surface methodology Assessment of cutting force and surface roughness in LM6/SiC usingresponse Technology Journal of Applied Researchand

(1)

Availableonlineatwww.sciencedirect.com

Journal of Applied Research and Technology

www.jart.ccadet.unam.mx JournalofAppliedResearchandTechnology15(2017)283–296

Original

Assessment of cutting force and surface roughness in LM6/SiC _p using response surface methodology

Aezhisai Vallavi Muthusamy Subramanian

^a,∗

, Mohan Das Gandhi Nachimuthu

^b

, Velmurugan Cinnasamy

^c

aDepartmentofMechanicalEngineering,KumaraguruCollegeofTechnology,Tamilnadu,India

bKalaignarKarunanidhiInstituteofTechnology,Tamilnadu,India

cDepartmentofMechanicalEngineering,KumaraguruCollegeofTechnology,Tamilnadu,India Received22June2016;accepted9January2017

Abstract

Metalmatrixcompositesarereplacingthetraditionalmaterialworldbecauseoftheirsuperiormechanicalproperties.Consequently,thenecessity foraccuratemachiningofthecompositehasincreaseddrastically.Themajorproblemswhilemachiningmetalmatrixcompositesaresurface roughnessand cutting force.Thepresentworkfocuses on thestudyof thecutting conditionswhichinfluence thecutting forceandsurface roughnessintermsofspindlespeed,feedrate,axial,radialdepthofcutandweightpercentageofsiliconcarbideparticle(SiC_p).Centralcomposite rotatablesecondorderresponsesurfacemethodologyhasbeenemployedtocreatethemathematicalmodel.Theadequacyofthemodelhasbeen verifiedusinganalysisofvariance.Thedirectandinteractioneffectsoftheprocessparametershavebeenstudiedtokeepthecuttingforceand surfaceroughnessminimal.

Keywords:Cuttingforce;Surfaceroughness;Metalmatrixcomposites;Endmilling;Responsesurfacemethodology

1. Introduction

Millingisametalremovalprocessoffeedingtheworkagainst arotatingmultipointcutter.Theratioofthemetalremovalrate israpidasthecutterrotatesatahighspeedandhasmanycut- tingedges.Itisemployedtoproduceslots,pockets,precision moldsand dies,automotive andaerospace components. Cut- tingforcecausedeflectionsofthepartormachinestructureand supplyenergytothemachiningsystemwhichresultsinexces- sivecuttingtemperatureorunstablevibrations.Thereasonfor analyzingcuttingforceofthemachinetoolistoestimatetool forcesthatmustberesistedbymachinetoolcomponents,bear- ingloads,jigsandfixturesandpowerrequirement.Tosatisfythe demandsonproductcapabilitiesandfunctions,naturalmaterials

∗Correspondingauthor.

E-mailaddress:[email protected](A.V.MuthusamySubramanian).

PeerReviewundertheresponsibilityofUniversidadNacionalAutónomade México.

are insufficient and hence metal matrix composites (MMC) areconsidered. MMCconsists ofatleasttwocomponents,of whichoneisametalmatrixandtheotherisareinforcement.

Themetalmatrixisgenerallyanalloy.MMChasbeenstudied mainly for the aerospace industry andspace component due to a higher specific modulus, higher specific strength, lower coefficient of thermalexpansion, better wear resistance, bet- terpropertiesatelevatedtemperature.Recently,electronicand automotiveindustrieshavebeenconcentratingthe composites (Rosso,2006).

ThecuttingforcemodelinturningofLM6/SiCpmetalmatrix composite hasbeenproposedbyJoardar,Das,Sutradhar,and Singh(2014).

Thecuttingspeed,depthofcutandweightpercentageofsili- concarbidewerespecifiedasmachiningparameters.Sequential approach in face centralcomposite design saves the number ofexperimentationsneeded.Theauthorpointedthattangential cuttingforceandradialcuttingforcearemoresensitivetocut- tingspeed.Jeyakumar,Marimuthu,andRamachandran(2013), carriedoutthemathematicalmodelofcuttingforce,toolwear

http://dx.doi.org/10.1016/j.jart.2017.01.013

1665-6423/©2017UniversidadNacionalAutónomadeMéxico,CentrodeCienciasAplicadasyDesarrolloTecnológico.Thisisanopenaccessarticleunderthe CCBY-NC-NDlicense(http://creativecommons.org/licenses/by-nc-nd/4.0/).

(2)

284 A.V.MuthusamySubramanianetal./JournalofAppliedResearchandTechnology15(2017)283–296

andsurfaceroughnessduringendmillingofAl6061/SiCpcom- positeunderdrycondition.Thecuttingforcecomponentinthe z-directionwasatsignificantlyhighermagnitudesthanthatof thex-direction.Theresultsindicatedthatthedepthofcutwas thedominantfactoraffectingtoolwear,cuttingforce,andsur- faceroughness.PalanikumarandMuniaraj(2014)employedthe responsesurfacemethodtoanalyzethethrustforceinthedrilling ofhybridmetalmatrixcompositesbycoatedcarbidedrill.The burrformed becauseof thrust reducedthe holequality. They demonstratedthatthethrustforceandburrformationweremost influencedbythefeedrate.Kalla,Sheikh-Ahmad,andTwomey (2010)putforwardamechanisticmodelingtechniquetopredict thecuttingforcebytransformingspecificcuttingenergiesfrom orthogonaltoobliquecuttinginhelicalendmilling.Theycon- cludedthatproposedmodelwasreasonablefor unidirectional butlessdesirableformulti-directionalcomposites.

