关 键 词：

罩圈落料 冲孔 复合 设计 冲压 模具 三维 Creo

资源描述：

喜欢这套资料就充值下载吧。。。资源目录里展示的都可在线预览哦。。。下载后都有，，请放心下载，，文件全都包含在内，图纸为CAD格式可编辑，【有疑问咨询QQ:414951605 或 1304139763】

内容简介：

Full Length ArticleThe effect of SiC powder mixing electrical discharge machining onwhite layer thickness, heat flux and fatigue life of AISI D2 die steelAhmed Al-Khazrajia, Samir Ali Amina, Saad Mahmood Alib,*aMechanical Engineering Department, University of Technology, District 906 Alsnaa St., Baghdad, IraqbMechanical Engineering Department, University of Karbala, Karbala 56001, IraqA R T I C L EI N F OArticle history:Received 5 September 2015Accepted 23 January 2016Available onlineKeywords:EDMPMEDMSilicone carbide powderRSMFEMWLTTotal heat fluxFatigue lifeA B S T R A C TThis paper deals with studying the effect of powder mixing electrical discharge machining (PMEDM) pa-rameters using copper and graphite electrodes on the white layer thickness (WLT), the total heat fluxgenerated and the fatigue life. Response surface methodology (RSM) was used to plan and design theexperimental work matrices for two groups of experiments: for the first EDM group, kerosene dielec-tric was used alone, whereas the second was treated by adding the SiC micro powders mixing to dielectricfluid (PMEDM). The total heat flux generated and fatigue lives after EDM and PMEDM models were de-veloped by FEM using ANSYS 15.0 software. The graphite electrodes gave a total heat flux higher thancopper electrodes by 82.4%, while using the SiC powder and graphite electrodes gave a higher total heatflux than copper electrodes by 91.5%. The lowest WLT values of 5.0 m and 5.57 m are reached at a highcurrent and low current with low pulse on time using the copper and graphite electrodes and the SiCpowder, respectively. This means that there is an improvement in WLT by 134% and 110%, respectively,when compared with the use of same electrodes and kerosene dielectric alone. The graphite electrodeswith PMEDM and SiC powder improved the experimental fatigue safety factor by 7.30% compared withthe use of copper electrodes and by 14.61% and 18.61% compared with results using the kerosene di-electric alone with copper and graphite electrodes, respectively.1. IntroductionThe mixing of a suitable abrasives and metallic material in powderform into the dielectric fluid is one of the latest advancement forimproving the innovations and enhancement capabilities of EDMprocess 1. This process is called powder mixed EDM (PMEDM).In this process, the electrically conductive powder particles are mixedin the dielectric fluid, which reduces the particlesinsulating strengthand increase the spark gap distance between the tool and workpieceto spread the electric discharge uniformly in all directions. As a result,the process becomes more stable, thereby improving materialremoval rate (MRR) and surface finish 2,3.Razak et al. 4 conducted investigations on powder concentra-tion and powder particles size of silicon carbide PMEDM based onTaguchi orthogonal array. The outcomes are expected to be capableof increasing MRR, improving surface finish, reducing TWR, and re-ducing machining time and cost.EDM is a thermal erosion process 5. Instantaneous tempera-ture rise during the machining process changes the physicalproperties of the machining surface layer, resulting in the pres-ence of residual stress, which is one of the key factors affecting themachined surface quality and its functional performance.Abhijeetsinh and Kapil 6 studied the influence of process pa-rameters and electrode shape configuration on AISI 316 Stainlesssteel workpiece material and pure copper as electrode material. De-tailed analysis of structural features of machined surface is doneby using Scanning Electron Microscope (SEM) to understand themode of heat affected zone (HAZ), recast layer thickness and microcracks, which alternatively affects structure of machined workpieceand hence tool life.The high magnitude thermal residual stresses, which are createdon the upper layer of workpiece surface due to rapid solidificationof EDM processes, influence the component properties. They in-crease with increasing pulse energy 7,8. EDM components arecommonly applied in high temperature, high-stress, and high-fatigue-load environments. Under such conditions, the cracks on themachined surface act as stress raisers and lead to a considerableThe authors confirm that this work has not been published previously (exceptas part of academic thesis), that it is not under consideration for publication else-where, that its publication is approved by all authors and that it will not be publishedelsewhere in the same form, in English or in any other language, without the writtenconsent of the Publisher.Peer review under responsibility of Karabuk University.* Corresponding author. Tel.: +96 478 03700877.E-mail address: smaengg (S.M. Ali)./10.1016/j.jestch.2016.01.014Engineering Science and Technology, an International JournalContents lists available at ScienceDirectEngineering Science and Technology,an International Journaljournal homepage: /locate/jestchPress: Karabuk University, Press UnitISSN (Printed) : 1302-0056ISSN (Online) : 2215-0986ISSN (E-Mail) : 1308-2043Available online at ScienceDirectHOSTED BY19 (2016)14001415April 201623 BY-NC-ND license (/licenses/by-nc-nd/4.0/).This is an open access article under the CC Karabuk University. Publishing services by Elsevier B.V. 2215-0986/?Karabuk University. Publishing services by Elsevier B.V.This is an open access article under the CC BY-NC-ND license (/licenses/by-nc-nd/4.0/).20162016reduction in the fatigue life of the component. Accordingly, thecurrent study conducts an experimental investigation of the powdermixing method, which does not require post-treatment process-ing to identify the optimal EDM machining parameters that suppressthe formation of cracks in the recast layer for longest lives underdifferent fatigue loads.Mehdi et al. 9 used Response surface methodology (RSM) toanalyze the effect of EDM input parameters for machining AlMg2Si composite material on microstructure changes. The results showthat voltage and current, and pulse on time are the most signifi-cant factors on microstructure of machined surfaced.A considerable amount of work has been reported by the re-searchers on the measurement of EDM performance on theperformance parameters with various DOE and optimization tech-niquesespeciallyutilizingtheResponsesurfacemethodology1015.The objective of this study concerns the effects of SiC powdermixing (PMEDM) parameters on the total heat flux generated, whitelayer thickness (WLT) and fatigue life properties of the selected AISIdie steel workpiece, and developing numerical models for verify-ing the experimental results by using the response surfacemethodology (RSM) and the finite element method (FEM) withANSYS 15.0 software.The adoption of the current work goals were intended withinother targets to reach the best combination parameters because thistype of die and mold steel owns its global importance in engineer-ing industries as well as the significance of EDM processes.2. Experimental workThe workpiece material was prepared for chemical composi-tion, mechanical properties tests and Rockwell hardness tests onthe bases of ASTM-77 steel standard for mechanical testing of steelproducts 16. The workpiece specimens were prepared withdimensions 89.9 30 4.25 mm, according to requirement of theAvery type 7305 plain bending fatigue testing machine, as shownin Fig. 1.The average values of chemical composition of the selectedworkpiece material and the equivalent values given according toASTM A 681-76 standard specification for alloy and die steels 17are listed in Table 1. The results of tensile test and Rockwell hard-ness tests are given in Table 2. The chemical composition of thecopper electrodes are listed in Table 3. The used kerosene dielec-tric was tested according to the ASTM (D 3699-90) industry standardspecifications for kerosene. The chemical composition of the siliconcarbide powder was tested for compositions by using the X-ray dif-fraction tester.Two types of electrode materials were selected (Copper andGraphite) with dimension of 30 24 24 mm. The copper elec-trode material was examined for chemical composition propertiesusing the X-Met 3000TX HORIZONT metal analyzer. Both selectedelectrode materials have their own extensive industrial applica-tions. Copper is more widely used, while graphite is also used inindustrial and researches areas. The current work will attempt toprove the high role of graphite in improving the performance re-sponses of the process. There are other electrode materials used,but with less applications, like aluminum and even steels.SiC powder was chosen as a mixed powder with kerosene di-electric because it has many important properties such as highmelting point, strength and electrical resistivity. The average grainsize for silicon carbide powder is 95.502 m.The powder with this particle size has been used to increase themicro shooting effect of its solid particles and consequently in-creases the material removal rates, and improves surface residualstresses and other properties of machined workpiece in additionto the cheapness price, which makes the process of mixing very suit-able in economic terms.The EDM parameters are: the gap voltage VP(140 V), the pulseon time duration period Ton(40 and 120 s), the pulse off time du-ration period Toff(14 and 40 s), the pulse current IP(8 and 22 A),the duty factor (= 75%), two sides flashing pressure = 0.73 bar (10.3psi.) and the SiC powder mixing concentration (0 and 5 g/L).In this work, for optimum machining performance measures, se-lecting the proper combination of machining parameters is animportant task. Generally, the machining parameters were se-lected on the basis of data provided by the EDM manufactures andresearch works. Such data were not available for the AISI D2 die steel,especially when using the graphite electrodes, the SiC and graph-ite mixing powder.Fig. 1. The Avery Denison plain bending fatigue testing machine type 7305, England.Table 1The chemical compositions of workpiece material.SampleC %Si %Mn %P %S %Cr %Mo %Ni %Co %Cu %V %Fe %Tested (Average)1.510.1740.2640.0140.00312.710.5550.1580.01370.0990.306Bal.Standard AISI D21.40 to 1.600.60 max.0.60 max.0.03 max.0.03 max.11.00 to 3.000.70 to 1.201.00 Max.1.10 Max.Bal.A. Al-Khazraji et al./Engineering Science and Technology, an International Journal 19 (2016) 140014151401In this work, the analytical and statistical response surface meth-odology (RSM) method was used to select best combination ofprocess parameters for an optimum machining performance andgiving the beat design of experiment DOE matrix with minimumnumbers of experiments. RSM analyze the experimental data by plot-ting Interaction graphs, residual plots for accuracy and responsecurves 12. A BoxBehnken design as a type of response surfacedesign was used, which does not contain an embedded factorial orfractional factorial design to determine the best responses conditions.For the purpose of powder mixing with dielectric fluid (PMEDM),a stainless steel container (of about 30 liters volume and overall di-mensions of 400 300 230 mm with a cover of 400 230 mm and3 mm thickness) was fabricated, which contained kerosene dielec-tric pump, an electric motor (300 RPM) connected to a mixer andstainless steel impellers, a workpiece clamping fixture, valves andpipe accessories. For the power supply, an AC/DC converter for drivingthe special kerosene pump was installed in an electrical board madeespecially for this work. This board contained also a pressure gauge(one bar capacity) as shown in Fig. 2.In this work, two groups were planned, each containing 22 ex-periments using a new set of workpiece and electrode in eachexperiment. The first 11 experiments in each group were ma-chined by using the copper electrodes, while the last 