1. AbstractIn thisinvestigation, the effect of four controllable input variables of EDM namely: peakcurrent, pulse on time, pulse off time and gap voltage on surface roughness(Ra), is investigated. A Box-Behnken design matrix is used to conduct experimentson Mild Steel using Copper electrode. RSM is used on experimental data to modelthe response. Analysis of variance at 95% confidence interval is performed andsignificant coefficients are obtained. From the ANOVA results it is observedthat peak current and pulse on time are most significant factors. Optimizationof input parameters is done to obtain minimum Surface Roughness (SR).
2. Introduction2.1 Principle of EDMElectricalDischarge Machining is a non-traditional machining process in which the metalis removing from the work piece due to erosion case by rapidly recurring sparkdischarge taking place between the tool and work piece.
There are three phasesoccurred in electrical discharge machining process. In the initial phase theignition breaks down the high voltage to low. Then the peak current increasesthe energy and the material is removed from the workpiece. In the last phase,plasma channel collapses and flushing flushes away the removed particles. InEDM the component produced is the exact replica of the electrode.
Through EDM complexshaped products are manufactured that cannot be produced by conventionalmethod. In this process both the workpiece and the tool have no physicalcontact with each other. Both are immersed in the dielectric liquid which alsoact as coolant.
2.2 Machining Parameters ofEDM Animportant consideration in Electrical Discharge machining (EDM), like other machiningoperations, is the selection of machining parameters/conditions. The important inputparameters that are affecting performance parameters in EDM are (1) Pulse ONtime (2) Pulse OFF time (3) Arc gap (4) Discharge current (5) Voltage (6) Polarity(7) Duty cycle (8) Diameter of electrode (9) Dielectric fluid2.3 Performance Parameters ofEDM The performanceof the electrical Discharge Machining is measured by the parameters as Overcut(OC), Material Removal Rate (MRR), Tool Wear Rate (TWR) and Surface Roughness (SR).3. Literature SurveyBalasubramanian and Senthilvelan 1 conducted the experiments on EN8and D3 steel materials using Cast Copper and Sintered Powder Metallurgy Copper(P/M Copper) electrodes on EDM. He used the RSM to analyze and optimize theinput parameters. Singh and Singh 2 analysed the effect of differentmaterials on surface roughness.
The result shows that the surface roughnessincreases with increasing pulsed current and pulse time. Santoki and Ashwin 3 reviewed the development done inthe EDM & the effects of machining parameters on performance parameterswith various DOE & Optimization Techniques. Kumar, Sivakumar 4 conducted the experiments on AISID3 steel using Taguchi’s L18 OA on EDM. The effect of input parametersincluding pulse on time, pulse off time, current and voltage on the materialremoval rate (MRR) and surface roughness (Ra) was analysed and then optimized.
Bhaumik and Maity 5 performed the electrode discharge machiningof AISI 304 stainless steel by using the tungsten carbide electrode in order toanalyse the effect of peak current, pulse on time, gap voltage, duty cycle on surfaceroughness (Ra). The effect of significant process parameters on the responsehas been studied. Then regression analysis is done and mathematical model iscreated to describe the correlation of parameters. The result shows that themost influential parameter for surface roughness is peak current.
Pradhan and Biswas 6 designed a face centred centralcomposite design matrix and used it to conduct the experiments on AISI D2 toolsteel with copper electrode. RSM is used on experimental data to make theregression model. It is found that discharge current and pulse duration aresignificant factors.
Torres, Luis 7 studied the behaviour of inputparameters of current intensity supplied by the generator (I), duty cycle (?),pulse time (ti), and polarity on INCONEL 600 alloy using electrical dischargemachining (EDM). The experimental results confirm that positive polarity leadsto higher MRR whereas negative polarity leads to lower Ra values. Khan, Rahman 8 studied the surface finishcharacteristics of the machined surface in EDM on Ti-5Al-2.5Sn titanium alloy.central composite design is used to analyze the effects of peak current,pulse-on time, pulse-off time, servo voltage and electrode material. The resultshows that surface roughness (SR) increases with peak current and pulse-on timeand decreases with servo voltage. Besides, the effect of the process parameterson surface roughness depends on electrode material.4.
