An
alternative
evaluation
method
for
friction
condition
in
cold
forging
by
ring
with
boss
compression
test
Chengliang
Hu
a,
Hengan
Ou
b,∗,
Zhen
Zhao
a,∗aInstituteofFormingTechnology&Equipment,ShanghaiJiaotongUniversity,Shanghai200030,China
bDepartmentofMechanical,MaterialsandManufactureEngineering,UniversityofNottingham,NottinghamNG72RD,UK
a
r
t
i
c
l
e
i
n
f
o
Articlehistory:
Received6February2014 Receivedinrevisedform 27September2014 Accepted12April2015 Availableonline22April2015
Keywords:
Friction RCT RCT-B
Calibrationcurves Frictiondifference
a
b
s
t
r
a
c
t
Ringcompressiontest(RCT)isaverypopularmethodusedtoquantitativelyevaluatefrictionconditions atthetool–workpieceinterfacebymeasuringthevariationsoftheinnerdiameteroftheringinmetal forming.TherearemanypossibilitiesformeasuringtheinnerdiametervariationsinRCTbecauseof non-uniformdeformationoftheinnerholeduringthetest.Suchnon-uniformdeformationoftheinnerhole causesdifficultiesinprecisemeasurementoftheinnerdiameterandhencetheaccuracyofthederived frictioncoefficient.ToavoidthedisadvantageindimensionmeasurementoftheconventionalRCT,an alternativemethodforevaluatingfrictionconditionsincoldforgingnamedringwithbosscompression test(RCT-B)isproposed.BytheintroductionoftheRCT-Bconcept,finiteelementsimulationresultsunder differentfrictionconditionswereobtained.Resultsshowedthattheshapeoftheouterbossremains sta-bleduringthecompressiondeformationandallowsthediameteroftheouterbosstobemeasuredmore easilyandaccurately.ThecalibrationcurvesoftheRCT-BconceptwereconstructedbyusingFE simula-tion,whichcovertherangeoffrictionconditionsincoldforgingprocess.Finally,theRCT-Bmethodwas successfullyappliedtodeterminethefrictionfactorsoffourdifferentlubricatingconditionsin compres-sionofaluminumrings.Furthermore,thephenomenawithdifferentlubricatingconditionsbetweenthe upperandlowerdie–workpieceinterfaceswerealsostudiedusingbothsimulationandexperimental testing.Theresultsshowthatitispossibletoquantitativelyassessthedifferenceoffrictionconditions attheupperandlowerdie–workpieceinterfacesbysimplycheckingtheinclinedangleoftheouterboss withtheRCT-Bmethod.
©2015TheAuthors.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense (http://creativecommons.org/licenses/by/4.0/).
1. Introduction
1.1. Historicaldevelopment
Inmetalformingprocesses,frictionalwaystakesplaceatthe
interfacebetweentheworkpieceandthedieortool.Friction
con-ditionhasasignificanteffectonthematerialflow.Ononehand,
becauseofhighfrictioncondition,underfillingofmaterialintothe
diecavityoccurs.Ontheotherhand,severelocalfriction
condi-tioncouldbe beneficial tothefilling of thediecavity. Friction
playsanimportantroleinthefinallyrequiredformingload,i.e.,
higher frictionresults in largertotal formingload required for
∗ Corresponding authorsat: Instituteof Forming Technology&Equipment,
SchoolofMaterialsScienceandEngineering,ShanghaiJiaotongUniversity,1954 HuashanRoad,Shanghai,200030,China.Tel.:+862152580156/+441158467391; fax:+862162826575.
E-mailaddresses:[email protected](H.Ou),[email protected],
[email protected](Z.Zhao).
completingnecessarydeformationinametalformingprocess.
Fric-tionalsodirectlyinfluencesthesurfacequalityoftheformedparts
andexcessivefrictionleadstowear,pick-uporgallingofthetool
surfaceandhencereducestoollife.Therefore,frictionisan
impor-tantparameterinmetalformingandshouldbepreciselymeasured
andcarefullycontrolledforformingqualifiedproductsatlowcost.
Theringcompressiontest(RCT)hasproventobeoneofthemost
popularandcommonlyusedmethodsforquantitativeevaluation
offrictionconditionsinbulkformingoperations(RaoandSivaram,
1993).TheringcompressiontestwasfirstlyusedbyKunogi(1956)
asaqualitativemethodofcomparinglubricantsforcoldextrusion.
Inthetest,aflatring-shapedspecimeniscompressedto50%
reduc-tionusingflattool.Ifthelubricanthasagoodlubricationpropertyto
ensureverylowfrictioncoefficient,theinnerdiameter(ID)expands
duringthedeformation,andwhenthefrictioncoefficientishighthe
IDshrinks,asshowninFig.1.Itisquiteeasytorecognizethe
dif-ferentfrictionconditionsofdifferentlubricantsbyobservingthe
IDvariationduetoitssensitivitytofriction.
