Seminar on Engine ering Mathematics @ 2008 Engineering Mathematics Group Editors: Am.mi Zaharim et al. TOOL LIFE ASSESSMENT USING FRACTIONAL FACTORIAL METIIOD IN TURNING OF TITANIUM ALLOY WITH CBN CUTTING TOOL Y. Burhanuddin, C.H. Che Haron, J.A. Ghani, A. K. A. M. Ihsan, G.A. Ibrahim,
A. Yasir, N. El-
Maghriby
Department of Mechanical and
Material
Engineering Universiti Kebangsaan Malaysia
43600
Bangi, Selangor, Malaysia e-mail : chase@vlsi.eng.
ukmy ABSTRACT Tool life is the
one of the most important parameters in the machining research area. Most researchers have dealt the effect of cutting variables on tool life by the one-variable-at-a-time method. This approach needs a separate set of tests for each combination of cutting condition and cutting tool. The approach required large amount of cost and cannot consider the combined effect of cutting conditions on response.
This research developed tool life model
which take account the combined effect of cutting variables using design of experiment including cutting speed, feed rate, and
depth of cut. The
effects
of
cutting variables
are
investigated by
the
application
of
fractional
factorial design method. The first-
order
of
tool life
mathematical model
are
formulated to predict tool life with a power form equation using cutting parameters. The
cuffing tests are conducted with Cubic Boron Nitride (CBN) as cutting tool when turning of titanium in dry condition. Keyword: machinability, tool life, cutting condition, fractional factorial, predicted equation INTRODUCTION Machinability data of titanium is usually obtained from the experimental work. The material is machined by varying only one cutting variable at a time while holding the other variables constant and known as one-variable-at-a-time (OVAT). This method is time consuming and requiring in large cost. The condition become much worse when the combination of work piece and cutting tool materials are expensive. The most experimenters have been using statistical design of experiment (DOE) despite of using one-factor-at-a-tir,ne to save cost and time.
Statistical design of
experiment
is the process of planning the experiments so that
the
appropriate data
should
be collected which may be analyzed by statistical methods resulting in
Ed.ited by: Azomi hharim et al. sswmtuhuocdedirynealatsu1lsrtohsnpainintingvidteTlseiisttigasfaneptaeieusesimbdl,seoa^femnteoetd'+smrviagatnk(eTeif,iic-pc6aurnAetttdrin-idc4gctvitod)o.nerspIitnthtihs,faunttnoeacocfetfliesogcsntearotoyhmretooecptrBe"ym.Nuaptnitlooodnyosltcohloiefoenodn.rei.titoTicnhasel rveasluiTdlhteathneodbdojaebtcajetiuvcnetiivfooefrcmthoinetycsluatusniddoynrseisd-u(rcMoeoentshttgaJobtomritsahelrtnyhuermgtuogeorr)l.olDifreeesfiiflgoiinmao"etfnoetrx.pceBriNmecnutttcinagn EXPERIMENTAL METHOD Design of Experiment fmdu1mfaipeuIlusnrefseeelaaaesxoliltexvavonlseercpcadpeeaIi'wiidAfrtrtTenmgealflmaocisst2orsninhychcrniueefnqmsit'it'peftuasoemtiohfem-ATeeoefiloxrlsr,el2alrenhsim,anpa.sfwkacnpseffnteloToa'ratt(usplehuahrlsTo2nhcl(fiemlxeftmceifghrtawehdtcaaprote.bfaceerifiecfenororrtaeteunathistdu2rdnarciokeroitcik-ktrmeatealr-ostion,t9isninrrsrefwdforree.sdeargdseonifmnotrsc)netasltstahrftrcp-toeiereoioircgorosaei.tidmrannrlhrmasnctsgelifddJtgoeacnateiopsaurd*seuddlctcitiaosdiosg{srumalteany,ehyo,itnltrixreihn.xirreiyfwn"lgpaimrtaiupcost*a.huhhee;ne,riploaifla"teTeerrar-"dufioaiideahrenmbrmnre;ser)cinren'xm;Gt;.eoedteeslh;ia.