Experimental investigation of effective parameters on productivity improvement of the EDM process for corrosion-resistant metals

Inconel 625 superalloy and stainless steel 304 are known for their significant corrosion resistance along with their high hardness and strength. Therefore, they are used in a wide range of industries, including oil and gas and nuclear. Electrical discharge machining is among the most widely used processes for machining of these metals. However, this process has limitations, such as low material removal efficiency, high surface roughness, and the formation of a recast layer. Therefore, in this study, the effective parameters on increasing the material removal efficiency and reducing recast layer thickness are investigated. These parameters include the dielectric fluid, electrode material, discharge current, and pulse duration. After performing the test matrix, the effect of each of the input parameters on the material removal rate, surface roughness, and thickness of the recast layer is evaluated using the ANOVA method. The results of this analysis showed that the type of dielectric fluid and the presence of silver oxide nanoparticles have a significant effect on output variables. When using sunflower oil fluid containing nanoparticles and the silver electrode, the recast layer and surface roughness are reduced, while the average material removal rate increases by 40% compared to the traditional mode. Also, due to the biodegradability of deionized water and sunflower oil fluids, the environmental sustainability of the process in this study is increased and while increasing productivity, it leads to the sustainable development of the EDM process.


Introduction
Electrical discharge machining (EDM) is one of the most applied processes in production of metal parts with high hardness and abrasion resistance. This process also has special applications in the molding and machining industry of parts with high sensitivity to shear forces and mechanical stresses [1]. Inconel 625 superalloy and stainless steel 304 are known for their high hardness and abrasion resistance as well as their resistance to corrosion at high temperatures and resistance to fatigue and creep. They are, thus, widely used in manufacturing parts needed in various industries, such as oil and gas, automotive, and defense. Therefore, EDM process is one of the best methods for machining of these metals [2].
Functional physics of the EDM process is such that the two metal electrodes, one in the form of a machining geometry and the other in the workpiece, are immersed in the dielectric fluid. By starting the device and sending electrical pulses from the power supply, an electric field is created at the closest distance between the two electrodes, which causes dielectric ionization and the formation of a plasma channel. Through this channel, electrons and ions are then exchanged between the electrode and the workpiece by high-frequency sparks, leading to surface melting of the workpiece and chip away. In this process, there is no contact between the electrode and workpiece, and sparks are formed at a small distance called a gap [3].
Due to the thermoelectric nature of EDM, which is based on sparking and surface melting, a phenomenon called recast layer occurs on the surface of the workpiece. The formation of this layer occurs in the period between the end and the resumption of the sparking current and causes the refreezing of part of the molten material on the worked surface of the workpiece [4]. The formation of this hard and brittle layer reduces the material removal efficiency, increases the surface roughness, and creates defects such as cavities and surface cracks in the workpiece. In general, various parameters such as the pulse duration and the properties of dielectric fluid affect the formation of the recast layer. Therefore, correct adjustment of machining operational parameters as well as the use of scientific innovations plays an important role in reducing the thickness of the recast layer, increasing the functional efficiency, and increasing the quality of the machined work pieces [5,6].
Numerous studies have been performed on investigation, optimization, and prediction of parameters affecting the increase of EDM efficiency and reduction of the recast layer [7,8]. In the work of Dong et al., water in oil nanoemulsion and kerosene were used as the dielectric fluid. Also, the intensity of the current and the pulse duration are operational parameters. The results of this study showed that due to the high thermal and electrical conductivity of ionized water relative to kerosene, the surface crack density and surface roughness rate improve when using this fluid [9]. In the work of Valaki et al., Jatropha Curcas vegetable oil and neem oil have been used as dielectric fluids. The results of this study showed that when using Jatropha Curcas oil and decreasing the pulse offtime, while increasing the material removal rate, the surface hardness of the samples also increases. Also, when using neem oil, the removal rate is 22% and the surface morphology is 17% improvement [10]. Nas and Akıncıoğlu investigated the performance of electrical discharge machining using Taguchi method on a nickel-based cryogenic superalloy. In this research, output variables, such as surface roughness (SR) and material removal rate (MRR), were studied and optimized to achieve the maximum wear amount and minimum surface roughness [11]. Also, Akıncıoğlu studied the electrical discharge machining process on titanium (Gr2) alloy. The effects of functional parameters, such as tool wear rate, material removal rate, and surface roughness, were investigated using the Taguchi method and variance analysis (ANOVA). In this research, quadratic regression analysis was used to show the correlation between experimental results and predicted values. According to the results of this research, the most effective factor for material removal rate, tool wear rate, and surface roughness was amperage at 99.66%, 99.56%, and 81.12%, respectively. The best value for average surface roughness was determined to be 3.29 μm obtained at 120-μs pulse on-time, 8-A discharge current, and 40-μs pulse off-time [12].