To determine the friction coefficient for hot-working
pro-cess, the RCT was further developed by introducing a set of
http://dx.doi.org/10.1016/j.jmatprotec.2015.04.010
(a)
(b)
Fig.1.Ringcompressiontest:(a)lowfriction;(b)highfriction.
calibrationcurves by Male and Cockcroft (1964–1965). For the
calibrationwork,thering specimensof 6:3:2were compressed
understickingandzerofrictionconditions.Analyticaldatafrom
Schroederand Webster(1949) which were extracted fromthe
experimental results of compression of disc-shaped specimens
wereusedintheintermediatefrictionconditions.Byconducting
thetestusingarangeofmaterialswithdifferentdeformationand
lubricantsatelevatedtemperatures,RCTappearstobeareliable
methodindetectingthevariationsoffrictionconditionsduring
bulkplasticdeformation.
Based ontheassumptionofuniformdeformation aswellas
conformanceofMisesruleandconstantshearfrictionfactor,the
first satisfactoryanalysis of thecompression of a flat ringwas
madebyAvitzur(1964)throughanoptimumupperboundmethod
andlaterverifiedbyHawkyardandJohnson(1967)usingastress
analysisapproach,andthecurrentpositionoftheneutralsurface,
whichdividestheregionsofinwardandoutwardflowsof
mate-rial,wasdetermined.Bydividingthewholedeformationprocess
intoaseriesofsmalldeformationincrements,MaleandDepierre
(1970)providedapossiblemeansforthecalibrationofdifferent
ringgeometriesbycomputersolution.However,theactual
fric-tionfactorobtainedbyusingthetheorytoanalyzeexperimentally
determinedshapechangesappearstoproduceadegreeof
overes-timation.
To minimize the discrepancy between the theoretical and
experimentalresults, thebulging effectcaused bynon-uniform
deformation should be further considered. By using a simple
paraboliccurvewithonlyoneuncertaincoefficienttorepresent
theprofile of thebulge,Liu (1972)theoretically calculated the
stress and strain components at the outerequator based on a
supposedvelocityfielddescribedbyatrigonometricfunctionand
presentedasetofcalibrationcurvesfor6:3:2ringgeometry.By
introducingaconstantaveragevalueofstressateachstepof
rel-ativesmalldeformation,LeeandAltan(1972)computedanother
setofcalibrationcurvesinconsiderationofbulging.Byadoptingan
assumedvelocityfielddescribedbyexponentialfunction,Depierre
andGurney(1974)developedareliablemethodfortreating
exper-imentalresults fromRCTwith andwithout bulgeformation so
astoobtainaquantitativeevaluationoffrictionfactor,inwhich
a paraboliccurve withtwo unknowncoefficients wereusedto
describethebulgeprofile.Comparedtothecurvesdevelopedby
Maleand Depierre(1970),thecalibrationcurves consideringof
bulgingareclosertotheexperimentalresults.
Tofacilitatetheapplication ofRCTinindustry,Abdul(1981)
attemptedtodevelopacalibrationchart,whichprovidesfriction
factorlinesandheightreductionlinesagainstthegraduatedaxis
ofIDchange,withouttheconsiderationofthebulgingeffect.To
considertheactualnon-constantfrictionshearstressdistributions
duringRCT,Wang(2001)proposedanewwaytorecreate
calibra-tioncurvesbyusingthenon-constantfrictionstressexpressedby
afunctionoftheF-coefficient,andtheobtainedcalibrationcurves
showedagoodagreementwithexperimentaldata.
Tosimulatethestateoflowandhighlevelsofnormalpressure
onthetool–workpieceinterfaceofmetalforming,twoalternative
ringcompressiontestswithconcave-shapedandconvex-shaped
cross-sectionintheringgeometrieswereputforwardasa
com-plementaryapproachofRCT.Thecorrespondingcalibrationcurves
wereobtainedusingFEManalysisandverifiedbytheexperimental
results(Petersenetal.,1998;Tanetal.,1998).
1.2. DimensionalmeasurementofRCT
ThepopularityofRCTcanbeattributedtoitspractical
conve-nienceandthefrictionconditioncanbequantifiedbymeasuring
thedimensionalchangeoftheinnerdiameter.Theaccuracyinthe
measurementsofthespecimendimensionsbeforeandafter
test-inghasasignificanteffectontheaccuracyoftheresultsoffriction.
Originally aimingattheslightlynon-circularshape upon
defor-mation,anenlargedtracingofeachspecimenwasmadeusinga
profileprojector,theareaofthecentralholewasfoundbymeans
ofaplanimeter,andthemeandiameterwascalculatedassuming
thatthiswastheareaofatruecircle,andtheincurrederrorwas
foundtobe1in250(MaleandCockcroft,1964–1965).
