Effect of machining parameters and heat treatment on the

EffectofmachiningparametersandheattreatmentontheresidualstressdistributionintitaniumalloyIMI-834

B.R.Sridhar∗,G.Devananda,K.Ramachandra,RamarajaBhat

GasTurbineResearchEstablishment,Bangalore560093,India

Abstract

TheresidualstressvariationintitaniumalloyIMI-834asafunctionofdepthfollowingmillingatdifferentfeeds,speedsanddepthsofcutwasdeterminedbyastrain-gaugetechniqueinvolvingblind-holedrilling.Theresidualstresseswerefoundtobecompressiveinnatureandtobedependentuponthemillingparameters.Heattreatmentwasfoundtorelievetheresidualstresses,thedegreeofstressreliefbeingfoundtoincreasewithincreaseintemperature.Optimumtemperaturesweredeterminedatwhichsignificantrelaxationoccurredwithoutadverselyaffectingthemicrostructureandmechanicalbehaviourofthematerial.©2003ElsevierB.V.Allrightsreserved.

Keywords:IMI-834;Strain-gaugetechnique;Residualstress

1.Introduction

IMI-834isanear-␣titaniumalloywiththenominalcom-positionTi–5.8Al–4.5Sn–4Zr–0.7Nb–0.5Mo–0.40Si–0.06C.Thisisacreepresistantalloywithaservicetemperatureupto600◦Candisprimarilyintendedforgasturbineaero-engineapplicationssuchasdiscs,ringsandblades.Residualstressesplayanimportantroleindeterminingthefatiguelifeofacomponent.Thiscallsforrelievingthesestressesinsuchcaseswhereanadverseeffectisseenonthefatiguelife.

Thispaperhighlightsthevariationinthemagnitudeanddistributionofresidualstresseslockedupinthematerialfollowingamillingoperationatdifferentspeeds,feedsanddepthsofcut.Theeffectsofvaryingtemperatureinrelievingthesestresseshavealsobeendiscussedinthepaper.

Residualstressesweremeasuredbythestandardstrain-gaugetechniqueinvolvingblind-holedrilling[1–7].Themethodcallsfortheuseofastrain-gaugerosetteofthesize25mm×25mm(shownschematicallyinFig.1)suppliedbyMicro-measurementGroup,USA[5].Ifε1,ε2andε3arethevaluesoftherelievedstrainnotedfromstraingauges1,2and3,respectively,ofthestraingaugerosetteusedduringblind-holedrilling,theprincipalresidualstressesσmaxandσminarecalculatedasbelow:

σmax

ε3+ε1=−

4Aε3+ε1

+4A

󰀁󰀁

(ε3−ε1)2+(ε3+ε1−2ε2)24B

(1)(2)(3)

σmin=

(ε3−ε1)2+(ε3+ε1−2ε2)24B

tan2α=where

ε3+ε1−2ε2

ε3−ε1

¯b

B=−

2E

(1+ν)a¯A=−

2E

and(4)

inwhichαistheanglemadebyσmaxwithrespecttogauge

¯arethe1,νthePoisson’sratioofthematerialanda¯andb

uniformstresscoefficientsdeterminedbySchajerbyfiniteelementstudies,andavailableasafunctionofZ/D(holedepth/holediameter).

Theresidualstressesinthecuttingdirection(σc)andnor-maldirection(σn)werecalculatedusingtheexpressions:σc+σn=σmax+σmin

σc−σn=(σmax−σmin)cos2θ

whereθistheanglemadebyσcwithσmax.2.Experimentalprocedures

(5)(6)

Correspondingauthor.

E-mailaddress:sridharbr@mail.gtre.org(B.R.Sridhar).

0924-0136/03/$–seefrontmatter©2003ElsevierB.V.Allrightsreserved.doi:10.1016/S0924-0136(03)00612-5

∗

Hot-rolled,solution-treatedandgroundpickled50mmdiameterbarsweremachinedinto38mm×25mm×15mm

B.R.Sridharetal./JournalofMaterialsProcessingTechnology139(2003)628–634629

Fig.1.Strain-gaugerosettearrangementformeasuringresidualstresses.

rectangulartest-pieces.Inordertoremovetheresidualstressesinducedduringthefabricationofspecimens,thetest-pieceswerestressrelievedbykeepingthemat600◦Cfor1hinavacuumfurnacefollowedbyanargongasquench.