*pGi"nnoo"Jqrntefee;liftg*rrun.tlenfrdawbrrnitasccamntendpcaoeulrotso"sy"eetamta.-ln*iouai,-ltgspnoesryrur'ndbanenvtiituvi"4ipdgmineseussrce.nin"ottsirr'bssisuphlm/s/*,si2Luogoao"eloo.Tapai"mnfrf"sa"rrftobarms6naerprole;obvlgatyicdprner;ro."Potuarlrt";ritrtalhsutrerJch,;cJlthe"t"itpoo.se?nheirtrna*s;geohewseniJtpidrtnrsoreseld-ttIimiee"ytrsunea-pdncsld.siiedaarbbee,eiiuiaf.gvrrynrltrltryeeactehe"yhnaitfahntemapnlooseciserrcgoarirrdut.toenetamafinneorhntrmksyaediaeanaennehcrebnfratdncmiisactlginfiyedioeltlnfocouohegeoxsnatsfmnseddvocaprirttgraoratiiecm_onodsiosnrwgnntrioengnofiaeomiaritrtsdsbsnrnthyhdttihifeelhgaleaeueo2yeonleanlsrty_lfftl a mctdaoeaknpecTtshhsihdinoieinsifrntecogsdututiaandcasdycneaodpuuercsneneBtsdspNtehotnhnetcseoevsn.aifmtrreaTiuanchrbtt.teliaeoTnrsnheeael"oaluttiofsoaonuvclsa.thiorriiir"*appia,trisrob"senuaetswrnofetdfeaebdunscuiagteinsunJ;ofGootm"llf,oiul&-wc;irp'rir*,n'ie"ndgrrui:inmdgicneurgnenstinsppgmoeneefostdhre"cosefdea.eandrIde,t T: C(IPf d) * (1) cre'aaqltneu=sab-te(iomwznmirsiist/treteohnvbeti)ant,ointaohenreddlilfofderogemapintr.hitsmThmaoinyfiuclcoteuflost"r,(mqv-r,**"f)r*iora.enitdsnt.pode*pcrttr,ieeveolyit"ht,einccju,ttwrtirn,.gxt,"sfpreeahdrfeesc.(omTnhs/metainentqs)u.,aTftehioeisnd y lnT : InC + w ln V + y In;f+ x ln d (2) 230 Tool Life Assessment I]sing Fracti.onal Factorial Method in Turning of Titanium Alloy with CBN Cutting Tool This equation actually represents
a regression model of the two-factorial experiment,
!: b&o* bfit * bzxz* b$t (3) xw;h:erIeDyVi,sxtzhe: mInefa,sxujre=d Itnoodl,,libfeo :toIna Clogaanrdithb1m,ibc2 asncdaleb,: Xaroe :theI (mduodmemlypavraamriaebtelers),.
In the present study, the parameters of
Equation (3) will be
estimated by using a computer package.
To develop the first-order model,
2a-1 fractional factorial design will be used.
A design consisting of
8
experiments was conducted.
TABLE 1 Level designation of different process variables Level V(m/min) F(mm/rev) D (mm)
CBN content
-1
(Low) 180 0.05 0.1 Low 1(Hieh) 280 0.25 0.5 Hieh Workpiece Material A
'150 mm diameter
x
300 mm long
bar
of Ti-6Al-4v was
used for the trials. Chemical composition and physical properties of
Ti-6Al-4V
are
shown
in
Table
2
and3. TABLE 2
Chemical compositions of Ti-6Al-4V
CFe
o
AI 0.21 04.00302 0.16 0.003 6.28 4.27 v Sn 6 TABLE 3
Physical properties of Ti-6Al-4V Characteristics Melting point (K) Density
(xl03
kg/m3) Young's modulus (MPa) Modulus of
rigidiry
(MPa) Poisson's ratio Specific heat
(K
(kg.K) Thermal conductivity (W(m.K) Ti-
6At
-4V 1813 - 1923
Cqz 1 t3,lg0 44,100 0.3 - 0.33 0.56 7.54 Machine and cutting inserts
The machining trials were carried out on a
Cincinnati Avenger 200T
CNC lathe. A
MCLNR 2020K09 tool holder was used to provide an 85o cutting edge angle and 5"-rake angles. The cutting tools used were Kennametal grade KD08l and 231 Edited by: Azami Za.harim et al.
