Keywords

1 Introduction

Wire electrical discharge machining (WEDM) has widely used for machining of high strength and conductive materials for generating an intricate profile that would be very difficult to produce using any conventional machining. The machining process is used in the automobile, aerospace and tool manufacturing industries where close tolerance and dimensional accuracy are required. WEDM is a non-traditional machining process in which material is being removed from the workpiece by a series of sparks between workpiece and wire separated by a dielectric fluid without any contact. The dielectric fluid in machining not only washes detritus produced between them, but it restricts them to become an obstacle for the next spark. It also acts as a cooling agent for the cutting zone. When a potential difference has established between wire and workpiece, then the emission of electron starts. These electrons collide with the dielectric fluid that breaks it into ions and an electron that forms a plasma. A high amount of heat is generated that melts the material nearby sparking zone and evaporated [1]. The current intensity is a more dominant factor on surface roughness, and also an excellent surface finish has been obtained at a low value of current intensity as well as pulse-on time [2,3,4]. Varun et al. [5] used grey relational analysis (GRA) coupled with genetic algorithm (GA) to study the effect of input parameters on the out responses (MRR) on EN353 alloy steel. Mathew et al. [6] optimized the parameters of WEDM on AISI 304 steel against the MRR and SR, and it was found that MRR and SR increased with an increase in servo voltage up to a specific range.

Durairaj et al. [7] study the analysis of WEDM on stainless steel to optimize the parameters against the performance characteristic such as kerf width and surface finish. Rebolo et al. [8] investigated the effect of EDM process parameters on the material removal rate and surface quality on high strength copper-beryllium alloys. They have developed the mathematical model using response surface methodology for MRR and SR to determine the relationship between process parameters and machining performance. An abrasive particle mixed with dielectric fluid shows better machining rates at the desired surface quality [9, 10]. However, the machining parameters have selected for experimental work on AISI 304 steel, and their effect on the output characteristics was studied.

2 Experimental Method

The experiments have executed on three-axis CNC controlled WEDM (ELECTRONICA ULTRA CUT 51), as shown in Fig. 1. A brass wire having a diameter of 0.25 mm is used as an electrode. De-ionized water has used as the dielectric medium for the experiment. The work material is AISI 304 steel of size 150 × 50 × 5 mm plate, and their constituents are shown in Table 1.

Fig. 1
figure 1

Photograph of WEDM

Full size image
Table 1 Chemical composition of AISI 304
Full size table

The working range of WEDM parameters was selected based on a trial experiment, and their levels are shown in Tables 2, 3 and 4. The machine variables have validated by performing confirmation experiment.

Table 2 WEDM parameters and their levels
Full size table
Table 3 L9 orthogonal array
Full size table
Table 4 Experimental results
Full size table

In the present research work, Taguchi L9 orthogonal array has used for the design of experiment (DOE) and three levels of each process parameters have selected.

The output response has been recorded in the form of MRR and SR. Machining time has been kept constant for each experiment. Surface roughness is measured using an apparatus called as Talysurf.

3 Results and Discussion

After experimentation, MRR and SR were calculated and arranged in a tabular form corresponding to each parameter combination. According to Taguchi technique, S/N ratio is the relation of signal to noise where the signal is desirable value and noise is undesirable value. As MRR value should be high for machining, and surface roughness value should be less for sound application. Since MRR should be maximum, then higher the better characteristic is used for the calculation of S/N ratio. The surface roughness should be minimum for machined parts, thus, lower the better characteristic is used for the calculation of S/N ratio, as shown in Table 5. The experimental results were analysed by ANOVA to find the significance of each process parameter on the material removal rate and surface roughness.

Table 5 S/N ratio value for MRR and SR
Full size table

3.1 Material Removal Rate

The results obtained by Taguchi analysis indicate that MRR increases with an increase in pulse-on time as the duration of sparks increases, the discharge energy increases with Ton. Similarly, it has observed that with an increase in pule-off time, the MRR decreases. There is no current supply to the electrode as time duration between two sparks is increased. Figure 2 shows the plot for the effect of the S/N ratio on MRR. Table 6 shows the S/N ratio of each parameter and its relation with MRR. From the responses, S/N ratio for MRR is calculated for larger the better by Taguchi analysis; the ranks are (1) pulse-on time (Ton), (2) pulse-off time (Toff) and (3) servo voltage (V).

Fig. 2
figure 2

Effect of parameters on MRR

Full size image
Table 6 Response for S/N ratios for MRR
Full size table

3.2 Analysis of Variance

Influence of process parameters on MRR and SR has calculated, and the contribution of each parameter on response has presented in Table 7. From Table 7, ANOVA shows the importance of selected input parameters on MRR. It has concluded that the MRR is influenced by Ton of 56.72%, Toff contributes of 19.60% and servo voltage (V) contributes of 6.89%.

Table 7 ANOVA results for MRR
Full size table

3.3 Surface Roughness (Ra)

The results obtained from Taguchi analysis for surface roughness (Ra) indicate that with an increase in pulse-on time, the surface roughness decreases. Similarly, an increase in pulse-off time, the Ra shows slight variation. Figure 3 shows the plot of S/N ratio effect on surface roughness. Table 8 shows the S/N ratio of each parameter and its relation with surface roughness. S/N ratio for Ra has calculated for smaller the better by Taguchi analysis; the ranks are (1) Ton, (2) Toff and (3) servo voltage (V).

Fig. 3
figure 3

Variation of S/N ratio with selected parameters for Ra

Full size image
Table 8 Response for S/N ratios for Ra
Full size table

The ANOVA constructed from the experimental data for surface roughness (Ra) shows that the percentage contribution of each control factor on Ra. From Table 9, it can conclude that surface roughness (Ra) is profoundly affected by Ton which is contributed by 95.71%, servo voltage contributes of 3.96% and Toff contributes of 0.32% with an error of 0.01%.

Table 9 ANOVA result for surface roughness
Full size table

3.4 Grey Relational Analysis

The most popular optimization techniques used for multiple responses in engineering problems is grey relational analysis. In GRA, multiple quality responses have been converted into a single grade. GRA used as the performance measurement characteristic where multiple parameters against the multiple responses have to be optimized.

Based on grey relational normalized and grey relation coefficient value, the grey relational grade has been calculated corresponding to each response value. From the grey relation grade (GRG), the value which nearer to one corresponding to that GRG process parameters is the optimum parameters for maximum MRR and minimum surface roughness value for the operation on AISI 304, which is depicted in Table 10.

Table 10 Optimized grey relational grade
Full size table

It has concluded that the experiment number 4 has the highest grey relation grade value, i.e., nearer to one, thus, a perfect combination of process parameters for MRR and SR.

4 Conclusions

The experiments were conducted based on the L9 orthogonal array on AISI 304 steel using WEDM. The following conclusion has drawn from the present study:

  1. (a)

    The optimized input parameters for maximum MRR are pulse-on time of 110 µs, pulse-off time of 55 µs and servo voltage of 40 V.

  2. (b)

    From the analysis of variance for MRR, pulse-on time and pulse-off time is the significant factor.

  3. (c)

    From grey relational grade, it has observed that the optimum process parameters combinations are the pulse-on time of 105 µs, pulse-off time of 55 µs and servo voltage of 30 V for better MRR and Ra.