AsrecommendedbyWangandLenard(1992),theinner
diam-eterwasmeasuredintwodirectionsatrightangles,onboththe
specimenendsandatthemiddle.Theaveragewasusedto
calcu-latethepercentagechange.Theheightreductionpercentagewas
alsocalculatedfromtheaverageoffourmeasurements.Inthiscase,
onlycaliperwasneeded.
MeasurementsoftheheightandIDweremadebyutilizinga
dig-italcaliper(Petersenetal.,1998),whereasdetailsofthegeometrical
profileofthedeformedspecimenswereobtainedonatoolmaker’s
microscopeequippedwithanX–Ymicrometertableconnectedtoa
PC-baseddata-loggingsystem.Toallowdetailedcomparisonwith
thecomputedchangesingeometryforthecomplementaryRCT,
theexternalprofileofselectedspecimenswasdigitizedusingan
inductiveprobe.
Consideringthesoftpropertyoftheworkpiecematerial
‘Plas-ticine’,aftercompressiontheIDwasmeasuredbeforethespecimen
wasremovedfromthebottomplatentoavoidanychangeofshape.
Anaveragevaluewastakenfromthreemeasurementsonthetop
surfacefromthreearbitraryanglesacrossthecentreofthering
(Robinsonetal.,2004).
AsshowninFig.1,theIDattheflatfacesdiffersfromtheIDatthe
specimenmid-lineduetobarreling.Hartleyetal.(2007)proposed
amethodtomeasuretheIDbyplacingaball-bearingofknown
diameterontothespecimensothatitprojectsslightlyintotheinner
hole,andthistechniqueenablesmeasurementoftheIDattheflat
specimenfaces.FromthedimensionsdepictedinFig.2,allofthese
canbemeasuredaccuratelyusingastandardmicrometer,theIDis
simplycalculatedfrom
ID=2
D2
4 −
OH−h−D2
2 (1)As mentioned by Shahriari et al. (2010), a profile projector
equippedwithanX–YmicrometertableconnectedtoaPC-based
Fig.2.Diagramillustratingthemeasurementtechniquewithball-bearing(Hartley etal.,2007).
Fig.3.Cross-sectionofdifferentRCTspecimen(Macaskill,2012).
ball-bearingwerebothusedtodeterminethefrictioncoefficients
fortheNimonic115hotforgingprocessbyRCT.
For theconvex(high friction) profile,thesmallestID atthe
equatorcouldbeproperlymeasuredbyusingtheprofileprojector
andplanimeter.Fortheconcave(lowfriction)profile,theprofile
projectorcouldnotcatchthelargestIDattheequator.Although
themeasurementtechniquewithball-bearingmaybeused
conve-niently,theassumptionofthedeformedprofileasacircularshape
makesthetechniquenotsoprecise.Toobtaintheaveragevalue,the
measurementoftheIDbyusingacaliperisaneasybutnota
pre-ciseway,becauseitisdifficulttopositiontheequator.Togetmore
accuratevalueoftheIDusingasimplecaliper,thespecimenafter
deformationcouldbecutatacrosssectionthroughthe
axisym-metricz-axis,asshowninFig.3.However,itrequiresmoretimeto
completethecutting.
InadditiontothetwotypicaldeformedmodesshowninFig.1,
Abdul(1981)reportedthreeotherpossiblemodesof
inhomoge-neousdeformationduringhisexperimentalstudyatvariouslevels
offriction.Goetzetal.(1991)alsofoundthattheprofileoftheID
afterdeformationwasnotonlyconcaveorconvex,butalsostraight
orbuckled.Andinthepracticalexperimentalwork,therearemany
possibilitiesforthedeformedIDprofile,asshowninFig.3.This
indicatesthat theshapeof theIDof thering duringtestis not
souniform.Correspondingly,theshapeoftheouterdiameteralso
becomesunsteady.
Inthepresentwork,toovercomethedisadvantagein
measure-mentof theinner diameterinconventional RCT, analternative
methodtoevaluatethefrictionconditionnamelyringwithboss
compressiontest(RCT-B)wasproposed.Inthefollowingsections,
theRCT-Bconceptis firstdiscussed.InSection3,finiteelement
(FE)simulationsarecarriedoutfordetailedevaluationofmaterial
deformationsunderdifferentRCT-BfrictionconditionsandtheFE
resultsareusedtoconstructthecalibrationcurvesoftheRCT-B.