Pramanik, Zhang, and Arsecularatne (2006) built up a systematic model for predictingthe forces using Merchant’s analysis,sliplinefieldtheoryofplasticityandtheGriffiththeory offracture.Theystatedthattheforceduringthechipformation ishigherthantheploughingandparticlefracture.Sivasakthivel, VelMurugan,andSudhakaran(2010)broughtforwardacentral composite rotatable second orderresponse surfacemethodol- ogytodevelopamathematicalmodeltopredictcuttingforces intermsofhelixangle,axialdepthofcut,radialdepthofcut, feedrateandspindlespeedofAl6063ofhighspeedsteelend mill cutter. The experimental results showed that increase in feedrateandaxialdepthofcutreducestheinfeedandcrossfeed force.Babu,Selladurai,andShanmugam(2008)establishedthe effectsofcuttingparametersonthevariationsofcuttingforces of AlSiCp metalmatrixcomposite material.The authorcon- cludedthatcuttingforcesweresensitiveinthehighspeedand fullimmersioncondition.Theystatedthatwiththeincreasein thedepthofcut,cuttingforceintangentialdirectionincreases.

Seeman, Ganesan, Karthikeyan, and Velayudham (2010) studiedthetoolwearandsurfaceroughnessduringthemachin- ingof Almetalmatrixcomposite.Theresult showedthatthe tool wear was affected by BUE formation at low speed and better surfacefinish was achievedat lowfeed rate withhigh speed.AnandakrishnanandMahamani(2011)experimentally investigatedoninsituMMCusinguncoatedtungstencarbide turninginsert.Theyconcludedthatwiththeincreaseinthedepth ofcut,therateofflankwear,cuttingforceandsurfacerough- nessincrease.Palanisamy,Rajendran,andShanmugasundaram (2007)optimizedthemachiningparameterssuchasspeed,feed, depthofcutusinggeneticalgorithm,yieldtominimizemachin- ing time while considering the technological constraints as cuttingforce,toollifeinendmillingprocess.DavimandBaptist (2000)discussedtherelationshipbetweencuttingforceandtool wearwhenmachiningthecomposite.Indrilling,correlationwas obtainedbetweenevolutionofflankwearofthedrillsandthe feedforce.Similarcorrelationshavebeenobtainedbyturning betweentheevolutionofflankwearoftheinsertandthefeed anddepthforces.

Valarmathi, Palanikumar, and Latha (2013) analyzed the thrustforceforcompositepanel basedonTaguchi’sdesignof experiments and response surface methodology. The authors

have taken inputparameters such as spindle speed,feed rate andpointangle.Theirresultsindicatedthathighspindlespeed withlowfeedreducedthethrustforce.Arokiadass,Palaniradja, andAlagumoorthi(2012)studiedtheeffectsofspeed,feeddepth ofcutand%weightofsiliconcarbideontoolwearinmachining LM25AlalloyreinforcedwithSiCpinendmillingoperation.

Theyanalyzedtheprocessconstraintsonperformancecharac- teristicsusingANOVA.Thedecisionattainedbythemisthatthe spindlespeedandthecontentofSiCparetheinfluencingfactor ontoolwear.

Mahesh, Muthu, and Devadasan (2015) propounded a methodology tostudy the optimumcuttingparameter aiming to minimize the surface roughness through response surface methodology andgeneticalgorithm (GA).They analyzed the directandinteractioneffectofcuttingparameterusingdesign expertsoftware.Theconformitytesthaveshownthatthereisa goodagreementwithpredictedandobservedvalues.Makadia and Nanavati (2013) investigated the influence of machining parametersasfeedrate,toolnoseradius,cuttingspeedanddepth of cut of turningonthe surfaceroughnessof AISI410 steel.

Three level fullfactorialdesignof experiment hasbeen used toconducttheexperiment.Responsesurfaceoptimizationwas employedtogetoptimummachiningcondition.Shihab,Khan, Mohammed,andSiddiquee(2014),studiedthesurfaceintegrity intermsofsurfaceroughnessandmicrohardnessofthedryhard turningprocess.Threelevels,threeparametercentralcomposite designwasemployedtocollecttheexperimentaldata.Goodsur- faceintegrityhasbeenachievedbykeepingfeedrateanddepth ofcutaslowerlevel.

Raju,Janardhana,Kumar,andRao(2011),appliedthemul- tipleregression analysis andGeneticalgorithm topredict the surfaceroughnessunderdryandwetconditionusinghighspeed steelandcarbidetools.Feedratewasthesignificantfactorwhich effectsthesurfaceroughness.Throughexperiments,ithasalso beenfoundthatsurfaceroughnessdecreaseswiththeapplica- tionofcoolantbyusingacarbidetool.Klilckap,Cakir,Aksoy, andInan(2005)experimentallyinvestigatedthetoolwearand surfaceroughnessof5%SiCpAlMMCusinguncoatedandTiN coatedcuttingtoolsinturningprocess.Cuttingparameterswere consideredasspindlespeed,feedrateanddepthofcuttopredict thesurfaceroughnessandtoolwear.Theyfoundthatincrease in cutting speed, increases tool wear and surface roughness.

Theyalsonoticedthatbuiltupedgewasabsentwhilemachining directlycastmaterials.

Chen, Wang, andLee(2013)presented an optimalcutting parameterdesignofhighspeedcuttingoftheDIN1.2344tool steel. The orthogonal array with gray relation analysis was appliedtooptimizetheendmillingprocesswithmultipleper- formancecharacteristics.Theselectedcuttingparameterswere surfaceroughnessonreliefface,cuttingspeed,feedpertooth, axialcuttingdepthandradialcuttingdepth,whiletheconsid- eredperformancecharacteristicsaretoollifeandmetalremoval rate.Theresultsofconfirmationexperimentsrevealedthatgray relationanalysiscaneffectivelyacquireanoptimalcombination ofthecuttingparameters.

