temperatureapplications can be obtained by precipitation hardening, where the precipitates act topin dislocations and grain boundary motion. Pinning of dislocations helps improve thetensile strength while pinning of grain boundaries improves creep resistance 1,2,9.


15/136

8

The effect of precipitate hardening depends on several factors such as:

(1) Coherency strains 1,2 that exist between the matrix () and the precipitate (, ) .

(2) Antiphase boundary (APB) energy1,2. The APB represents the energy needed forthe dislocation to cut through the ordered precipitate. Cutting of dislocations createsdisorder between the matrix and the precipitate.

(3) Volume fraction of the secondary phase precipitates 1,2 (, ) present in the matrix.

(4) The average particle size of the precipitates 1,2.

Coherency Strains: Dislocations can either cut through precipitates (smallerprecipitates) or bypass the precipitates (larger precipitates) depending on their size. Asthe precipitate size changes, the strengthening obtained by precipitate hardening willalso change. Therefore, an increase in precipitates size increases the hardening effectas long as the precipitate remains coherent, due to coherent strains and ordering.However, once the precipitates become incoherent then increase in strength withincreasing particle size is given by the Orowan equation (Equation 1.1) 13.

1.1

where is the shearing rate, is the dislocation density, is the magnitude of theburgers vector (denotes the magnitude of lattice distortion by a dislocation), and isdislocation velocity. Strength can be seen to increase but is limited by Orowan bowing,where the dislocation will bypass the particle.13.

The shape of coherent particles depends on a balance between the elastic strainenergy associated with the lattice mismatch between precipitate and matrix and by theinterfacial energy of the particle-to-matrix boundary14. The elastic strain energy depends


16/136

9

on the shape, habit, and volume of the precipitates, while interfacial energy merelydepends on the surface area of the precipitates. For a matrix with similar latticeparameters, a spherical precipitate is formed, while for large differences in latticeparameters a cuboidal precipitate is usually formed 1.

Antiphase Boundaries: Similar to coherency strains, APBs are effective in preventingdislocation motion and act as a hardening mechanism. The general strengtheningequation for solid solution strengthening was estimated by the Fleischer model 15 and isgiven by equation 1.2.

1.2

where, shear stress, - shear modulus, - total strain and - concentration of

solutes. This equation can be used to estimate the approximate stress that can bewithstood by Inconel 718 material with different solute concentrations.

Volume Fraction and Particle Size: The effect of volume fraction and particle size ofprecipitates (which determines mean particle spacing) on mechanical properties is givenby the well-known Hall-Petch Relationship (Eq 1.3)16:

1.3where yield stress, materials constant for the starting stress for dislocation

movement, Petch Parameter (unpinning constant), average grain diameter

and/or particle spacing. The degree to which hardening of material occurs depends on

the materials Hall -Petch coefficient and the degree of grain-size / particle spacingrefinement possible in the material 17. This equation helps to correlate the effect ofprecipitate spacing and concentration, present in different microstructures, to thestrength of the materials.


17/136

10

1.2.6 Failure of Ni-based superalloy parts

During service, failure of Ni-based superalloy parts due to insufficient strength of thematerial to withstand the applied load is rare. Rather, overload failures occur whenfailure of another component causes a stress to be applied at a level that far exceedsthe initial design criteria. At high operating temperatures, creep can play an importantrole in the rupture life of a Ni-based turbine blade part 2, and for this reason directionallysolidified and single crystal blades were developed. Currently the most common modeof failure is cracking / rupture due to fatigue. Fatigue is defined as a progressivestructural damage that occurs when the material is subjected to a cyclic loading that isbelow the tensile yield stress limit. Failure due to fatigue is common as virtually all alloysand parts under cyclic loading show fatigue in service applications. Low-cycle fatigue

occurs in gas turbine disks, cases, and other structures highly loaded in tension but notin creep-rupture conditions. Fatigue life is related to many factors such as defects,surface finish, applied load, and the inherent non-cyclic strength properties of an alloy,etc. Low-cycle fatigue is related to the yield strength of the alloy, while high-cycle fatigueis related more to the ultimate strength of the material. Both are related to the amount ofresidual stress present in the material since the presence of a tensile stress is one ofthe prime contributors for fatigue to occur. Successful incorporation of residual stressmeasurements in life management models will require a good understanding of thesurface and sub-surface residual stress changes 18,19 since these values play a criticalrole in accurate determination of predicted life as a function of applied stress. In stressand strain-life calculations residual stress is usually regarded as a static stress that isadded to the mean stress during each load cycle 19,20. The introduction of residualcompressive stresses to offset either residual or applied tensile stresses is thus of majorinterest and concern when evaluating the life of any part, especially life-critical partssuch as used in jet turbine engines. A summary of how compressive stresses can be

introduced and measured is given in the next section.


18/136

11

1.3 Residual Stresses

Residual stresses can be defined as stresses that exist in the absence of any externalloading or thermal gradients within the material. Manufacturing processes are the mostcommon causes of residual stress. Virtually all manufacturing and fabrication processessuch as casting, welding, machining, molding, heat treatment, plastic deformationduring bending, rolling or forging introduce residual stresses into the manufacturedobject.

While residual stresses are introduced as a result of plastic deformation they penetrate

further into the sample than the deformed region. Consider the case of shot peening.

As the surface of the material is deformed it first yields elastically, up to the yield point,then plastic deformation begins. This means the elastic region is distributed into the

material and extends well into the sample, often hundreds of microns, while the plastic

deformation may be limited to a few tens of microns on the surface. After the ball

rebounds from the surface the material tries to regain the elastic deformation that has

occurred. However, since the deformation on the surface has essentially produced a

longer distance (that must somehow be accommodated as the material trues to recover

elastically) the elastic stress distributed throughout the depth of the part near thesurface is unable to fully relax. The material is thus left with a residual compressive

stress layer, which can be very beneficial in certain applications.

1.3.1 Compressive Stresses

Surface enhancement treatments are widely applied to fracture-critical metalliccomponents such as turbine disks of aircraft engines. These treatments significantlyimprove fatigue resistance of the components by introducing protective compressiveresidual stresses in the surface layer. Surface compressive stresses are known to bebeneficial to slowing crack initiation and propagation associated with the phenomenonof fatigue9,21.


19/136

12

Improvement of cyclic fatigue performance of engineering components has beenachieved by the application of various surface treatments such as shot peening, laserpeening, ion implantation or processing methods after a component has beenmanufactured by such processes as casting, welding, forming, machining, etc. Amongthe surface treatment methods shot peening, i.e., the rapid bombardment andsubsequent shallow deformation of the work piece surface by hard projectiles (e.g.,shot) has been widely used in industry. The shot peening process induces compressiveresidual stresses on the surface, thus improving the cyclic fatigue life and decreasingthe susceptibility of the component material to corrosion cracking22-30.

