Effect of process parameters on the electrical discharge machining of aluminum metal matrix composites through a response surface methodology approach

1 Introduction

Electrical discharge machining (EDM), an unconventional machining process, is used to machine any complex shapes in any conductive workpieces irrespective of their hardness. In the EDM process, the material is removed from the workpiece by means of repetitive sparks struck between the tool and the workpiece material [1–3]. The EDM technology is increasingly being used in the tool, die, and mold-making industries for machining heat-treated tool steels and advanced materials such as super alloys, ceramics, and hard composite materials [4]. Among the composite materials, aluminum (Al)-based metal matrix composites (MMC) are finding very wide applications [5, 6]. In the present scenario of the research, optimization plays a vital role since it gives suitable parameters and yields optimal performance. The mathematical model is used in optimization and it plays a crucial role for the selection of machining parameters. For the development of Al MMC, many researchers [7, 8] have carried out quite a large number of experimental works, which include techniques to improve the material removal rate (MRR), to reduce the electrode wear rate (EWR), to improve the surface roughness (SR), and to reduce the thickness of the recast layer formed. On the contrary, some researchers worked in optimizing the process parameters such as current, pulse on-time, pulse off-time, intensity, and flushing pressure for an optimal output [8, 9].

Among the Al MMCs, Al-silicon carbide (SiC) composites are being widely used in industrial applications [8–11]. Few researchers have determined the mathematical models for the MRR, EWR, and SR with the volume fraction of SiC [12–14]. Teimouri and Baseri [15] worked with a SPK cold work steel. In this work, the effects of rotary magnetic field and also the ultrasonic vibration of workpiece were studied on dry EDM process performance. The results showed a positive effect on the MRR and SR. By the application of ultrasonic vibration, the maximum MRR is possible; also, they developed a mathematical model to correlate the relation between the EDM input and output parameters. Bai et al. [16] worked on a powder mixed EDM process to improve the MRR. They reported that the MRR increased initially then decreased with the powder concentration. The empirical formula for the MRR was established based on the experimental results. In this work, the AISIH13 material was used for investigation. They reported that increasing current, voltage, and duty cycle values increased the MRR, but increasing pulse on-time value initially increased the MRR until a maximum value and then started to decrease. Hosseini Kalajahi et al. [17] formulated the thermal model and finite element simulation of EDM considering several factors such as the temperature-dependent material properties, shape and size of the heated zone, energy distribution factor, plasma flushing efficiency, and phase change to predict the thermal behavior and material removal mechanism.

In these models, the common parameters such as current, pulse on-time, intensity, and duty cycle were discussed. However, the size of the SiC is believed to have its influence on the responses. In this work, the mathematical models for the MRR, EWR, SR and cylindricity (CY) were developed, which are the function of current (A), pulse on-time (P_on), flushing pressure (P_r), and particle size (P_s). The mathematical models for the MRR, EWR, SR, and CY were used to find the effect of process parameters on the performance measures. The main objective was to optimize the process parameters using the desirability approach through response surface methodology (RSM).

2 Materials and methods

In this study, the experiments were planned with four factors and three levels. The preferred levels and the factors selected for the machining parameter are shown in Table 1.

The Sparkonix EDM machine shown in Figure 1 is used to perform the machining operations. The flushing of the dielectric fluid was carried out by a jet flushing system in which the dielectric fluid was pumped through the nozzle with the aid of a circulation system. Kerosene was used as the dielectric fluid. The Al 6061-15% SiC composite material was taken as a workpiece. The composite samples with different particle size (i.e., 15, 25, and 40 μm) were prepared by a stir casting process. A fine-machined copper rod of ∅12 mm was used as the electrode.

Figure 1

Electrical discharge machine setup.

Many factors are influencing the EDM process. In this experimental work, the factors such as peak current, pulse on-time, flushing pressure, and particle size were taken into account as the design factor. Twenty-seven experiments were conducted on Al 6061-SiC with three different particle sizes. Two trials of machining were performed for each set of parameters, and blind holes of ∅12 mm diameter and 5 mm depth were made. The machining time of each trial was measured. The tool and the workpiece were weighed before and after machining. Based on the wear weight, the MRR and EWR are calculated using Equations (1) and (2). Figures 2 and 3 show the SR tester electronic weight balance machine used in this work.

Figure 2

SR tester.

Figure 3

Electronic weight balance.