In the present work, anattempt ismade toinvestigate the effect of cutting variableson cutting forcein endmilling of

(3)

LM6aluminumalloycomposites.Themainaimistodevelop the mathematicalmodel for cuttingforceinterms of spindle speed,feedrate,axialdepthofcut,radialdepthofcutandvarious weightpercentageofSiCp(wt%ofSiCp)bytheresponsesurface method.Theinfeed,crossfeedandtrustforcearemeasuredin tangential,radial andaxial direction duringendmilling. The mathematicalmodelaidedustostudythedirectandinteraction effectsofeachparameter.

Veryfewauthors(Arokiadassetal.,2012)haveconsidered weightpercentageofsiliconcarbideas oneof theparameters toanalyzethewearcharacteristics.Mostoftheauthorsstudied theeffectofcuttingparameterssuchasspeed,feedanddepth ofcut,butdidnotembarkontheinvestigationbyvaryingthe weightpercentageofsiliconcarbideoncuttingforceandsurface roughness.Toolwearstronglycorrelatesthesurfaceroughness andcuttingforcewhilemachiningofcomposite.Inthecurrent work, experiments havebeen conducted withthe purpose of specificallyidentifyingtherelationshipbetweensurfacerough- nessandcuttingforce.Henceforth,theaimsofthispresentwork are

• ToimplementthefivefactorfivelevelDoE(DesignofExper- iments)techniqueinthemeasurementofcuttingforce(axial, radialandthrustforce)andsurfaceroughnessinmachining ofvariousweightpercentageofSiCp.

• Togeneratethemathematicalmodelforcuttingforceconsid- eringtheprocessparameterasSpindlespeed,Feedrate,Axial depthofcut,Radialdepthofcutandvaryingweightpercent- ageofSiCp.

• Tostudythedirectandinteractioneffectofprocessparame- ters.Specifically,todescribetherelationshipbetweensurface roughnessandcuttingforce.

2. Experimentalprocedure

2.1. Machiningparametersandresponsevariables

Machining parameters that are taken into account during machininginendmillingprocessarespindlespeed,feedrate, axialdepth of cut,radial depth of cutandweight percentage of SiCp. The responses are cutting force andsurface roughness(Ra).Thethreecomponentsofforceinmillingareinfeed force(Fx),crossfeedforce(Fy),thrustforce(Fz)actstangential, normalandparalleltothemachinetoolrespectively.

2.2. Limitsofmachiningparameters

Theupperandlowerlimitofeachprocessvariableshaveto beidentifiedfromtherelation.

Xi= 2(2X−(X_max+Xmin))

X_max−X_min (1)

whereXi is the required codedvalue of the variable X.X is anyvalueofavariablefromXmin toXmax.Theupperlimitof the processparameter is codedas 2, lowerlimit as −2. The intermediatevalueswerecalculatedfromEq.(1)andmachining parametersandtheirlevelswerepresentedinTable1.

Table1

Machiningparametersandtheirlevel.

Parameters Unit Notation FactorLevels

−2 −1 0 1 2

SpindleSpeed rpm N 1500 2000 2500 3000 3500

Feedrate mm/rev F 0.02 0.03 0.04 0.05 0.06

Axialdepthofcut mm X 1 1.5 2 2.5 3

Radialdepthofcut mm Y 1 1.5 2 2.5 3

Siliconcarbide wt% W 5 10 15 20 25

Table2

ChemicalcompositionofLM6aluminumalloy.

Elements Si Ni Pb Fe Mn Zn Cu Mg Sb Ti Al Percentage 10–13 0.1 0.1 0.6 0.5 0.1 0.1 0.1 0.05 0.2 Remaining

2.3. PreparationofMMCs

Aluminum metal matrix composite finds vast applications in the fields of construction, automotive and marine fields.

Especially,LM6isusedformarinecomponents,water-cooled manifolds and jackets, motor car and road transport fittings andpumpparts.LM6aluminumAlloychemicalcomposition isgiveninTable 2.Silicon carbideparticulatesof 30micron size wereusedas reinforcementmaterial.Theparticlesize is alsoanimportantfactor,whereiniftheparticlesizeincreases, thenthepropertywilldecreasebecauseofinterparticlespacing.

However,the additionof SiC particleswill increasethe wear resistance,hardness andbrittleness ofthe composite material (Dwivedi,Kumar,&Kumar,2010;Ozben,Kilickap,&Cakir, 2008). The Stir casting technique was employed inorder to preparecompositematerial.LM6Alloyispreheatedto750^◦C andkeptinagraphitecrucibleinanelectricinductionfurnace.

SiC particleswerepreheatedto350^◦C infurnaceinorderto releasemoistureandothertoxicgases.SiCparticlesofamount weighing5%,10%,15%,20%,25%wereslowlyaddedtothe moltenLM6aluminumalloyinsidecrucible.Mechanicalstir- rerat1250rpm isusedtopreparethe compositefor aperiod of15min.Themoltenmaterialwaspouredinsidethedieata temperatureof850^◦C.Castcompositeswereobtainedafterthe dieisallowedtocoolinstillairatroomtemperature.TheSEM micrographof5%,10%,15%,20%and25%siliconcarbidepar- ticlesintheparentmetalareshowninFigure1.Itindicatesthat thesiliconcarbideparticlesareevenlydistributedinthematrix material.

2.4. Experimentalsetup

Experiments were performed on the HAAS CNC verti- cal machining centerwith two carbide insert end mill cutter withthe diameter of 12mm andhas the following specifica- tion:tablelength1070mm,width230mm,maximumspindle speed 4000rpm, feedrate 5.1m/min and the power of spindle motor 5.6kW.The dimension of the work specimen was 100mm×100mm×25mm.

(4)

Fig.1.SEMmicrographof(a)castLM6/5wt%SiCp,(b)castLM6/10wt%SiCp,(c)castLM6/15wt%SiCp,(d)castLM6/20wt%SiCpand(e)castLM6/25wt%

SiCp.