1.3.2 NDE Measurement of Residual Stress As stated above, residual stress needs to be included in life prediction of manycomponents. It has been over 100 years since the first reporting of residual stressmeasurement 31 occurred. All early measurement techniques relied on destructivemeans where the part is cut to observe any dimension changes occurred in thecomponent 32. In recent decades non-destructive evaluation (NDE) diffractiontechniques, such as high energy neutron and synchrotron diffraction, have becomepowerful tools for residual stress studies. Unfortunately, an accurate nondestructivemeans of determining residual stress that can easily be applied in a manufacturingenvironment or routinely conducted as a part of regular maintenance of a componentremains elusive 18. Because of the size, availability and lack of portability of diffractiontechniques these measurements are primarily limited in application to parts in thedevelopment stage.

While X-ray diffraction (XRD) is the standard method for residual stress measurements,penetration to ~200 m in depth without surface material removal requires higher

energy X-ray sources, such as synchrotron sources or a dedicated laboratory system 33.Such instruments are expensive and, therefore, not routinely employed formanufacturing. What is desired is a process that enables accurate monitoring of


20/136

13

residual stress at life limiting critical locations using nondestructive evaluation (NDE)techniques. NDE of the surface and subsurface residual stresses is the key to fullyexploiting the benefits of the surface treatments, thereby extending the service lives ofthe components and ensuring system reliability 9.

Potential NDE techniques available for measuring residual stresses are 1) Ultrasonictesting 2) X-ray diffraction and 3) Electromagnetic methods. All these NDE techniqueshave been studied for residual stress profiling 18,29,30,34 . Ultrasonic testing can be used tomeasure bulk residual stresses 35; X-ray diffraction can be used to measure surfaceresidual stresses approximately 10-20 microns deep 36. Among electromagnetic

methods, the eddy current (EC) technique was identified as a leading candidate fornondestructive characterization of near-surface residual stress profiles 18,37-40 based onthe piezoresistivity effect, which refers to the stress-induced changes in electricalconductivity40. The basics of eddy-current measurement are discussed in the nextsection.

1.4 Nondestructive Evaluation

Nondestructive evaluation/ testing (NDE/ NDT) is used to examine an object, material orsystem without causing damage 41. NDE techniques can be divided into differenttechniques, each based on a scientific principle. The most commonly used NDEtechniques are electromagnetic testing, ultrasonic testing, magnetic particle testing,acoustic emission testing, acoustic resonance testing, infrared testing, dye penetranttesting, radiographic testing, and visual testing. Electromagnetic testing is the processof inducing electric currents and/or magnetic fields inside a test object and observingthe response. A defect in the test object may be detected where electromagneticinterference creates a measurable response, e.g. an eddy current measurement 41.


21/136

14

1.4.1 Eddy Current Measurements

The eddy current method consists of passing an alternating current through a coil so asto induce circulating currents, i.e. eddy currents, in an electrically conducting object inits vicinity. A basic eddy current device is shown schematically in Figure 1.542. Animpedance analyzer measures the impedance produced in the applied current thatresults due to the eddy currents within the object. Depending on the sample conditions(homogenous sample, cracks, etc.), the impedance signal will change. As thepenetration of the eddy current is determined by the frequency used, one can accuratelymeasure at which depth the sample loses its homogeneity. Eddy current tests can bemade on all materials that are electrically conducting. Applications include the sizing ofsurface and sub-surface cracks, measurement of the thickness of metallic plates and of

non-metallic coatings on metal substrates, assessment of corrosion and measurementsof electrical conductivities and permeabilities43.

Figure 1.5 Basic eddy current test setup.


22/136

15

Figure 1.6 A schematic diagram explaining the depth of penetration of eddy current.

Eddy currents are confined to the near surface region by the skin effect (Figure 1.6) 42.The depth that eddy currents penetrate into a material is affected by the frequency ofthe excitation current and the electrical conductivity and magnetic permeability of thespecimen. The depth of penetration decreases with increasing frequency and increasing

conductivity and magnetic permeability. Due to this skin effect, the eddy current methodcan be expected to accurately characterize surface modification at different depths inthe range of a few mm to hundreds of mm, have good spatial resolution, and providethe ability to scan large areas rapidly 43.

1.4.2 Limitations of Eddy Current Signals

While eddy current NDE has its own advantages, it also has its own limitations. Eddycurrent testing is extremely sensitive to surface variations and, therefore, requires asmooth flat surface when observations of small variations are required. A seconddrawback is that it is applicable to electrically conductive materials only. While eddycurrent testing can be used on both magnetic and non-magnetic material it is notreliable on some carbon steels for the detection of subsurface flaws as its depth of


23/136

16

penetration is limited and the measured signal depends on the frequency, cracktightness, material property and orientation of eddy current flow to the crack orlinearity42 of interest.

1.4.3 Factors Contributing to Eddy Current Signals

From the discussion above it should be clear that the impedance of an eddy currentprobe may be affected by number of factors including 1) frequency at which the eddycurrent signals are measured; 2) electrical conductivity and the magnetic permeability ofthe test objects of a material, which in turn depends on microstructure, e.g. grainstructure, presence of a second phase, work hardening, heat treatment etc.; 3) changes

in contact of the eddy current probe with the surface (i.e. lift-off) or fill-factor from probewobble (electromagnetic coupling in the case of rods, tubes, uneven surfaces); 4) thepresence of surface defects, such as cracks, and sub-surface defects such as voids andnon-metallic inclusions; 5) dimensional changes, e.g. thinning of the object that occursduring service due to wear or corrosion; 6) the presence of supports, e.g. brackets, thatmay be below the surface being examined; and 7) the presence of discontinuities suchas edges 31. These effects can be divided into two broad categories, namely, thoseinherent to the material itself (i.e. material conductivity, microstructure of the material,etc.) and those inherent with the quality of the measurement (e.g. lift-off noise,temperature of the sample when measured, etc.). These will be discussed in turn.