ThemachiningoperationswerecarriedoutonFN-2VHMTmillinglathe,5HPcapacity,usingaTMaxK-20facemillingcutterof50mmdiameterandfourTN-35-Mtitaniumnitridecoatedinserts(TPAN-1603PPN,WIDIA).ServoCut‘S’1:20solubleoilwasemployedasacuttingfluid.Thetest-piecesweremilledatdifferentmillingparametersbyvaryingthespeed(v,m/min),feed(f,mm/tooth/rev.)anddepthofcut(d,mm)asgiveninTable1.

Test-piecesmilledwiththeparameters,v=11m/min,f=0.056mm/tooth/rev.,d=0.25mmwerestressrelievedat400,500and600◦Cfor1hinavacuumfurnacefollowedbyanargongasquench.

Boththemachinedandheat-treatedtest-piecesweresub-jectedtoresidualstressmeasurementbyblind-holedrilling.

Table1

MillingparametersSl.no.v(m/min)f(mm/tooth/rev.)d(mm)111.000.0560.250211.000.1000.250356.000.0560.250456.000.1000.250511.000.0562.000611.000.1002.000756.000.0562.000856.000.1002.000935.000.0751.12510

35.00

0.075

1.125

Anairturbineassemblywasusedtodrillablindholeatthecentreofthestrain-gaugerosette(Fig.1)usedintheexper-iment.Drillingwascarriedoutatincrementaldepthsandateachdepththerelievedstrains,ε1,ε2andε3(ofthestraingauges1,2and3fromthestrain-gaugerosette)werenoted.ThesestrainswereusedintheevaluationofresidualstressesbytheuseofEqs.(1)–(6).

3.Results

Figs.2–9presentplotsoftheresidualstressdistributionasfunctionsofdepthbelowthesurfacefordifferentmillingparametersemployedinthepresentstudy.Negativevaluesofthemaximumandminimumresidualstressesindicatethattheresidualstresseswerecompressiveinnature.Themaximumstressesinthecuttingandnormaldirectionswerefoundtobeatadepthofabout0.1mmbelowthesurface.3.1.Effectoffeed(constantvandd)

Atlowspeed(11m/min)theresidualstresseswerefoundtodecreasewithincreaseoffeed(cf.Figs.2and3).However,athighspeed(56m/min)theywerefoundtoincreasewithincreaseinfeed(cf.Figs.4and5;Figs.8and9).Thepeakresidualstresswasalsoobservedtoslightlyshifttowardsthesurfacewithdecreaseinfeed(cf.Figs.2and3;Figs.4and5;Figs.6and7).

3.2.Effectofcuttingvelocity(constantfandd)

Themagnitudeofthecompressivestressesincreasedwithincreaseincuttingvelocity(cf.Figs.3and5;Figs.7and9).However,themagnitudewasfoundtodecreasewithincreaseincuttingvelocityatlowvaluesoffeedanddepthofcut(cf.Figs.2and4).Also,withincreaseincuttingspeed,thepeakresidualstresswasfoundtoshifttothesurface(Figs.3and5;Figs.6and8).Further,inafewcasesthedepthoftheresiduallystressedlayerincreasedwithincreaseinthemagnitudeoftheresidualstress.

3.3.Effectofdepthofcut(constantvandf)

Themagnitudeoftheresidualstressesdecreasedwithincreaseindepthofcut(cf.Figs.2and6;Figs.3and7;Figs.4and8;Figs.5and9).Thepeakresidualstresswasfoundtoshiftslightlytowardsthesurfacewithincreaseindepthofcut.Noobvioustrendwasseenpertainingtothedepthoftheresiduallystressedlayer.3.4.Effectofheattreatment

Figs.10–12revealtheeffectofheattreatmentatdifferenttemperatures(400–600◦C)onthemagnitudeanddistribu-tionoftheresidualstresses.Itwasfoundthatthemagnitudeaswellasthedepthoftheresiduallystressedlayerdecreased

630B.R.Sridharetal./JournalofMaterialsProcessingTechnology139(2003)628–634

Fig.2.Variationofresidualstressesduetoverticalfacemillingasafunctionofdepth.IMI-834;cuttingspeed,11m/min;feed,0.056mm/tooth/rev.;depthofcut,0.25mm.

withincreaseintemperature.Thedegreeofrelaxationwas60%at400◦C,75%at500◦Cand90%at600◦C.There-laxationratewasveryhighintheinitial0.25mmofdepthbutthereafteritdecreased,thedegreeofrelaxationbeingalmostlinearwithdepth(Figs.13and14).