11 experimentswere done by using the graphite electrodes. The specimens afterEDM machining with the used copper and graphite electrodes forboth groups are shown in Fig. 3.For studying of the influence of EDM parameters and electrodematerial types on the surface recast white layers micro defects, theoptical stereo microscope (OSM) was used. The selected speci-mens were prepared after grinding, polishing and etching processesfor surface topography examination. The white layer thickness (WLT)inspections were implemented by using the Optical microscope(OM).3. Modeling and simulation of the heat flux using FEMDue to the random, high complexity and uncertainty nature ofEDM, the following assumptions have been considered to make theproblem mathematically feasible for all proposed EDM and PMEDMmodels unless otherwise specified:1 An axisymmetric model has been considered.2 Workpiece materials are homogeneous and isotropic in nature18.3 The material properties of the workpiece and tool are temper-ature dependent.4 Density and element shape are not affected 19.5 EDM discharge spark channel is considered as a uniform cylin-drical column shape.6 The heat source is assumed to have Gaussian distribution of heatflux on the surface of the workpiece material during pulse ontime period 20,21.7 Temperature analysis is considered to be of transient type 18,22.8 The channel diameter is approximately between 10 m and100 m, thus the electrode can be considered as a semi-infinitebody.9 The magnitude of the heat flux incident on the electrodes is in-dependent of the affected surface profile.Many authors have considered uniformly distributed and thehemi-spherical disc heat source within a spark 23,24. This as-sumption is far from reality for thermal modeling in EDM. However,Dibitonto et al. 20, Eubank et al. 25 and Bhattacharya et al. 26have shown that the Gaussian heat distribution is more realisticthan disc heat source for modeling the heat input in EDM. Thisfact is evidenced from the actual shape of a crater formed duringEDM.Table 2The mechanical properties of the selected materials.SampleUltimate tensilestress N/mm2Yield strengthN/mm2Elongation%HardnessHRBAverage704.25415.2518.12590.75Table 3The chemical compositions of copper electrodes material.Zn %Pb %Si %Mn %P %S %Sn %Al %Ni %Sb %Fe %Cu %0.0060.0010.0110.00020.0050.0020.00050.0070.0040.0050.00799.96Fig. 2. The (CNC) EDM machine with all the fabricated accessories.A. Al-Khazraji et al./Engineering Science and Technology, an International Journal 19 (2016) 140014151402The intensity of heat imparted to the workpiece surface, whichis denoted by Qw, is a function of instantaneous spark radius (r).The Gaussian distribution of heat applied at the axis of a spark fora single discharge with a maximum radius (R), and then the heatflux Qw(r) at radius r of the system is given by the following rela-tion 18:Qw rRw Vb Ip KnR( )=4 574 5.e-rR(1)The rate of energy incident on the workpiece is equal to the rateof energy supplied, which is equal to Rw*VP*IP, where Rw is the energypercentage fraction of heat input to the workpiece, Vbis the break-down voltage (different from the applied voltage), R is the sparkradius in m, and IPis the discharge current.The value of Rw has been determined by Shuvra et al. 19,Dibitonto et al. 20 and Patel et al. 21 who have suggested thata constant fraction of the total power is transferred to the elec-trodes. They have used the value of Rw as 8% as the percentage ofheat input absorbed by the workpiece for their theoretical work ofconventional EDM, and the same value was used in this work. TheRw values during PMEDM in the present model have been assumedto be 9% of the total heat lost in the workpiece. Shankar et al. 22have calculated that about 18% is absorbed by the cathode and therest is discharged to the dielectric fluid.The breakdown voltage value (Vb) is taken as 20 V, while forPMEDM it is about 20%30% lower than that for EDM, i.e., Vb= 15 V.The spark radius (R) was taken as 15 m, and the heat flux for AISID2 die steel isQw r( )= 6802MW m Kfor various values of dis-charge current. The spark radii for PMEDM processes were takento be 40% larger than traditional EDM.The new parameter (Kn), which takes into account the effect ofsuspended powder particles on the spark frequency and break-down voltage, was also calculated. The value of Kn depends uponthe type of powder and the powder properties, such as shape, size,concentration etc. The values of Kn factor were experimentally es-timated for EDM and PMEDM machining groups as given in Table 4using the results of material removal rates which were experimen-tally calculated.Marafona and Chousal 27 suggested an equation called theequivalent heat input radius R(t) or the radius of plasma channel(m), which is dependent on the current intensity (IP) and pulse onduration (Ton), thereby:R tITPon( )=2 040 430 44.(2)The total heat flux generated and absorbed by the workpiece,electrode and dielectric can be calculated for machining area is about500 mm2and the effective geometric machining area is about 0.75%of the above total machining area and about 10% of the total numbersof discharge sparks generated instantaneously during one pulse onperiod of time, then:Thetotalnumber of dischargesparks SNTotalmachiningarea()=20% 0 75.Areaof onedischargespark(3)The total heat fluxes for all EDM experiments were calculatedusing equation (1), but this equation needs to modify as:Qw r( )=4 574 5.