Material Conductivity

The conductivity of a material has a direct effect on the eddy current density: the greaterthe conductivity of a material, the greater the density of eddy currents near the surface.Conductivity in turn is affected by microstructural factors such as 1) conductivity of thepure (base) metal; (2) the conductivity and volume fractions of the phases within thematrix; (3) the scattering of electrons by small (nanometer-sized) precipitates/zones;and (4) alloying atoms dissolved in the matrix phase as stated by Blitz et al 43. All ofthese factors come into play when one considers precipitation hardening of an Al-Cu


24/136

17

alloy. This is shown in Figure 1.7. Upon heating a solutionized sample, while thematerial is still in the pre-precipitation period, one observes a decrease of conductivityfollowed by an increase as the precipitates starts growing in their size. This might bedue to the coherency strains produced by the precipitates; another possibility is thatabnormal scattering zones result when the precipitates size is of the same order as thewavelength of the electrons (~ 10) 44-50. As solute atoms then leave solution andprecipitate to harden the alloy a corresponding change in conductivity is seen again. Assolute atoms leave the matrix, strain is released, resulting in increased conductivity.However, this must be balanced by the fact that a new, second phase is being created.Thus phase creation has two effects on conductivity a) the phase itself can be expectedto have a different conductivity than the matrix, either greater, or less than; b) new

interfaces are created as the phase precipitates. The actual conductivity change is abalance of these competing factors.

Figure 1.7. Variation of aluminum conductivity with heat treatment51.

The entire progression from solution > hardened > overaged > re-solutionizing heattreatment can be monitored using conductivity / eddy current measurements and is


25/136

18

often monitored using simple hardness tests. The above example shows that whilemicrostructural changes can be monitored using eddy current measurements, the exactnature of the change occurring is unknown. A summary of studies that have employededdy currents to study / monitor various microstructural effects is given in the nextsection.

Microstructure

There have been numerous studies to evaluate microstructure changes using the eddycurrent method 52. Eddy current testing has been used to characterize microstructuralchanges such as grain size and hardness changes after thermal treatments based on

measurement of conductivity and magnetic permeability changes 53. Other examplesinclude measurement of pearlite percentage in plain carbon steels and ductile castirons54,55, surface carbon content of carburized steels 56, measurement of case depth ofcase hardened steel rods 57 and the effect of mechanical micro-hardness on impedancevariations58.

Evaluation of decarburizing depth of steels with a martensitic base microstructure hasbeen investigated using harmonic analysis 59 and magnetic Barkhausen noise (MBN)emission60,61. An investigation was made on magnetic properties as well as completeeddy current responses of decarburized steel parts. The difference in magneticproperties of the decarburized zone (with a higher percentage of ferrite microstructure)and the core of the specimen (martensite) is the basis of the eddy current method s capability to determine the decarburized depth of steel rods.

Note that the studies above do not seek to separate the various factors that could becontributing to the measured signal. For example, in precipitation hardening, the overallsignal is a sum of the conductivity increase that occurs as solute atoms leave thematrix, and the conductivity decrease that results as precipitation of a second phase


26/136

19

occurs, bringing with it a different conductivity for that phase and introducing strain thataffects the conductivity of the parent matrix. Similarly, decarburization of a surfaceresults not only in a change of microstructure from martensite to ferrite but also arelease of the residual stress inherent in the formation of the martensite. Thus, theoperating assumption is that the microstructure being studied is relatively uniform andthat the conductivity differences that exist between the various phases present in thestructure are either irrelevant or minor. When measuring the amount of decarburizationand its effect on pearlite content of a steel, for example, similar normalization treatmentson the samples is assumed to place residual stress at its minimum value and, moreimportantly, similar for all. In this study it is reported that eddy current outputs are notaffected by grain size as the average grain size for the core of steel samples is similar.

Thus, the ferrite-pearlite content change is the main factor that affects the eddy currentoutputs (impedance signals) 52.

If one wishes to know the exact contribution of any particular microstructure on the eddycurrent response, the problem becomes significantly more difficult. Such studiesrequire not only a measurement of the eddy current response, but also a carefulmonitoring of the microstructural changes that can occur during any surface treatments.Such detailed studies require expertise in both eddy current testing methods andknowledge of material microstructure along with material characterization techniques,such as scanning and transmission electron microscopy. Even in the heavily studied Fe-C system no studies of this type exist since multi-phase systems where themicrostructure can vary greatly as a function of heat treatment are especially difficult toanalyze.

The same lack of microstructural data as related to eddy-current response exists in Ni-based superalloy systems, where numerous precipitates can occur. Currently nostudies of eddy current response as a function of microstructure exist in these complex


27/136

20

systems. The study of eddy current response as a function of microstructure on Ni-based superalloy systems forms the main focus of this thesis.

Permeability

Permeability of the material has a significant influence on the eddy current response. Itis not uncommon for the permeability to vary greatly within a metal part due to localizedstresses, heating effects etc. The disadvantage of inspecting ferro-magnetic materials isthat permeability changes generally have a greater effect on the eddy current responsethan conductivity variations. Crack detection becomes difficult when permeabilitychanges randomly. To account for heterogeneity (such as cracks, non- uniformity in

microstructure) existing in the material due to the manufacturing process, a referencesample is commonly used, which is obtained from the same manufacturing processalong with the testing sample 62.

Several studies have been conducted to take magnetism into account in eddy currentanalyses. For example, Uzal et al. calculated the impedance of a cylindrical air-cored

probe over layered metallic materials whose conductivity and permeability variedcontinuously as a function of the depth 63. Also, Kasai et al. have used magnetization tocancel external magnetism 64. In this study, an external magnetic field is applied to themagnetic sample which is being tested, by using a C-core probe. The external magneticfield cancels the magnetic effect of the testing sample. By cancelling out the magnetismeffect, the eddy current responses corresponding to the conductivity of testing sampleswere studied.

While permeability can have a large effect, it is important to note that it causes problemsonly in materials that are ferro-magnetic. Thus, all irons and many steels are susceptibleto permeability problems as are certain Ni-based alloys. For example, Inconel 600displays this problem. The alloy studied in this thesis, Inconel 718, is a non-


28/136

21

ferromagnetic Ni-based alloy. Thus permeability is not expected to play a role in theanalyses conducted below.

Residual Stress

The presence of residual stress in a material can also affect the conductivity of thematerial and thus, affects the eddy current response obtained from the material.Residual stress and the conductivity of the material are related by the positivepiezoresistivity effect. The piezoresistivity effect is defined as the change in electricalconductivity of the material when mechanical stress is applied. A review of research byJavier et al showed that many authors have measured residual stress using eddy

current techniques 51. Coils can detect small stress variations in ferromagnetic steelsdue to the magnetic-elastic effect 65. Stress can be measured based on the changes inthe impedance of an electromagnetic coil 66. The impedance change occurs due tovariations in the electrical conductivity and the magnetic permeability of the test piece incomparison to a calibrated sample.