4.Discussion

TheresultsobtainedforIMI-834subsequenttoamillingoperationwereconsistentwiththoseobtainedfortitaniumalloys,IMI-318(Ti–6Al–4V)andIMI-685(Ti–6Al–5Zr–0.5Mo–0.25Si)inthemachinedandpolishedcondition[6].Shiftofthepeakresidualstresstothesurfaceindicatedanincreaseinthedegreeofcoldwork[8,9].How-

ever,thisiscontradictedbyadecreaseinthepeakresidualstress,whichisalsosupposedtoincreasewithincreaseinthelevelofcoldworkinthesurfacelayers[8,9].Theseindicatethatthemillingparameterssuchasspeed,feedanddepthofcuthavearoletoplayindeterminingtheextentofcoldworkduringamachiningoperation.Theeffectofthemachiningparametersontheresultingresidualstressandotherresponsevariableshasbeenstudiedtheoreticallybydifferentauthors[10,11].Accordingly,thedependenceoftheresidualstressesonthemillingparameterscouldbeexplainedbythefollowinglinearequation:σ=b0+b1v+b2f+b3d

(7)

InEq.(7),coefficientsb0,b1,b2andb3canbedeterminedbyexperimentaldatarelatingtheresidualstressestothemillingparameters.Asperthisequationtheresidualstressesvary

Fig.3.Variationofresidualstressesduetoverticalfacemillingasafunc-tionofdepth.IMI-834;cuttingspeed,11m/min;feed,0.1mm/tooth/rev.;depthofcut,0.25mm.Fig.4.Variationofresidualstressesduetoverticalfacemillingasafunctionofdepth.IMI-834;speed,56m/min;feed,0.056mm/tooth/rev.;depthofcut,0.25mm.

B.R.Sridharetal./JournalofMaterialsProcessingTechnology139(2003)628–634631

Fig.5.Variationofresidualstressesduetoverticalfacemillingasafunctionofdepth.IMI-834;cuttingspeed,56m/min;feed,0.1mm/tooth/rev.;depthofcut,0.25mm.

linearlywithrespecttoanyparameterprovidedthatthereisconstancyoftheothertwo.Forexample,atthesamefanddthemagnitudeofresidualstressincreasesordecreases,respectively,withincreaseordecreaseofv.However,asex-plainedearlierandaspertheresidualstressdatapresentedinTable2,thistrendwasnotshownconsistently.Thisin-dicatedthatthemillingparameters,inadditiontoactingin-dependently,arelikelytointeractwitheachotherindeter-miningtheresidualstressstateofthematerial.Addingsomeinteractiveterms[10,11],Eq.(7)couldbemodifiedas:σ=b0+b1v+b2f+b3d+b4vf+b5vd+b6fd+b7vfd

(8)

Eq.(8)wassolvedusingthegenerateddatafortheresidualstressesinthecutting(σct)andnormal(σnt)directionsandrelationshipswereobtainedasgivenbelow:

σct=−1784.2+24.9v+17392.6f+438.1d−393vf

−3.4vd−4593.8fd+69.7vfd

σnt=−1782.5+18.5v+16221.1f+3.8d

−293.9vf−vd−3797fd+38.6vfd

(10)(9)

TheresidualstressesinthecuttingandnormaldirectionswereevaluatedusingEqs.(9)and(10),beinghighlighted

Fig.6.Variationofresidualstressesforverticalfacemillingasafunctionofdepth.IMI-834;cuttingspeed,11m/min;feed,0.056/tooth/rev.;depthofcut,2.0mm.Fig.7.Variationofresidualstressesduetoverticalfacemillingasafunc-tionofdepth.IMI-834;cuttingspeed,11m/min;feed,0.1mm/tooth/rev.;depthofcut,2.0mm.

632B.R.Sridharetal./JournalofMaterialsProcessingTechnology139(2003)628–634

Fig.8.Variationofresidualstressesduetoverticalfacemillingasafunc-tionofdepth.IMI-834;cuttingspeed,56m/min;feed,0.056mm/tooth/rev.;depthofcut,2.0mm.

inTables3and4.Experimentalandcalculatedvalueswerefoundtomatchreasonablywithminimalerrors.

Residualstressrelaxationintitaniumwasfoundtostartatabout280◦C[12].Stressrelaxationshavebeenassoci-atedwiththeonsetofmicro-plasticdeformationsresultingintherearrangement/annihilationofdislocations[13].Anyprocessessuchasprecipitationscanadverselyaffectthesephenomenaandreducetherelaxationrates.