Theworkpiecewasrigidlymountedonaspeciallydesigned vicewhichcontainssensorstomeasurethecuttingforces.The cutting force of the tool point in X, Y and Z directions are measured as infeed, crossfeed and thrust force respectively, using three axis milling tool dynamo-meter made bySyscon

instruments whichis pre-calibratedto1kgf.Sensing element bondedwithstraingaugeisused tosense theseforces.These forces fromstrain gauges are processedwithinstrumentation amplifiersthenconvertedintodigitaldataby8bitADC.This digitaloutputhasbeensendoutseriallythroughRS232Ctostore

(5)

Fig.2.ExperimentalsetupinHaasverticalmachiningcenter.

thedatainthedataacquisitionsoftware(MSofficeonenote).The acquireddataaretabulatedtopredictthemathematicalmodel.

ArithmeticaverageroughnessRa,rootmeansquaredevia- tionRq,maximumheightoftheprofileRy,tenpointheightof irregularitiesRz,maximumprofilepeakheightRparesomeof theroughnessparameterstocharacterizethesurfaceroughness.

Amongthese,Raisthecommonlyused roughnessparameter andbestsuitedformonitoringtheconsistencyofthemachining process.Raistheaverageabsolutedeviationoftheworkpiece fromthecenterline andthesameisused inthisstudy. After machiningofeachworkpiece,surfaceroughnessvalueismea- suredusingMITUTOYOSJ-201surfaceroughnesstesterasa cutofflengthof2.5mm.Foreachspecimenthreesetsoftrails aremeasuredandgraphsaresaved.Averagesurfaceroughness ofeachspecimenhavebeennotedintermsofRa.Figure2shows theexperimentalsetupusedforconductingexperiments.

2.5. Generationofmathematicalmodel

Theregressionprocedurewasusedforthedevelopmentof themathematicalmodeltopredictthe cuttingforce. Thesec- ond orderpolynomial (Bhattacharyya,1998) representingthe responsesurfacefor“k”factorsisgivenbyEq.(2)

y=a0+^k

i=1

aixi+^k

i=1

aiix²_i +^k

i<j

aijxixj+εi (2)

wherea0=freeterm,ai=lineartermcoefficient,aii=quadratic termcoefficient, aij=interactionterm coefficientandε=error term.TheresponseYisexpressedasafunctionofprocessparam- etersspindlespeed(N),feedrate(Z),axial(X)andradialdepth ofcut(Y),wt%ofSiCP(W)showninEq.(3)

Y =Φ(Niu,Ziu,Xiu,Yiu,Wiu)+eu (3) whereΦ=response surface, eu=residual, u=no. of observa- tionsinthefactorialexperimentandiurepresentsthelevelofthe ithfactorintheu^th observation.Whenthemathematicalform

of Φ isunknown, thisfunctionisapproximated satisfactorily withintheexperimentalregionbypolynomialsintermsofpro- cessparametervariables.BoxandHunterproposedthecentral compositerotatable designfor fittingasecondorderresponse surfacebasedonthecriterionofrotatability.Thecentralcompos- iterotatabledesignforfiveparameterswithfivelevelsconsistsof 32experimentswiththecombinationofsixteenfactorialdesign points (lie atthe vertices of the regular polyhedral), ten star points(toformspherewithαradius,consistingofequallyspaced pointsfrom the center) and six replicatedcenterpoints (also knownasaxialpointswhichprovideroughlyequalprecisionof standarderror).Thedesignmatrixwithmeasuredandpredicted valueofcuttingforceandsurfaceroughnesshasbeenshownin Table3.TheMiniTabstatisticalsoftware(Version15)package hasbeenusedtodeveloptheresponseequationsandevaluatethe coefficientvalues.Thissoftwareisalsousedtoperformthedata analyses(Montgomery,2009).The statistical empirical equa- tionhasbeendevelopedusingonlythesignificantcoefficients andisshowninEqs.(4)–(7).R2 valuefor infeed,crossfeed andthrustforceinthex, yandzdirectionsareapproachesto unityasindicatedinTable4.BecauseofhighR²values,good correlationsexistbetweenthe experimentalandthepredicted values.

F_X=114.673−27.203N+16.24F+23.528X+12.21Y +29.986W+3.124N²−0.839F²+7.309X² +8.452Y²+5.945NF −1.219NX+8.057NY

−4.141NW+6.479FX+10.079FY +5.846FW +4.244XY+11.232XW+5.562YW (4)

F_Y =95.13−14.422N+19.637F+18.874X+5.398Y +30.792W+1.46N²+6.67X²+1.369Y²+4.357NF +7.193NY−9.554NW+3.248FX+5.351FY +3.21FW+4.861XY+8.329XW+6.242YW (5)

F_z=90.466−18.722N+12.568F+18.602X+5.894Y +29.97W+8.517X²+4.489Y²+2.395W² +7.626NF+11.235NY−6.229NW+3.809FX +5.278FY+4.487FW+6.437XY +12.794XW

+8.543YW (6)

Ra=5.628−0.62N+0.29F+0.45X−0.07Y+0.52W +0.019N²+0.64F²−0.85X²+0.37Y²−0.46W² +0.57NF+0.06NX+0.19NY+0.7NW−0.48FX +0.09FY+0.07FW−0.1XY −0.68XW+1.1YW

(7)

(6)

Table3

Designmatrixwithmeasuredandpredictedvaluesofcuttingforceandsurfaceroughness.

