Wedm Process Optimization for Machining Characteristcis of AISI 52100 Grade Alloy Steel

,


Introduction
Wire electric discharge machining (WEDM) is an electrothermal machining process in which material removal takes place by the erosion of metal particles removed from the workpiece by high-intensity electric sparks carried out on the metallic wire [1,2]. In EDM, both the workpiece and the cutting tool (wire electrode) are submerged in the di-electric fuid [3]. Te di-electric fuid is constantly removed and fltered to erode the particles, and a thin electrode wire is fed into the workpiece. Te toolpath motion of the Wire EDM is similar to a CNC machine operated through a computer program for the pre-defned three-dimensional shape [4]. Te wire-cut EDM is popularly used in the fabrication of aerospace, automobile, and tooling components in precision. Te EDM is capable of efciently making complex geometrical parts, hard to machine materials which are quite difcult in the conventional machining processes [5]. Henceforth, this technology is a very promising area of research in the past few decades and has been explored by numerous authors for parameter optimization.
Wire EDM has been successfully employed on diferent materials, viz. metals, ceramics, and composites in the past [6,7]. Diferent materials for the electrodes/wire, viz. brass, copper, and composite wires [8][9][10], have also been surveyed by diferent researchers in the past for achieving better cutting rates and surface quality in wire EDM. Similarly, works are presented on the optimization of process parameters to obtain minimum tool wear and better surface quality. However, the performance of the wire EDM process still needs advancement due to sluggish material cutting rate, the inefciency of producing sharp corners, high cost, and lack of large-scale production capacity associated with the wire-EDM process [11].
Ming et al. [12] presented an assessment of the research trends in EDM techniques by considering process parameters, energy distribution, and discharge energy along with the mechanical indicators like metal removal, tool wear ratio, surface topography, etc., and contrasted EDM with conventional machining processes. Te environmental and economic impacts were also deliberated by the author, along with the applications. Kumar et al. [13] also reviewed the utilization of the EDM technique for surface modifcations and its phenomenon of surface composition after the EDM operation to explore its properties. Hou et al. [14] performed wire EDM on Nickel-titanium shape memory alloy to explore the surface characteristics, including surface damage, shape recovering capability, and hardness characteristics. Te authors reported a decrease in roughness from 2.79 µm to 0.12 µm.
Paulson et al. [15] investigated titanium super-alloy by expending the Taguchi method for attaining the highest material removal along with minimum surface roughness. Te peak current increase directly infuences the above two output parameters, and the same trend has been observed for pulse-of time. Te optimum peak current of 3A was reported from the grey relational analysis, and the pulse-on and pulse-of time were achieved to be 30 µs and 9µs, respectively. Tombul et al. [16] presented a comparative analysis of wire-EDM of steel at diferent heights by using brass wire and coated electrodes of three distinct diameters for comparison. 12 specimens were fabricated and tested for surface roughness of the specimen. It was concluded that the cutting rate is considerably increased using 0.4mm wire at the cost of increased surface roughness. Kadam et al. [17] studied the roundness error in wire EDM for three metals, viz. H13, EN24, and SS316 by varying cutting speeds from 4.80-7.87mm/s with other parameters staying fxed. It was concluded that the roundness error is more evident at lower cutting speed and vice-versa. In another study, Liao et al. [18] tried to attain good surface roughness in wire-EDM by controlling the parameters at the fnal stage of machining. A good surface roughness of 0.22 µm is achieved after optimizing the control parameters of the machine. A study by Shabgard et al. [19] has revealed the parametric efect of input variables on WEDM cut surface recast layer thickness and heated zone of Inconel 617 materials. From the study, it is outlined that the pulse current