- .Rw Vb Ip KnReSN Machiningareaof the wrRo orkpiece(4)Group (1)Group (2)Fig. 3. The specimens and the used copper and graphite electrodes for groups (1 and 2) experiments after EDM and PMEDM machining.Table 4The average values of material removal rate (MRR) and the experimentally estimated values of Kn factor for all EDM and PMEDM machining groups.Exp.no.Type ofelectrodePulse ontime Ton(s)Pulse offtime Toff(s)Pulsecurrent (A)Material removal rate averagevalues (MRR) mm3/minAverage values ofKn coefficientGroup 1 Av.Group 2 Av.Group 1Group 21.Copper1204088.678919.63681.005.052.Copper120402227.288840.72081.003.423.Copper401487.644512.20381.004.164.Copper40142216.434132.78421.004.795.Graphite1204087.018526.24210.759.026.Graphite120402234.491371.07671.404.477.Graphite401489.751712.35341.243.408.Graphite40142224.748065.21851.825.26A. Al-Khazraji et al./Engineering Science and Technology, an International Journal 19 (2016) 140014151403The results for the total heat fluxes entering the workpiece, theelectrode and the dielectric fluid during the on-time for all PMEDMexperiments are given in Table 5. The theoretical and numericaltotal heat flux generated by the using copper and graphite elec-trodes and FEM and ANSYS solutions for experimental group (2) isgiven in Table 6. The percentage increase in the value of theexperimental total heat flux using SiC mixed powder (group 2)compared with that in group (1) using kerosene dielectric alone isgiven in Table 7.The maximum total heat flux generated by the discharge pro-cesses and obtained by the FEM and ANSYS solutions and simulationsafter EDM machining using copper and graphite electrodes for group(2) are given in Fig. 4. In the present study, a three-dimensionalaxisymmetric model is developed to predict the total heat flux inAISI D2 die steel workpiece using the transient thermal FE Analy-sis ANSYS software.The three-dimensional meshed domain models of the workpiece,the electrode and the used kerosene dielectric were done by usingthe mapped meshing technique with a triangle element patternshape with more elements mapped toward the heat-affected regionsfor better study. The total elements and nodes numbers are 54,407and 46,664, respectively. The numbers of contact elements are31,444 and the solid elements are 22,963. The smoothness ofelements is of medium size and their minimum edge length is2.513m.Fig. 4 shows two simulations models for maximum values of inputEDM parameters using copper and graphite electrodes, respective-ly. The right figures represent the total heat flux generated form ofthe electrode, the workpiece and the kerosene dielectric after hidingthe stainless steel container. The figures in the left show the thermalmodel of the electrode and workpiece after the hiding of kero-sene dielectric with the values of maximum heat flux generated andthe input EDM and PMEDM processes parameters and also the veri-fied percentage errors between the experimental thermal modelscompared with the theoretical calculations.Three levels factorial response surface methodology (RSM) andthe Design-Expert 9.0 software were used to analyze the obtainedtotal heat power for each parametric subgroup, the analysis of resultsfor both experimental groups using copper and graphite elec-trodes are shown in Figs. 5 and 6.The Model F-value of 31,212.03 implies that the model is signif-icant.The“PredR-Squared”of0.9997isinreasonableagreementwiththe “Adj R-Squared” of 1.0000; i.e. the difference is less than 0.2.The predicted equation of the total heat flux generated forPMEDM experimental group (2), using copper electrodes and ker-osene dielectric with SiC powder mixing, is:Table 5The total heat flux fraction values absorbed by the workpieces, the electrodes and kerosene dielectric for group (2).Exp.no.Type ofelectrodePulse on durationTon(s)Pulse off durationToff(s)Pulsecurrent (A)Heat flux percentage fraction absorbed by the MW/m2WorkpieceElectrodeKerosene dielectric1.Copper120408157.40354.151437.062.Copper1204022122.82276.351121.353.Copper40148342.26770.093124.834.Copper401422451.541015.974122.565.Graphite120408210.88474.481925.336.Graphite1204022277.56624.512534.127.Graphite40148346.88780.483167.018.Graphite401422902.452030.518239.37Table 6The experimental and numerical total heat flux (power) generated by PMEDM processes using kerosene dielectric with SiC mixed powder.Exp.no.Type ofelectrodePulse ontime Ton(s)Pulse offtime Toff(s)Pulsecurrent (A)Experimental totalheat flux MW/m2Numerical totalheat flux MW/m2Error in numericalmode %1.Copper1204081948.612026.60+4.02.Copper12040221520.521579.90+3.93.Copper401484237.184414.70+4.24.Copper4014225590.075826.60+4.25.Graphite1204082610.692374.309.06.Graphite12040223436.193118.009.37.Graphite401484294.373891.109.38.Graphite40142211,172.3310,087.009.7Table 7The experimental heat flux (power) generated by EDM (group 1) and PMEDM (group 2) processes.Exp.no.Type ofelectrodePulse ontime Ton(s)Pulse offtime Toff(s)Pulsecurrent (A)Experimental total heatflux MW/m2Percentage increasein total heat flux (%)Group (1)Group (2)1.Copper120408434.621948.61+348.42.Copper1204022531.881520.52+185.93.Copper401481218.504237.18+247.74.Copper4014221396.125590.07+300.45.Graphite120408346.262610.69+654.06.Graphite1204022919.623436.19+273.77.Graphite401481511.004294.37+184.28.Graphite4014222541.0011,172.33+339.7A. Al-Khazraji et al./Engineering Science and Technology, an International Journal 19 (2016) 140014151404TotalheatpowerPulsecurrent= +4986 48607125 0314344 22331. ()Pulseontime Ton(5)And, using graphite electrodes, the equation is:TotalheatpowerPulsecurrent= +7447 41107125 0314344 22331. ()Pulseontime Ton(6)The total heat flux generated by the discharge processes for ex-perimental group (2), using the copper electrodes and SiC mixedpowder reached the maximum value as 5.827*109 W/m2at a pulsecurrent value of 22 A and a pulse on time of 40 s. While, whenusing graphite electrodes, the total heat flux values reached themaximum value as 1.009*1010 W/m2at the same input current andtime on period.This means that the use of graphite electrodes and the kero-sene dielectric with SiC mixed powder yields higher total heat fluxvalues by 91.5% when compared with the use of copper electrodesand by 285.3% and 602.7% more when using copper and graphiteelectrodes and the kerosene dielectric alone, respectively.The high total heat flux levels are obtained when using the graph-ite electrodes and the SiC mixed powder