Section4summarizestheproceduresforboththeRCT-BandRCT
experimentssoastovalidatetheRCT-Bmethodascomparedtothe
Fig.4. TheringgeometryofRCT-B.
conventionalRCTapproach.Detailedevaluationanddiscussionof
thefrictionconditionsobtainedfromboththeRCT-BandRCT
meth-odsaregiveninSection5.Thisisfollowedbytheconclusionsdrawn
fromthisresearchinSection6.
2. Theconceptofringwithbosscompressiontest(RCT-B)
Toachieveauniformlydeformedshape,theringwithanouter
bossisproposedforthecompressiontestbasedonthefollowing
considerations:(1)Theouterbosswouldnotbedeformedduring
thecompressionprocessbecauseitdoesnotcontactwiththedies;
(2)Theouterbosswouldbearigidzoneexpandingwiththeouter
diameterbecauseitisnotaffectedbythedeformationofthering.
The ring geometry 6:3:2 used by Male and Cockcroft
(1964–1965)iscommonlyusedforobtainingexperimental
cali-brationcurvesbecauseofpracticalconvenience(RaoandSivaram,
1993).Theproposedringgeometrywithbossisdesignedbasedon
thiscommonlyusedRCTgeometry.AsshowninFig.4,theratio
ofouterdiametertoinnerdiameterandtoheightis6:3:2,andthe
heightandwidthoftheouterbosshavethesamevalueof0.4.Based
ontheaboveconsiderationsofmaterialdeformationoftheRCT-B
specimenundercompression,itisassumedthatthevariationsof
thebosswouldbesufficientlysensitivetodifferentfriction
condi-tions.SimilartotheconventionalRCTapproach,itisnecessaryto
constructasetofcalibrationcurvesforthemeasurementoffriction
conditions.FEmethodmaybeusedtoevaluatematerial
deforma-tionoftheRCT-Bspecimenunderdifferentfrictionconditionsand
toconstructthecalibrationcurvesoftheproposedRCT-Bmethod.
3. Finiteelementsimulation
3.1. Analysisofdeformationprocess
Inthis work,aluminum materialAl6082wasselectedasthe
materialfortheRCT-B.Toobtainmaterialpropertiestensiletests
werecarriedout,andthetruestressandstraincurvewasmodeled
usingthesimple powerlaw=Kεn.Afterregressionanalysisof
experimentaldata,theflowstressconstantKandthestrain
hard-eningexponentnwerecalculatedtobeK=191.88andn=0.181,
respectively.Thusthespecificmaterialmodelcouldbedetermined
andadoptedfortheFEmodel.
2Daxisymmetricmodelwithshearfrictionlawwasusedto
sim-ulatetheRCT-Btestsatroomtemperature.Thesimulationresultsof
deformationprocessunderdifferentfrictionconditionsareshown
inFig.5.Whenthefrictionfactoris0.0,thematerialuniformlyflows
outwardandtheinnerdiameterkeepsstraightwithheight
reduc-tion.Whenthefrictionfactoris0.1,thematerialflowsoutward
andaconcaveprofileofIDisobtained.Whenthefrictionfactoris
0.3,thematerialflowsinwardandaconvexprofileofIDisformed.
AllthesephenomenaontheIDdeformationarequitesimilartothe
conventionalRCT.Thissuggeststhattheouterbossdoesnotchange
Fig.5.SimulationresultsofdeformationduringRCT-Bunderdifferentfrictioncondition:(a)frictionfactorm=0.0,heightreductionRH=30%;(b)m=0.1,RH=30%;(b)m=0.3,
RH=30%;(e)m=0.0,RH=50%;(f)m=0.1,RH=50%;(g)m=0.3,RH=50%;(e)m=0.0,RH=60%;(f)m=0.1,RH=60%;(g)m=0.3,RH=60%.
Whatismoreimportantisthatthereisalmostnodeformation
attheouterbossofthering,asshowninFig.5.Theequivalentstrain
inthatareaisverysmallduringthedeformationprocess.Evenat
veryhighfrictionconditionssuchasm=0.7andm=1.0,theshape
oftheouterbossstillkeepsunchangedasarigidzone,asshown
inFig.6.DuringtheRCT-Btest,theouterbossexpandswiththe
outerdiameterandthereisnotanybulgingorbarrelingeffectat
theouterboss.Therefore,theouterdiameterofthespecimencan
bemeasuredquiteeasilyandprecisely.
3.2. Constructionofcalibrationcurves
From the literature mentioned above, discrepancies exist
betweenthereportedcalibrationcurvesbecauseofmanyfactors
thataffectthefinalshape ofthecalibrationcurves. Goetzetal.