In the work of Shabgard and Khosrozadeh, the effect of combining different nanopowders in dielectric fluid on the performance of EDM process as well as output parameters such as material removal rate, surface roughness, and integrity have been studied [13,14]. In one of the studies performed on Inconel 718 superalloy, the effect of machining parameters was investigated by adding copper nanoparticles to kerosene. The results of this study show that with increasing discharge current, the material removal rate increases, and the surface finish decreases, which of course is accompanied by a decrease in surface crack density. Also, with increasing the pulse on-time, the material removal rate decreases and the surface finish increases, which, of course, is accompanied by an increase in the surface crack density [15]. In the work of Kumar et al. by adding alumina nanopowder to dielectric fluid, the machining performance has been investigated. The results showed that the nanopowder-mixed dielectric medium gives better surface finish and higher metal removal rate as compared to traditional dielectric. Also, atomic force microscopy (AFM) and field emission scanning electron microscopy (FESEM) investigation of the machined surface reveal that the presence of microcrack, micro-hole, and tensile residual stress decrease substantially during the process [16]. In another study, Hourmand et al. analyzed the electrical discharge machining process on Al-Mg2Si metal matrix composite. Output variables include material removal rate and microstructure texture. The results of this research revealed that among all interactions, the discharge current-voltage and discharge current-pulse on-time interactions have the most significant effects on MRR. An analysis of the Al-Mg2Si microstructure demonstrated that discharge current, pulse on-time, and voltage have significant impacts on the microstructure, size of craters, and profile of the machined surface. Moreover, a decrease in spark energy leads to less microstructural change and better surface finish [17].
In all the mentioned research works: (1) Often, the performance of machining has been evaluated by examining a new fluid, which has caused insufficient information about the simultaneous effect of other fluids on the performance of the process.
(2) In the field of dielectric additive nanoparticles, aluminum oxide, graphite oxide, and copper oxide nanoparticles are often used, which have almost the same function. (3) In most studies on the qualitative parameters of surface health, only the surface roughness has been investigated, and the thickness of the recast layer has not been considered. (4) Most EDM machining research works have been done using copper or aluminum electrodes.
Therefore, in the present study, in order to increase machining efficiency and reduce the destructive environmental effects of kerosene fossil fluid, the effects of the most important parameters affecting the output variables, namely, material removal efficiency, surface roughness, and thickness of recast layer, have been investigated. Finally, new input variables are introduced which, while increasing machining efficiency and quality of manufactured parts, reduce waste and increase environmental sustainability and facilitate sustainable development of the EDM process.

Workpiece materials and equipment
The workpieces used in this research are in the form of sheets with a thickness of 6 mm and include Inconel 625 and stainless steel 304. Chemical compositions of the two materials are shown in Table 1.
The dielectric fluid, as the interface between the tool and the workpiece, is responsible for facilitating the formation of the plasma channel as well as washing the arc site. Fluids used in this study include kerosene, deionized water, and sunflower oil. From each fluid, two simple species and   current  5  10  20  2  Pulse duration  100  200  450  3  Dielectric fluid  Kerosene  Sun oil  450  4 Electrode material Copper Silver 5 Nano particle Exist Absent 6 Workpiece material Inconel 625 Ss 304   combined with additive nanoparticles in the amount of 1.5 L were prepared (Fig. 1). Silver oxide nanoparticles with dimensions of 50 nm were added at a rate of 5 g per liter of fluid. A suction pump and a hand-made metal tub are used to inject and transfer each fluid to the location of the arc pool. The electrode, as the interface between the power supply and the workpiece, is responsible for transmitting electrical and thermal current as well as establishing the sparking current. In this research, two materials, namely, copper and silver, are used as electrodes. Silver, as the material with the highest coefficient of electrical and thermal conductivity, has the best properties as the electrode material. The use of this material is one of the innovation indicators of the present study. The copper electrode is a stepped rebar with a diameter of 12 mm and a length of 60 mm. In the case of silver electrode and due to the high cost of this material, a rebar with a diameter of 12 and a length of 25 mm is used. Also, for easier installation of this electrode in the tool holder, a sequence is considered for it (Fig. 2).