Recently, Blodgett, Nagy and Yu38,67,68

showed that the apparent conductivity of anickel-based alloy increased after shot peening. The observations by Blodgett, Nagyand Yu have cast new light on the feasibility of using eddy currents for quantitativelymeasuring residual stress of nickel-based alloys, which are extensively used in theaviation industry. Yu and Nagy have developed empirical models to relate the measuredapparent electric conductivity change (AECC) to the residual stress 69,70.

Residual stress may be developed by various surface treatment processes such asshotpeening, laser peening etc. Shot peening a surface causes plastic deformation,which induces a residual compressive stress on the surface, balanced by a residualtensile stress in the interior. The plastic deformation causes the increase in resistivity of


29/136

22

the material66. The surface compressive stresses confer resistance to metal fatigue andto some forms of stress corrosion70, which is the reason for peening in the first place.

Surface enhancement treatment by shot peening has been widely used in industrialapplications, especially for aircraft engine components, to produce compressive residualstresses. Typical peening processes use small shots of a few hundred micrometers indiameter forcefully impinging on component surfaces, resulting in compressive residualstress from the surface to a depth range of a few hundred micrometers nominally.Compressive surface residual stress is useful for improving crack initiation resistancethat prolongs service life of the part. To implement this highly desirable maintenance

strategy, an in-service nondestructive method is needed to monitor the residual stressstate of parts periodically, so that appropriate maintenance actions can be taken whenresidual-stress protection is lost (relaxing of residual stress occurs), by either replacingor re-treating the part 18.

Based on the results of 69,70, the effects of piezo-resistivity on the eddy-current responsecan be accounted for, however, the effect of microstructure is still not understood. Thisresearch work focuses on examining and understanding the effects of microstructure oneddy current response of shot peened nickel based superalloys. By understanding theeffects of microstructure, in relation to the eddy-current response obtained from thematerial as the microstructure varies, it should be possible to differentiate betweenmicrostructural effects and the effects of stress alone. Thus, microstructureobservations coupled with eddy current measurements are used in this study. Theparticular type of eddy-current measurements employed, namely, swept-field

measurements, are discussed in the next section.
http://en.wikipedia.org/wiki/Fatigue_(material)http://en.wikipedia.org/wiki/Corrosionhttp://en.wikipedia.org/wiki/Corrosionhttp://en.wikipedia.org/wiki/Corrosionhttp://en.wikipedia.org/wiki/Fatigue_(material)


30/136

23

1.5 Swept Frequency Eddy Current (SFEC) Measurements

1.5.1 Theory

Conventional eddy current measurements are performed under 10 MHz with the

smallest penetration depth of around 200 m for typical aircraft engine materials.However, there is a strong desire to determine residual stress profiles in shot-peened

engine components within 200 m from the surface. Thus higher frequency operation

with smaller penetration depths is needed 39.

A swept frequency eddy current (SFEC) system which can operate up to 50 MHz has

been developed for electromagnetic nondestructive characterization of residual stressesin shot peened aerospace materials such as nickel-based superalloys with typicalconductivities of one to several percent International Annealed Copper Standard (IACS) 39. In this approach, shot-peened surfaces are regarded as modified surface layers ofvarying conductivity, and the conductivity deviation profile is determined by inversion ofthe SFEC data 39. The instrument used is described below.

1.5.2 Instrument

The SFEC measurement system employed (shown in Figure 2.11) consists of a pair ofclosely matched printed-circuit-board coils driven by laboratory instrument undersoftware control. This provides improved sensitivity and high frequency performancecompared to conventional coils, so that swept frequency EC measurements up to 50

MHz can be made to achieve a skin depth of 80 m for nickel-based superalloys 39.

SFEC measurements up to 50 MHz or higher are prone to liftoff noise and spuriousinstrumentation effects39,40,72. This problem was circumvented here by the use of theliftoff-normalized vertical component EC signalV EXP (V-component signal, j denotes theimaginary part of the signal) defined in Equation 1.3


31/136

24

1.3

where H is the horizontal component of the signal, S T, S R, and S L are experimental,complex-valued EC signals, and correspond to, respectively, the test signal, referencesignal, and the extra lift off signal. The detailed experimental procedure for this device isdiscussed in section 2.9.

Figure 1.8 shows an example of horizontal and vertical component signals obtainedduring 5 rounds of measurement from a solutionized Inconel 718 sample. Examinationof Figure 1.8a shows the horizontal component of the signal while Figure 1.8b showsthe vertical component. The horizontal component is dependent on lift-off as it ismeasured in the direction parallel to lift-off. However, the vertical component does notdepend on the lift-off variation. This is easily seen in Fig 1.8b by conducting repetitivemeasurements.

a. b.

Figure 1.8 Eddy current signals obtained during repeated measurements from asolutionized Inconel 718 sample. a) horizontal component b) vertical component.Observed variation is due to lift-off variation.


32/136

25

Hence, the use of V EXP alone offers the advantages that it helps suppress the liftoffnoise, while avoiding mismatch due to variation across different instruments used in the

frequency range studied. It thereby provides the basis of model-based inversion ofconductivity profiles by allowing direct comparison ofV EXP with the theoretical verticalcomponent signal V TH defined in Equation 1.5.

1.5

where Z R, Z L, and Z T are the corresponding theoretical coil impedances calculated forthe reference, liftoff, and test configurations, respectively 39,72,73. An additional advantageis that V TH thus defined takes a particularly simple form under the approximation thatholds valid under our measurement conditions used. Specifically, when the relativeconductivity changes are sufficiently smaller than unity, and when the outer diameter ofthe detection coil (assumed a cylindrical air-cored coil) is much larger than any other

length parameters, such as the skin depth or coil liftoff , V TH can take an approximateform shown in Equation 1.6 where denotes the small conductivity deviation

relative to the reference conductivity Ref as a function of depth z .

1.6

Somewhat remarkably, Equation 1.5 indicates that, under the aforementionedconditions, the V-component signal is entirely independent of the coil parameters suchas its dimensions and number of windings 73. This allows us to use multiple coilsoptimized for EC measurements in different frequency bands, while yielding continuousbroad-band EC spectra so that both the near-surface conductivity profile and the bulk

Ref )( z

Ref )()( z z


33/136

26

conductivity can be determined by model-based inversion 57. The profiling capabilityoffers the opportunity to detect any pre-existing spatial variations of bulk conductivitythat are commonly found in forged components, so that their effects on the detected ECsignals can be separated from those induced by surface treatments.

By measuring SFEC signals on samples with different microstructures, and differentshot peening intensities with the same microstructure, one can quantify the effect ofmicrostructure and residual stress on the eddy current signals obtained. However, caremust be taken to ensure that signal variation due to instrumental factors does notoverwhelm the signal due to the effect being studied. This requires careful consideration

of the test methodology, in particular addressing concerns due to surface roughnessthat can produce lift-off. This is discussed in the next section.