Relaxationsobservedat400,500and600◦CinIMI-834wereconsistentwiththegenerallyobservedtrend.TitaniumalloyIMI-834issimilartotitaniumalloyIMI-685thatwasfoundtobepronetosilicideprecipitation[14].Theincreaseinstressrelaxationwithincreaseintemperatureindicatedthatnodeleteriousprecipitationssuchassilicidesoccurred

Fig.9.Variationofresidualstressesduetoverticalfacemillingasafunc-tionofdepth.IMI-834;cuttingspeed,56m/min;feed,0.1mm/tooth/rev.;depthofcut,2.0mm.Fig.10.Variationofresidualstressesforverticalfacemilling:cuttingspeed,11m/min;feed,0.056/tooth/rev.;depthofcut,0.25mm.Stressrelievedat400◦C.

Fig.11.Variationofresidualstressesforverticalfacemilling:cuttingspeed,11m/min;feed,0.056mm/tooth/rev.;depthofcut,0.25mm.Stressrelievedat500◦C.

Fig.12.Variationofresidualstressesforverticalfacemilling:cuttingspeed,11m/min;feed,0.056mm/tooth/rev.;depthofcut,0.25mm.Stressrelievedat600◦C.

B.R.Sridharetal./JournalofMaterialsProcessingTechnology139(2003)628–634633

Fig.13.Variationofresidualstressinthecuttingdirection,σcforverticalfacemillingasafunctionofdepthatdifferenttemperaturescuttingspeed,11m/min;feed,0.056mm/tooth/rev.;depthofcut,0.25mm.

Fig.14.Variationoftherelaxationofresidualstressinthecuttingdirection,σcasafunctionofdepthatdifferenttemperatures.

Table2

PeakresidualstressesfordifferentmillingparametersSl.no.1a2a3b4b1c5c8d4d1e7e6f4f

ab

v(m/min)11.0056.0011.0056.0011.0011.0056.0056.0011.0011.0056.0056.00

f(mm/tooth/rev.)0.0560.0560.1000.1000.0560.1000.0560.1000.0560.0560.1000.100

d(mm)0.2500.2502.0002.0000.2500.2502.0002.0000.2502.0000.2502.000

σc(MPa)−738.72−604.02−174.31−502.44−738.72−205.93−238.17−502.44−738.72−412.49−815.57−502.44

σn(MPa)−806.04−701.04−198.56−429.74−806.04−271.82−274.53−429.74−806.04−473.09−730.30−429.74

Withincreaseinvatlowfandd,thepeakresidualstresseshavedecreased.Withincreaseinvathighfandd,thepeakresidualstresseshaveincreased.cWithincreaseinfatlowvandd,thepeakresidualstresseshavedecreased.dWithincreaseinfathighvandd,thepeakresidualstresseshaveincreased.eWithincreaseindatlowvandf,thepeakresidualstresseshavedecreased.fWithincreaseindathighvandf,thepeakresidualstresseshavedecreased.

Table3

Peakresidualstresses(σc)inthecuttingdirectiona,bSl.no.123456710

ab

v(m/min)11.0011.0056.0056.0011.0011.0056.0056.0035.0035.00

f(mm/tooth/rev.)0.0560.1000.0560.1000.0560.1000.0560.1000.0750.075

d(mm)0.2500.2500.2500.2502.0002.0002.0002.0001.1251.125

σc(MPa)−738.72−205.93−604.02−815.57−412.49−174.31−238.17−502.44−445.93−410.81

σct(MPa)−732.12−199.15−597.58−808.19−405.93−167.61−231.79−495.43−463.61−463.61

(σc−σct)(MPa)−6.60−6.78−6.44−7.37−6.56−6.70−6.37−7.00−17.68−52.80

(σc−σct)2(MPa)243.55.9741.4754.3243.0344.40.5849.00312.582787.84

Sumofsquarederror=3463.24.

σct=−1784.2+24.9v+17392.6f+438.1d−393vf−3.4vd−4593.8fd+69.7vfd.