95% CI for the Mean
The pooled standard deviation is used to calculate the intervals.

95% CI for the Mean
The pooled standard deviation is used to calculate the intervals.

Experimental procedure
In this study, to investigate the effect of six input variables, namely, discharge current, pulse duration, workpiece material, electrode material, additive nanoparticles, and dielectric fluid type, on changes in three output variables, namely, material removal rate, surface roughness, and reset layer thickness, Taguchi method and an L36 test matrix are used. The input variables and the values of each of them are listed in Table 2.

Results and discussions
After performing the experiments, the values of each output parameter are measured and recorded in Table 3. To ensure the validity of the experimental data as well as to check the repeatability of the experiments, each row of the test table is repeated twice, and the average value is considered. The purpose of these experiments is to investigate the effect of input variables on the material removal rate, surface roughness, and thickness of the recast layer. Thus, the ANOVA method is used to evaluate the effects of the input variables on the experimental outputs.

Material removal rate
The material removal rate is one of the most important output variables of the EDM process and indicates the amount of material harvested per unit time. Increasing the value of this variable increases the machining efficiency and production speed of parts [18]. Material removal tests based on the Taguchi design table are performed on both Inconel 625 and stainless steel 304, and the results are listed in Table 3, and Fig. 3 shows the experimental setup for material removal when using copper and silver electrodes.
Analysis of variance is performed to evaluate the effects of six input parameters on the material removal rate, and the results are shown in Table 4. In this analysis, the error value is taken as α = 0.5. This means that with a 95% confidence level, parameters with a P-value of less than 0.05 are effective inputs and parameters with a P-value of greater than 0.05 are ineffective inputs. Therefore, the effective parameters on material removal efficiency from the most effective to the least effective are the type of dielectric fluid, addition of nanoparticles, electrode material, discharge current, pulse duration, and workpiece material. Figure 4 shows the effects of each of the six input parameters on the material removal rate. In part a, it is clear that by increasing the discharge current, the amount of heat energy transferred to the workpiece surface increases and thus leads to an increase in the material removal rate. In part b, with increasing pulse duration, first, the material removal efficiency increases and after passing the range of 200 μs due to unstable and non-uniform conditions, this rate decreases. In part c, due to the short breakdown voltage and the short pulse off time of sunflower oil, the durability of the plasma channel and effective sparking is increased, and the highest material removal rate is achieved. On the other hand, due to the higher thermal conductivity of ionized water than kerosene, energy wastage in ionized water is higher, and thus, the material removal efficiency is reduced. This is in line with results reported in the literature [16]. In part d, due to the higher electrical and thermal conductivity of silver as compared to copper, the material removal rate of the silver electrode is higher than that of the copper electrode. In part e, it is evident that when adding silver oxide nanoparticles, due to the reduction of dielectric strength and expansion of gap distance and better gap washing, the material removal rate increases up to 40%. Part f shows that due to the lower melting point and higher thermal conductivity of Inconel 625 when compared to stainless steel 304, the material removal rate of Inconel is higher. This is in line with results reported in the literature [19].