1.5.3 Instrumental Factors Affecting Swept Frequency Eddy Current

Measurements

SFEC measurements will be affected by the same material factors that affect all eddycurrent measurements, as discussed in section 1.4.2. In addition, there are instrumentalfactors that must be considered when evaluating any eddy current response, such aslift-off and frequency.

Lift-Off: The lift-off is the impedance change that occurs when there is variation in thedistance between the inspection coil probe and the test piece. Lift-off variations can becaused by varying coating thicknesses and irregular surfaces encountered during

movement of the probe across a surface 74. In many applications, eddy currentmeasurements are adversely affected by lift-off 75. Lift-off is often considered a noisesource and it is especially undesirable in defect detection. For example, the measuredsignal in the impedance plane due to lift-off could occur in the same direction as that


34/136

27

due to a crack, thereby concealing the crack response. Therefore, the distance betweenthe probe and metal must be as constant as possible in order to avoid lift-off noise 51.

There are methods for lift-off compensation when eddy currents are used in order todetect cracks and lift-off becomes an undesired variable. Yin et al. researched dualexcitation frequencies and coil design to minimize the lift-off effect 75. Tian et al. haveresearched the reduction of lift-off effects via normalization techniques 74. In this work,lift-off is overcome by using the normalization technique as explained in section 1.5.1.

Frequency: Eddy current response is strongly affected by the frequency chosen for theinvestigation. This is the factor controlled by the operator which is chosen based on theapplication needed. Higher frequencies are chosen for surface characterization whilelower frequencies are chosen for bulk characterization. By choosing multiplefrequencies covering low to high frequencies (100 kHz to 50 MHz), different depths ( ~100 m to 500 m) for a Ni- based superalloy sample can be examined 76.

1.6 Measurement of Stress Induced by Shot Peening

X-ray and neutron diffraction methods currently are the only two standard methodsconsidered as being reliable for residual stress measurements associated with shotpeening. However, conventional XRD methods achieve relatively low penetration depth

(


35/136

28

a new measurement technique to measure residual stress using eddy current NDE 38. Inthis technique, the measured probe coil impedance is evaluated for an apparent eddycurrent conductivity (AECC). Abu- Nabah et. al 77stated that At a given frequency,

AECC is defined as the electrical conductivity of an equivalent homogeneous, non-magnetic, smooth and flat specimen placed at a properly chosen distance from the coilthat would produce the same complex electrical coil impedance as the inhomogeneousspecimen under study 77. The surface of the specimen is aligned with the scanningplane of the probe and the probe is adjusted to a constant normal lift-off distance ( =

0.1 mm). The complex impedance plane is then rotated by changing the phase angleso that the lift-off direction appears horizontal, and the vertical component of theimpedance variation is used to assess the apparent eddy current conductivity. The

adverse effects of inevitable lift-off variations during the scanning are effectivelyreduced by the choice of phase angle.

Nagys group performed eddy current measurement on three different materials:

Waspaloy, IN 100, and Ti-6Al-4V. These samples were shot peened at different Almenintensities such as 4A, 8A, 12A and 16A (Almen strips are SAE 1070 steel used tomeasure the intensity of shotpeening by using an arc measurement gauge and theintensity is measured by using a calibration graph of Almen strip arc curves vs theexposure time of shot peening). The measured AECC conductivity was found to beproportional to peening intensity. The AECC measured was inverted to obtain thecorresponding conductivity profile using the Cheng-Dodd-Deeds approximation. In this

study, Nagys group used a single coil to measure eddy current signals for the completefrequency range (1 MHz -100 MHz). Eddy current signals obtained were found todepend on the inductive effect and capacitive effect of the coil used. Lift-off effects were

observed at high frequencies like 50 MHz. The obtained apparent eddy currentconductivity is not reduced significantly by increasing dislocation density and othermicrostructural defects due to cold work.


36/136

29

Hillmann et. al 78 observed that microstructure of the material may affect the eddycurrent signal obtained. Difference in eddy current signals obtained from shotpeeneedsolutionized samples and shotpeeneed precipitate hardened samples of Inconel 718were observed. Both the samples were shot peened at the same shotpeening intensity.This led to the conclusion that microstructure of the sample may affect the signal apartfrom the residual stress induced in the samples.

Based on both these observations, two important factors need to be taken into accountwhile measuring residual stress of the samples using the eddy current technique,namely, 1) lift-off (instrument factor) and 2) microstructure (material factor). In this

thesis work, the main focus is on microstructure of the material. Lift-off studies wereperformed by another member of the same research group, and their results are usedextensively in this thesis. Details of this earlier work are given in the bibliography39.Briefly, the lift-off factor, which was a major drawback in the studies performed by

Nagys group, was overcome by using the normalized V-component and approximationtheory. This acts to eliminate any instrument dependent factors such as specific coildesign, and measurement dependent factors such as lift-off.

By using the swept frequency eddy current technique developed here 39, combined withusing normalized V-component and approximation theory, the ability to measure theeddy current signals from the frequency of 100 kHz to 50 MHz is possible. The obtainedSFEC signals can be inverted into corresponding conductivity profiles using Cheng-Dodd-Deeds theory and approximation theory. By using Inconel 718 (a Ni- basedsuperalloy) samples of various heat treatments which are solutionized, under-aged,

peak aged and over-aged combined with different shotpeening intensities like 4A, 8Aand 12A will enable study of the microstructure dependency of the material and residualstress induced in the material.


37/136

30

1.7 Problem Statement

This thesis is the first attempt made to separate the effects the microstructure andresidual stress (induced by shot peening) using eddy current measurement technique inthe Nickel based superalloy Inconel 718. This problem is approached by understandingthe microstructure of the material at different heat treatments and the effects of shotpeening at different shot peening intensities.

This dissertation discusses how the eddy current signals are affected by microstructureof the samples, and a systematic investigation has been carried out to investigate theseeffects. Specifically, the effects of precipitation as a function of heat treatment and

damage due to shot-peening on the eddy current signal were studied. A set of Inconel718 samples, heat treated to possess different amounts of precipitation, were chosenfor the study. This was done for several reasons. Firstly, Inconel 718 is a major alloyusing in turbine engines, and thus is of great importance. Secondly, Inconel 718 is non-ferromagnetic, which means the permeability factor may be excluded from the eddycurrent responses measured in this investigation. Thirdly, residual stresses in Inconel718 have been measured using commercial XRD techniques and residual stress

profiles have been measured using surface layer removal79-82

, giving some idea of thelevel of stresses that may be encountered.