Run CodedFactor Cuttingforce(N) SurfaceRoughnessRa(␮m)

X1 X2 X3 X4 X5 InfeedforceFx(N) CrossfeedforceFy(N) ThrustforceFz(N) Observed Predicted %error Observed Predicted %oferror Observed Predicted %oferror Observed Predicted %oferror

1 −1 −1 −1 −1 1 153.712 153.016 0.45 121.644 122.717 −0.88 131.454 131.389 0.08 4.21 4.281 −1.68

2 1 −1 −1 −1 −1 59.306 58.352 1.61 44.145 44.751 −1.37 48.069 47.931 0.07 1.21 1.203 0.57

3 −1 1 −1 −1 −1 105.802 105.824 −0.02 83.385 84.529 −1.37 93.164 93.388 −0.21 5.62 5.537 1.47

4 1 1 −1 −1 1 87.644 87.708 −0.07 87 88.875 −2.16 73.575 74.966 −1.02 5.36 5.303 1.06

5 −1 −1 1 −1 −1 136.306 135.626 0.50 88.29 82.459 6.60 101.043 101.776 −0.74 8.92 9.075 −1.73

6 1 −1 1 −1 1 111.04 110.398 0.58 88.29 89.853 −1.77 84.366 86.078 −1.44 3.59 3.619 −0.80

7 −1 1 1 −1 1 228.273 228.61 −0.15 194.238 193.679 0.29 187.371 189.609 −4.20 3.26 3.258 0.12

8 1 1 1 −1 −1 84.462 84.538 −0.09 85.347 86.985 −1.92 65.555 67.531 −1.30 5.28 5.229 0.96

9 −1 −1 −1 1 −1 99.302 98.578 0.73 39.101 41.089 −5.08 60.543 59.454 0.66 2.83 2.807 0.81

10 1 −1 −1 1 1 88.46 87.774 0.78 66.2 68.907 −4.09 71.613 71.692 −0.06 5.18 5.209 −0.55

11 −1 1 −1 1 1 182.991 183.282 −0.16 149.112 152.373 −2.19 135.742 136.159 −0.57 6.67 6.791 −1.81

12 1 1 −1 1 −1 98.962 98.994 −0.03 80.018 82.799 −3.48 75.679 76.025 −0.26 2.04 1.981 3.17

13 −1 −1 1 1 1 211.7 211.288 0.19 168.55 168.819 −0.16 181.485 182.411 −1.68 6.07 6.073 −0.04

14 1 −1 1 1 −1 77.47 76.8 0.86 58.86 61.341 −4.22 57.879 58.545 −0.39 2.46 2.423 1.50

15 −1 1 1 1 −1 162.845 163.152 −0.19 106.4 106.743 −0.32 94.189 95.410 −1.15 4.23 4.197 0.78

16 1 1 1 1 1 239.22 239.564 −0.14 194.39 198.145 −1.93 205.082 207.276 −4.50 6.70 6.683 0.25

17 −2 0 0 0 0 182.265 181.575 0.38 126.234 129.814 −2.84 124.084 124.354 −0.34 6.92 6.925 0.02

18 2 0 0 0 0 73.013 72.763 0.34 74.405 72.126 3.06 51.012 49.466 0.79 4.48 4.517 −0.82

19 0 −2 0 0 0 77.574 78.837 −1.63 49.05 55.856 −13.88 62.258 65.330 −1.91 7.62 7.767 −1.92

20 0 2 0 0 0 146 143.797 1.51 130.29 134.404 −3.16 116.121 115.602 0.60 8.81 8.703 1.21

21 0 0 −2 0 0 96.991 96.853 0.14 85.347 84.062 1.51 85.347 87.330 −1.69 1.31 1.303 0.53

22 0 0 2 0 0 191.763 190.965 0.42 156.96 159.558 −1.66 165 161.738 5.38 3.10 3.003 3.12

23 0 0 0 −2 0 124.295 124.061 0.19 86.138 89.81 −4.26 98.1 96.634 1.44 4.24 4.139 2.54

24 0 0 0 2 0 173.607 172.901 0.41 113.772 111.402 2.08 120.024 120.210 −0.22 3.98 3.907 1.83

25 0 0 0 0 −2 56.805 54.701 3.70 29.43 33.546 −13.99 39.507 40.106 −0.24 2.77 2.751 0.68

26 0 0 0 0 2 177.346 174.645 1.52 154.998 156.714 −1.11 161.865 159.986 3.04 4.79 4.739 1.06

27 0 0 0 0 0 113.00 114.673 −1.48 97.309 95.13 2.24 88.29 90.466 −1.92 5.60 5.625 −0.44

28 0 0 0 0 0 114.85 114.673 0.15 91.423 95.13 −4.05 90.252 90.466 −0.19 5.60 5.625 −0.44

29 0 0 0 0 0 114.803 114.673 0.11 95.347 95.13 0.23 90.252 90.466 −0.19 5.61 5.625 −0.26

30 0 0 0 0 0 114.92 114.673 0.21 94.366 95.13 −0.81 91.233 90.466 0.70 5.68 5.625 0.96

31 0 0 0 0 0 114.803 114.673 0.11 96.328 95.13 1.24 92.214 90.466 1.61 5.63 5.625 0.08

32 0 0 0 0 0 114.721 114.673 0.04 97.309 95.13 2.24 89.271 90.466 −1.07 5.66 5.625 0.61

Table4

Summaryofregressionanalysis.

Responses Svalue R2% AdjustedR2%

Infeedforce(Fx)(N) 1.20327 99.98 99.94 Crossfeedforce(FY)(N) 3.44895 99.80 99.44 Thrustforce(Fz)(N) 2.23813 99.90 99.72 Surfaceroughness(Ra)(␮m) 0.03312 99.99 99.97

2.6. Checkingtheadequacyofthemodel

Adequacy of theexperimentaldesign hasbeenverified by Analysisofvariance(ANOVA).InANOVAanalysislackoffit, FandRratiohavebeenusedtochecktheadequacyofamodel.

Tables5and6indicatestheANOVAofregressionparametersof

Table5

Analysisofvarianceforcuttingforce.