Surface roughness
The surface quality of the machined parts is an important variable in the EDM process. One of the most important indicators for measuring surface quality is surface roughness. Reducing the surface roughness is considered to be a main goal in all machining processes. Increasing the surface roughness destroys the surface morphology of the workpiece and causes surface defects such as cracks and cavities. This phenomenon can reduce the mechanical properties and lead to the premature failure of the workpiece [14]. In this study, after complete execution of the material removal operation, surface roughness of the samples is measured based on Ra criterion using a Taylor Hobson surface roughness tester (Surtronic model) with an accuracy of 0.01 μm. The results are listed in Table 3. The results of analysis of variance of the variables affecting the surface roughness are listed in Table 5. The parameters affecting the surface roughness from the most effective to the least effective are type of dielectric fluid, additive nanoparticles, discharge current, electrode material, pulse duration, and workpiece material. Figure 5 shows the effects of each of the six input parameters on the surface roughness. In part a, it is clear that with increasing discharge current, the amount of energy and heat stress transferred to the workpiece surface increases and results in an increase in surface roughness. In part b, with increasing pulse on-time and due to decreasing pulse offtime, surface roughness increases at first, but then decreases due to unstable and non-uniform conditions. In part c, due to the high thermal conductivity and the expansion of the radius of plasma channel when using ionized water, the density of surface cracks and the amount of surface roughness are reduced. In addition, the emission rate of toxic and harmful gases is reduced when using ionized water compared to kerosene, and the environmental compatibility of the process is increased. This is in line with results reported in the literature [20]. In part d, it is proved that by increasing the electrical and thermal conductivity of silver compared to copper, sparks are more evenly distributed on the surface of

95% CI for the Mean
The pooled standard deviation is used to calculate the intervals.

95% CI for the Mean
The pooled standard deviation is used to calculate the intervals.

95% CI for the Mean
The pooled standard deviation is used to calculate the intervals.

95% CI for the Mean
The pooled standard deviation is used to calculate the intervals. Fig. 5 Analysis of the effects of input parameters on machined surface roughness the samples, which reduces the peak values of the surface and improves its morphology. Part e shows that the addition of silver oxide nanoparticles to the dielectric fluid leads to expansion of plasma channel, and the sparking frequency is improved. In this case, the energy of the sparks and the intensity of electric field will have a more uniform distribution. Also, by reducing the fracture strength of the dielectric fluid, electrical discharge begins at larger gap intervals, which leads to better washing of the gap space and reduces the possibility of harmful pulses such as shortcuts, leading to reduced roughness and improved surface morphology. This is in line with results reported in the literature [20]. Part f shows that due to the higher density of Inconel 625 compared to stainless steel 304 and also its lower melting point and higher thermal conductivity, the surface roughness of Inconel 625 is higher than that of stainless steel 304.

Recast layer
Due to the thermoelectric nature of the material removal mechanism in the EDM process, a heat-affected area is created on the workpiece surface. The most important part of this area that forms on the surface of the workpiece is called the recast layer. This layer is another indicator of surface health, which consists of the re-freezing of molten material on the surface of the workpiece. When molten material erupts out of the molten pool at the end of the pulse ontime, some of this material re-freezes in the form of bumps and surface appendages on the workpiece. As a result, a hard and brittle layer is formed on the surface of the workpiece, which has microcracks due to the non-uniformity of the forming phases. Therefore, the formation of the recast layer, in addition to reducing machining efficiency, causes surface defects and reduces the mechanical properties of the workpiece. Therefore, in order to increase the operational efficiency of the process and achieve the appropriate surface quality in the parts, the recast layer should be reduced as much as possible [21].
After completing the experimental design table, each sample is cut and prepared using the wire cut process. Then, by performing metallographic operations, including polishing, felting, and etching, the samples are prepared for analysis using a scanning electron microscope (SEM). The SEM tests are performed by a VP 1450 device manufactured by LEO-Germany. According to the results of ANOVA analysis, it can be stated that the parameters affecting the recast layer thickness from the most effective to the least effective are type of dielectric fluid, additive nanoparticles, discharge current, electrode material, pulse duration, and workpiece material. The results of analysis of variance of the variables affecting on the recast layer are listed in Table 6. Figure 6 shows the effects of each of the six input parameters on the recast layer thickness. Examples of SEM images of recast layers obtained in experiments are shown in Fig. 7.
In Fig. 6, part a shows that the thickness of the recast layer increases with increasing discharge current. From the view of the impact mechanism, it can be stated that with the increase of discharge current and thermal energy on the workpiece surface, the volume of material removal increases. This phenomenon causes a larger volume of molten material to be harvested and frozen on the surface of the workpiece. In part b, with increasing pulse duration and due to decreasing off-time of sparks and lack of proper washing of the gap space, first, the thickness of the recast layer increases and then decreases with decreasing material removal efficiency. This is in line with results reported in the literature [21]. In part c, as mentioned in the previous sections, sunflower oil fluid has the highest material removal efficiency and surface roughness, and deionized water has the lowest material removal efficiency and surface roughness. This has caused the deionized water fluid to have the lowest recast layer thickness and the sunflower oil fluid to have the highest recast layer thickness. This is in line with results reported in the literature [21]. In part d, it is evident that the silver electrode has a more uniform mechanism in material removal than copper and naturally has a smaller thickness in the recast layer. Part e proves that the addition of nanoparticles changes the electrical properties of the dielectric fluid and thus reduces the spark delay time. This reduces the off-time of the pulses and thus reduces the opportunity for the molten material to re-freeze on the surface of the workpiece. Part f shows that due to the higher material removal efficiency of Inconel 625 than stainless steel 304 (the reasons for which were mentioned in the previous sections), the thickness of the recast layer in this material is greater than steel. This is in line with results reported in the literature [19].
Another important factor affecting the recast layer is the carbon content of the workpiece material. As reported in the literature [22], higher carbon content leads to an increased thickness of the recast layer during the EDM process. Here,

Conclusions
In this research, in order to increase the material removal efficiency and the quality of machined parts and improve the environmental sustainability of the EDM process, the effects of six input variables, including discharge current, pulse duration, dielectric fluid, electrode material, and additive nanoparticles, on Inconel 625 and stainless steel 304 were investigated. According to the experimental results and statistical analysis, when using sunflower oil fluid with silver oxide nanoparticles and the silver electrode, the material removal rate increases by an average of 40% compared to the traditional case. This improvement is 30% for stainless steel. As a result, these variables can replace the traditional input variables, namely, kerosene

Recast layer
Recast layer

Recast layer
Recast layer

Recast layer
Micro-void dielectric and copper electrode. Analysis of surface roughness on the machined samples showed that ionized water fluid with nanoparticles and silver electrode has the lowest surface roughness. In this case, while achieving acceptable material removal efficiency, the average surface roughness is reduced by up to 100% compared to the traditional case. Investigation of the thickness of the recast layer in produced parts proved that in test conditions with the lowest surface roughness, the thickness of the recast layer is decreased. Therefore, the minimum thickness of the recast layer is obtained when using ionized water fluid with nanoparticles and silver electrode. It should also be noted that when using sunflower oil fluid with nanoparticles and silver electrode, while increasing the material removal efficiency by 40%, the thickness of the recast layer is reduced compared to traditional variables. Analysis of results proved that the EDM process with the new variables introduced in this study, while increasing the efficiency of material removal and surface quality indicators of machined parts, has a high compatibility with environmental indicators that leads to reduced emissions and toxic waste and facilitates sustainable development of the EDM process.
Author contribution All the authors contributed to the study conception, design, material preparation, data collection, and analysis. The first draft of the manuscript was written by Mohammad Reza Saberi, and all the authors commented on previous versions of the manuscript. All the authors read and approved the final manuscript.
Funding The research of the corresponding author is supported by a grant from Ferdowsi University of Mashhad (N.3/51627).

Competing interests
The authors declare no competing interests. Fig. 7 Thickness of recast layer for different input parameters; a dielectric: water, electrode: silver and not nanoparticles; b dielectric: water, electrode: silver and with nanoparticles; c dielectric: kerosene, electrode: silver and not nanoparticles; d dielectric: kerosene, electrode: silver and with nanoparticles; e dielectric: sunflower oil, electrode: silver and not nanoparticles; f dielectric: sunflower oil, electrode: silver and with nanoparticles ◂