as both own low thermaland electrical conductivity and high electrical resistivity com-pared with copper electrodes. Therefore, when high pulse currentpassage through the gap between the electrode and the workpiece,a plasma channel with high thermal energy is generated, but cannotable to transform them into fusion state, because of the high miltingpoints of both materials. Then this plasma channel moves to thesurface of the workpiece and starts the main effect of the abrasivepowder working on enlarged the electrodes gaps and give moreThe total heat fluxW/mPulse Current (A)Pulse on duration (s)Type of electrodeExp. No.2240Copper4.Total heat flux = 5.827E+009 W/m(+4.2%)2240Graphite8.Total heat flux = 1.009E+010 W/m (-9.7%) Fig. 4. The total modeled heat flux generated by the PMEDM experiments using the SiC powder, the pulse current (22 A) and the pulse on time (40 s).A. Al-Khazraji et al./Engineering Science and Technology, an International Journal 19 (2016) 140014151405arranged discharge plasma channels consequentially an increase inthe amount of metal removal rate were caused by increase of themelting and eroded areas within the heating periods.4. Calculation of the white layer thickness (WLT)The average of three values of workpieces WLT measurementsafter EDM machining for both experimental groups using the twogroups is given in Table 8.The Model F-value of 7.33 implies that the model is significant.The “Pred R-Squared” of 0.9998 is in reasonable agreement with the“Adj R-Squared” of 1.0000; i.e. the difference is less than 0.2. Thepredicted equation of WLT for EDM machining using copper elec-trodes and SiC mixed powder obtained from using the three levelsfactorial response surface methodology (RSM) and the Design-Expert 9.0 software is:WLTPulsecurrentPulseontime Ton= +4 401790 0232140 055125.( ()(7)And, for graphite electrodes, the equation is:WLTPulsecurrentPulseontime Ton= +5 176790 0232140 055125.( ()(8)gTotal heat power (MW/m)Design Points11172.31520.51X1 = A: Pulse current X2 = B: Pulse on time TonActual FactorC: Type of electrode = Copper Group (2) / EDM + Kerosene dielectric +SiC powder mixing810121416182022406080100120Total heat power (MW/m)A: Pulse current (A)s(noTemi tnoes luP:B2000300040005000Fig. 5. The total heat flux (power) generated by the EDM machining using copper electrodes and SiC powder mixing.gTotal heat power (MW/m)Design Points11172.31520.51X1 = A: Pulse current X2 = B: Pulse on time TonActual FactorC: Type of electrode = GraphiteGroup (2) / EDM + Kerosene dielectric +SiC powder mixing810121416182022406080100120Total heat power (MW/m)A: Pulse current (A)B : P u ls e o n tim e T o n ( s )40006000800010000Fig. 6. The total heat flux (power) generated by the EDM using graphite electrodes and SiC powder mixing.Table 8Calculation the average values of the white layer thickness (WLT) after EDM and PMEDM machining.Exp.no.Type ofelectrodePulse onTon(s)Pulse offToff(s)Pulsecurrent (A)Average WLTGroup (1) (m)Average WLTGroup (2) (m)1.Copper12040833.348.902.Copper120402226.6713.343.Copper4014813.349.404.Copper40142211.675.005.Graphite12040820.0013.676.Graphite12040228.3411.107.Graphite4014815.005.578.Graphite40142215.009.40A. Al-Khazraji et al./Engineering Science and Technology, an International Journal 19 (2016) 140014151406The average of WLT for after PMEDM machining using the copperand graphite electrodes is shown in Figs. 7 and 8, respectively. Forexperimental group (2) using the SiC mixed powder and copper elec-trodes, the WLT reached the minimum value as 5.0 m at a currentvalue of 22 A and a pulse time of 40 s, whereas when using thegraphite electrodes, the WLT reached the minimum value as 5.57 mat the input current 8 A and with the same time on a period of 40 s.This means that there is improvement by 134% and 67% when com-pared with using of copper and graphite electrodes, and usingkerosene dielectric alone, respectively.The lower levels of the obtained WLT when using the copper elec-trodes and the abrasive SiC have high hardness work on removalof the melted white layer at high rates by micro shot peening processwith a short period of time on/off and then with high pulse current.This is because a high sparks heat power produced with a good ar-rangements of the plasma channels and the use of dielectric flushingfrom the both sides of cutting zone. The micro shot peening processrequired a longer period of time when using the graphite elec-trodes because, they producing a lower discharge heat power.The images in Tables 9 and 10 show the macro graphic and mi-crostructure of the heat affected zones (HAZ) for the workpiecessurfaces for each subgroup input parameters after EDM and PMEDMprocesses (groups 1 and 2) using copper and graphite electrodes.These images reveal that the craters sizes increased with the in-crease in the pulse current and the time on duration, and theyreached their minimum sizes, especially when using the graphiteelectrodes where higher plasma discharge heat power wasgenerated.The construction of these microscopic layers and HAZ for ex-perimental group (2) indicates that lower layer thicknesses anddefects exist due to the high abrasive and micro shot peening prop-erties of the silicon carbide powder and the plasma dischargepressure of the process as well as the high role of the dielectric flush-ing of the formed craters from both sides of the cutting areas. Withhigh pulse current value and time on duration, the defects sizes wereincreased in these layers, especially when using the copper elec-trodes due to the high thermal energy generated. These figures showthe discontinuity and uncompleted formation of the white recastlayer, especially when using the copper electrodes due to the highheat power produced and the abrasive properties of the SiC powdermaterials.5. Modeling and simulation of the fatigue life using FEMIn the present work, the three-dimensional analysis and mod-eling for EDM of AISI D2 die steel was studied. The damage modelwas developed to study the influence of different parameters on theworkpiece fatigue life. The objective of fatigue analysis is to explainthe characterizing capability of a material component to serve thedesigned cycles loads during its lifetime 28.The main assumptions of the models are as showing below:810121416182022406080100120WLT (m)A: Pulse current (A)B : P u ls e o n tim e T o n ( s )681012gWLT (m)Design Points13.675X1 = A: Pulse current X2 = B: Pulse on time TonActual FactorC: Type of electrode = Copper Fig. 7. The WLT for experimental group (2) after PMEDM using copper electrodes.Group (2) / EDM + Kerosene dielectric +SiC powder mixing810121416182022406080100120WLT (m)A: Pulse current (A)B : P u ls e o n tim e T o n ( s )681012gWLT (m)Design Points13.675X1 = A: Pulse current X2 = B: Pulse on time TonActual FactorC: Type of electrode = GraphiteFig. 