Experimental Details4.1 ProcedureIn this work, experimentsare performed by electrical discharge machining on the mild steel workpiece bycopper electrode in different machining conditions. Experiments are designed inorder to investigate the effect of different input EDM parameters namelydischarge current, pulse ON time, pulse OFF time and gap voltage on the EDMoutput parameter of the interest namely surface roughness Ra.4.2 Machine ToolA Neu-ar M50die sinking EDM machine is used.
It has Heidenhain EDM controller and 9 setscoordinate memory. It contains Built-in origin and mold centre settingfunction. It has the lowest machining depth display function. And electrodeconsumption offset function. Also a built-in residue proofing feature to driveout carbon residue is present. It is equipped with high precision Heidenhain 1?mlinear encoder. It has durable wear-resistant Teflon Slippery track in “Vshape” and “horizontal track”.
It also contains multiple fire- proof detectionsystem.Figure 1- Electrical Discharge Machine Model M504.3 Material4.3.1 Workpiece A square plateof Mild steel having dimensions 20*12*0.5cm was taken.
This has 14.30 g/cm3 ofdensity, 1240 HV10 hardness, 2597°C of melting point and 420 kgf/mm2 ofcompressive strength.4.3.2 Electrode A Copperelectrode of cylindrical shape with 10 mm diameter and 100 mm length undernegative polarity was axially mounted within mild steel workpiece. Theproperties of Cu electrode used in this work are the following: melting pointof 3500°C and density of 12.6 g/cm3.
4.4 DOEIn order toreduce the number of experiments due to limited resources, Box-Behnken designis used. Four factors each at three levels are taken. One block is made havingtwenty-four factorial points and six centre points, so the total number ofexperiments are 30. Design-Expert 7.0, 2005 is used to make the randomizedesign. Machining was carried out to remove approximately 0.
5mm from the topsurface. The different levels of factor considered for this study areillustrated in Table 1. Table 1 – Factors and Levels S.
No Input parameters Level Unit -1 0 1 1 Peak Current 1 5 9 Amp 2 Pulse ON time 50 75 100 Microsec 3 Pulse OFF time 80 100 120 Microsec 4 Gap Voltage 40 50 60 Volts 4.5 Measurement of ResponseRa is a measureof the surface finish quality of a product. It is defined as the arithmeticvalue of the profile from the centreline along the length.
Afterperforming the machining operation, the surface roughness of each cut ismeasured using a portable stylus type profilometer, Roughness measurement isdone in the traverse direction on the workpiece and the values of Ra parameterare recorded.Table 2 – Design Matrix Run Ip Ton Toff V Ra 1 1 50 120 40 2.39 2 1 50 80 40 2.15 3 9 100 120 60 7.
08 4 9 100 80 60 7.64 5 9 100 80 40 7.43 6 5 75 100 50 5.
41 7 9 50 80 40 6.23 8 1 100 120 40 2.09 9 1 100 80 40 1.65 10 1 100 80 60 1.74 11 9 100 120 40 8.66 12 9 50 120 40 6.
01 13 5 75 100 50 5.22 14 1 50 80 60 2.11 15 9 50 120 60 6.24 16 1 100 120 60 2.15 17 5 75 100 50 5.
29 18 1 50 120 60 2.45 19 5 75 100 50 5.36 20 9 50 80 60 5.83 21 9 75 100 50 6.48 22 5 75 100 50 5.6 23 5 100 100 50 5.81 24 5 75 100 50 5.
53 25 1 75 100 50 1.98 26 5 75 80 50 5.54 27 5 75 100 40 5.
97 28 5 75 100 60 5.52 29 5 50 100 50 4.77 30 5 75 120 50 5.77 5. Result and Discussion5.1ANOVAExperiments are conducted to analyze the effect of machiningparameters on surface roughness.
Design Expert Software was used to find outthe relationship between the input factors and the response Ra. To decide the degree of the regression model, the R2 andR2 adjusted values are summarized in Table 3 for various models. Thetable shows that quadratic model is best with R2 = 99% Therefore,the quadratic model is considered for regression analysis.
Table 3 – R2 and R2 adj test for surface roughness regressionmodel Source Std. Dev. R-Squared Adjusted R-Squared Linear 0.73 0.89 0.87 2FI 0.
69 0.92 0.88 Quadratic 0.28 0.