Depending on the cutting
condition
and wear rate, machining was stopped at
{c9B1i2{0cdoensteignnta.teAdlClNoGfAth1e20e4x0p8e5r1im02e0n,tsinwoerdreerctoonidnvuecstetigdatienthderyinfcluoenndciteiono.f various interval of time varying from 5 sec to I
min to record the wear of the
measured with a Mitutoyo microscope. The machining was stopped when an
insert' Flank wear
was
considered as the criteria
of
tool
wear
and the wear was
average flank wear greater than 0.30
rnm or fracturing happened. RESULTS AND DISCUSSIONS Table 4
shows the experimental conditions together with the measured tool life
probability plot is obtained (Figure l). The figure ,ho*, all of ihe a"to., values. These results are inputted to Design Expert 6.0.10 and half-normar The increasing of these factors will decrease the tool life. The feed rate has the interaction (cutting depth-CBN content) should have the effect to cutting to',o6l1li1fe'". CBN content. strongest effect to the
tool life and followed by
cutting speed,
depth of cut
and TABLE 4 Experimental conditions and results Run cutung.p""o m/min I 180 2 280 3 180 4 280 5 180 6 28A 7 180 8 280 F actors Feed rate Cutting depth f mm/rev dmm 0.05 0.10 0.05 0. l0 0.25 0.10 0.25 0.10 0.05 0.s0 0.05 0.50 0.25 0.50 0.25 0.50 CBN content Low High High Low High Low Low Tool life (sec) . n:_ 1200 630 640 60 1200 150 60 36 ef^f^eTchlseroeafftceur,ttianngalsypsiesedof(vra),riafneceed (rAatNeoQvl)A,i)epwtahs oapipcluietd(dto) ca"anldcuclaBteNthceomnteanint together with their two-level intbraction effects-on tool life. 'in eNova outpui cut-CBN content are significant. interaction. Table 5 shows that all of the factors and the interaction of depth of apnedrctehnet clealvceullawteads Fusreadtiofsorarteessthinogwnthien Tsiagbnlieficsanfocreeaocftrthseig-mnainin"*etffeefcfetscta.nTdhethe5 th-eTvhaeriRa-bsiqliutyariendtostoaltislitfiec (inTd)i.caTthees cthaalctuthlaetifoirnsta-olsrdoeirndmicoadteel tehxaptlathinesm9o9d.ge5l h%avoef and the plots of the residuals versus the predicted ,.rponr" ffr
tool life are shown
an adequate signal to noise ratio.
The normal probability plots of the residuals
232 Tool Life Assessnent Using Fractional Factorial Method in Twning of Titanium Alloy with CBN Cutting Tool
in Figures
2
and
3
,respectively. A check on the plots in Figures
2 and3
revealed
that the residuals
generally
fall on a straight line implying that the errors are distributed normally. This implies that the
models
proposed
are
adequate and there is no reason to suspect any violation of the
independence
or constant variance assumption. It
means
that the
proposed
model
using fractional factorial design is suitable for running the experiment.