8. The WLT for experimental group (2) after PMEDM using graphite electrodes.A. Al-Khazraji et al./Engineering Science and Technology, an International Journal 19 (2016) 140014151407Table 9The white layer thickness (WLT) and HAZ microstructures for EDM group (1) using kerosenedielectric alone.The white layer microstructure (optical microscope X300)Pulse Current (A)Pulse on duration Ton (s)Type of electrodeExp. No.8120Copper 1.X40X1.522120Copper2.X40X1.5840Copper3.X40X1.52240Copper4.X40X1.58120Graphite5.X40X1.522120Graphite6X40X1.5840Graphite7.X40X1.52240Graphite8.X40X1.5A. Al-Khazraji et al./Engineering Science and Technology, an International Journal 19 (2016) 140014151408Table 10The WLT and HAZ microstructures for PMEDM group (2) using kerosene dielectric and SiCmixed powder.The white layer microstructure (optical microscope X300)Pulse Current (A)Pulse on duration (s)Type of electrodeExp. No.8120Copper 1.X40X1.522120Copper2.X40X1.5840Copper3.X40X1.52240Copper4.X40X1.58120Graphite5.X40X1.522120Graphite6X40X1.5840Graphite7.X40X1.52240Graphite8.X40X1.51409A. Al-Khazraji et al./Engineering Science and Technology, an International Journal 19 (2016) 140014151 The EDM and PMEDM damaged layers cannot carry any loadcycles during fatigue. Therefore, the fatigue damage beginsbeneath the damaged layer, not at the top surface of the specimen.2 Since the residual stresses are produced at the damaged layer,their effects on the fatigue life are ignored.3 The size effect is not considered, because the thickness of thedamaged layer is small.4 The crack growth is determined primarily by both the levels ofthe actual stress and the electrical pulse-induced damage.This study used the stress life fatigue modeling analysis by usingthe ANSYS software 15.0. The cyclic fatigue of the material was ob-tained from completely reversed, constant amplitude tests with aload ratio of 1 at room temperature. The Goodman mean stresscorrection theory is a good choice for hard materials. The flat bendingworkpieces were used for the fatigue test. The test frequency of themachine was maintained constant at 60 Hz. All the fatigue limitswere defined at the stress level of 106 loading cycles. The von Misesstress theory, which takes the sign of the largest absolute princi-pal stress, will be used to compare against the experimental stressvalue.The Multiphysics, static structural models domain loads, in-cludes the environment temperature, the fixing supported and theloading force. The fatigue strength concentration factor (Kf), whichis equal to 1 and 0.72 for flat as received specimens and for PMEDMmachining workpieces 29, respectively, was set. The experimen-tal fatigue results after EDM and PMEDM machining are given inTable 11 for the two groups.The three-dimensional meshed domain models for the workpiece,using the triangle surface meshed mapping technique where theelement pattern shapes meshing, were created with more ele-ments mapped toward the machined regions for better study. Thetotal elements and nodes numbers for establishing S/N curve forthe flat as received specimens are 134,280 and 195,027, respec-tively and for EDM and PMEDM machining workpieces specimensare 83,824 and 123,748, respectively.The experimental average values of fatigue stress at 106cycles andnumerical fatigue safety factor values for group (2) are given inTable 12. The fatigue safety factor values were calculated at 106cyclesof the experimental result respect to the fatigue stress at 106cyclesof the as-received material, which is equal to 270 MPa. The percent-age enhancement in the value of experimental fatigue stresses at 106cycles using SiC mixed powder (group 2) is compared with using ker-osene dielectric alone (group 1) as shown in Table 13.The maximum fatigue life and safety factor obtained by the FEMand PMEDM with ANSYS solutions and simulations using the copperand graphite electrodes using the pulse current of 8 A and pulseon time of 40 s are given in Fig. 9. Each simulation model in thisfigure shows two models for each of input parameters of PMEDMsub-group. The right figures represent the numerical modeled fatiguesafety factor. The figures in the left show the fatigue life models simu-lations with the values of fatigue stresses at 106cycles obtained fromTable 11The experimental fatigue life results after EDM and PMEDM machining.Exp.no.Type ofelectrodePulse ontime Ton(s)Pulsecurrent (A)Pulse offtime Toff(s)For group (1)For group (2)Applied stress() (MPa)No. of cycles tofailure (1000)Applied stress() (MPa)No. of cycles tofailure (1000)1.Copper120840350.00100.250350.00138.2502.Copper120840300.00239.750300.00304.0003.Copper120840230.001260.500250.001225.0004.Copper1202240350.0061.000350.0085.5005.Copper1202240300.00133.500300.00188.0006.Copper1202240215.001273.250230.001304.0007.Copper40814350.0084.250350.00125.7508.Copper40814300.00199.750300.00264.2509.Copper40814220.001157.500240.001207.00010.Copper402214350.0056.250350.0074.00011.Copper402214210.001212.500220.001196.25012.Graphite120840350.0094.500350.00183.50013.Graphite120840300.00214.750300.00436.25014.Graphite120840220.001319.000270.001164.00015.Graphite1202240350.0045.250350.00113.25016.Graphite1202240300.0087.250300.00258.75017.Graphite1202240200.001063.750250.001226.00018.Graphite40814350.0070.250350.00154.00019.Graphite40814300.00164.750300.00352.25020.Graphite40814215.001201.500260.001263.50021.Graphite402214350.0051.250350.00106.00022.Graphite402214200.001188.500240.001271.500Table 