(1991)examinedtheeffectsofstrainrate,innerdiameterprofile
andringgeometryuponringcurvesusingtherigidviscoplasticFE
programALPIDandexperimentaldata.Anderssonetal.(1996)
eval-uatedtheeffectoftheheat-transfercoefficientinring-compression
testsusingacommerciallyavailableFEprogramDEFORM,andthe
effectcanbeconsideredsmall ascompared totheinaccuracies
involvedinthecalibrationcurves.Byinvestigatingtheeffectsof
materialproperties,strain-ratesensitivity,andthebarrelingeffect
onthecalibrationcurvesbyusingFEsimulation,SofuogluandRasty
(1999)concludedthattheuseofageneralizedcalibrationcurves
regardlessofthematerialtypeandtestconditionsmustbeavoided.
Itwasrecommendedthatspecificcalibrationcurvesforaspecific
materialbeusedtoobtainreliabledataoffrictioncoefficientsunder
thespecifictestconditions.Forthesereasons,thecorresponding
calibrationcurvesforthenewRCT-Bconceptandtheconventional
RCTarebothconstructedbyFEmethod.
ToobtainthecalibrationcurvesofthenewRCT-Bconcept,the
displacementofthenodeatthemiddleoftheouterbossandone
nodeatthetopedgeareextractedfromtheFEsimulationresults
toquantifythedeformationoftheringwithbossspecimen.The
reductionintheouterdiameterofthebossafterdeformationis
calculatedasROD=(DB1−DB0)/D0andthereductioninheightis
cal-culatedasRH=(H0−H1)/H0,asshowninFig.7.Finally,thespecific
calibrationcurvescanbecreatedasshowninFig.8.
Tomakenecessarycomparisons,thecalibrationcurvesforthe
conventionalRCTarecreatedbasedonthereductionintheinner
diameteratthemiddleofthespecimen,asshowninFig.9.
Fig.7.Reductioninouterdiameter.
0 10 20 30 40 50 60
0 10 20 30 40 50 60 0.60-1.00 0.50 0.20 0.15 0.10 0.40 0.30 0.05 0.00 Re du ct io n i n ou ter d ia me ter of boss( %)
Reduction in height(%)
Fig.8. CalibrationcurvesfortheRCT-B.
0 10 20 30 40 50 60
-30 -20 -10 0 10 20 30 40 0.20 0.15 0.10 0.05 0.00 Re du cti on i n i n te rna l di ameter (%)
Reduction in height (%)
0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
Fig.9.CalibrationcurvesoftheRCTapproach.
Itisobservedthatofthechangeoftheouterdiameteroftheboss
inRCT-BislesssensitivethanthatoftheinnerdiameterintheRCT,
andwhenthefrictionfactorislargerthan0.5itwouldbedifficult
fortheRCT-Btoquantitativelydeterminethefrictioncondition.
However,inmostcoldforgingprocessesthefrictionfactorwould
belessthan0.5,sotheRCT-Bmethodcouldbeusedtomeasurethe
frictionconditions.
Forfurtherstudy,FEsimulationswerecarriedoutwithdifferent
frictionfactorsbetweenthetopandbottomdie–workpiece
inter-faces.AsshowninFig.10,becauseofthefrictiondifferencethe
inclinedanglecanbeobservedeasily,asshowninFig.10dandf).
Asdiscussedabove,thesensitivitytotheinterfacialfrictionofthe
RCT-Breduceswhenthefrictionfactorbecomeslargerthan0.5.
Therefore,theinclinedanglesunderthesamefrictiondifference
betweenthetopandbottomsurfacesathigherfrictionconditions
(Fig.10i–k)areobviouslysmallerthantheresultsatlowerfriction
conditions(Fig.10a,b,gandh)asexpected.
AsshowninFig.10k,thefrictionfactorsatthetopandbottom
die–workpieceinterfacesaresupposedtobe0.70and0.75,
respec-tively,andtheinclinedangleissmall.However,whenthereduction
inheightisincreasedfrom60%to65%,theinclinedanglebecomes
larger.Theinclinedangleoftheouterbossalsoincreaseswiththe
increaseofdeformationduringtheRCT-BasshowninFig.10kand
l.
Basedontheaboveanalysis,theRCT-Bcanbeusedtoevaluate
thelubricationperformance(frictioncondition)byusingdifferent
lubricantsbetweenthetopandbottomdie–workpieceinterfacesin
onesimpletest.Accordingtothesimulationresults,ifthefriction
differenceismorethan0.02atlowerfrictionconditionandismore
than0.05atlargerfrictioncondition,itispossibletodistinguishthe
differencebysimplyinvestigatingtheinclinedangleoftheouter
bossofthering.
4. Experimentalworks
TheworkpiecesshowninFig.11weremachinedfromannealed
Al6082bar.Fortheringworkpieceswithouterboss,thenominal
dimensionsoftheouterdiameter,innerdiameter,height,aswell
astheheightandwidthoftheouterbossare18.00mm,9.00mm,
6.00mm,1.20mmand1.20mm,asshowninFig.4.Forthe
con-ventionalring workpiece,the nominaldimensionsof theouter
diameter,innerdiameterandheightare18.00mm,9.00mmand
6.00mm,respectively.