Microstructure of the samples was studied using various microscopy techniques likeScanning electron microscopy and Transmission electron microscopy. Microstructure ofthe samples was quantified in terms of their secondary phase precipitates. This wasapproached by using the micrographs obtained from the above mentioned materialscharacterization techniques. At various levels of heat treatment the size and volumepercentage of strengthening precipitates can be expected to change. Therefore, thevolume fraction of the secondary phase precipitates present in the Inconel 718 sampleschosen for the study needs to be determined with some degree of reliability in order torelate microstructure to eddy current response. The correlation of eddy current signals


38/136

31

with the corresponding samples precipitates (quantified), helped to determine theimpact of microstructure on the eddy current signals apart from residual stress.

Samples were shot peened at three different intensities and swept frequency eddycurrent signals were measured before and after shot peening. For the eddy currentmeasurements different frequencies were used so that different depths of the samplecould be explored, encompassing both the shot peened layer, the center bulk of thesample, and the bottom region where an edge is present due to the nearby free surface.

The lift-off effect was observed in the eddy current measurement technique used byNagy group 38 and it was more pronounced at higher frequencies above 10 MHz. Thislift-off effect was cancelled out by using the normalized V-component signal, which isnot affected by the lift-off39. The other problem that might be faced is edge effects whileusing eddy current technique. In this study, comparison of the free edge results to thebulk and shot peened results allowed edge effects to be accounted for by making surethe measurements are always performed in the center of the sample to avoid edgeeffects. By comparing the swept frequency eddy current signals to the extensivemicrostructure characterization allowed a calibration curve to be obtained relatingconductivity profiles to microstructure.

1.8 References

1 Geddes, B., Leon, H. & Huang, X. Superalloys: Alloying and Performance .(ASM International, 2010).

2 Donachie, M. J. & Donachie, S. J. Superalloys: a technical guide . (ASMInternational (OH), 2002).

3 C.T.Sims. in Superalloys II (ed N.S.Stoloff C.T.Sims, W.C.Hagel) Ch.Superalloys: Genesis and Character, 3-26 (Wiley-Interscience, John Wileyand Sons, 1987).


39/136

32

4 Holt, R. & Wallace, W. Impurities and trace elements in nickel-basesuperalloys. International Metals Reviews 21 , 1-24 (1976).

5 High Temp Metals , , Last accessed on June12,2013.

6 Material Datasheets , (2004), Last accessed onJune 12, 2013.

7 Madeleine, D. C.,The microstructure of superalloys, Gordan and BreachScience Publishers, The Netherlands ( 1997).

8 Kumar, P. Role of niobium and tantalum in superalloys. Advances in HighTemperature Structural Materials and Protective Coatings , 34-53 (1994).

9 Randy, B. in 9th International symposium on Superalloys, Superalloys: A

Primer and History (The Minerals, Metals & Materials Society, 2000).10 N.S.Stoloff. in Superalloys II (ed N.S.Stoloff C.T.Sims, W.C.Hagel) Ch.

Fundamentals of Strengthening, 61-96 (Wiley-Interscience, John Wiley andSons, 1987).

11 Saied Azadian,Liu-Ying Wei,Richard Warren, Materials Characterization, 2004, 53(4), pp. 7-16,License Number 3166600235080, License date 6-12-2013.

12 Thomas G. Digges, S. J. R., and Glenn W. Geil. Heat Treatment andProperties of Iron and Steel. (United States Department of Commerce,National Bureau of Standards Monograph 88, 1966).

13 Orowan, E. Dislocations in Metals, AIME, Newyork, 1954.14 Ko, H., Paik, K., Park, L., Kim, Y. & Tundermann, J. Influence of rhenium on

the microstructures and mechanical properties of a mechanically alloyedoxide dispersion-strengthened nickel-base superalloy. Journal of materialsscience 33 , 3361-3370 (1998).

15 Dieter, G. E. & Bacon, D. Mechanical metallurgy . Vol. 3 (McGraw-Hill NewYork, 1986).

16 Courtney, T. H. Mechanical behavior of materials . (McGraw-Hill, 1990).17 Davis, J. R. Nickel, cobalt, and their alloys . Vol. 8 (ASM International (OH),

2000).


40/136

33

18 John, R., Larsen, J. M., Buchanan, D. J. & Ashbaugh, N. E. Incorporatingresidual stresses in life prediction of turbine engine disks. (DTIC Document,2003).

19 Bannantine, J. A., Comer, J. & Handrock, J. Fundamentals of Metal Fatigue Analysis. 1990. Prentice-Hall, Inc. BIOGRAPHIES Eric A. Ware GuidantCorporation 4100 , 55112-55798.

20 Morrow, J. Fatigue design handbook. Advances in engineering 4 , 21-29(1968).

21 Azadian, S., Wei, L. Y. & Warren, R. Delta phase precipitation in Inconel 718.Materials characterization 53 , 7-16 (2004).

22 Zafred, P. R.(EP Patent 0,218,354, 1990).

23 Lari, A. N.-. in First International Conference on Shot Peening 1-21(Pergamon Press Ltd, Paris, France, 1981).

24 Marsh, K. J. Shot peening: techniques and applications. EngineeringMaterials Advisory Service Ltd.(United Kingdom), 1993 , 320 (1993).

25 De Los Rios, E., Walley, A., Milan, M. & Hammersley, G. Fatigue crackinitiation and propagation on shot-peened surfaces in A316 stainless steel.International journal of fatigue 17 , 493-499 (1995).

26 Hammond, D. & Meguid, S. Crack propagation in the presence of shot-peening residual stresses. Engineering fracture mechanics 37 , 373-387(1990).

27 G.S. Was, R. M. P., M.C.Frabolot. in First International Conference on ShotPeening , 445-451 (Pergamon Press Ltd, Paris, France, 1981).

28 Luo, W., Noble, B. & Waterhouse, R. The Interaction Between Shot-Peeningand Heat Treatment on the Fatigue and Fretting-Fatigue Properties of theHigh Strength Aluminium Alloy 7075.Impact Surface Treatment. Elsevier

Applied Science Publishers , 57-67 (1986).29 Oshida, Y. & Daly, J. Fatigue damage evaluation of shot peened high

strength aluminum alloy . (Elsevier Applied: New York, NY, USA, 1990).30 Al-Obaid, Y. A Rudimentary An. alysis of Improving Fatigue Life of Metals by

Shot-Peening. (1990).


41/136

34

31 E. Heyn , O. B. On Stresses in cold drawn metal. International Journal ofMetallography 1 , 16-50 (1911).

32 Ueda, Y. in Handbook of Measurement of Residual Stresses, Chapter 4.Sectioning Methods, (Society for Experimental Mechanics, 1996).