99 0.98 Table 4 – Result of the ANOVA table for surface roughness(Before elimination) Source Sum of Square Df Mean Square F Value p-value Model 117.06 14.00 8.36 103.65 0.0001 A-Peak Current 102.
20 1.00 102.20 1266.89 0.0001 B-Ton 2.05 1.00 2.
05 25.37 0.0001 C-Toff 0.35 1.00 0.35 4.37 0.0539 D-Gap Voltage 0.
18 1.00 0.18 2.
28 0.1517 AB 3.97 1.00 3.97 49.21 0.
0001 AC 0.02 1.00 0.02 0.25 0.6231 AD 0.
18 1.00 0.18 2.
27 0.1530 BC 0.04 1.00 0.04 0.44 0.5192 BD 0.07 1.
00 0.07 0.89 0.
3612 CD 0.07 1.00 0.
07 0.92 0.3526 A^2 3.
95 1.00 3.95 48.96 0.0001 B^2 0.
08 1.00 0.08 0.
98 0.3380 C^2 0.09 1.00 0.09 1.
16 0.2977 D^2 0.20 1.
00 0.20 2.52 0.1330 Residual 1.21 15.
00 0.08 Lack of Fit 1.11 10.00 0.11 5.37 0.
0387 Pure Error 0.10 5.00 0.02 Cor Total 118.27 29.00 Table 4 is an ANOVA summary which shows the F and P values fordifferent terms. The results show that in main effects ‘Voltage’ isinsignificant.
Also all the quadratic terms except A2 and AB, areinsignificant. Thus, these terms are eliminated for the further analysis.After elimination of insignificant terms, ANOVA is performed. Theresult of ANOVA is summarized in Table 5.
After elimination of non-significantterms, the values of R2 and R2adj are 98.1% and 97.8%, respectively. The mainand interaction effects, that are significant, are Ip, Ton, Toff, Ip2, andIp×Ton. Table 5 – The ANOVA table for the fitted model Source Sum of Square Df Mean Square F Value p-value Model 116.07 5.00 23.21 253.
78 0.0001 A-Peak Current 102.20 1.
00 102.20 1117.25 0.
0001 B-Ton 2.05 1.00 2.05 22.
38 0.0001 C- Toff 0.35 1.
00 0.35 3.86 0.0612 AB 3.
97 1.00 3.97 43.40 0.
0001 A^2 7.50 1.00 7.50 82.
03 0.0001 Residual 2.20 24.
00 0.09 Lack of Fit 2.09 19.
00 0.11 5.34 0.0361 Pure Error 0.10 5.
00 0.02 Cor Total 118.27 29.00 From this analysis, the simplest model obtained is stated in thefollowing equation.Ra = 1.605 + 0.8601*Ip – 0.
0114*Ton +0.007*Toff – 0.063Ip2+0.005*Ip*TonNormal probability plot of the residuals is displayed in Fig. 2. Itcan be seen that the residuals are almost falling on a straight line, whichindicates that the errors are normally distributed. Figure 2- Normal probability plot of residualsFigure 3 – Predicted vs. experimental surface roughnessFig.
3 depicts the comparison of experimental observations versethe predicted response values. It can be examined that the regression modellikely fits the experimental values.5.2 Influence of Input ParameterOn ResponseFigure 4 – Effect of factors on RaFig. 4 depictsthe plots of main effects on Ra.
The plot shows that peak current is the mostinfluential factor. As the peak current increases the surface roughness increasesrapidly. Also with the increase in Ton, Ra also increases. Trend for Toff issame i.e. Ra increases with the increase in Toff.
5.3 Model GraphFig. 5 represents contour plot and response surface for SurfaceRoughness in relation to input parameters of peak current and pulse ON time. Itcan be concluded that the at any value of Ton, the Ra increases rapidly withthe increase in Ip. Hence, in order to obtain minimum Ra, peak current shouldbe at low level (1A) and pulse on time on (50?s).Fig. 6, depicts the contour plot and response surface for Ra inrelation to Ip and Toff, where Ton remains constant at the level of 75?s.
Itcan be seen that, when Ip increases Ra also increases. However, Ra drops slowlydecreases with the increase in Toff at lower Ip, and at higher Ip Ra increaseswith Toff. However, the influence of Toff on Ra is very low as compared to Ipand Ton.Figure 5 – Contour & Response surface plot depicting theeffect of Ip and Ton on RaFigure 6 – Contour & Response surface plot depicting theeffect of Ip and Toff on RaFigure 7 – Contour & Response surface plot depicting theeffect of Ton and Toff on Ra Finally, Fig.7 depicts the contour plot and response surface for Rain relation to Ton and Toff, where Ip remains constant at the level of 5 A.From these plots, it can be concluded for the given range of experimentsconducted for this test, that peak current and pulse ON time are directlyproportional to the Ra and for pulse OFF time the effect is very less ascompared to the other parameters.6.
OPTIMIZATIONEDM is a usefuland valuable tool tom make complex shape parts that cannot be machined bytraditional machining processes. In order to increase the quality and rate ofproduction, the process parameters have to be optimised. Also, especially incase of EDM, it is very essential to optimize the input parameters to yieldminimum SR. In single objective optimization only one solution has beenobtained. it has been observed that low Peak current, low Pulse on time, lowpulse off time and marginal Voltage gives minimum Surface roughness.Table 6 – Optimization Table S. No Input /output Parameters Optimized value Units 1 Peak current 1 Amps 2 Pulse on time 100 Microsec 3 Pulse off time 80 Kg/Cm2 4 Voltage 56 V 5 Surface Roughness 1.77801 ?m 7.
ConclusionIn this studythe effect of most significant input parameters of EDM on the surface roughnesshas been studied for Mild steel. A Box-Behnken design with factors of dischargecurrent, pulse on time pulse off time and gape voltage, is used forexperimentation. The ranges of these parameters are chose from the literaturereview. Using the Response Surface Methodology (RSM), the regression mode(quadratic) is formed using the Design-Expert 7.0 software. The analysisshows that the output response is significantly affected by the inputparameters of discharge current, pulse on time, pulse off time and gap voltagewith 95% confidence interval. The results also reveal that the value ofdischarge current, pulse on time and pulse off time should be set as low aspossible, in order to get a good surface finish on mild steel. For the bestsetting of Ra, the discharge current of 1 A, pulse on time of 100?s and offtime should be 50?s, which yields the best value Ra of 1.
778?m. The regressionmodel developed for surface roughness can be effectively used for the optimalselection of input parameters in EDM to achieve good surface finish for Mildsteel workpiece. These findings will be helpful to manufacturing engineers inselecting the appropriate parametric combinations for EDM processes toaccomplish desired levels of Ra.
References1. Balasubramanian, P. and T.Senthilvelan, Optimization of machining parameters in EDM process using castand sintered copper electrodes. Procedia Materials Science, 2014. 6: p. 1292-1302.2.
Singh, H. and E. Singh, Examination ofSurface Roughness Using Different Machining Parameter in EDM. 2012.
3. Santoki, P.N. and P. Ashwin, Areview–status of recent developments and effect of machining parameters onperformance parameters in EDM. Int.
J. Innov. Emerg. Res.
Eng, 2015. 2(1): p. 32-41.4.
Kumar, M.S., et al., ‘ParametersOptimisation of Wire Electrical Discharge Machining on AISI D3 Steel withDifferent Thickness’. International Journal of Applied Engineering Research,2015. 10(62): p. 2015.
5. Bhaumik, M. and K. Maity, Effect ofmachining parameter on the surface roughness of AISI 304 in silicon carbidepowder mixed EDM. Decision Science Letters, 2017.
6(3): p. 261-268.6. Pradhan, M.
and C. Biswas, Effect ofprocess parameters on surface roughness in EDM of tool steel by responsesurface methodology. International Journal of Precision Technology, 2011. 2(1): p. 64-80.7. Torres, A., C.
Luis, and I. Puertas,Analysis of the influence of EDM parameters on surface finish, material removalrate, and electrode wear of an INCONEL 600 alloy. The International Journal ofAdvanced Manufacturing Technology, 2015. 80(1-4):p. 123-140.8. Khan, M.
Rahman, and K.Kadirgama, An experimental investigation on surface finish in die-sinking EDMof Ti-5Al-2.5 Sn. The International Journal of Advanced ManufacturingTechnology, 2015.