Twocylindricaldieswith100.00mmdiameterweremachined
and hardened by heat treatment, and then the diesurface
re-hardeningwascarriedoutbyTufftrideprocess.Theworkingsurface
of thedie wastreated by lapping and polishing processes and
themirrorsurface witha surfaceroughnessof Ra0.05mwas
achieved,asshowninFig.11c.
Threetypesoflubricatingoilswereadoptedasthelubricantfor
thetests,andthemainpropertiesareshowninTable1.Duringeach
test,thedieworkingsurfaceswerecleanedupwithalcoholatfirst,
thenalubricatingoilwasbrushedonthemasevenaspossible,after
thataworkpiecewasputinthemiddleandthecompressiontest
wasperformed.
Allthetestswereconductedona250kNInstron5985
mate-rialtestingmachineandthecompressionvelocitywascontrolled
at1mm/minbyaBluehill2system.Thedimensionsofworkpiece
beforeandaftercompressionwerecheckedusingadigitalvernier
caliper.DuringtheRCT-Btests,theinitialouterdiameterofboss,
outerdiameterandheightofeachworkpieceweremeasured,and
theouterdiameter ofboss and theheightwerealso measured
aftertest,eachdimensionwasobtainedfromtheaverageofthree
measurements.DuringtheconventionalRCTtests,theinitialinner
diameterandheightofeachworkpieceweremeasured, andthe
innerdiameterandheightwerealsomeasuredaccordingtothe
recommendationfromWangandLenard(1992),eachdimension
Fig.10.Deformationresultswithdifferentfrictionfactorbetweenthetopandbottomdie:(a)RH=60%;(b)RH=60%;(c)RH=60%;(b)RH=60%;(d)RH=60%;(h)RH=60%;(i)
RH=60%;(j)RH=60%;(k)RH=60%;(l)RH=65%.
Fig.11. Workpiecesanddiesforexperimentaltests.
5. Resultsanddiscussion
5.1. Determinationoffrictionfactors
Fig.12showsthedeformationconditionoftheringworkpieces
withouterbossaftercompressionatdifferentheightreductions.
Itcanbeobservedthattheworkpiecesaftercompressionunder
thedrydiesstillkeptincircularshapeevenwhenthereduction
inheightreached58%.Andasexpectedaccordingtothe
simula-tionresultsdiscussedabove,littledeformationtookplaceatthe
outerboss.Thisconfirmsthattheshapeoftheouterbossremains
undeformedduringthetest.
Bycalculationofthereductionoftheouterdiameterofboss
andthereductionofheight,thefrictionfactorsunderfourdifferent
lubricatingconditionsincludingtheuseofthreelubricantsanddry
contactbetweentheworkpieceanddiesweredeterminedbased
onthecalibrationcurvespresentedinFig.8.AsobservedinFig.13,
thefrictionfactorswithlubricatingoilA,BandCare0.21,0.20and
0.22,respectively.Thefrictionfactorunderdrycontactcondition
is0.24.
FortheconventionalRCTcases,thereductionoftheinner
diam-eter and thereduction of heightwere alsocalculated, and the
corresponding data are shown in Fig. 14. From the calibration
curves,thefrictionfactorsfromtheconventionalRCTcaseswith
Table1
Lubricatingoilsusinginthetest.
No. Lubricatingoil Density@15◦C/kg/m3 Kinematicviscosity@40◦C/mm2/s Kinematicviscosity@100◦C/mm2/s Viscosityindex
A RandoHD68 – 64.6 8.4 98
B MagnaBD68 880 68 8.7 99
Fig.12.Theringworkpiecewithouterbossaftercompressionunderthedrydies.
0 60 6 0
5 0
4 10
20 30 40 50 60
0.00
0.05 0.15 0.20 0.30 0.40 0.50
Reduction in height(%)
Re
du
ctio
n in ou
ter
diameter
wi
th
b
o
ss(%
)
L-B (m=0.20) L-A (m=0.21)
Dry (m=0.24) L-C (m=0.22)
Fig.13.ExperimentaldatapointsonthecalibrationchartoftheRCT-Btests.
0 6 0
5 0
4 -30
-20 -10 0 10 20 30
Dry (m=0.28) L-B (m=0.21) L-A (m=0.23)
L-C (m=0.25)
Re
du
cti
on
in
inn
er
d
iam
ete
r(%)
Reduction in height (%)
0.20
0.15
0.10
0.05
0.00 0.30 0.40 0.50
Fig.14.ExperimentaldatapointsonthecalibrationchartoftheconventionalRCT tests.