12The experimental average values of fatigue stress at 106cycles and fatigue safety factor for group (2) PMEDM machining with SiC powder mixing.Exp. no.Type ofelectrodePulse ontime Ton(s)Pulse off timeToff(s)Pulsecurrent (A)Fatigue stress at106cycles (Mpa)Fatigue safetyfactor experimentFatigue safetyfactor numericError innumeric mode %1.Copper1204082570.950.97+2.12.Copper12040222400.890.854.53.Copper401482470.920.911.14.Copper4014222270.840.787.15.Graphite1204082751.021.13+10.86.Graphite12040222560.950.96+1.17.Graphite401482670.991.06+7.18.Graphite4014222480.920.911.1A. Al-Khazraji et al./Engineering Science and Technology, an International Journal 19 (2016) 140014151410Table 13The experimental values of fatigue stress at 106cycles for EDM group (1) and PMEDM group (2).Exp.no.Type ofelectrodePulse on timeTon(s)Pulse offtime Toff(s)Pulsecurrent (A)Experimental fatigue stressat 106cycles (Mpa)Percentage enhancementin fatigue stress (%)Group (1)Group (2)1.Copper120408240257+7.12.Copper1204022225240+6.73.Copper40148227247+8.84.Copper401422215227+5.65.Graphite120408232275+17.76.Graphite1204022203256+26.17.Graphite40148223267+19.78.Graphite401422207248+20.1Fatigue Safety FactorPulse Current (A)Pulse on duration Ton (s)Type of electrodeExp. No.Fatigue Safety Factor = 0.91840Copper3.Fatigue life=1.21E+6Cycles /b=247MPa / F=163NFatigue Safety Factor = 1.06840Graphite7.Fatigue life=1.26E+6Cycles /b=267MPa / F=151NFig. 9. The FEM Models for maximum fatigue life and safety factor of group (2) with SiC powder mixing (PMEDM) machining.1411A. Al-Khazraji et al./Engineering Science and Technology, an International Journal 19 (2016) 14001415the S/N curves of each experimental sub-group, the input EDMprocess parameters and the model loading force.The S/N fatigue stress at 106cycles curves for both groups afterEDM and PMEDM machining is shown in Figs. 10 and 11 usingpulse currents 8 A and 22 A, respectively. These figures show thatfatigue life increasing with decreasing the pulse current and in-creasing the pulse on duration time, and copper electrodes gavefatigue life values higher than graphite electrodes, after EDM ex-perimental group (1). Whereas, graphite electrodes gave fatiguelife values higher than copper electrodes after PMEDM experimen-tal group (2).The three levels factorial response surface methodology (RSM)and the Design-Expert 9.0 software were used to analyze the ob-tained fatigue safety factor for each parametric subgroup.The Model F-value of 444.81 implies that the model is signifi-cant. The “Pred R-Squared” of 0.9991 is in reasonable agreementwith the “Adj R-Squared” of 1.0000; i.e. the difference is less than0.2.The predicted equation of the fatigue safety factor for PMEDMexperimental group (2) using kerosene dielectric with SiC mixingpowder and copper electrodes is:FatigueSafetyFactorPulsecurrent= +0 967680 0101797 81250.E EPulseontime Ton()004(9)And, using graphite electrodes, the equation is:FatigueSafetyFactorPulsecurrent= +1 105180 0101797 81250.E EPulseontime Ton()004(10)The predicted equation of the fatigue stress at 106cycles ob-tained by using copper electrodes is: 200 220 240 260 280 300 320 340 360 380 40000 1000000 Fatigue stress MPa Number of cycles Log (Nf) SN curve for D2 die steel as received Copper + Kerosene / 8A -120 s Copper + Kerosene / 8A - 40 s Graphite + Kerosene / 8A -120 s Graphite + Kerosene / 8A - 40 s Copper + Kerosene + SiC Powder / 8A -120 s Copper + Kerosene + SiC Powder / 8A - 40 s Graphite + Kerosene + SiC Powder / 8A -120 s Graphite + Kerosene + SiC Powder / 8A - 40 s Fig. 10. The S/N curves of experimental groups (1) and (2) after EDM and PMEDM using the SiC powder mixing and pulse current (8 A).200220240260280300320340360380400001000000Fatigue stress MPaNumber of cycles Log (Nf)SN curve for D2 die steel as receivedCopper + Kerosene /22A -120 sCopper + Kerosene / 22A - 40 sGraphite + Kerosene /22A -120 sGraphite + Kerosene / 22A - 40 sCopper + Kerosene + SiC Powder / 22A -120 s Copper + Kerosene + SiC Powder /22A - 40 s Graphite + Kerosene + SiC Powder / 22A -120 s Graphite + Kerosene + SiC Powder / 22A - 40 s Fig. 11. The S/N curves for experimental groups (1) and (2) after EDM and PMEDM using the SiC powder mixing and pulse current (22 A).1412A. Al-Khazraji et al./Engineering Science and Technology, an International Journal 19 (2016) 14001415Fatiguestress atcyclesPulsecurrent ti106253 089291 33929= +.m mePulseon time Ton+()0 12187.(11)And, using graphite electrodes, the equation is:FatiguestressatcyclesPulsecurrent ti106271 839291 33929= +.mmePulseontime Ton+()0 12187.(12)The analysis of results for fatigue safety factor for experimen-tal group (2) using the copper and graphite electrodes are shownin Figs. 12 and 13, respectively. While, the fatigue stresses at 106cycles for all experimental groups using the copper and graphiteelectrodes are shown in Figs. 14 and 15, respectively.The fatigue safety factor using the copper electrodes for exper-imental group (2) increases with the decrease in pulse current valuesand the increase in pulse on duration time, reaching its maximumvalue as 0.97, experimentally (0.95) at a current value of 8A and pulsetime of 120s. While, when using the graphite electrodes, the fatiguesafety factor values reached their maximum values as 1.13, exper-imentally (1.02) at the same input current and pulse on time period,as shown in Fig. 13.This means that after PMEDM processes, the use of graphite elec-trodes and the kerosene dielectric with SiC mixed powdergives higher experimental fatigue