634B.R.Sridharetal./JournalofMaterialsProcessingTechnology139(2003)628–634

Table4

Peakresidualstresses(σn)inthenormaldirectiona,bSl.no.123456710

ab

v(m/min)11.0011.0056.0056.0011.0011.0056.0056.0035.0035.00

f(mm/tooth/rev.)0.0560.1000.0560.1000.0560.1000.0560.1000.0750.075

d(mm)0.2500.2500.2500.2502.0002.0002.0002.0001.1251.125

σn−806.04−271.82−701.04−730.30−473.09−198.56−274.53−429.74−471.91−480.70

σnt(MPa)−804.18−269.82−699.30−727.86−471.28−196.63−272.94−427.60−496.46−496.46

(σn−σnt)(MPa)−1.85−1.99−1.73−2.44−1.81−1.93−1.59−2.1424.5515.76

(σn−σnt)2(MPa)23.423.962.995.953.283.722.534.58602.70248.38

Sumofsquarederror=881.51.

σnt=−1782.5+18.5v+16221.1f+3.8d−293.9vf−vd−3797fd+38.6vfd.

atthestressrelievingtemperaturesemployed.Greaterre-laxationatthesurfacecomparedtothecore(Fig.14)couldbeattributedtotheeffectofgreaterresidualstressgradientsatthesurfacethanatthecore[11,15].Majorcontributiontotheoverallstressrelaxationappearedtooriginatefromtheinitial200to300␮mofthesurfacelayers.

Acknowledgements

TheauthorsexpresstheirsincerethankstotheDirector,GTRE,forextendingfacilitiesandgivingencouragementthroughoutthecourseofthepresentwork.References

[1]J.Mathar,Trans.ASME56(4)(1934)249–254.

[2]R.A.Kelsey,Proc.Soc.Exp.StressAnal.14(1)(1956)181–194.[3]R.J.Rendler,I.Vigness,Proc.Soc.Exp.StressAnal.23(2)(1966)

577–586.

[4]G.S.Schajer,J.Eng.Mater.Technol.110(4)(1988)338–349.[5]TechnicalBulletin304-FandInstructionManualforRS-200-01

MillingGuide,MeasurementsGroupInc.,Raleigh,USA,1988.[6]B.R.Sridhar,W.G.Nafde,K.A.Padmanabhan,J.Mater.Sci.27

(1992)5783–5788.

[7]AnnualBookofASTMStandards,E837-45,ASTM,Philadelphia,

1987,pp.991–997.

[8]B.R.Sridhar,K.Ramachandra,K.A.Padmanabhan,J.Mater.Sci.

31(1996)4381–4385.

[9]H.Wohlfarht,in:H.O.Fuchs(Ed.),ProceedingsoftheSecond

InternationalConferenceonShotPeening(ICSP-2),AmericanShotPeeningSociety,NewJersey,1984,pp.306–315.

[10]E.HarringtonJr.,IndustrialQualityControl21(10)(1965)494–498.[11]G.Derringer,R.Suich,J.Qual.Technol.12(4)(1980)214–219.[12]O.Vohringer,T.Hirsch,E.Macherauch,Titaniumscienceandtech-nology,in:G.Luetjering,W.Zwicker,W.Bunk(Eds.),FifthInter-nationalConferenceonTitanium,DeutscheGesellschaftfurMetal-lkundeE.V.,vol.4,Oberusel,1984,pp.2203–2210.

[13]O.Vohringer,AdvancesinSurfaceTreatmentTechnology,Applica-tions.

[14]D.F.Neal,P.A.Blenkinsop,TitaniumandTitaniumAlloys,in:

J.C.Williams,A.Belov(Eds.),ProceedingsofThirdInternationalConferenceonTitanium,vol.3,PlenumPress,NewYork,1982,pp.2003–2014.

[15]A.L.Esquivel,R.Evans,Exp.Mech.8(1968)496.

5.Conclusions

1.Residualstressesduetothemillingoperationsforthemillingparametersemployedwerecompressiveinnature.

2.Thevariationinthemagnitudesoftheresidualstressesdidnotrevealanydefinitetrendwithrespecttothemillingparameters.

3.Alinearrelationshipcouldnotexplainthevariationoftheresidualstresseswithrespecttothemillingparame-ters,butapolynomialrelationshipinvolvingbothlinearandinteractivetermscouldexplainthevariationswithreasonableaccuracy.

4.Stressrelievingtreatmentsledtotherelaxationofresidualstressesalmostlinearlywithdepthfromthesurface.

5.Therelaxationratesincreasedwithincreasingtempera-tureanddidnotrevealanyevidencefordeleteriouspre-cipitationsuchasofsilicides.

6.Thegreaterrelaxationratesatthesurfacecomparedtothecorecouldbeattributedtothesteeperresidualstressgradientsatthesurfacethanatthecore.

7.Themajorcontributiontotheoverallstressrelaxationappearedtooriginatefromthesurfacelayers.