Fig.15.Theringworkpieceaftercompressionunderthedrydies.
lubricatingoilA,BandCcouldbedeterminedas0.23,0.21and 0.25,respectively,andthefrictionfactoratdrycontactcondition canbedeterminedas0.28.
ComparingthetestresultsobtainedbyboththeRCT-BandRCT methods,themeasurementsoffourdifferentfrictionconditions fromtheRCT-Bcasesareinagoodagreementwiththatfromthe conventionalRCTmethodproducingamaximumerrorlessthan 15%inthedrycontactcondition,asshowninTable2.Ingeneral,
thespecificfrictionfactordeterminedbytheRCTmethodislarger
thanthatfromtheRCT-Bmethod.AsshowninFig.15,theinner
diameteroftheconventionalRCTringisnotcircularandinsteadit
becomesanovalshapeespeciallyatlargerreductionofheight.The
minimuminnerdiameterduetotheovalshapetoacertaindegree
contributestotherelativelargerreductionofinnerdiameterand
hencelargerfrictionfactorvalue.
5.2. Evaluationoflubricationconditions
Toevaluatethefrictionsunderdifferentlubricatingconditions
atthetopandbottomsurfacesbetweentheworkpieceanddies,
onlythelowerdiewasbrushedwiththebestlubricatingoilMagna
BD68andfollowedbycompressionsofthreeworkpieceswithouter
boss.AsshowninFig.16,theworkpiecesremainincircularshape
basically,andtheouterbossisinclinedtotheupperdieunderdry
condition,andtheinclinedangleoftheouterbossincreaseswith
thereductionofheightduringthetest.Theseexperimentalresults
areingoodagreementwiththesimulationresultsasdiscussedin
Section4,whichfurtherprovesthatthedifferenceinlubrication
conditionscanberecognizedbycheckingtheinclinedangleofthe
outerbossofthering.
Table2
Frictionfactorsdeterminedbydifferenttests.
No. Lubricatingoil Frictionfactorm(RCTtest) Frictionfactorm(RCT-Btest) Error(%)
A RandoHD68 0.23 0.21 8.7
B MagnaBD68 0.21 0.20 4.8
C Magna100 0.25 0.22 12.0
Fig.16.Theringworkpiecewithouterbossaftercompressionwithdifferentlubricatingconditionsbetweentheupperandlowerdie–workpieceinterfaces.
6. Conclusions
(1)Analternativequantitativeevaluationmethodforthefriction conditionsincoldforgingbyringwithbosscompressiontest wasproposedinordertoaddressthedifficultiesencountered inthemeasurementoftheinnerdiameterinconventionalring compressiontest.
(2)FortheRCT-Btest,theproposedproportionofringgeometry oftheouterdiameter,innerdiameter,heightandtheheighto andwidthoftheouterbosswasdesignedtobe6:3:2:0.4:0.4. Thecompressionprocessesoftheringwithbossunderdifferent frictionconditionsweresimulatedandtheresultsshowedthat theshapeoftheouterbossremainsundeformedduringthe ringdeformation.Theundeformedouterbosswasverifiedby theRCT-Bexperimentalresultstoallowtheouterdiameterof bosstobemeasuredeasilyandprecisely,whichenablesmore accuratemeasurementoffrictionconditions.
(3)ThecalibrationcurvesofthenewRCT-Bconceptandthe con-ventionalRCTmethodsarebothconstructedbyFEsimulations. AlthoughthechangeoftheouterdiameterofbossinRCT-Bis lesssensitivethanthatoftheinnerdiameterinthe conven-tionalRCT,itisstillsufficientforaccuratepredictionoffriction conditionwhenthefrictionfactorislessthan0.5.Therefore theRCT-Bmethodcanbeusedtomeasuretherangeoffriction conditionsincoldforgingprocess.
(4)TheRCT-Bmethodwassuccessfullyappliedtodeterminethe friction factors of four differentlubricating conditions with aluminumworkpieces.TheRCT-Bmeasurementresultswere verifiedbyconductingexperimentaltestsusingthe conven-tionalRCTmethod.ThetestingresultsfromboththeRCT-Band conventionalRCTwereina goodagreementintermsofthe measuredfrictionfactorswithamaximumerrorlessthan15%. (5)FEsimulationswithdifferentfrictionfactorsbetweenthetop andbottomdie–workpieceinterfacesarecarriedout.Theouter bossisinclinedtothediesurfaceofgreaterfriction,andthe inclinedangleoftheouterbossbecomeslargerwithgreater frictiondifferencebetweentheinterfaces.Theinclinedangle tendstoincreasewiththeincreaseofdeformationduringthe RCT-BasobservedbyRCT-Btestingresults.Thisobservation leadstotheconclusionthatthedifferenceoflubrication con-ditionscanbequantitativelyevaluatedbysimplyinvestigating theinclinedangleoftheouterbossunderringcompression conditions.