Using the above equations, the relationship between the process variables and the performance measures are established using the Design Expert 7.0.0 software. The MRR, EWR, SR, and CY are the performance measures, which are important in the aspect of increasing the productivity and improving the quality. The Mitutoyo Shimadzu electronic weight balance model BL 220H (L) of capacity 220 g and with the readability of 0.001 g was used to weigh the workpiece and electrodes. A Mitutoyo SR tester was used to measure the SR of the workpieces measured on the sides of the machined hole for three places and the average value is observed. The CY value was measured using the coordinate measuring machine (CMM) Checkmaster with the least count of 0.001 μm as shown in Figure 4. The CY is an actual volumetric assessment measured through the sides of the machined hole on three different regions. The specifications of EDM and CMM are shown in Table 2. The design matrix and measured outputs are shown in Table 3.

Figure 4

Coordinate measuring machine.

3 Response surface methodology

From the set of process parameters and the output responses measured, the mathematical model for the MRR, EWR, SR, and CY was developed and analyzed using RSM. The RSM is a collection of the mathematical and statistical technique that is useful for the modeling and analysis of the problem in which a response of interest is influenced by several variables and the objective is to optimize the parameters. It is an empirical modeling approach for determining the relationship between the process parameters and performance measures with various desired criteria and searching the significance of the parameters on the performance measures. RSM is utilized to accurately describe and identify the influence of the interaction of different independent variables on the response when they are varied simultaneously [18, 19]. RSM is also an empirical technique devoted to the evolution of the relationship between the groups of controlled experimental factors and the observed results of one or more selected criteria.

The effect of process parameters of the machining process on the response variables in the MRR, EWR, and SR is summarized and the empirical relations are formulated. The Design Expert 7.0.0 software has constructed a nonlinear regression expression using the experimental data. The nonlinear estimation is a general fitting procedure that will estimate the kind of relationship between a dependent (response) variable and a list of independent variables. The regression model is expressed as

Here, Y is the desired response and f is the response function. In the procedure of the analysis, the approximation of Y was proposed using the fitted second-order polynomial regression model, which is called as the quadratic model. The quadratic model was studied for the interactive effects of combinative factors on the performance evaluations [11]. The quadratic model of Y can be written as follows:

where β₀ is constant and β₁, β₂, and β_ij are the coefficients of the linear, quadratic, and cross-product terms, respectively. x_i is a coded variable corresponding to the machining parameters.

The optimization for the multiple response work was carried out through the desirability approach in the Design Expert 7.0.0 software, which is a mathematical method to find the optimum. The desirability of each performance measure is calculated as follows:

The desirability converts the performance measures into a dimensionless value between 0 and 1, where 0 means the performance measures are unacceptable and 1 means it is acceptable.

If any of the parameters or performance measure falls outside their desirability range, the overall function becomes zero. For simultaneous optimization, each parameter must have a low and high value assigned to each goal. The parameters should be in the range of lower and upper limits. The performance measure will always be included in the optimization, at their design range by default, or as a maximum, minimum of target goal. The meanings of the goal parameters are as follows: If the goal is to be maximum, d_i=0 if response<low value, d_i=1 if response>high value, and 0≤d_i≤1 as response varies from low to high; if the goal is to be minimum, d_i=1 if response<low value, d_i=0 if response>high value, and 1≥d_i≥0 as response varies from low to high; if the goal is target, d_i=0 if response<low value, d_i=0 if response>high value, 0≤d_i≤1 as response varies from low to target, and 1≥d_i≥0 as response varies from target to high; the range is d_i=0 if response<low value, d_i=0 if response>high value, and d_i=1 as response varies from low to high, where D is the objective function, d is the desired range, n is the number of response, and i is the machining parameters.

4 Proposed mathematical models

The mathematical models for the responses viz. MRR, EWR, and SR are discussed. The models proved that the effect of particle size has a significance on the responses compared with the other responses.

5 Effect of process parameters

The effect of parameters such as peak current, pulse on-time, flushing pressure, and different particle size on the output responses (i.e., MRR, EWR, SR, and CY) is discussed using graphs generated by the Design Expert 7.0.0 software for the experimental data.

5.1 Material removal rate

Figure 5 shows the interactive effect of current and particle size on the MRR. When the particle size is low, the MRR increases as the current is increased. This is also true for the larger value of particle size, but the rate of increase is less compared to that for the lower value of particle size. This can be visualized by this slope of the curve for different particle sizes. This may be because the binding area between the matrix and this requires a large energy for the removal. This may occur due to the large binding area between the matrix and reinforcement of large particle size.