33 Al-Shorman, M. Y. Nondestructive residual strain measurement using highenergy x-ray diffraction Ph.D. thesis, Iowa State University, (2008).

34 Campbell, F. C. Elements of metallurgy and engineering alloys . (ASMInternational (OH), 2008).

35 Kudryavtsev, Y. & Kleiman, J. Measurement of Residual Stresses in WeldedElements and Structures by Ultrasonic Method. International Institute ofWelding. IIW Document XIII-2339-10 (2010).

36 Bruno, G. & Dunn, B. Surface and bulk residual stress in Ti6Al4V weldedaerospace tanks. Journal of pressure vessel technology 126 , 284-292 (2004).

37 Zilberstein, V. et al. Residual and applied stress estimation from directionalmagnetic permeability measurements with MWM sensors. Journal of pressurevessel technology 124 , 375-381 (2002).

38 Blodgett, M. P. & Nagy, P. B. Eddy current assessment of near-surfaceresidual stress in shot-peened nickel-base superalloys. Journal ofNondestructive Evaluation 23 , 107-123 (2004).

39 Shen, Y., Lee, C., Lo, C., Nakagawa, N. & Frishman, A. Conductivity profiledetermination by eddy current for shot-peened superalloy surfaces towardresidual stress assessment. Journal of Applied Physics 101 , 014907-014907-014910 (2007).

40 Abu-Nabah, B. A., Yu, F., Hassan, W. T., Blodgett, M. P. & Nagy, P. B. Eddycurrent residual stress profiling in surface-treated engine alloys.Nondestructive Testing and Evaluation 24 , 209-232 (2009).

41 Cartz, L. ,Nondestructive Testing, ASM International, OH (1995).42 The Colloboration for NDT/NDE Education led by the Center for NDE, Iowa

State University, www.ndt-ed.org. Last accessed on June 20,2013.43 Blitz, J. Electrical and magnetic methods of non-destructive testing . Vol. 3

(Springer, 1997).


42/136

35

44 Starink, M. & Li, X. A model for the electrical conductivity of peak-aged andoveraged Al-Zn-Mg-Cu alloys.Metallurgical and Materials Transactions A 34 ,899-911 (2003).

45 Federighi, T. & Passari, L. Anomalous increase of resistivity during ageing ofaluminium-silver alloys. Acta metallurgica 7 , 422-424 (1959).

46 Rossiter, P. L. The electrical resistivity of metals and alloys . (CambridgeUniversity Press, 1991).

47 Hillel, A., Edwards, J. & Wilkes, P. Theory of the resistivity and Hall effect inalloys during Guinier-Preston zone formation. Philosophical Magazine 32 ,189-209 (1975).

48 Hillel, A. & Edwards, J. The ageing of Al-Zn theoretical analysis of the

resistivity data. Philosophical Magazine 35 , 1231-1237 (1977).49 Luiggi, N., Simon, J. & Guyot, P. Residual resistivity during clustering in Al

Zn solid solutions. Acta Metallurgica 28 , 1115-1122 (1980).50 Guyot, P. P. & Bauer, A. Calcul de la rsistivit et du pouvoir

thermolectrique des alliages en cours de dmixtion. Philosophical MagazineB 46 , 345-355 (1982).

51 Garca-Martn, J., Gmez-Gil, J. & Vzquez-Snchez, E. Non-destructive

techniques based on eddy current testing. Sensors 11 , 2525-2565 (2011).52 Kahrobaee, S., Kashefi, M. & Saheb Alam, A. Magnetic NDT Technology for

characterization of decarburizing depth. Surface and Coatings Technology

205 , 4083-4088 (2011).53 Zergoug, M., Lebaili, S., Boudjellal, H. & Benchaala, A. Relation between

mechanical microhardness and impedance variations in eddy current testing.NDT & E International 37 , 65-72 (2004).

54 Khan, S., Ali, F., Khan, A. N. & Iqbal, M. Pearlite determination in plain carbonsteel by eddy current method. Journal of Materials Processing Technology 200 , 316-318 (2008).

55 Konoplyuk, S., Abe, T., Uchimoto, T., Takagi, T. & Kurosawa, M.Characterization of ductile cast iron by eddy current method. NDT & EInternational 38 , 623-626 (2005).


43/136

36

56 Sheikh Amiri, M. & Kashefi, M. Application of eddy current nondestructivemethod for determination of surface carbon content in carburized steels. NDT& E International 42 , 618-621 (2009).

57 Sun, H., Bowler, J., Bowler, N. & Johnson, M., AIP Conference Proceedings615, pp 1561- 1568 (2001), doi :http://dx.doi.org/10.1063/1.1472979.

58 Lo, C., Frishman, A. M., Shen, Y. & Nakagawa, N. Self-Calibrated SweptFrequency Eddy Current Measurements for Surface Characterization. Studiesin Applied Electromagnetics and Mechanics 32 , 99 (2009).

59 Mercier, D., Lesage, J., Decoopman, X. & Chicot, D. Eddy currents andhardness testing for evaluation of steel decarburizing. NDT & E International

39 , 652-660 (2006).

60 Vaidyanathan, S., Moorthy, V., Jayakumar, T. & Raj, B. Evaluation ofinduction hardened case depth through microstructural characterisation usingmagnetic Barkhausen emission technique. Materials science and technology

16 , 202-208 (2000).61 Blaow, M., Evans, J. & Shaw, B. Surface decarburisation of steel detected by

magnetic barkhausen emission. Journal of materials science 40 , 5517-5520(2005).

62 Tian, G., Zhao, Z. & Baines, R. The research of inhomogeneity in eddycurrent sensors. Sensors and Actuators A: Physical 69 , 148-151 (1998).

63 Uzal, E. & Rose, J. The impedance of eddy current probes above layeredmetals whose conductivity and permeability vary continuously. Magnetics,IEEE Transactions on 29 , 1869-1873 (1993).

64 Kasai, N., Ogawa, S., Oikawa, T., Sekine, K. & Hasegawa, K. Detection ofcarburization in ethylene pyrolysis furnace tubes by a C core probe withmagnetization. Journal of Nondestructive Evaluation 29 , 175-180 (2010).

65 Ricken, W., Schoenekess, H. & Becker, W. J. Improved multi-sensor for forcemeasurement of pre-stressed steel cables by means of the eddy currenttechnique. Sensors and Actuators A: Physical 129 , 80-85 (2006).

66 Ricken, W., Liu, J. & Becker, W. J. GMR and eddy current sensor in use ofstress measurement. Sensors and Actuators A: Physical 91 , 42-45 (2001).