safety factor values by 7.37%compared with the use of copper electrodes and higher by 14.61%and 18.61% when compared with the results of group (1) using thecopper and graphite electrodes, respectively.The reason behind obtaining higher fatigue safety factor isbecause the use of low pulse current generates lower thermal energy,which cannot work to make a large metallurgical changes in the crys-talline structure of the workpiece surface. Although the graphiteelectrode generates thermal energy more than that of copper, itworks with the longer pulse time on annealing the workpiece surfaceand on reducing the creation of martensitic structure, assisted bythe use of hard SiC particles powder that increases the removal rateof this layer by increasing erosive rates, and all that will lead to re-ducing the fatigue failure thereby increasing the fatigue life.The fatigue stresses at 106cycles after PMEDM experimentalgroup (2) using the copper electrodes increase with the decreasein pulse current and the increase in pulse on duration time, reach-ing the maximum value as 257 MPa at a current value of 8 A andpulse on time 120 s. Whereas, when using the graphite elec-trodes, these fatigue stresses values reached their maximum valueas 275 MPa at the same input current and pulse on period time, i.e.improved by 7% as shown in Fig. 15.This means that the use of graphite electrodes and the kero-sene dielectric with SiC powder mixing yields higher fatigue stressesat 106cycles values by 7.00% when compared with the use of copperelectrodes and gives a higher fatigue life than the situation whenworking without mixing powder by 14.58% and 18.54% using thecopper and graphite electrodes, respectively.Fatigue Safety FactorDesign Points1.130.78X1 = A: Pulse current X2 = B: Pulse on time TonActual FactorC: Type of electrode = Copper 810121416182022406080100120Fatigue Safety FactorA: Pulse current (A)B : P u ls e o n tim e T o n ( s )0.80.850.90.95Fig. 12. The fatigue safety factor after EDM using copper electrodes and SiC mixed powder.Fatigue Safety FactorDesign Points1.130.78X1 = A: Pulse current X2 = B: Pulse on time TonActual FactorC: Type of electrode = Graphite 810121416182022406080100120Fatigue Safety FactorA: Pulse current (A)B : P u ls e o n tim e T o n ( s )0.9511.051.1Fig. 13. The fatigue safety factor after EDM using graphite electrodes and SiC mixed powder.1413A. Al-Khazraji et al./Engineering Science and Technology, an International Journal 19 (2016) 14001415The stress values are equal to the ratios 0.95 and 1.02 for copperand graphite electrodes, respectively, compared with the fatiguestresses at 106 cycles for the as-received material.Higher fatigue safety factor levels were obtained when using thecopper electrodes with low current and high pulse on time periodas lower heat discharges energy at the gap produced between theelectrode and the workpiece. The longer time on period with thehigh abrasive effect of SiC powder work as micro shot peening ofthe machining surfaces, producing less brittle carbides formation,fewer defects and lower white layer thickness.Also, the better surface roughness (SR) and white layer thick-ness (WLT) workpieces gave longer fatigue lives as less micro cracksand other surface defects were observed with less stress risers andstress concentration areas, helped eliminate the creation of cracksinitiation and propagation, and then increased the dynamic servicelives of the components.And, all these factors are strengthening the workpiece againstfatigue failure, and then longer lives were obtained.6. Conclusions1 The total heat flux values were increased with the increasing ofpulse current values up to 22 A and the decreasing of pulse onduration time to 40 s.2 The graphite electrodes gave a total heat flux higher than copperelectrodes by 82.4%, while using the SiC powder and graphiteelectrodes gave a higher total heat flux than copper electrodesby 91.5%, and by 285.3% and 602.7% more than using the copperand graphite electrodes and the kerosene dielectric alone,respectively.3 The lowest WLT values of 5.0 m and 5.57 m were reached ata high current and low current with low pulse on time using thecopper and graphite electrodes and the SiC powder, respective-ly. This means that there is an improvement by 134% and 67%when compared with using the copper and graphite electrodesand kerosene dielectric alone, respectively.4 The fatigue safety factor as compared with as-received materi-al and fatigue stresses increased with the decrease of pulsecurrent, and increase of pulse on time.5 Using graphite electrodes with PMEDM and SiC powder yieldedexperimental fatigue safety factor values of 1.02, which is higherby 7.30% when compared with the use of copper electrodes andhigher by 14.61% and 18.61% when compared with results ofusing the kerosene dielectric alone with copper and graphite elec-trodes, respectively.6 The use of graphite electrodes with SiC powder gave fatiguestresses at 106cycles as 275 MPa, which is higher by 7.00% whencompared with the use of copper electrodes, and yielded a higherfatigue life than when working without mixing powder by 14.58%and 18.54% using the copper and graphite electrodes, respectively.7 All predicted fatigue stresses to the fatigue stresses at 106cyclesfor the as-received material ratio are close with those results offatigue safety factors for the same input parameters, and thisFatigue stress at 10E6 cyclesDesign Points275227X1 = A: Pulse current X2 = B: Pulse on time TonActual FactorC: Type of electrode = Copper 810121416182022406080100120Fatigue stress at 10E6 cyclesA: Pulse current (A)B : P u ls e o n tim e T o n ( s )230240250Fig. 14. The fatigue stresses at 106cycles after EDM using copper electrodes and SiC mixed powder.Fatigue stress at 10E6 cyclesDesign Points275227X1 = A: Pulse current X2 = B: Pulse on time TonActual FactorC: Type of electrode = Graphite810121416182022406080100120Fatigue stress at 10E6 cyclesA: