Skip to content
BY 4.0 license Open Access Published by De Gruyter October 1, 2019

Microstructural analysis and multi response optimization of WEDM of Inconel 825 using RSM based desirability approach

Pawan Kumar, Meenu Gupta and Vineet Kumar

Abstract

Increasing demand of aerospace industry for more heat resistant and tough material have open up the possibility of the use of Inconel 825 for making of combustor casing and turbine blades. Because of its robust nature, Inconel 825 is a difficult-to-cut material with conventional methods. Wire-cut electrical discharge machining (WEDM), a non traditional method uses thermoelectric erosion principle to produce intricate shape and profiles of such difficult-to-cut material. In this study, various operating parameters of WEDM are optimized using desirability approach and microstructural behavior at optimum combinations was studied. Input parameters viz. pulse-on time, pulse-off time, peak current, spark gap voltage, wire tension, wire feed and performance has been measured in term of material removal rate, surface roughness and wire wear ratio. It has been observed that at 110 machine unit pulse-on time (Ton), 35 machine unit pulse-off time (Toff), 46 volt gap voltage (SV), 120 ampere peak current (IP), 11 machine unit wire tension (WT) and 5 m/min wire feed (WF), the values obtained for material removal rate (MRR), surface roughness (SR) and wire wear ratio (WWR) were 27.691mm2/min, 2.721 μmand 0.117 respectively. Scanning electron microscopy, energy dispersive spectrograph and X-ray diffraction analysis has also been carried out to study the surface characterization. Comparatively less numbers of cracks, pockmarks, craters, and pulled out material were found on work specimen surface and wire electrode surface under standardized conditions, thus maintaining the surface integrity of the machined surface.

1 Introduction

In the field of manufacturing technologies to achieve better design configuration, use of new and difficult-to-machine materials demands the development of economically more feasible processes. Superalloys, due to their high-temperature mechanical strength and high corrosion resistance properties, are currently being used in technologically advanced industries such as marine, space, automobiles and other applications [1, 2]. Such properties are mainly significant to the aerospace industry for manufacturing of turbine disks, blades, combustors and casing [3]. Nickel-based superalloys are the most multifaceted widely used alloy for the hottest parts and constitute over 50% of the weight of advanced aircraft engines. Among Nickel based superalloys, Inconel series possesses high temperature corrosion resistance, oxidation resistance and creep resistance properties which are important for aerospace industry for manufacturing of turbine disks, blades, combustors and casing [4]. Among the various Inconel series, Inconel 825 possesses superior mechanical properties, provides resistance to chloride pitting as well as resistance to a variety of oxidizing atmospheres than other Inconel alloys [5].

Inconel 825 has an austenitic structure that imparts high ductility and work hardening properties to this superalloy resulting in a gummy machining behavior similar to that of steels. In addition, because of high temperature resistant and less thermal conductivity, Inconel 825 remains strong at the temperature of chip formation during machining. Inconel 825 also contains abrasive titanium and aluminium particles. All these factors make Inconel 825, a difficult to machine material even in comparison to steels resulting in increased cost to the manufacturer [6]. During machining, Inconel 825 has a propensity

to weld the cutting tool that causes tool failure [7, 8]. Because of high toughness and hardness, the machining of this hard material with conventional method is very difficult [9].Moreover, high abrasive particles in its microstructure, tendency forming chip to weld the tool and built up edges (BUE) make it more difficult to machine [10]. Generation of high temperature is also a problem due to abrasion between tool and material [11].

WEDM is non-traditional machining technique works on spark erosion principal in which a discrete form of spark is generated in between the gap of wire and workpiece ranges from 0.025 mm to 0.5 mm. By the use of non-conventional technique in machining of superalloys a revolutionary change has been found in tools. WEDM is best alternative of making intricate shape and profile with better surface finish and accurate dimensions of hard and tough material like Inconel 825 [12].

The performance characteristics of WEDM is material removal rate (MRR), surface roughness (SR) which are influenced by numerous machining characteristics such as pulse-on time (Ton), pulse-off time (Toff), peak current (IP), gap voltage (SV), wire tension (WT) and wire feed (WF) [13]. Optimum machining parameters selection of WEDM, avoid the wire breakage problem and surface roughness because improperly selected parameters reduce productivity and quality requirements. To establish the relationship between output and input parameters a suitable modeling and optimization technique is necessary.

Generally, the values of input parameters are selected on the basis of machine manuals but it does not ensure optimal machining performances. Conventional experimental design methods require a large amount of experimental data for determining the best combination of parameters, which is expensive and prolonged. To overcome these problems, researchers applied factorial design named Design of Experiment. Taguchi’s methodology has been used in various studies to obtain best combination of parameters for desired machining performances [14, 15, 16, 17, 18]. Response surface methodology is a better approach than taguchi method as it can help to find the nearest range of machining parameters beyond or within the range of level of factors. Also, RSMis a more mathematical approach as it identifies the interaction between the variables rather than investigating the effect of individual parameters [19].

Goyal [20] investigated the effect of process parameters of WEDM on Inconel 625 and found that pulse-on time, tool electrode and current intensity are the significant parameters that affect the MRR and SR. Kumar et al. [21] optimized the machining parameters i.e. current 1 A, pulse-on-time 0.98 μs, pulse-off-time 0.03 μs, tool material 0.31 and the powder (suspended particles) 0.64 on machining of Inconel 800 using Powder mixed EDM by employing desirability based RSM approach. Amongst the nickel based superalloy, many reports have been published on the WEDM cutting of Inconel 718 [10, 18, 22, 23, 24]. A thorough scrutiny of the previous investigations suggests that despite of having superior properties over other nickel based alloys, much less work has been done on Inconel 825. While there are some reports on ECM/EDM of Inconel 825 [25, 26], only one author had focused on standardization of machining characteristics of WEDM using Inconel 825 [5]. Rajyalakshmi and Venkata Ramaiah [5] had optimized the multiple response measures of WEDM on Inconel 825 by consideration of Taguchi grey relational analysis to obtain improved material removal rate, surface roughness and spark gap. The substantial improvement in the output responses were observed under optimal set of conditions.

Moreover, surface integrity investigation is required to determine the final performance of the machined product. Since WEDM process occurs at very high temperature (8000-12000C), it is having significant impact on the surface integrity of the work piece. Surface parameters including surface morphology, residual stress surface roughness, microstructure, and micro-hardness are crucial in determining the final performance of the machined specimen. The formation of heat affected zone, micro-cracks, porosity etc. remains a big problem in material microstructure processed by WEDM[10, 27]. Therefore, analysis of surface integrity is required to determine the final performance of the machined product. Thus, after critically analyzing the literature, it was revealed that most of the work reported in the literature has been focused on machining of work materials such as ceramics, titanium and its alloys etc. using WEDM or EDM. There is limited work done on the machining of Inconel 825 by using WEDM. Moreover, mostly two machining characteristics such as MRR and surface roughness (SR) have been investigated in literature. No investigation work is available involving wire wear ratio as output response in WEDM of Inconel 825. So, wire wear ratio can be considered as an output response of the WEDM process. Therefore, in this study, to find the optimal performance of WEDM a suitable modeling and optimization technique, RSM, is used to establish the relational ship between performance characteristics and controllable input parameters. Multi-objective optimization was carried out using RSM based desirability approach. In addition, surface integrity studies of the finally machined sample was also performed.

Table 1

Chemical, physical and mechanical properties of Inconel 825

Chemical composition of Inconel 825
ElementNiFeCrMoCuTiCPSMg
Content (%)38-462219.5-23.52.5-3.51.5-3.00.6-1.20.050.0020.0031.0 max
maxmaxmax
Physical properties of Inconel 825
PropertyDensityMelting PointCoefficient of expansionModulus of rigidityModulus of elasticity
Metric8.14 g/cm31370-1400 C14.0 m/m/C (20–100 C)75.9 N/mm2196 MPa
Mechanical properties of Inconel 825
PropertyAlloy stateTensile strengthMPaYield strength MPaElongation A 5 %Brinell hardness HB
Metric82566233830≤ 200

2 Materials and methods

2.1 Work material

Inconel 825 (150mm× 150mm× 10 mm) was used as work material for the present study. Table 1 shows the chemical, physical and mechanical properties of work material. All experiments were performed using CNC WEDM machine tool (ELECTRA SPRINT CUT 734) in advanced manufacturing laboratory of Mechanical Engineering Department, N.I.T., Kurukshetra, Haryana, India. The workpiece was mounted with the help of a fixture on the machine table. The electrode material used was a single-strand plain brass wire of diameter 0.25 mm. Wire was continuously fed through the workpiece by a microprocessor. The upper head supplied the fresh wire under tension through the workpiece and lower head received the used wire after machining. A suitable gap of 0.025 - 0.5 mm was maintained between wire and the workpiece (Figure 1).

Figure 1 Mechanism of material removal

Figure 1

Mechanism of material removal

2.2 Selection of input and output parameters

Six parameters i.e. pulse-on time (Ton), pulse-off time (Toff), peak current (IP), gap voltage (SV), wire tension (WT) and wire feed (WF) were chosen as input parameters as suggested from literature search. The machining performance was measured in terms of material removal rate (MRR), surface roughness (SR) and wire wear ratio (WWR). During experimentation, it was reported that due to constant wire diameter the kerf width varies negligibly. Consequently, the MRR was calculated by taking equation 1 into consideration:

(1)MRR(mm2/min)=cuttingspeed(mm/min)×thicknessofmaterial(mm)

The surface roughness (Ra) of machined specimen is measured in μm using Accretech’s surfcomflex instrument SJ-301. Least count of the instrument is 0.8 mm.

A sampling length of 5 mm was selected for measurement. Wire wear ratio is measured by using initial wire weight and used wire weight. It is the ratio of difference between initial wire weight and used wire weight to the initial wire weight. Used wire weight for one specimen is directly measured by using weighing machine. The least count the weighing machine is 0.02g. Fresh wire weight used for one specimen is measured by calculating the length of used wire. One meter length of fresh wire was of 0.400 mg. The fresh wire weight is calculated by multiplying the length of used wire to the weight of the one meter fresh wire.

Some factors which are likely to affect the performance measures viz. extent and profile of the product (rectangular), conductivity of the dielectric (20 mho), temperature of the dielectric (28C), pulse peak voltage setting (110 V), work piece thickness (15 mm), wire feed setting (6 m/min), servo feed setting, wire type (0.25mm diameter brass) and angle of cut (vertical) are kept constant.

2.3 Design of experiments

Central composite design (CCD) at α value of ± 2 was employed using Design expert software (version 9.0.7, Statease) to optimize the levels of significant variables. Table 2 shows the coded and actual values of the variables. MRR, SR and WWR were taken as response parameters. Response data were fed and analyzed by ANOVA to generate 3D plots indicating the optimum conditions and interaction among these factors. Regression equation 2 was used to fill the data by multiple regression procedure. The developed empirical statistical model showed the relation between measured characteristics and input parameters of the experiment

Table 2

Experimental factors and their levels for machining process

S. No.SymbolParameterUnitsLevel (−2)Level 2 (−1)Level 3 (0)Level 4 (1)Level 5 (2)
1APulse-on time (Ton)MU*107109111113115
2BPulse-off time (Toff )MU*3235384144
3CGap Voltage (SV)V*4246505458
4DPeak Current (IP)A*110120130140150
5EWire Tension (WT)MU*89101112
6FWire feed rate (WF)m/min45678

  1. *MU – machine Unit; V-Voltage; A-Ampere

(2)y=β0+i=1kβiXi+i=1kj=1,i<jkβijXiXj+i=1kβiiXi2+ϵ

where, y is the predicted response [MRR, SR and WWR], xi, xj are the independent variables, β0 is intercept coefficient, βi are the regression coefficients of zero order, βij is the squared coefficients. The quality of fitting by the polynomial model equation was expressed using coefficient of determination R2.

The individual desirability index (d) for each response is calculated. Desirability function approach is extensively used for single and multiple quality characteristics problems. The optimization process searches the optimum value of MRR, SR and WWR by maximization of MRR while minimizing the SR and WWR. If the output response is required to be maximized then desirability index (d) is calculated by using the equation 3.

(3)d=0y<LyLTLrLyT1y>T

If the output response (y) is required to be minimized then desirability index (d) is calculated by using the equation 4.

(4)d=1y<TUyUTrLyT0y>U

If the output response (y) is required to achieve a particular target T then desirability index (d) is calculated by using the equation 5.

(5)d=0y<LyLTLr1LyTUyUTr2TyU0y>U

Where d is the desirability index, L is the lower value, U is the upper value, and T is the target value. Desirability function (d) assigns number between 0 and 1. For d = 0 representing the completely undesirable value and d = 1 representing the completely desirable value.

Global desirability (D) is the combination of individual desirability for each response and can be calculated by using the equation 6.

(6)D=d1×d2dm1/m

Where, m is the number of responses.

2.4 Prediction of optimum values for maximum MRR, minimum SR and WWR

After getting the model equation that explains the process, it was used for optimization of the process parameters using the numerical optimization option of the software. Criteria were set for each independent variable and the response (dependent variable). The independent variables were kept in the range used by the experimental set up. The response for MRR was set to maximum and for SR and WWR, set to minimum. Response data was generated both for single response and multiple response optimizations. A solution was generated with predicted levels of the independent variables and predicted maximum MRR, minimum SR and minimum WWR.

2.5 Validation experiments

To ensure the validity of the chosen quadratic model, experiments were designed using the predicted optimum values of the parameters from equation 1. The responses were measured and compared with the predicted value. Experiments were conducted in triplicates and the data presented as mean ± SD.

2.6 Scanning electron micrographs of Inconel 825 specimens

All measurement related to surface micrograph were performed on JEOL scanning electron Microscope (Model 6100, JEOL, USA); a profile measuring microscope that determine the surface microstructures, formation of recast layer and heat-affected zone of the work material machined with WEDM. Etching process of machined sample was carried out with krolls reagent (2% (v/v) hydrofluoric acid, 10% (v/v) nitric acid). Then, the samples were cleaned using acetone (CH3)2CO.

2.7 Elemental and phase analysis of machined Inconel 825

The very high temperatures used in WEDM process immediately liquefy and vaporize the work material from the surface during the discharge process. The reactions between work specimen, wire electrode and dielectric resulted in the formation of different compounds. Energy Dispersive spectrograph (EDS) analysis was used to measure the element composition of machined surface. X-ray diffraction studies was carried out to analyze the phases of WEDM machined Inconel 825

3 Results and discussion

A variety of materials can be machined by using WEDM for instance titanium alloys, steels, and superalloys but steel is the material which has been studied in majority. Inconel 825 is a nickel based super alloy that has found wide range of applications owing to its high temperature stability. For optimal machining performance, standardization of important machining parameters viz. pulse-on time, pulse-off time, spark gap, peak current, voltage, wire tension, wire feed is required [9]. Many workers had used various strategies to optimize WEDM parameters for materials like steel, aluminum, Inconel 718 etc. [10, 15, 16, 17] but there are very few reports on the use of WEDM for the machining of Inconel 825 [4]. Therefore, this study was focused on the standardization of WEDM parameters for the machining of Inconel 825. Response surface methodology is a statistical method for analyzing the outcome of multiple input parameters on responses. In this study RSM has been employed to study and optimize the individual and interactive effect of operating parameters on the cutting of Inconel 825. Three significant parameters i.e. material removal rate (MRR), surface roughness (SR) and wire wear ratio (WWR)was taken as output parameters as these responses have crucial effect on surface integrity of Inconel 825. A total of 52 experiments were conducted as suggested by the software and results were fed and analyzed by ANOVA as shown in Table 3. In all the studies done so far, machining performance is optimized in such a way so that maximum material removal rate can be achieved with little damage to material surface.

Table 3

Central composite design with actual responses

RunTONTOFFSVIPWTWFMRRSRWWR
10.0000.0000.0000.0000.000−2.37821.272.6180.21966
20.0000.0000.0000.0000.0000.00020.442.3060.01531
3−1.000−1.0001.0001.0001.0001.00018.592.4530.01347
41.000−1.0001.000−1.0001.0001.00027.792.9670.0041
51.0001.0001.000−1.000−1.0001.00020.632.7660.01908
61.000−1.0001.000−1.000−1.000−1.00027.692.8670.01425
71.0001.000−1.0001.000−1.0001.00020.062.7110.00337
80.0000.0000.0000.0000.0000.00021.972.6730.13256
9−1.0001.000−1.000−1.000−1.0001.00017.012.7830.01848
10−1.000−1.0001.0001.000−1.000−1.00021.342.8440.15013
111.0001.000−1.0001.0001.000−1.00031.812.7930.00976
12−1.0001.0001.0001.000−1.0001.00016.172.7160.05423
13−1.0001.000−1.0001.000−1.000−1.00020.232.7320.02236
141.000−1.000−1.000−1.000−1.0001.00027.632.6780.00032
150.0000.0000.0000.0000.0000.00021.892.9110.03677
160.0000.0000.0000.0000.0000.00020.412.8750.03255
17−1.000−1.000−1.0001.000−1.0001.00028.942.4420.07354
182.3780.0000.0000.0000.0000.00031.253.320.01093
191.000−1.000−1.0001.0001.0001.00034.722.840.04773
200.000−2.3780.0000.0000.0000.00030.872.6130.01845
210.0002.3780.0000.0000.0000.00017.192.6670.01298
221.000−1.0001.0001.0001.000−1.00031.423.2710.02028
23−1.0001.0001.0001.0001.000−1.00018.552.5140.00291
240.0000.0000.0000.0000.0000.00035.932.9630.01573
251.0001.000−1.000−1.0001.0001.00028.953.2990.15061
26−1.000−1.000−1.000−1.000−1.000−1.00025.792.7210.20167
27−1.000−1.0001.000−1.0001.000−1.00019.922.4890.14103
280.0000.0000.0000.000−2.3780.00024.092.7670.00394
290.0000.0000.0002.3780.0000.00027.082.6840.01292
30−1.000−1.0001.000−1.000−1.0001.00020.212.6810.02832
311.0001.0001.0001.0001.0001.00027.682.7420.04154
320.0000.0000.0000.0002.3780.00026.052.7920.00823
330.0000.0000.0000.0000.0000.00024.593.0740.00545
34−1.000−1.000−1.0001.0001.000−1.00025.472.6220.01644
350.0000.0000.0000.0000.0000.00023.942.4610.01049
360.0000.000−2.3780.0000.0000.00025.572.5940.00661
371.0001.0001.0001.000−1.000−1.00033.282.8920.12435
38−1.0001.000−1.0001.0001.0001.00020.522.4150.00404
39−1.0001.0001.000−1.000−1.000−1.00016.72.5210.04631
401.0001.0001.000−1.0001.000−1.00024.462.2630.03279
41−1.0001.0001.000−1.0001.0001.00016.712.3180.01062
42−1.000−1.000−1.000−1.0001.0001.00023.142.4850.08701
430.0000.0002.3780.0000.0000.00019.612.7340.01789
441.0001.000−1.000−1.000−1.000−1.00028.392.8050.00616
451.000−1.000−1.000−1.0001.000−1.00034.562.8270.00815
46−2.3780.0000.0000.0000.0000.00014.72.8440.03203
470.0000.0000.0000.0000.0002.37825.772.940.01582
481.000−1.000−1.0001.000−1.000−1.00036.142.8330.02691
491.000−1.0001.0001.000−1.0001.00036.132.8420.01832
500.0000.0000.0000.0000.0000.00029.482.8530.00853
51−1.0001.000−1.000−1.0001.000−1.00023.672.640.1215
520.0000.0000.000−2.3780.0000.00028.082.7480.08998

3.1 Analysis of variance for MRR

By applying multiple regression analysis on the experimental data, a predictive two factor polynomial equation was constructed to describe the correlation between MRR and the six process parameters as follows:

(7)Final Equation in Terms of Coded Factors:MRR=25.30+3.73A2.80B1.59C+0.40D+0.75E0.29F1.14AB+0.93AC+0.68AD+0.16AE0.23AF+0.093BC+0.10BD+1.74BE0.24BF+0.13CD1.26CE+0.39CF+0.084DE+0.41DF+0.024EF

Where, R is the response (MRR, mm2 min−1), A, B, C, D, E and F were the coded values of pulse-on time, pulse-off time, gap voltage, peak current, wire tension and wire feed rate respectively. The analysis of variance for the response surface model is summarized in Table 4. The Model F-value of 33.67 implies that the model is significant. The p-values <0.05 indicated that the linear (A, B, C, E) and interactive (AB, AC, AD, BE, CE) terms had quite significant influence on material removal rate.

Table 4

Analysis of variance (ANOVA) for response MRR

SourceSum of SquaresdfMean SquareF Valuep-value Prob > F
Model1334.742163.5633.67< 0.0001significant
A-Pulse-on603.231603.23319.54< 0.0001
Time
B-Pulse-off338.791338.79179.46< 0.0001
Time
C-Gap109.231109.2357.86< 0.0001
Voltage
D-Peak7.0417.043.730.0630
Currrent
E-Wire24.66124.6613.060.0011
Tension
F-Wire Feed3.6913.691.950.1725
AB41.72141.7222.10< 0.0001
AC27.71127.7114.680.0006
AD14.58114.587.720.0093
AE0.7910.790.420.5216
AF1.6611.660.880.3564
BC0.2710.270.150.7060
BD0.3310.330.180.6779
BE96.95196.9551.36< 0.0001
BF1.8911.891.000.3248
CD0.5410.540.280.5982
CE50.95150.9526.99< 0.0001
CF4.9814.982.640.1149
DE0.2210.220.120.7326
DF5.4815.482.900.0988
EF0.01810.0189.561E-0030.9228
Residual56.63301.89
Lack of Fit33.79231.470.450.9298not significant
Pure Error22.8573.26
Cor Total1391.3751
Std. Dev.1.37R-Squared0.9593
Mean25.30Adj R-Squared0.9308
C.V. %5.43Pred R-Squared0.9137
PRESS120.04Adeq Precision21.727

The percentage contribution of A, B, C, E, AB, AC, AD, BE, CE for MRR is 45.19, 25.38, 8.18, 1.84, 3.12, 2.07, 1.09, 7.26 and 3.81% respectively which are calculated from Table 4 by dividing the each variable sum of square term by “model” sum of square. The lack of fit was found to be not significant. The p-value for lack of fit was 0.9298, indicating that this model adequately fit into the data. The determination coefficient R2 (0.9593) indicated that the predicted and experimental values had perfect coherence with each other. The value of adjusted R2 (0.9308) suggested that the variation of 93.08% in the MRR was attributed to the independent variables and 6.92% of the total variation could not be explained by the model.

3.1.1 Individual effect of Process parameters on MRR

Effect of individual process parameters on MRR is determined by perturbation graph (Figure 2). It was observed that pulse-on time (A), peak current (D) and wire tension (E) has significant effects on MRR during machining of Inconel 825 by WEDM. The steep curve of Ton,WT, IP showed that MRR is highly sensitive to pulse-on time (A), followed by Wire tension (E) and peak current (D) when compared with other controllable factors such as wire feed (F), gap voltage (C) and pulse-of time (B).

Figure 2 Perturbation plot showing the effect of individual parameters on MRR

Figure 2

Perturbation plot showing the effect of individual parameters on MRR

3.1.2 Interactive effect of process parameters on MRR

The interaction between two variables is presented in Figure 3 which was generated by the pair-wise combination of the two factors while keeping the other one at its optimum level. The three dimensional plot between pulse-on time and pulse-off time (AB), pulse-on time and gap voltage (AC), pulse-on time and peak current (AD), pulse-on time and wire tension (AE), pulse-off time and wire tension (BE) and peak current and wire tension (DE), peak current and wire feed (DF), wire tension and wire feed (EF) with MRR are shown in Figure 3 (a-h) respectively.

Figure 3 Three dimensional plot of combined effects of (a) pulse-on time and pulse-off time (b) pulse-on time and gap voltage (c) pulse-on time and peak current (d) pulse-on time and wire tension (e) pulse-off time and wire tension (f) and gap voltage and wire tension (g), peak current and wire tension (h), peak current and wire feed (i), wire tension and wire feed on MRR when other factors were kept constant

Figure 3

Three dimensional plot of combined effects of (a) pulse-on time and pulse-off time (b) pulse-on time and gap voltage (c) pulse-on time and peak current (d) pulse-on time and wire tension (e) pulse-off time and wire tension (f) and gap voltage and wire tension (g), peak current and wire tension (h), peak current and wire feed (i), wire tension and wire feed on MRR when other factors were kept constant

From the figure it was observed that pulse-on time had significant positive effect on MRR, as MRR increased from 19mm2/min to 33 mm2/min approximately with an increase in value of pulse-on time from 109 MU to 113 MU (Figure 3a) while pulse-off time had negative effect on MRR, as MRR decreased with increases in value of pulse-off time from 35 MU to 41 MU (Figure 3a). However, when applied in combination, the MRR value decreases. At high value of Ton, high discharge energy is produced leading to melting of more work material from the surface. High intense heat produced in plasma zone accelerates the erosion process which results in increase in the material removal rate [28, 29].

Similarly, gap voltage was shown to have negative effect on MRR (Figure 3b). The spark gap voltage is theoretical voltage difference between wire electrode and workpiece during erosion [30]. In combination pulse-on time and gap voltage reduced the MRR. MRR will be maximum at 113 MU pulse-on time and low gap voltage i.e. 46 V to 48 V (Figure 3b). The reason can be attributed due to the fact that high value of gap voltage increases the gap between two successive sparks and less discharge energy is produced between the gaps resulted in decreased MRR [31, 32]. Similar results have been reported during the WEDM machining of Ti50Ni50-xCux alloy using both brass wire and zinc coated brass wires [33].

It is evident from Figure 3c that MRR increased with increase in the value of pulse-on time and peak current as discharge energy is the product of pulse-on time and peak current which in turn improve MRR [34].MRR is maximum (30 mm2/min) at 113 MU pulse-on time and 140 amp peak current (Figure 3c). At higher value of IP the gap condition become unstable and to stable the gap condition it is necessary to reduce the value of peak current. Similarly, MRR is maximum (29mm2/min) at interactive effect of pulse-on time (113 MU) and wire tension (11 kg-f) (Figure 3d) while wire tension in combination with pulse-off time and peak current decreases MRR (Figure 3e-3f). Wire tension is the gram equivalent load with which the continuously fed wire is kept under tension so that it remains straight between the wire and workpiece. Saini et al. [34] also analyzed the effect of wire tension on MRR in the machining of Ti-6Al-4V Alloy. They observed that MRR decreases with increase in wire tension. Wire feed have little positive effect on MRR thus, resulting in little increase in MRR in combination with peak current and wire tension (Figure 3g-3h). Similar results had been observed by Patel et al. [35].

3.2 Analysis of variance for surface roughness (SR)

By applying multiple regression analysis on the experimental data, a predictive two factor polynomial equation was constructed to describe the correlation between SR and the six process parameters as follows: Final Equation in Terms of Coded Factors:

(8)Final Equation in Terms of Coded Factors:SR=2.68+0.12A7.032E003B+0.019C+0.012D0.022E8.676E003F0.041AB1.219E003AC+0.022AD+0.053AE+0.039AF0.078BC+6.781E003BD0.066BE+0.049BF+0.067CD0.053CE+0.011CF+0.028DE0.078DF+0.014EF

The Model F-value of 37.20 implies the model is significant. In this case A, C, E, AB, AD, AE, AF, BC, BE, BF, CD, CE, DE, DF are significant model terms for SR with their contribution percentage of 36.87, 0.837, 0.346, 3.016, 0.893, 5.083, 2.681, 11.173, 7.821, 4.357, 7.821, 0.212, 1.396 and 10.614% respectively which are calculated from Table 4 by dividing the each variable sum of square term by “model” sum of square. Lack-of fit value of 0.1642 implies that it is not significant relative to pure error. The determination coefficient for SR is found to be 0.9630 which shows that the factorial model can explain the variation in the surface roughness up to the extent of 96.30%. On the basis of the high values of the determination coefficient, it can be said that the proposed model is adequate in representing the process. The other R2 statistics, the Pred R2 (0.8491), is in good agreement with the Adj R2 (0.9371).

3.2.1 Individual effect of Process parameters on surface roughness (SR)

Surface roughness is the variation or irregularity of a machined surface from its ideal atomic value. Roughness of surface must be minimum for high quality material [11]. Therefore, the effect of each parameter on surface roughness of Inconel 825 was determined by perturbation plot (Figure 4). The steep slope curves and relatively flat line was observed which shows that surface roughness is highly sensitive to pulse-on time (A) and peak current (C) and less sensitive to wire tension (E), wire feed (F).

Figure 4 Perturbation plot showing the effect of individual parameters on SR

Figure 4

Perturbation plot showing the effect of individual parameters on SR

3.2.2 Interactive effect of process parameters on SR

The two variables interaction when the others were kept at its optimum value is for SR is presented in Figure 5 were generated on the basis of two factors combination while keeping the other one at its optimum level. The interaction plot between pulse-on time and pulse-off time (AB), pulse-on time and peak current (AD), pulse-on time and wire tension (AE), pulse-on time and wire feed (AF), pulse-off time and gap voltage (BC), pulse-off time and wire tension (BE), pulse-off time and wire feed (BF), gap voltage and peak current (CD), gap voltage and wire tension (CE), peak current and wire tension (DE) peak current and wire feed (DF), wire tension and wire feed (EF) are shown in Figure (5a-5k) respectively.

Figure 5 Three dimensional plot of combined effects of (A) pulse-on time and pulse-off time (B) pulse-on time and peak current (C) pulse-on time and wire tension (D) pulse-on time and wire feed (E) pulse-off time and gap voltage (F) pulse-off time and wire tension (G) pulse-off time and wire feed (H) gap voltage and peak current (I) gap voltage and wire tension (J) peak current and wire tension (K) peak current and wire feed (L) wire tension and wire feed on SR when other factors were kept constant

Figure 5

Three dimensional plot of combined effects of (A) pulse-on time and pulse-off time (B) pulse-on time and peak current (C) pulse-on time and wire tension (D) pulse-on time and wire feed (E) pulse-off time and gap voltage (F) pulse-off time and wire tension (G) pulse-off time and wire feed (H) gap voltage and peak current (I) gap voltage and wire tension (J) peak current and wire tension (K) peak current and wire feed (L) wire tension and wire feed on SR when other factors were kept constant

From the Figure 5a-5b, it was observed that surface roughness increases upto 2.756 μm at combination of pulse-on time and pulse-off time. The reason can be attributed to the fact that pulse-on time is the prominent factor to increase the discharge energy. At high value of pulse-on time duration of current flowing increases and high discharge energy produced between two electrodes which in turn increase the surface roughness [29]. At low value of pulse-on time, the time duration of current flowing decreases and at high value of pulse-off time, the time interval between two sparks increases so that discharge energy produced was less and less temperature generated between two electrodes which in turn improve the surface roughness [36].

At high value of pulse-on time 113 machine unit, surface roughness increases upto 2.838 μm with increase in the value of peak current from 120 A to 140 A. as shown in fig 5b. Minimum surface roughness occurred (2.569 μm) at low value of pulse-on time i.e. 111 to 109 machine unit and low value of peak current 125 to 120 A (Figure 5b). At high value of peak current, the pulse discharge energy increases which in turn increase the cutting rate as well as increase the surface roughness and WWR [37, 38].

Wire tension is significant process parameter for surface roughness and affects the geometry of specimen cut from the workpiece. Surface roughness decrease with increase of wire tension from 9 to 11 machine unit and when applied in combination with pulse-on time the surface roughness increases upto 2.834 μm as shown in Figure 5c. It has been observed that increase in wire tension will reduce the vibrations of wire and cause reduction in surface roughness resulting in improved quality of the machined surface [32]. Wire feed has little effect on surface roughness. Increase in wire deed resulted in less increase (Figure 5d). At high value of wire feed with pulse-on time, the surface roughness increases upto 2.835 μm because at high value of wire feed, more wire moves on wire guides through the geometry of the workpiece resulting in large area of the wire get exposed to sparking zone forming large craters and pockmarks on the work surface which in turn increase the surface roughness [39].

Figure 5e-5f shows the interactive effect of pulse-off time with gap voltage and wire tension. The surface roughness decreases upto 2.618 μm and 2.590μm respectively as shown in Figure 5e-5f. At high value of gap voltage spark gap becomes wider which leads to less numbers of sparks per unit time and slow down the cutting rate as a result less numbers of craters form on work surface [40]. High value of pulse-off time increases the gap between two successive sparks which in turn decrease the surface roughness [41]. Wire feed has negative significant effect on surface roughness (2.746 to 2.714 μm) with pulse-off time whereas peak current has positive influence on surface roughness (2.717 to 2.777 μm) when applied in combination with gap voltage (Figure 5g-5h). At high value of peak current the pulse discharge energy increases which leads to increase in the cutting rate as well as increase in the surface roughness.

3.3 Analysis of variance for wire wear ratio (WWR)

By applying multiple regression analysis on the experimental data, a predictive two factor polynomial equation was constructed to describe the correlation between WWR and the six process parameters as follows:

(9)Final Equation in Terms of Coded Factors:WWR=0.0287.211E003A2.300E003B+1.777E003C6.534E003D1.851E004E6.258E003F+0.017AB+5.031E004AC+3.858E003AD+6.473E003AE+0.011AF2.647E003BC3.798E003BD+6.251E003BE+6.696E003BF+0.014CD0.011CE9.720E003CF0.011DE5.424E003DFm,./+0.013EF

The F-value of model is 23.07 indicates that model is significant (Table 5). In this case A, D, F, AB, AD, AE, AF, BD, BE, BF, CD, CE, CF, DE, DF, EF are significant model terms for WWR with their contribution percentage of 4.69, 3.85, 3.53, 19.51, 0.99, 2.79, 7.47, 0.96, 2.60, 2.98, 13.59, 8.38, 6.29, 7.97, 1.96 and 11.26% respectively, which are calculated from Table 4 by dividing the each variable sum of square term by “model” sum of square. Lack-of fit value of 0.4856 implies that it is not significant relative to pure error. The determination coefficient for WWR is found to be 0.9417 which shows that the quadratic model can explain the variation in the wire wear ratio up to the extent of 94.17%. On the basis of the high values of the determination coefficient, it can be said that the proposed model is adequate in representing the process. The other R2 statistics, the Pred R2 (0.8340), is in good agreement with the Adj R2 (0.9009).

Table 5

Analysis of variance (ANOVA) for response SR

SourceSum of SquaresdfMean SquareF Valuep-value Prob > F
Model1727.36210.08537.20< 0.0001Significant
A-Pulse-on637.261637.26288.17< 0.0001
Time
B-Pulse-off2.0612.060.930.3417
Time
C-Gap14.74114.746.670.0149
Voltage
D-Peak5.9915.992.710.1104
Currrent
E-Wire20.14120.149.110.0052
Tension
F-Wire Feed3.1413.141.420.2426
AB51.93151.9323.48< 0.0001
AC0.04610.0460.0210.8865
AD15.14115.146.850.0138
AE87.77187.7739.69< 0.0001
AF46.24146.2420.91< 0.0001
BC188.111188.1185.06< 0.0001
BD1.4211.420.640.4295
BE134.481134.4860.81< 0.0001
BF75.10175.1033.96< 0.0001
CD137.811137.8162.32< 0.0001
CE85.32185.3238.58< 0.0001
CF3.6713.671.660.2076
DE24.24124.2410.960.0024
DF186.301186.3084.25< 0.0001
EF6.4616.462.920.0978
Residual66.34302.21
Lack of Fit57.81232.512.060.1642not significant
Pure Error8.5371.22
Cor Total1793.7151
Std. Dev.1.49R-Squared0.9630
Mean16.72Adj R-Squared0.9371
C.V. %8.89Pred R-Squared0.8491
PRESS270.62Adeq Precision35.840

3.3.1 Individual effect of Process parameters on WWR

Figure 6 shows the Perturbation plot of WWR. From the figure it was clear that wire wear ratio is highly sensitive to pulse-on time (A), peak current (D) and wire feed (F), as increase in these factors results in decrease in WWR while increase in pulse-off time (B), gap voltage (C) and wire tension (E) increase WWR.

Figure 6 Perturbation plot showing the effect of individual parameters on WWR

Figure 6

Perturbation plot showing the effect of individual parameters on WWR

3.3.2 Interactive effect of process parameters on WWR

Based on analysis of variance, the interactions between peak current and pulse-on time (AD), pulse-on time and wire tension (AE), pulse-on time and wire feed (AF), pulse-off time and peak current (BD), gap voltage and wire tension (CE), gap voltage and wire feed (CF), peak current and wire tension (DE) peak current and wire feed (DF) which contribute significantly for WWR are shown in Figure 7 (a-g) respectively.

Figure 7 Three dimensional plot of combined effects of (A) pulse-on time and peak current (B) pulse-on time and wire tension (C) pulse-on time and wire feed (D) pulse-off time and peak current (E) gap voltage and wire tension (F), gap voltage and wire feed (G) peak current and wire tension (H) peak current and wire feed on WWR when other factors were kept constant

Figure 7

Three dimensional plot of combined effects of (A) pulse-on time and peak current (B) pulse-on time and wire tension (C) pulse-on time and wire feed (D) pulse-off time and peak current (E) gap voltage and wire tension (F), gap voltage and wire feed (G) peak current and wire tension (H) peak current and wire feed on WWR when other factors were kept constant

From the interaction plot, it was observed that wire wear ratio set at its minimum value i.e. 0.0217 at low value of pulse-on time and peak current. The highest increment of WWR i.e. 0.0443 was observed at high value of pulse-on time (113 machine unit) and peak current (140 A) (Figure 7a). It was observed experimentally that high value of pulse-on time increases the discharge energy which leads to formation of large craters on the tool surface and wear out the wire electrode rapidly. Similar report has been found from the interaction plot (Figure 7b) of pulse-on time and wire tension. When pulse-on time and wire tension is set at highest value the wire wear ratio is recorded as 0.0354. The higher wire wear ratio i.e. 0.0332 is also accompanied at highest value of pulse-on time and wire feed as shown in Figure 7c. In contrast, at high value of pulse-off time, rate of spark generated get decreased and decreased value of WWR was observed i.e. 0.0219 (Figure 7d). It was observed from the Figure 7e that increase value of peak current results in a corresponding increase of wire wear ratio from 0.0235 to 0.0310 due to intense thermal effect on the surface of wire electrode and more material eroded from the wire electrode in the form of craters [37]. In contrast, at high value of pulse-off time, rate of spark generated get decreased and WWR of 0.019 was observed as minimum. Similarly, when gap voltage is set up at high level 54 V and wire tension is set up 11 MU, the wire wear ratio of 0.02 was recorded as minimum. At high value of wire tension fewer craters were observed on wire electrode surface so less wire wear ratio was recorded [32].

3.4 Multi response optimization using desirability approach

After optimizing the each output responses individually, multi-response optimization was carried out for the response variables MRR, SR and WWR. Desirability function approach is extensively used for single and multiple quality characteristics problems [42]. It is based on the idea that the "quality" of a product or process which is outside the desired limit is completely undesirable for multiple quality characteristics. The method results in operating conditionswhich provides the "most desirable" output response characteristics. Numerical optimization option of design expert software was used to carry out this. The values of six input variables were kept in range and the response factor MRR was set to maximum while SR was set to minimum. The value of WWR is target to minimum value. The lower and upper limit of each variable was taken individually and the weight of each variable and response was set to their default value of 1. Because the target of the present study is to increase the cutting rate while maintaining the minimum surface damage, the importance of MRR, SR and WWR were taken as 3.

Experiments were performed under predicted conditions as given by the software. It was observed that at 110 MU Ton, 35 MU Toff , 46 V SV, 120A IP, 11 MU WT and 5 m/min WF, the values obtained for MRR, SR and WWR were 27.691, 2.721 and 0.117 respectively which are near to the predicted values (Table 6) with an error of less than 5%. Thus, the model was successfully validated.

3.5 Microstructure analysis of worked specimen surface and wire electrode surface

To study the effect of machining on work surface, scanning electron micrograph was performed for Inconel 825 machined with WEDM and wire electrode used. The surface topography includes micro-cracks, craters, debris, spherical debris, pockmarks, heat affected zone and recast layer. The experiments which were showing maximum MRR (run 48) (Figure 8), minimum SR (run 40) (Figure 9) and minimum WWR (run 14) (Figure 10) were chosen for SEM analysis. Material cut with finally optimized values was also scanned using surface analysis (Figure 11).

Figure 8 Scanning electron micrographs of work material for MRR

Figure 8

Scanning electron micrographs of work material for MRR

Figure 9 Scanning electron micrographs of work material for SR

Figure 9

Scanning electron micrographs of work material for SR

Figure 10 Scanning electron micrographs of work material for WWR

Figure 10

Scanning electron micrographs of work material for WWR

Figure 11 Scanning electron micrographs of work material under optimized conditions

Figure 11

Scanning electron micrographs of work material under optimized conditions

3.5.1 Effect of operating parameters on machined surface for maximum MRR

It was observed that at maximum value of MRR i.e. 36.14 mm2/min (Exp. 48), deep and large craters were formed due to high value of pulse-on time (113MU) and high value of peak current (140 A) (Figure 8a). The reason can be attributed to the fact that at high value of pulse-on time and discharge energy, intense heat transfers towards the work surface which causes melting and evaporation of more work material from the surface resulting in larger numbers of pockmarks, craters and spherical debris [10, 27]. During machining some amount of the melted material was carried away by deionized water and material extruded in pulled-out shape formed on work surface. Rest of the material re-solidified in the form of spherical or irregular shapes of debris on work surface and because of the dramatic temperature changes in the workpiece a multi layered surface is created named as recast layer, heat affected zone and the transformed layer as shown in Figure 8(d). From the Figure 8(d) it was observed that due to rapid heating and quenching heat affected zone is formed below the recast layer and due to thermal residual stresses some cracks produced near the zone. These spherical nodules were formed due to high value of pulse-on time and high value of peak current which also resulted in the martensite structure.

3.5.2 Effect of operating parameters on machined surface of minimum SR

On the other hand, fewer craters and less micro-crack and fewer pockmarks were observed on the machined specimen surface when run 40 showing minimum surface roughness was analyzed through SEM (Figure 9). The reason can be because peak current at its low value and high value of pulse-off time increase the time interval between two sparks resulting in less material removal from the surface and improved smoothness. Low value of peak current transfers less intense heat towards the surface of the sample [37, 38]. Moreover, high value of pulse-off time increase the time interval between the spark thus leads to improve the surface roughness [43]. It was observed from the Figure 9(d) that the recast layer and heat affected zone tends to decrease with parallel decrease of peak current and increase of pulse-off time. The change in grain structure from the base structure is less apparent as compared to Figure 8(d).

3.5.3 Effect of operating parameters on machined surface of minimum WWR

Similarly, minimum WWR was achieved from exp 14 which is identified as, pulse-on time 113 MU, pulse-off time 35 MU, spark gap voltage 46 V, peak current 120 amp, wire tension 9 MU , wire feed 7 m/min. SEM micrographs of the specimen showed less numbers of craters, cracks and debris on wire electrode surface (Figure 10). The reason can be because when peak current and spark gap voltage are set at lowest level (120amp and 46 V) the wire wear ratio reaches its minimum value so fewer craters and fewer microcracks were created on the surface of wire electrode [44].

3.5.4 Effect of operating parameters on machined surface under optimized condition

Under optimized conditions of process parameters, multiple quality characteristics i.e. MRR, SR and WWR get slightly improved. It was observed from the SEM Micrograph, that at optimal combination of input parameters i.e. 111 MU Ton, 35 MU Toff , 46V SV, 140A IP, 9 MU WT and 6 m/min WF comparatively less numbers of cracks, pockmarks, craters, and pulled out material were found on work specimen surface and wire electrode surface (Figure 11) as compared to under unoptimized conditions. The reason can be because at low value of pulse-on time and gap voltage less heat transfer toward the workpiece surface which causes melting and evaporation of less work material from the surface in comparison to single quality characteristics.

Table 6

Analysis of variance (ANOVA) for response surface WWR

SourceSum of SquaresdfMean SquareF Valuep-value Prob > F
Model21328.33211015.6323.07< 0.0001Significant
A-Pulse-on999.311999.3122.70< 0.0001
Time
B-Pulse-off101.691101.692.310.1390
Time
C-Gap60.71160.711.380.2495
Voltage
D-Peak820.391820.3918.640.0002
Currrent
E-Wire0.6610.660.0150.9035
Tension
F-Wire Feed752.511752.5117.090.0003
AB4156.6814156.6894.43< 0.0001
AC3.5913.590.0820.7770
AD211.261211.264.800.0364
AE594.901594.9013.510.0009
AF1592.6111592.6136.18< 0.0001
BC99.47199.472.260.1432
BD204.741204.744.650.0392
BE554.701554.7012.600.0013
BF636.621636.6214.460.0007
CD2895.6612895.6665.78< 0.0001
CE1786.3211786.3240.58< 0.0001
CF1341.3711341.3730.47< 0.0001
DE1698.4411698.4438.58< 0.0001
DF417.751417.759.490.0044
EF2398.9312398.9354.50< 0.0001
Residual1320.623044.02
Lack of Fit1033.562344.941.100.4856not significant
Pure Error287.07741.01
Cor Total22648.9551
Std. Dev.6.63R-Squared0.9417
Mean81.13Adj R-Squared0.9009
C.V. %8.18Pred R-Squared0.8340
PRESS3759.52Adeq Precision24.877

3.6 Elemental and phase analysis of machined sample

From the EDX analysis of work specimen, it was observed that some metal particles are present on the work surface other than base material because during WEDM process some of the debris cannot be flushed from the sparking gap so that rest part of debris get deposited on the surface as compounded form (Figure 12). EDX analysis showed that significant amount of elements migrated from wire electrode to workpiece. Fe (Ferrous), Cr (Chromium), Cu (Copper), Ni (Nickel) was observed at highest peak in the spectrum of EDX which shows that an appreciable amount of these elements are migrated to the surface of the workpiece. Migration of the elements depends upon the value of pulse-on time and gap voltage. At high value of pulse-on time and gap voltage, spark energy increases which causes more melting and evaporation of material [45].

Figure 12 EDX analysis of work material

Figure 12

EDX analysis of work material

Some residual of the C (Carbon), Cu (Copper) and O (Oxygen) elements were also observed on the machined sample because of decomposition of dielectric, resolidification of the wire electrode and mixing of the debris at elevated temperature.

Table 7

Validation of predicted model

Type ofObjectiveOptimization parametersResponseResponseDesirability
optimizationTonToffSVIPWTWF(Predicted)(Experimental)
Single responseMaximize11335541409735.84134.0120.986
MRR
Single responseMinimize SR1093546140972.0722.1521.000
Single responseMinimize WWR1093551132950.1100.1161.000
Multi response &Minimize SR,110354612011528.166,27.691 2.7210.715
Maximize MRRWWR2.700,0.117
0.111

The debris that gets deposited in compounded form on the work surface was examined by X-ray diffraction analysis. Different compounds were formed due to migration of workpiece elements, tool electrode elements and decomposition of deionized water. Different phases were analyzed by transfer of the tool and dielectric elements on the work surface by using X’ Pert High score plus (Figure 13). It was observed from the Figure 13 that copper dioxides (CuO2), Molybdenum dioxides (MoO2) and ferric oxide (FeO) were found as compounded form on the surface of the specimen.

Figure 13 XRD analysis of work material

Figure 13

XRD analysis of work material

4 Conclusions

In this study, multi objective optimization was carried out to study the effect of WEDM machining parameters on machine performance using RSM based desirability approach and microstructure analysis at optimum combinations of process parameters was performed.

  • It was concluded that pulse-on time, gap voltage and peak current have significant positive effect on increasing MRR while increase in pulse-off time resulted in decreased SR.

  • Comparatively less numbers of cracks, pockmarks, craters, and pulled out material were found on work specimen surface and wire electrode surface as compared to under unoptimized conditions.

  • EDX analysis showed that significant amount of elements migrated from wire electrode to workpiece.

  • Copper dioxides (CuO2), Molybdenum dioxides (MoO2) and ferric oxide (FeO) were found as compounded form on the surface of the specimen.

  • The optimum parametric combination obtained from the present study will be advantageous for working on high strength, high thermal conductivity and low melting point materials like nickel alloys even at higher temperatures.


Tel.: +91-9996260406

  1. Conflict of interest

    Declaration of Conflicting Interests: The author states that there is no conflict of interest.

  2. Authors’ Contribution: All the authors have contributed to the manuscript. First author is responsible for carrying out all the practical work and writing of manuscript. Second and third author have verified the content and calculations made in the manuscript and provided all the technical support for the work.

References

[1] Akca E., Gursel A., A review on Superalloys and IN718 Nickel-Based INCONEL Superalloy, Period. Eng. Nat. Sci., 2013, 3(1), 15-27.10.21533/pen.v3i1.43Search in Google Scholar

[2] Palakudtewar R.K., Gaikwad S.V., Dry Machining of Superalloys: Diflculties and Remedies, Int. J. Sci. Res., 2014, 3(7), 277-282.Search in Google Scholar

[3] Payal H., Maheshwari S., Bharti P.S., Process Modeling of Electric Discharge Machining of Inconel 825 Using Artificial Neural Network, World Acad. Sci. Eng. Technol., 2017, 11(3), 562-566.10.1007/s41870-018-0102-7Search in Google Scholar

[4] Kumar P. Meenu, Kumar V., Optimization of Process Parameters for WEDM of Inconel 825 Using Grey Relational Analysis, Decis., Sci. Lett., 2018, 7(4), 405-41610.5267/j.dsl.2018.1.006Search in Google Scholar

[5] Rajyalakshmi G., Venkata Ramaiah P., Multiple process parameter optimization ofwire electrical discharge machining on Inconel 825 using Taguchi grey relational analysis, Int. J. Adv. Manuf. Tech. 2013, 69, 1249-1262.10.1007/s00170-013-5081-zSearch in Google Scholar

[6] Pleta A., Mears L., Cutting Force Investigation of Trochoidal Milling in Nickel-Based Superalloy, Procedia Manuf., 2016, 5, 1348-1356.10.1016/j.promfg.2016.08.105Search in Google Scholar

[7] Liao Y.S., Lin H.M., Wang J.H., Behaviors of end milling Inconel 718 superalloy by cemented carbide tools, J. Mater. Proc. Technol., 2008, 201(1-3), 460-465.10.1016/j.jmatprotec.2007.11.176Search in Google Scholar

[8] Ulutan D., Ozel T., Machining induced surface integrity in titanium and nickel alloys: A review, Int. J. Mach. Tool Manu., 2011, 51, 250-280.10.1016/j.ijmachtools.2010.11.003Search in Google Scholar

[9] Goswami A., Kumar J., Investigation of surface integrity, material removal rate and wire wear ratio for WEDM of Nimonic 80A alloy using GRA and Taguchi method, J. Eng. Sci. Technol., 2014, 1, 173-184.10.1016/j.jestch.2014.05.002Search in Google Scholar

[10] Singh A., Anandita S.,Gangopadhyay S., Microstructural analysis and multi response optimization during ECM of Inconel 825 using hybrid approach, Mater. Manuf. Proc., 2015, 30(7), 842-851.10.1080/10426914.2014.973575Search in Google Scholar

[11] Rahul Datta S., Biswal B.B., Mahapatra S.S., A Novel Satisfaction Function and Distance-Based Approach for Machining Performance Optimization during Electro-Discharge Machining on Super Alloy Inconel 718, Arabian J. Sci. Eng. 2017, 42, 1999-2020.10.1007/s13369-017-2422-5Search in Google Scholar

[12] Rajyalakshmi G., Venkata P.A., A parametric optimization using Taguchi method: effect of WEDM parameters on surface rough-nessmachining on Inconel 825, Elixir Int. J., 2012, 43, 6669-6674.Search in Google Scholar

[13] Aggarwal V., Singh S., Garg R.K., Parametric modeling and optimization for wire electrical discharge machining of Inconel 718 using response surface methodology, Int. J. Adv. Manuf. Tech., 2015, 79, 31-47.10.1007/s00170-015-6797-8Search in Google Scholar

[14] Mahapatra S.S., Patnaik A., Optimization of wire electrical discharge machining (WEDM) process parameters using Taguchi method, Int. J. Adv. Manuf. Tech., 2007, 34, 911-925.10.1007/s00170-006-0672-6Search in Google Scholar

[15] Chhabra S., Gupta D., Yadav P., Optimizing the Process Parameters of WEDM using Taguchi Technique on P-21 die steel, Int. J. Innov. Res. Sci. Eng. Technol., 2014, 3(7), 14815-14821.Search in Google Scholar

[16] Patra D.R., Rout I.S., Sahoo M., Optimization of WEDM parameters using Taguchi method for higher material removal rate on EN31 steel, Int. J. Eng. Res. Appl., 2015, 5(6), 57-62.Search in Google Scholar

[17] Singh V.K., Singh S., Multi-objective optimization using Taguchi Based Grey Relational Analysis for Wire EDM of Inconel 625, J. Mater. Sci. Mech. Eng., 2015, 2, 38-42.Search in Google Scholar

[18] Dabade U.A., Karidkar S.S., Analysis of response variables in WEDM of Inconel 718 using Taguchi technique, Procedia CIRP, 2016, 41, 886-891.10.1016/j.procir.2016.01.026Search in Google Scholar