TABLE 5 ANOVA for Selected Factorial Model
[Partial sum
of
squares] Source
Sum of Squares
Model 1.73986 A 6.183E5 B 7.10485 C t.469Es D 1.958E5 CD t.290Es Residual 2644.0 Cor Total tJ4tE6 DF 5 I 1 1 I I 2 7 Mean F Prob > F Square Value 3.478E5 6. l 83Es 7.104E5 1.469E5 1.958E5 1.29085 1322.0 263.05 0.0038 significant 467.68 0.0021 537.39 0.0019 l11.ll 0.0089 148.09 0.0067 97.60 0.0101 Figure 5 shows plot of cutting depth-CBN content interaction on tool life. In this study is found that the higher CBN content will prolong the tool life. It is according to Ezugwu (2005) statement the increasing of CBN content will reduc{ notch wear. The change of cutting depth is more significant at low CBN content than at high CBN content. The proposed predicting equation model The first-order equation model is obtained using ANOVA. Design Expert software have separated the tool life model into two models are the tool life model at low CBN content and the tool life model at high CBN content. For the low CBN content, the first order model as following: ln Tcat{ :1n2487.05 - 5.56 ln V -2980Lnf - 1312.5 ln d (4) Whereas for the high CBN content is: ln Tcau:1n2365.05 - 5.56 ln V - 2980lnf - 42.5 ln d (s) The both of equations will construct the tool life response curves as in Figure 6 below. 233 Edited by: Azarni hhafim et al. CONCLUSION In this work, it has been shown that partial factorial design of experiments can be applied for modeling the tool life when turning Titanium 6Al-4V with CBN cutting tool. The design is very helpful in the running of expensive cutting tool- material combinations. The work showed that the cutting depth
is the most significant factor
to tool life,
followed by cutting speed and
feed rate.
The
lower depth
of
cut is more significant to tool life
at the cutting speed
range
of
180 -280
m/min.
It showed that
there is no
contribution of CBN content to tool life g ! 6 o s o d zo E - G 3'.60 lEfiectl FIGURE 1 Half-normal probability plot of results 5!a b & o s a zb E 0 Residual FIGURE 2 Normal plots of residuals for tool life data 234 Tool Life Assessment Using Fractional Factorial Method in Turning of Titanium Alloy with CBN Cutting Tool 12.5 ido d (l {2.5 -x r2.(n 3$,$ 62,1.6 9i5 q) Predicted FIGURE 3 Plot of residuals and predicted responses 125 'iod"- t o -12.5 '-,f' 123{5674 Run Numbsr FIGURE 4 Plot residuals and run number 1200 contsnt 65' $ srssoa 1 -168-18 0.r0 0.8 0.35 0.17 0.El Depth FIGURE 5 Plot interaction between depth of cut and CBN content 235 EfudbY: ,4zamihharim et al' 1200 857.S53 515.S06 1 73.85S 5 -rse,ras o 255.00 FIGURE 6 3D plot of tool life response ACKNOWLEDGEMENT Tanhde Ianufothromrsatwioonufldolrik;;e;"totidthathnkisthewMoarklauynsdiaenr pMroinjeiscttryIRoPfA.sc0ie3n'0c2e''0T2'e0c0h6n2o-lEogAy' REFERENCES Arbizu, I.p. & p6rez, C.J.L. 2003. Surface roug-Jh.ness prediction by factorial design of experiments in turning processes. Mat. Processing Tech' 143- 144| 390-396 che-Haron, c.H. 2001. Tool life and surface integnty in turning titanium alloy' .[ ChoudMohafutr.eyP,xrpIo.eAcer.ism&si"nfrglt-,BTfe*;cu,hd-tiu1"1r,n8M:in2.3gA1.-h21i3g979h8.stTreonogl-tlihfesptereedli(c2ti9on0 mBoHdNel)b.yJ,deMsiagtn. Proce s sing Tech. 77 : 3 19 -326' D.C. Montgomery. 1991. Design and analysis of .Experiments' John wiley and Sons: New York. E.M. Trent. 1991 . Metal cutting.3rd ed. Butterworth-Heinemann: oxford' Ezugwu, E.o. Bonney, J. & Yamane, Y. 2003. An overview of the machinability of aero engine aioyr. J. Mat. processing Tech.l34:233-253. 236