and duration happen to be the key aspects of the development of RTL and HAZ. Similarly, Kumar et al. [20] analyzed the surface topography of Inconel 825 material in the WEDM workpiece and witnessed a reducing trend of the surface fnish with an increase in the duration of the pulse. Te study also ensured that the formation of surface crack, recast layer, and HAZ has relation to pulse duration and peaks current. Te study was extended by Anish Kumar et al. [21], who scrutinized the infuence of controlling parameters on MRR and surface roughness during wire EDM of CP-Ti G2 biocompatible material using a desirability' function hybrid with a machine learning algorithm. On the same grounds, the wire EDM was successfully employed and tested for optimal machining of Carbon fber reinforced plastic [22], Te employment of the wire-EDM process has been lucratively done on diferent types of ceramics also by researchers in the past. Grigoriev [23] presented a detailed review on the EDM of diferent ceramics, viz. ZrO 2, Al 2 O 3 , SiAlON, Si 3 N 4 , AlN, etc., via comparison. Te authors established diferent surface quality and other mechanical and electrical aspects of the machined ceramics of diferent conductivity. For diferent electrodes and di-electric mediums, maximum depth and operation time was summarized for all diverse range of ceramics. In another study by Lok and Lee [7], advanced ceramics viz. Sialon and Al 2 O 3 -TiC were tested lucratively utilizing EDM. Te diferent machining conditions were scrutinized for comparison of the material removal rate and surface integrity. Banu et al. [24] reported an EDM for nonconductive ceramic viz. Zirconium Oxide (ZrO 2 ) has tungsten as an electrode and discussed the hardness and machining properties under diferent input conditions. Similarly, many other studies have been reported in the literature for diferent conductive and non-conductive ceramics with EDM [25][26][27][28].
Besides the experimental studies, multi-objective parametric optimizations have been carried out for the wire-EDM process by the authors using the Taguchi approach, desirability, grey relational analysis, etc. Kumar and Nishasoms [29] studied the operating parameters in wire EDM of Al (6082) and tungsten carbide composite employing multiobjective optimization using desirability function analysis. Likewise, Singh and Singh [30] utilized the Taguchi approach to achieve maximum cutting rate using cryogenictreated D-3 steel and reported 50% less tool wear for cryogenically treated copper wire as compared to conventional wire at 4A and 8A current. A similar trend has been achieved for the brass electrode. Few more studies are available in the literature stating the optimal process parameters for wire EDM process obtained using Taguchi, desirability and grey relational analysis, etc. [31][32][33][34][35][36][37]. Bayraktar and Kara [38,39] studied the cryogenic treatment condition for cutting tool material and milling operation for the optimization of surface roughness parameters.
From the literature review it may be observed that several parameters of the wire-edm process afects the machining of object. Some of them are as mentioned below in Table 1 with reference details.
1.1. Scope/Novelty for the present study. From the literature review, it has been observed that diferent researchers have worked on the wire-EDM processing of diferent grades of steel and other materials. Te literature study reveals that the input processing parameters of wire-EDM exhibited a large efect on output responses of the process. But hitherto, little has been explored for the wire-EDM machining of AISI 52100 grade of bearing steel by varying 6 diferent processing parameters of wire-EDM at a time. Terefore, the present study investigates the efect of machining parameters of wire-EDM setup for AISI 52100 grade of steel [41]. Te present study takes six factors of Wire-EDM process at a time in consideration. Te six diferent parameters of the machining are used viz: Pulse on Time (T on ), Wire Feed (WF), Spark Gap Set Voltage (SV), Wire Tension (WT), Pulse-of-Time (T of ), and Servo Feed (SF) of three diferent level for optimization of the processing conditions for improving the machining desirability. Te level of parameters and number of factors makes the study better with improved design of experiment for optimization.