[19] Raykundaliya D.P., Shanubhogue A., Comparison Study: Taguchi Methodology vis.-a-vis. Response Surface Methodology through a case study of accelerated failure in Spin-on-Filter, Int. Adv. Res. J. Sci. Eng. Technol., 2015, 2, 1-5.10.17148/IARJSET.2015.2301Search in Google Scholar

[20] Goyal A.J., Investigation of material removal rate and surface roughness during wire electrical discharge machining (WEDM) of Inconel 625 super alloy by cryogenic treated tool electrode, King Saud Univ.- Sci., 2017, 29(4), 528-535.10.1016/j.jksus.2017.06.005Search in Google Scholar

[21] Kumar S., Dhingra A.K., Kumar S., Parametric optimization of powder mixed electrical discharge machining for nickel based superalloy inconel-800 using response surface methodology, Mech. Adv. Mater. Modern Pro., 2017, 3(7), 1-17.10.1186/s40759-017-0022-4Search in Google Scholar

[22] Bijeta Nayak B., Sankar Mahapatra S., Optimization of WEDM process parameters using deep cryo-treated Inconel 718 as work material, Int. J. Eng. Sci. Technol., 2015, 19, 161-170.10.1016/j.jestch.2015.06.009Search in Google Scholar

[23] Li L., Li Z.Y., Wei X.T., Cheng X., Machining Characteristics of Inconel 718 by Sinking-EDM and Wire-EDM, Mater. Manuf. Proc., 2015, 30(8),968-973.10.1080/10426914.2014.973579Search in Google Scholar

[24] Yusoff Y., Mohd Zain A., Sharif S. et al., Potential ANN prediction model for multi-performances WEDM on Inconel 718, Neural Comp. Appl., 2018, 30(7), 2113-2127.10.1007/s00521-016-2796-4Search in Google Scholar

[25] Mohanty A., Gangadharudu T., Gangopadhyay S., Experimental Investigation and Analysis of EDM Characteristics of Inconel 825, Mater. Manuf. Proc., 2014, 29(5), 540-549.10.1080/10426914.2014.901536Search in Google Scholar

[26] Shen Y., Liu Y., Dong H., Zhang K., Lv L., Zhang X., Wu X., Zheng C., Ji R., Surface integrity of Inconel 718 in high-speed electrical discharge machining milling using air dielectric, Int. J. Adv. Manuf. Technol., 2017, 90, 691-698.10.1007/s00170-016-9332-7Search in Google Scholar

[27] Kumar P., Gupta M., Kumar V., Surface integrity analysis of WEDMed specimen of Inconel 825 superalloy, Int. J. Data Net. Sci., 2018, 210.5267/j.ijdns.2018.8.001Search in Google Scholar

[28] Ramakrishnan R., Karunamoorthy L., Multi response optimization of wire EDM operations using robust design of experiments, Int. J. Adv. Manuf. Tech., 2006, 29(1-2), 105-112.10.1007/s00170-004-2496-6Search in Google Scholar

[29] Manjaiah M, Narendranath S, Basavarajappa S, Gaitonde VN., Wire electric discharge machining characteristics of titanium nickel shape memory alloy, Trans. Nonferr. Met. Soc. China, 2014, 24, 3201-3209.10.1016/S1003-6326(14)63461-0Search in Google Scholar

[30] Singh B., Misra JP., A critical review of wire electric discharge machining, In: DAAAM International Scientific Book, Chapter 23, 249-266, B. Katalinic (Ed.), Published by DAAAM International, ISBN 978-3-902734-09-9, ISSN 1726-9687, Vienna, Austria, DOI: 10.2507/daaam.scibook.2016.23, 2016.10.2507/daaam.scibook.2016.23Search in Google Scholar

[31] Habib S.S., Study of the parameters in electrical discharge machining through response surface methodology approach, App. Math. Model. 2009, 33, 4397-4407.10.1016/j.apm.2009.03.021Search in Google Scholar

[32] Arikatla S.P., Mannan K.T., Krishnaiah A., Parametric Optimization in Wire Electrical Discharge Machining of Titanium Alloy Using Response Surface Methodology, Mat. Today: Proc. 2017, 4(2A), 1434-1441.Search in Google Scholar

[33] Saini P.K., Verma M., Experimental Investigation of Wire-EDM Process Parameters on MRR of Ti-6al-4v Alloy, Int. J. Innov. Technol. Explor. Eng. 2014, 4, 2278-3075.Search in Google Scholar

[34] Somashekhar K.P., Ramachandran N. Mathew J., Optimization of Material Removal Rate in Micro-EDM Using Artificial Neural Network and Genetic Algorithms, J. Mater. Manuf., Proc. 2010, 25(6), 467-475.10.1080/10426910903365760Search in Google Scholar

[35] Patel P.R., Patel B.B., Patel V.A., Effect of Machining Parameters on Surface Roughness for 6063 Al-Tic (5 & 10 %) Metal Matrix Composite Using RSM, Int. J. Res. Eng. Technol., 2013, 2, 65-71.10.15623/ijret.2013.0201013Search in Google Scholar

[36] Nahak B., Yusufzai M.Z.K., Vashista M., Correlation between surface integrity of EDMed high carbon high chromium die steel with Barkhausen Noise parameters, Int. J. Appl. Eng. Res., 2017, 12, 5709-5714.Search in Google Scholar

[37] Bobbili R., Madhu V., Gogia A.K., Multi response optimization of wire-EDM process parameters of ballistic grade aluminium alloy, Int. J. Eng. Sci. Technol., 2015, 18, 720-726.10.1016/j.jestch.2015.05.004Search in Google Scholar

[38] Singh J., Singh R., Kumar R., Optimization of Micro WEDM process parameters formachining on TI-6AL-4V Alloy, Int. J. Innov. Res. Sci. Technol., 2016, 2(11), 701-706.Search in Google Scholar

[39] Goswami A., Kumar J., Investigation of surface integrity, material removal rate and wire wear ratio for WEDM of Nimonic 80A alloy using GRA and Taguchi method, Int. J. Eng. Sci. Technol., 2014, 17(4), 173-184.10.1016/j.jestch.2014.05.002Search in Google Scholar

[40] Guo Y.,Wang L., Zhang G., Hou P.,Multi-response optimization of the electrical discharge machining of insulating Zirconia, Mater. Manuf. Proc., 2016, 32(3), 294-301.10.1080/10426914.2016.1176180Search in Google Scholar

[41] Holmberg J., Wretland A., Berglund J., Beno T., Surface integrity after post processing of EDM processed Inconel 718 shaft, Int. J. Adv. Manuf. Technol., 2017, 96, 1429-1443.10.1007/s00170-017-1342-6Search in Google Scholar

[42] Myers R.H., Montgomery D.C., Anderson-Cook C.M., 2009, Response Surface Methodology: Process and Product Optimization Using Designed Experiments, John Wiley and Sons, Hoboken.Search in Google Scholar

[43] Pradhan M.K., Biswas C.K., Modeling and Analysis of process parameters on Surface Roughness in EDM of AISI D2 tool Steel by RSM Approach, Int. J. Mech. Mechatron. Eng. 2009, 3, 1132-1137.Search in Google Scholar

[44] Chalisgaonkar R., Kumar J., Multi-response optimization and modeling of trim cut WEDM operation of commercially pure titanium (CPTi) considering multiple user’s preferences, Int. J. Eng. Sci. Technol., 2015, 18(2), 125-134.10.1016/j.jestch.2014.10.006Search in Google Scholar

[45] Khosrozadeh B., Shabgard M., Effects of hybrid electrical discharge machining processes on surface integrity and residual stresses of Ti-6Al-4V titanium alloy, Int. J. Adv. Manuf. Technol., 2017, 93, 1999-2011.10.1007/s00170-017-0601-xSearch in Google Scholar

Received: 2019-02-05
Accepted: 2019-07-10
Published Online: 2019-10-01

© 2019 P. Kumar et al., published by De Gruyter

This work is licensed under the Creative Commons Attribution 4.0 Public License.