Source Infeedforce(Fx)(N) Crossfeedforce(FY)(N)

DOF SS Adjustedmeansquare Fvalue Pvalue DOF Sumofsquare Adjustedmeansquare Fvalue Pvalue

Regression 20 73732.6 3686.6 2546.28 0.000 20 53241.7 2662.1 277.75 0.000

Linear 5 62533.4 12506.7 8638.11 0.000 5 46251.8 99250.4 965.15 0.000

Square 5 3651.2 730.2 504.36 0.000 5 1488.3 297.7 31.06 0.000

Interaction 10 7548 754.8 521.33 0.000 10 5501.6 550.2 57.4 0.000

Residualerror 11 15.9 1.4 11 105.4 9.6

Lack-of-fit 6 13.1 2.2 3.94 0.077 6 80.4 13.4 2.68 0.15

Pureerror 5 2.8 0.6 5 25 5

Total 31 73748.5 31 53347.2

(7)

Table6

Analysisofvarianceforcuttingforceandsurfaceroughness.

Source Thrustforce(Fz)(N) Surfaceroughness(Ra)(␮m)

DOF SS Adjustedmeansquare Fvalue Pvalue DOF Sumofsquare Adjustedmeansquare Fvalue Pvalue

Regression 20 54707.12735.4 546.07 0.000 20 114.255 5.7127 5207.74 0.000

Linear 5 42898 8579.6 1712.7 0.000 5 22.944 4.5887 4183.09 0.000

Square 5 2788.2 557.6 166.46 0.000 5 46.102 9.2204 8405.35 0.000

Interaction 10 9020.8 902.1 180.08 0.000 10 45.209 4.5209 4121.25 0.000

ResidualError 11 55.1 5 11 0. 012 0.0125

Lack-of-Fit 6 45.5 7.6 3.94 0.077 6 0.006 0.0076 0.96 0.528

PureError 5 9.6 1.9 5 0.006 0.0011

Total 31 54762.2 31 114.267

thepredictedsurfaceresponsemodelforcuttingforces(infeed, crossfeed and thrust force) andsurface roughness. The cal- culatedvaluesofF-ratio forlackoffitarecomparedwiththe standardvaluesofF-ratiocorrespondingtotheirdegreesoffree- dom.ThestandardpercentagepointofFdistributionfor95%

confidencelevelis4.95.FromTables5and6theobtainedFval- ueswere3.94,2.68,3.94,0.96forinfeed,crossfeed,trustforce andsurfaceroughnessrespectively, smaller thanthe standard value,indicatingthatthemodelisadequate.Itisalsofoundthat fromthePvalues,cuttingforceandsurfaceroughnessmodels linear,squareandinteractioneffectsaresignificant.

2.7. Validationofthemodel

Thevalidityofthemodelischeckedforthelevelsofparame- terswhichhasnotbeenincludedintheexperimentaldesign.The resultsofconformityareshowninTable7.Thetestresultshows that developedmodel computesthe cuttingforceandsurface roughnesswithreasonableaccuracy.Thedifferenceinpercent- agebetweenexperimental andpredictedvalue relative tothe predictedvalueisestimatedasthepercentageerrorintheregres- sionmodel.Theerrorbetweentheexperimentalandpredicted valueislessthan5%,whichconfirmsthevalidityofthemodel.

3. Resultsanddiscussion

Amathematicalmodel,asindicatedinEqs.(4)–(7),hasbeen developedtopredict theinfeed force,crossfeed force,thrust forceandsurfaceroughnessinrelationwithspindlespeed,feed rate,axialdepthofcut,radialdepthofcutandwt%SiCp.The influenceof theprocessparametersonthe cuttingforcehave beenstudiedusingthedevelopedmodel.Directandinteraction effectsoftheprocessparametersondifferentcuttingforceswere plottedinFigures3–7.

3.1. Directeffectsofmachiningparametersoncutting forcesandsurfaceroughness

3.1.1. Effectofspindlespeed

Theeffectofspindlespeedoninfeed,crossfeedforce,thrust forceandsurface roughnessis showninFigure 3. It isclear that anincreaseinspindle speedresulted indecreasing trend incuttingforcesinallthe threedirections.Whenthespindle

1.5 2.5 3.5 4.5 5.5 6.5 7.5

0 50 100 150 200

1000 1500 2000 2500 3000 3500

Surface roughness (Ra)

Cutting force (N)

Spindle speed (N)

Fx Fy Fz Ra

Fig.3.Directeffectofspindlespeed.

0 2 4 6 8 10 12

50 80 110 140

0.01 0.03 0.05 0.07

Cutting force (N)

Feed rate (mm/rev)

Fx Fy Fz Ra

Fig.4.Directeffectoffeedrate.

speedisincreased,thechip-toolcontactlengthgotdecreased, thematerialgotheatedup,andsoftenedthematrix.Hencethe cuttingforce inendmilling process isreduced. Further, it is inferredthat highcuttingforceisproducedatextremelyslow speeds(Stephenson&Agapiou,2006).Withtheincreaseinspin- dle speed, the surface roughness is on the decline. At lower spindlespeed,theunstablelargerbuilt-up-edgeisformedand thechipfractureisproducedwhichleadstotheroughsurface.

Whereasathighercuttingspeed,thebuilt-up-edgeformationis

(8)

290A.V.MuthusamySubramanianetal./JournalofAppliedResearchandTechnology15(2017)283–296

Table7

Resultofconformity.