Figure 5

Interactive effect of current and particle size on MRR.

From Figure 6, it is clear that the effect of flushing pressure on the MRR is the same for both lower particle size and larger particle size. It is clear that the MRR increases with the increase in flushing pressure. It is also noted that, for any flushing pressure, the effect of flushing pressure on the MRR is the same. It is observed that the MRR decreases with an increase in particle size. The larger particle size leads to a larger binding area and also a larger energy is required for the removal.

Figure 6

Interactive effect of flushing pressure and particle size on MRR.

5.2 Electrode wear rate

Figure 7 shows the interactive effect of current and particle size on the EWR. It is observed that the EWR increases with both current and particle size. At a lower particle size, the EWR is less compared to that in a larger particle size. Thus, the larger particle size is increased as an erosion effect. When both particle size and flushing pressure become larger, the EWR increases largely as shown in Figure 8. SiC particles are removed from the workpiece, which flows in the dielectric fluid around the soft electrode and erode the electrode. This is the reason for the larger EWR in the process.

Figure 7

Interactive effect of current and particle size on EWR.

Figure 8

Interactive effect of flushing pressure and particle size on EWR.

5.3 Surface roughness

It is obvious that the increase in the particle size and peak current leads to the increase in the SR value. This interactive effect is shown in Figure 9. From Figure 10, it is clear that the effect of flushing pressure has less significant on the SR. When the flushing pressure increases, there is slight decrease in the SR. The effect of particle size has larger impact on the SR. When the particle size increases, the SR is decreases largely.

Figure 9

Interactive effect of current and particle size on SR.

Figure 10

Interactive effect of flushing pressure and particle size on SR.

5.4 Cylindricity

Figure 11 shows the interactive effect of current and particle size on the CY. It is observed that the CY increases in a larger manner for increasing current and particle size. At a lower particle size, the CY is less compared to that in larger particle size. The larger particle size required a maximum erosion effect, which leads to an increase in the CY. The larger particle size and flushing pressure increased the CY as shown in Figure 12. The larger particles are removed from the workpiece, which is carried by the dielectric flow, and it affects the machined hole. This is the reason for the larger CY in the process.

Figure 11

Interactive effect of current and particle size on CY.

Figure 12

Interactive effect of flushing pressure and particle size on CY.

6 Confirmation experiment

After the selection of the optimal design parameters, the next step is to predict and verify the improvements of the quality characteristics for the machining of Al-SiC MMC by the EDM. The predicted optimum values are obtained from the Design Expert 7.0.0 software with a desirability value of 0.513.

Based on the experimental results for the maximum MRR, minimum EWR, minimum SR, and minimum CY, the Design Expert 7.0.0 software provides the predicted machining parameters for the optimum performance measures. A confirmation experiment was carried out through the predicted parameters and the performance measures are observed in Table 5.

7 Conclusions

In this work, EDM was conducted on a Al-SiC MMC with different SiC particle sizes using a copper electrode of ∅12 mm. The experiment was designed using the Design Expert 7.0.0 software. Using the experimental data, the mathematical models for the MRR, EWR, SR, and CY were developed. The peak current is found to be the most significant parameters in the models for the MRR and EWR. The optimization was carried out by the desirability approach. The 3D plots were generated for analyzing the effect of particle size using the Design Expert 7.0.0 software. Based on the analysis, the following conclusions were derived.

When the particle size is high, the MRR increases with current. It is due to the larger binding area between the matrix and reinforcement particles. The MRR is less for the MMC having larger particles. The MRR is slightly decreased when both particle size and current increased. For any particle size, the effect of flushing pressure is same. That is, when the flushing pressure increases, the MRR also increases. However, the MRR is not maximum at the larger level of particle size and flushing pressure.

The EWR largely increases with the increase in current. Meanwhile, it is slightly increased when the size of the particle is increased. The EWR is largely affected with the increasing current and particle size. The EWR is significantly increased with the increase in the particle size. For higher particle size and flushing pressure values, the EWR is maximum. It is due to the low hardness of the electrode. It is noted that the EWR is minimum at a lower particle size compared to that in a larger particle size.

The effect of particle size has a larger impact on SR. When the particle size increases, the SR decreases largely.

It is observed that the CY increased in a larger manner with the increased current and particle size.