44/136

37

67 Yu, F., Blodgett, M. & Nagy, P. Eddy Current Assessment of Near-SurfaceResidual Stress in Shot-Peened Inhomogeneous Nickel-Base Superalloys.Journal of Nondestructive Evaluation 25 , 16-27, doi:10.1007/s10921-006-0003-8 (2006).

68 Blodgett, M. P. & Nagy, P. B., AIP Conference Proceedings 700, pp 1216-1223 (2003), doi: http://dx.doi.org/10.1063/1.1711756.

69 Yu, F. & Nagy, P. B. Numerical method for calculating the apparent eddycurrent conductivity loss on randomly rough surfaces. Journal of applied

physics 95 , 8340-8351 (2004).70 Yu, F. & Nagy, P. B. Simple analytical approximations for eddy current

profiling of the near-surface residual stress in shot-peened metals. Journal of

applied physics 96 , 1257-1266 (2004).71 Wick, C. Tool and Manufacturing Engineers Handbook: Materials, finishing

and coating . Vol. 3 (SME, 1985).72 Nakagawa, N., Lee, C. & Shen, Y, AIP Conference Proceedings 820, pp.

1418- 1425 (2005), doi: http://dx.doi.org/10.1063/1.2184690.

73 Lee, C., Shen, Y., Lo, C. & Nakagawa, N., AIP Conference Proceedings 894, pp. 1213-1220 (2006), doi: http://dx.doi.org/10.1063/1.2718104..

74 Tian, G. Y., Li, Y. & Mandache, C. Study of lift-off invariance for pulsed eddy-current signals. Magnetics, IEEE Transactions on 45 , 184-191 (2009).

75 Yin, W., Binns, R., Dickinson, S., Davis, C. & Peyton, A. in Instrumentationand Measurement Technology Conference, 2005. IMTC 2005. Proceedings of

the IEEE. 1779-1784 (IEEE).76 Hansen, J. The eddy current inspection method. Insight 46 , 279 (2004).77 Abu-Nabah, B. A., Hassan, W. T., Ryan, D., Blodgett, M. P. & Nagy, P. B.

The Effect of Hardness on Eddy Current Residual Stress Profiling in Shot-

Peened Nickel Alloys. Journal of Nondestructive Evaluation 29 , 143-153(2010).

78 Hillmann, S., Heuer, H., Baron, H. U., Bamberg, J.,Yashan, A., andMeyendor, N., AIP Conference Proceedings 1096, pp. 1349-1355 (2008), doi:http://dx.doi.org/10.1063/1.3114113.


45/136

38

79 Chang, H., Schoenig, F. C. & Soules, J. A. Eddy current offers a powerful toolfor investigating residual stress and other metallurgical properties. Materialsevaluation 57 , 1257-1260 (1999).

80 Arunachalam, R., Mannan, M. & Spowage, A. Residual stress and surfaceroughness when facing age hardened Inconel 718 with CBN and ceramiccutting tools. International Journal of Machine Tools and Manufacture 44 ,879-887 (2004).

81 Lavella, M. & Berruti, T. Residual Stress on Inconel 718 Turbine Componentsafter Machining. Key Engineering Materials 417 , 601-604 (2010).

82 Cai, D. et al. Precipitation and residual stress relaxation kinetics in shot-peened Inconel 718. Journal of Materials Engineering and Performance 15 ,

614-617, doi:10.1361/105994906x124613 (2006).


46/136

39

CHAPTER 2: EXPERIMENTAL PROCEDURE

2.1 Sample Preparation

A series of Inconel 718 samples were chosen for the study. Six samples of sizes 0.250x 3.00 x 5.00 solution heat treated (1040 C) for 1 hour and air cooled and one sample

of size 0.250 x 3.00 x 5.00 stock Inconel 718 per ASM 5596 spec ification waspurchased from Advanced Alloys, Deer Park, New York. The samples were cut into

0.250 X 1.5 X 2.5 coupons using electric discharge machining (EDM). Different heattreatment conditions were selected using the time temperature-transformation (TTT)diagram1 shown in Figure 2.1 to produce various microstructures with different amounts

and sizes of secondary precipitates.

Figure 2.1 Time-Temperature-Transformation for Inconel 718 1. Red color dots indicatethe sample conditions chosen for the investigation. Red line for condition V indicatesthat condition V is subjected to double aging at two temperatures indicated by red dots.


47/136

40

Sample I (AR) was kept in the as-received condition. To ensure a homogeneous

starting microstructure all the other samples were again solutionized at 1024 C for 0.5

hour using a controlled atmosphere GCA box furnace and then furnace cooled toensure homogeneity. Thus, Sample II (SHT) is designated as the solution heat treatedsample although it in actually was solutionized twice. The remaining samples wereaged under different conditions as given by Table 2.1. In accordance with the TTTdiagram of Figure 2.1, Samples III and IV are expected to be underaged, and aredesignated UA1 and UA2, respectively. Sample V is expected to be in the peak agedcondition and is designated PA, while Sample VI and VII should be overaged and aredesignated OA1 and OA2, respectively. It should be noted that Sample V underwent atwo-step aging process, being held at initially at 718 C then furnace cooled to 621 C.

This means that Sample V actually was held in two different regions of the TTT curveshown in Figure 2.1. This is reflected on the figure by showing two points for Sample V

joined by a line.

Table 2.1 Heat-treatment conditions for the set of Inconel 718 samples 1.

Sample Heat Treatment

I (AR) As-received (solutionized)

II (SHT) Solutionized at 1024C/0.5hr

III (UA1) Solutionized, aged at 620C/10hrs

IV (UA2) Solutionized, aged at 680C/50hrs

V (PA)Solutionized, aged at 718C/8hrs,furnace-cooled to 621C, aged at

621C/8hrs

VI (OA1) Solutionized, aged at 850C/10hrs

VII (OA2) Solutionized, aged at 900C/20hrs


48/136

41

2.2 Shot peening

Each heat-treated sample was divided into four quarters (Figure 2.2) for the shotpeening study. One quarter was kept in the pristine condition and was used as thereference for swept frequency eddy current (SFEC) measurements, while the otherthree were shot peened at Almen intensities of 4A, 8A and 12A with 100% coverage

using ceramic shot (Zirshot Z850 , Saint-Gobain ZirPro) with average diameter of 0.8

mm. Shot peening was conducted using a Trinc Dry blast instrument, standard model24/BP. The shot peen intensities were measured using standard shot peening control

A strips bought from Electronics, Inc.

Figure 2.2 Shot peened Inconel 718 sample with different Almen intensities (4A, 8A and12A).

2.3 Electrical Conductiv