S.No. Controlfactors Infeedforce(N) Crossfeedforce(N) Thrustforce(N) Surfaceroughness(␮m)

N F X Y Z Observedvalue Predictedvalue %Error Observedvalue Predictedvalue %Error Observedvalue Predictedvalue %Error Observedvalue Predictedvalue %Error

1 −2 −2 1 1 2 266 265.21 0.3 227.2 229.66 −1.1 255.13 255.71 −1.48 7.143 7.217 −1.03

2 0 1 2 1 0 256.68 258.55 0.43 208.9 207.53 0.66 211.99 210.46 3.24 2.472 2.482 −0.40

3 1 1 2 2 1 380.81 375.51 1.41 290.2 291.48 −0.4 332.43 332.65 −0.73 3.089 3.092 −0.09

4 2 2 −2 −2 0 39.46 40.26 −3 70.53 68.71 2.57 56.78 56.154 0.36 4.216 4.244 −0.66

(9)

0 1 2 3 4 5 6 7

0 50 100 150 200 250

0.5 1.5 2.5 3.5

Cutting force (N)

Axial depth of cut(mm)

Fx Fy Fz Ra

Fig.5.Effectofaxialdepthofcut.

0 1 2 3 4 5 6 7

20 70 120 170

0.5 1.5 2.5 3.5

Cutting force (N)

Radial depth of cut (mm)

Fx Fy Fz Ra

Fig.6.Effectofradialdepthofcut.

0 1 2 3 4 5 6 7

20 70 120 170 220

0 10 20 30

Cutting force (N)

W (wt %)

Fx Fy Fz Ra

Fig.7.Effectofwt%ofSiCp.

disappearingandthe chipfracture isreduced,andhencesur- faceroughnessisdecreased.Thesameinferencewasderivedby Seemanetal.(2010).Theminimumcuttingforceandsurface roughnessisobtainedat3500rpmofspindlespeed.

3.1.2. Effectoffeedrate

FromFigure4,itisevidentthatthefeedratehasanimportant effectonthecuttingforceandsurfaceroughness.Thefeedrate isdirectlypropositionaltothecuttingforceintangential,radial andaxialdirection.The reasonisthat theareaofcontactand loadcarriedbythecutterincreasesbecauseoftheincreaseinthe feedrate.Increasingthefeedrateincreasestheroughnessofthe machinedsurface.Atahigher feedrate,thematerialremoval rateisalsohighwhichresultsinthecreationoffrictionbetween workandtool,leadingtoapittingmark.Asimilarobservation wasmadebyDwivedietal.(2010).

3.1.3. Effectofaxialdepthofcut

Theincreaseintheaxialdepthofcutresultedinanincreased infeed,crossfeed,thrustforceandsurfaceroughness,asshown inFigure5.Whentheaxialdepthofcutincreases,morelength oftheflutegetsengagedwhichresultsinmorecuttingforcein therotating direction.Itisinferredfromthefigurethat,when the axialdepth ofcut rangesfrom1mm to2mmthe surface roughnessincreasesbecauseofthehigherchipthicknessofthe workmaterial.

3.1.4. Effectofradialdepthofcut

FromFigure6,itisevidentthattheradialdepthofcuthas thesignificanteffectonallthethreeforcesandsurfacerough- ness.Anincreaseintheradialdepthofcutduringthemilling operationresultedinincreasedcuttingforce.Thereasonbeing that,higher radialdepthof cutincreasesthevolumeof metal removal and increase the contact area of work piece, which resultedinincreased cuttingforceinrotational direction.The surfaceroughnessincreaseswiththeincreaseintheradialdepth ofcutduetoandincreaseinthevibrationandwear.

3.1.5. Effectofwt%ofsiliconcarbide

The effect of wt% of SiCp on cutting force is shown in Figure7.Whilethewt%of siliconcarbideparticleincreases, the cuttingforces andsurface roughnessalso increases.This isduetotheadditionofharderandstifferreinforcingmaterial intothesoftmetallicmaterialwhichbecomesmoredifficultto machine.Thecuttingforceandsurfaceroughnessisminimum atthelowestwt%ofSiCpcombination.Arokiadassetal.(2012) derivedthesameinferenceinhisresearchwork.

3.2. Interactiveeffectsofmachiningparametersoncutting force

The changeintheeffectof onevariable,when thesecond variableischangedfromoneleveltoanother,isknownasInter- action effect. The interaction effects of the processvariables areusefulinunderstandingtheprocessbehavior.Thetwo-way interactive effects of the processvariableswhichhavestrong interactionwithcuttingforceareshowninFigures8–14 3.2.1. Interactiveeffectsofspindlespeedandfeedratefor thrustforce

Figure 8 showsthe interactive effectof spindle speedand feedrate onthrust force. Ithasbeen inferred fromthe direct

(10)

Spindle speed (N) rpm

Feed rate (F) mm/rev

3500 3000

2500 2000

1500 0.06

0.05

0.04

0.03

0.02

Axial depth of cut (X) mm 2 Radial depth of cut (Y) mm 2

% wt SiCp(W) 15

Hold Values

>

– – – – – –

< 0

0 20

20 40

40 60

60 80

80 100

100 120

120 Thrust force(N)

Fig.8.Interactiveeffectofspindlespeedandfeedrateonthrustforce.

Radial depth of cut (Y) mm

3500 3000

2500 2000

1500 3.0

2.5

2.0

1.5

1.0

Feed rate (F) mm/rev 0.04 Axial depth of cut (X) mm 2

% wt SiCp(W) 15

Hold values

>

– – – –

< 50

50 75

75 100

100 125

125 150

150 Crossfeed (Fy) N

Fig.9.Interactiveeffectsofspindlespeedandradialdepthofcutforcrossfeedforce.

effectanalysisthatthespindlespeedhasanegativeeffectand feedrate hasapositiveeffectoninfeedforce, crossfeedforce andthrustforce.It isobservedthat thethrust forcedecreases withincreaseincuttingspeedfrom1500rpmto3500rpm.The trendgetssimilarforallthelevelsoffeedratefrom0.02mm/rev to0.06mm/rev.