Acknowledgements
ThisworkwassupportedbytheNationalNaturalScience Foun-dation of China (No. 51475294), FP7-Marie Curie Action IRSES
MatProFutureproject(No.318968)andtheEngineeringand Physi-calSciencesResearchCouncil(EPSRC,EP/L02084X/1).Theuseful discussion with Prof. Niels Bay from Technical University of Denmarkonthetoolinglappingandpolishingduringthe46thICFG PlenaryMeetinginParisshouldbementioned.Theauthorswishto thanktheMr.TomBussoftheFacultyofEngineeringatNottingham forhissupportinexperimentaltesting.
References
Abdul,N.A.,1981.Frictiondeterminationduringbulkplasticdeformationofmetals. CIRPAnn.—Manuf.Technol.30(1),143–146.
Andersson,K.,Kivivuori,S.,Korhonen,A.S.,1996.Effectoftheheat-transfer coeffi-cientinring-compressiontests.J.Mater.Process.Technol.62(1),10–13. Avitzur,B.,1964.Forgingofhollowdiscs.Isr.J.Technol.2(3),295–304.
Depierre,V.,Gurney,F.,1974.Amethodfordeterminationofconstantandvarying frictionfactorsduringringcompressiontests.J.Lubr.Technol.,Trans.ASME96, 482–488.
Goetz,R.L.,Jain,V.K.,Morgan,J.T.,Wierschke,M.W.,1991.Effectsofmaterialand processingconditionsuponringcalibrationcurves.Wear143(1),71–86. Hartley,R.S.,Cloete,T.J.,Nurick,G.N.,2007.Anexperimentalassessmentoffriction
effectsinthesplitHopkinsonpressurebarusingtheringcompressiontest.Int. J.ImpactEng.34(10),1705–1728.
Hawkyard,J.B.,Johnson,W.,1967.Ananalysisofthechangesingeometryofashort hollowcylinderduringaxialcompression.Int.J.Mech.Sci.9(4),163–182. Kunogi,M.,1956.Anewmethodofcoldextrusion.J.Sci.Res.Inst.(Tokyo)50,
215–246.
Lee,C.H.,Altan,T.,1972.Influenceofflowstressandfrictionuponmetalflowin upsetforgingofringsandcylinders.J.Eng.Ind.,Trans.ASME94,775–782. Liu,J.Y.,1972.Ananalysisofdeformationcharacteristicsandinterfacialfriction
conditionsinsimpleupsettingofrings.J.Eng.Ind.,Trans.ASME94,1149–1156. Macaskill,S.,2012.Theeffectoffrictiononthegeometricdeformationofaluminium foraringcompressiontest.In:FinalYearReport.TheUniversityofNottingham. Male,A.T.,Cockcroft,M.G.,1964–1965.Amethodforthedeterminationofthe coef-ficientoffrictionofmetalsunderconditionsofbulkplasticdeformation.J.Inst. Met.93,38–46.
Male,A.T.,Depierre,V.,1970.Thevalidityofmathematicalsolutionsfor determin-ingfrictionfromtheringcompressiontest.J.Lubr.Technol.,Trans.ASME92, 389–397.
Petersen,S.B.,Martins,P.A.F.,Bay,N.,1998.Analternativering-testgeometryforthe evaluationoffrictionunderlownormalpressure.J.Mater.Process.Technol.79 (1),14–24.
Rao,K.P.,Sivaram,K.,1993.Areviewofring-compressiontestingandapplicability ofthecalibrationcurves.J.Mater.Process.Technol.37(1),295–318. Robinson,T.,Ou,H.,Armstrong,C.G.,2004.Studyonringcompressiontestusing
physicalmodellingandFEsimulation.J.Mater.Process.Technol.153,54–59. Schroeder,W.,Webster,D.A.,1949.Press-forgingthinsections:effectoffriction,
areaandthicknessonpressurerequired.J.Appl.Mech.16,289–294. Shahriari,D.,Amiri,A.,Sadeghi,M.H.,2010.Studyonhotringcompressiontestof
Nimonic115superalloyusingexperimentalobservationsand3DFEM simula-tion.J.Mater.Eng.Perform.19(5),633–642.
Sofuoglu,H.,Rasty,J.,1999.Onthemeasurementoffrictioncoefficientutilizingthe ringcompressiontest.Tribol.Int.32(6),327–335.
Tan,X.,Martins,P.A.F.,Bay,N.,Zhang,W.,1998.Frictionstudiesatdifferentnormal pressureswithalternativering-compressiontests.J.Mater.Process.Technol. 80–91,292–297.
Wang,F.,Lenard,J.G.,1992.Anexperimentalstudyofinterfacialfriction-hotring compression.J.Eng.Mater.Technol.114,13–18.