Selection of workpiece material and method of machining.
Te study deals with machining of AISI 52100 grade of steel on the high precision machine of Ultracut-S2 5 axis computer numerically controlled (CNC) wire-EDM setup. Tere are several industrial usages of AISI 52100 grade steel viz. anti-friction bearings, punches taps, dies etc [42]. Six different parameters of the machining were used viz: Pulse on Time (T on ), Wire Feed (WF), Spark Gap Set Voltage (SV), Wire Tension (WT), Pulse-of-Time (T of ), and Servo Feed (SF) for optimization of the processing conditions for improving the machining desirability. For optimization, experimentation, and DOE selection, Taguchi l27 OA was used as a standard operating condition. Table 2 shows the Taguchi  l27 OA for diferent input parameters of the machining. Some other machining parameters were also used but as fxed conditions (as given in Table 3) in the study because of two reasons (a) previous literature shows that their efect is nominal of performance and (b) to have standard DOE only 6 parameters with three levels can be used in Taguchi l27 DOE. Figure 1 shows the methodology used for the optimization of the processing condition of wire-EDM. Figure 2 shows the machining setup available in the Lab with realtime conditions of machining.

Selection of Output responses to be measured with workpiece dimensional details.
Copper wire with a zinc coating (CuZn37) has been used for wire cutting/machining of the AISI 52100 metal. Te selection of the wire for cutting and machining has been done based on its properties, such as resistance to short circuits (due to zinc coating on copper surface) and high precision cutting without any breakage of the wire (due to desirable tensile strength of the wire 900-920 N/mm 2 ). Te cubical object of AISI 52100 steel has been chosen for the size of 20 * 20 * 20 mm (weight 62.4g). For  machining the initial setting of 0.035 mm of wire, an ofset gap was maintained. Te surface roughness value as an output characteristic has been measured as an average value from the mean surface roughness using the Surfcom 130 A instrument. Te average surface roughness value (Ra) has been observed by using cut of length of 0.8mm after each experiment. Based on ISO:4287 standard, measurement of surface roughness was carried out at room temperature. Similarly, the second output dimensional deviation was observed by using a digital dimensional measuring instrument (Height master; Make-Mitutoyo) with an accuracy level of 0.0001 mm, and equation number 1 was employed for calculating the dimensional deviation [43][44][45].

Processing conditions and output Analysis technique.
For machining of the AISI 52100, a metal water pressure of 10 Kg/cm 2 was used with other input conditions. Whereas deionized water with a conductivity value of 12 units has been employed for the wire-EDM process. For reliability study of the DOE and experiments, randomly 6 experiments (experiment number 1,5,10,15,20,25) were selected from the DOE and performed again with similar processing conditions. Two output parameters, surface roughness (Ra) and dimensional deviation (DD), which are important for the machined surfaces, have been selected as the output response for the present study. For analysis, part of the outputs obtained from the study MiniTab software 19.1 package tool has been used. Further, the best and the worst samples as per DOE have been selected for morphological analysis using SEM characterization. Te SEM photomicrographic image was used for 3D surface rendering to characterize the rendered image analysis of the samples.

Result and Discussion
3.1. Discussion for surface roughness and dimensional deviation. Te machined workpieces according to the standard methodology using Taguchi L27 OA have been successfully prepared and tested for diferent output characteristics. Te data set given in Table 4 shows the values for the surface roughness and dimensional deviations for the corresponding processing parameters. From   It may be observed that from the analysis of the processing condition for the best and the worst samples obtained by surface roughness testing. Figure 3a shows the trend obtained for the surface roughness analysis, which depicts the best nature for sample no 19 and the worst results for sample no 9. Tis may be because surface roughness value largely depends upon the spark time (pulse on time T on )), spark of time (T of ), and wire tension, as the other parameters was similar for both of the wire-EDM processing condition for sample 9 and 19 (see Table 4). As it may be observed that in processing condition 9 the spark of (T of ) time was 18µs which resulted in a lowering of temperature due to no current and ionization for a large time interval on the surface of the metal. Te pulse on time (T on ) interval was 1 µs which again reduced the surface temperature of the metal, and the high wire tension of the zinc-coated copper wire together resulted in better machining characteristics.
Similar results were obtained for the dimensional deviation characteristics as the lowest dimensional deviation of 0.02 was observed for sample 9, and the maximum dimensional deviation of 0.12 was observed for sample 19. High surface temperature and ionization of medium along with low wire tension resulted in poor surface characteristics of the wire-EDM machined AISI 52100 grade steel.
Te results obtained for the machined surface of AISI 52100 grade steel shows that processing condition 9 generates moderate surface temperature condition (due to low pulse on time (T on ): 1µs, and large pulse of time (T of ): 18 µs), and high wire tension of 1650 grams together contributed towards the best machining conditions for the wire-EDM process. Figure 3b shows the trends obtained for dimensional deviation of the wire-EDM processed AISI 52100 grade steel. Te web chart depicts that sample 9 held a low value for dimensional deviation as the wire tension was highest at 1650, and surface temperature conditions were moderate due to low pulse on and high pulse of timings.      Table 4 was further processed with Taguchi l27 OA using MiniTab 19.1 statistical tool for optimization of the processing parameters. Table 5 shows the analysis of variance (ANOVA) for the signal-tonoise (SN) ratios obtained by Taguchi analysis using surface roughness value data from Table 4. From Table 5, it has been ascertained that the factors pulse on time (Ton), pulse of time (Tof), and spark gap voltage (V) are the signifcant parameters as their p-value is less than 0.05. Whereas wire tension was also found to be a signifcant contributor in deciding the surface characteristics of the machined workpiece of AISI 52100 grade steel. Table 5 shows the rank Table for the SN ratio of the surface roughness values, and it may be ascertained that pulse on time (Ton) was ranked at 1 st , spark gap voltage came to be as 2 nd rank parameter, and pulse of (Tof) time was ranked 3 rd, and wire tension was ranked at 4 th position which signifcantly contributed towards the wire-EDM machining of AISI 52100 grade steel. Figure 4 shows the main efect plot for the SN ratios from which it can be observed that for the surface roughness low pulse on the value of 1 µs, high pulse of vale of 18 µs, wire feed of 8 mm/min, wire tension of 1250 grams, spark gap voltage of 30 V and servo feed of 6mm/min was the optimized condition for the surface roughness of machined workpiece. Figure 5 shows the four in one plot for SN ratios of surface roughness values, and it has been observed that the data obtained for the surface roughness value was normal as all the data lies near to the normal trend line as shown by the normal probability trend of Figure 5. Te Anova analysis of the selected Taguchi L25 OA has been performed as per the standard technique as performed by previous researchers [45,46].