3.2.2. Interactiveeffectsofspindlespeedandradialdepth ofcutforcrossfeedforce

Figure9indicatestheinteractiveeffectofspindlespeedand radialdepthofcutforcrossfeedforce.Ithasbeeninferredfrom thedirecteffectanalysisthatthespindlespeedhasthenegative effectandradial depth of cut hasa positiveeffect. It reveals

thatcrossfeedforceincreasewithincreaseinspindlespeedfor all the level. The trendgets same for the radial depth of cut from 1mm to2mm. The crossfeed force is minimalfor the 3500rpmofthespindlespeedand1mmoftheradialdepthof cut.

3.2.3. Interactiveeffectsofaxialdepthofcutandwt%of SiCpforInfeedforce

Figure10showstherelationbetweenaxialdepthofcutand wt%ofSiCponInfeedforce.Thedirecteffectanalysisindicates thatbothaxialdepthofcutandwt%ofSiCphaveapositiveeffect oninfeedforce,crossfeedforceandthrustforce.Itisobserved that axialdepthofcutincreasesresultedinincreaseof infeed

(11)

Axial depth of cut (X) mm

% wt SiCp(W)

3.0 2.5

2.0 1.5

1.0 25

20

15

10

5

Spindle speed (N) rpm 2500

Feed rate (F) mm 0.04

Radial depth of cut (Y) mm 2 Hold values

>

– – –

< 100

100 150

150 200

200 250

250 Infeed (Fx) N

Fig.10.Interactiveeffectsofaxialdepthofcutandwt%ofSiCpforinfeedforce.

Axial depth of cut (X) mm

0.06 0.05

0.04 0.03

0.02 3.0

2.5

2.0

1.5

1.0

Spindle speed (N) rpm 2500 Radial depth of cut (Y) mm 2

% wt SiCp(W) 15

Hold values

>

– – – –

< 80

80 120

120 160

160 200

200 240

240 Infeed (Fx) N

Fig.11.Interactiveeffectsoffeedrateandaxialdepthofcutforinfeedforce.

force.A similartrendis followedforall thelevels of weight percentageofsiliconcarbide.

3.2.4. Interactiveeffectsoffeedrateandaxialdepthofcut forinfeedforce

Figure11displaystheinteractiveeffectofthefeedrateand theaxialdepthofcutontheinfeedforce.Thedirecteffectanaly- sisshowsthatbothfeedrateandaxialdepthofcuthaveapositive effectonthecuttingforce,whichindicatesthatanincreasein the feed rate increases the infeed force from 0.02mm/rev to

0.06mm/revforallthelevelofaxialdepthofcut.Thecutting forceisminimumwhen0.02mm/revto0.03mm/revoftherate and1–2.3mmoftheaxialdepthofcut.

3.2.5. Interactiveeffectsofspindlespeedandfeedratefor surfaceroughness

Figure12showstheeffectofthespindlespeedandthefeed rateonsurfaceroughness.Theinferencefromthedirecteffect analysisisthatthespindlespeedhasanegativeeffectandthe feedratehasapositiveeffect.Itisalsoobservedthatanincrease

(12)

3500 3000

2500 2000

1500 0.06

0.05

0.04

0.03

0.02

Axial depth of cut (X) mm 2 Radial depth of cut (Y) mm 2

% wt SiCp 15

Hold values

>

– – – –

< 3

3 4

4 5

5 6

6 7

7 Roughness (Ra) µm

Surface

Fig.12.Interactiveeffectsofspindlespeedandfeedrateforsurfaceroughness.

% wt SiCp

0.06 0.05

0.04 0.03

0.02 25

20

15

10

5

Spindle Speed (N) rpm 2500 Axial depth of cut (X) mm 2 Radial depth of cut (Y) mm 2

Hold values

>

– – –

< 5

5 6

6 7

7 8

8 Roughness (Ra) µm

Surface

Fig.13.Interactiveeffectsoffeedrateandwt%ofSiCpforsurfaceroughness.

in the spindle speed resulted in a better surface finish. The sametrendfollowsforthechangeoflevelofthefeedratefrom 0.02mmto0.04mm.Itisobservedthatthesurfaceroughness isminimumat3500rpmof spindlespeedand0.02mm/revof feedrate

3.2.6. Interactiveeffectsoffeedrateandwt%ofSiCpfor surfaceroughness

Figure13describestheinteractiveeffectoffeedrateandwt%

of SiCp surfaceroughness. The interpretationfromthe direct effectanalysisis thatbothfeedrate andwt%of SiCp havea positiveeffectonsurfaceroughness.Thegraphindicatesthat, decrease in feedrate, results the better surface finish for the

changeofwt%SiCpfrom5%to25%.Thebettersurfacefinish canbemaintainedat5wt%SiCp.

3.2.7. Interactiveeffectsspindlespeedandwt%ofSiCpfor surfaceroughness

Figure14indicatestheinteractiveeffectofspindlespeedwt%

ofSiCponsurfaceroughness.Fromthedirecteffectanalysis,the spindlespeedhasanegativeeffectandwt%ofSiCphasapositive effect.Itinfersthat,theincreaseinspindlespeeddecreasesthe surfaceroughnessatalllevelsofweightpercentageofsilicon carbide.At3500rpmofspindlespeedand5wt%SiCpensures bettersurfacefinish.

surface methodology Assessment of cutting force and surface roughness in LM6/SiC usingresponse Technology Journal of Applied Researchand

Journal of Applied Research and Technology

Original

Assessment of cutting force and surface roughness in LM6/SiC p using response surface methodology

Aezhisai Vallavi Muthusamy Subramanian

, Mohan Das Gandhi Nachimuthu

, Velmurugan Cinnasamy

Assessment of cutting force and surface roughness in LM6/SiC _p using response surface methodology