Te calculation for the optimized/suggested set of processing parameters.
Further, for the calculation of the optimized value for the suggested setting by Taguchi l27 OA, the signal-to-noise ratios obtained by the analysis were used. Table 7 shows the SNRA values for the surface roughness and dimensional deviation as output response from Taguchi l27 OA analysis using MiniTab 19.1 statistical software package tool. Further, for the calculation of optimized values, (2) has been used. Te calculation for the surface roughness value has been shown below. From the calculation, it has been observed that the value for the optimized setting was obtained to be 1.080 Ra which was very near to the value obtained for sample number 9 (surface roughness value as per experiment: 1.02Ra (see Table 4). Tus, the optimized setting as suggested by Taguchi l27 OA, which was outside the selected DOE, can be taken as experimental condition 9 in Table 4. A similar calculation can be made for the optimized value of dimensional deviation output characteristics. (2) Y opt is the optimized value of peak elongation k � 4.868 (See Table 7) k A1 � 1.891 (See Table 6) k B3 � 4.103 (See Table 6) k c3 � 4.861 (See Table 6) k D1 � 4.528 (See Table 6) k E3 � 3.527 (See Table 6) k F2 � 4.761 (See Table 6) For smaller is the better case Whereas, the high pulse on time: 1.2 µs, low pulse of time of 11.5 µs, and low wire tension of 1450 grams corresponding to high-temperature conditions of machining resulted in poor surface characteristics due to large fuctuation in machined surface straightness. Figure 6a shows the SEM image for sample 9, and 6b shows the SEM image for sample 19 [46]. Further, the SEM photomicrographic images were processed using the open source software package tool

Conclusions
Te current work dealt with the machining of AISI 52100 grade steel on wire-EDM setup and investigated the efect of diferent input processing conditions of machining for geometrical and morphological characteristics. Te study involved 6 input processing parameters of wire-EDM viz: Pulse on Time (T on ), Wire Feed (WF), Spark Gap Set Voltage (SV), Wire Tension (WT), Pulse-of-Time (T of ), and Servo Feed (SF). Te study focused on the efect of these processing parameters on geometrical output (dimensional deviation) and morphological output (surface roughness). A standard set of DOE based on Taguchi l27 OA has been used for the optimization of the diferent input processing parameters, and the following are the conclusions [47] of the study. 3. Roughness value 1.02Ra and dimensional deviation: 0.02 can be concluded as the best values for the selected Taguchi L25 DOE model. Similarly, the surface roughness 2.98R a and dimensional deviation of 0.12 can be concluded as worst values for output characteristics. 4. According to Taguchi l27 OA analysis, the pulse on time (T on ) was ranked at 1 st , spark gap voltage came to be as 2 nd rank parameter and pulse of (T of ) time was ranked 3 rd, and wire tension was ranked at 4 th position which signifcantly contributed towards the wire-EDM machining of AISI 52100 grade steel. Whereas the other input parameters contributed least towards the output responses. 5. Te SEM and 3D rendered surface roughness analysis supported the observed trend and validated the reasons given for the observed behavior of surface roughness and dimensional deviation. Te surface roughness profle for sample 9 has shown a low surface roughness value of 29.85 nm. Similarly, sample 19 has shown a high surface roughness of 77.25 nm.

Data Availability
Te data used to support the fndings of this study are included within the article.

Disclosure
Tis study was performed as a part of the employment of the authors.

Conflicts of Interest
Te authors declare that there are no conficts of interest.