Prediction of material removal rate and surface roughness for wire electrical discharge machining of nickel using response surface methodology

Thangam Chinnadurai, Subbiah Arungalai Vendan^*

VIT University, Division, School of Electrical Engineering (SELECT), Industrial Automation and Instrumentation, Near Katpadi Road, Vellore, Tamil Nadu 632014, India

*Corresponding author: arungalaisv@yahoo.co.in

1. INTRODUCTIONTOP

WEDM (Wire Electrical Discharge Machining) is categorized as a sophisticated and advanced process in tooling and manufacturing industry. This process is however confined to machining materials or alloy that is electrically conductive regardless of the hardness. In WEDM, the rapid and repetitive spark discharges across the gap between the tool and the work piece facilitates the material removal. In WEDM process, spark discharge occurs between a small diameter wire electrode (usually less than 0.5 mm diameter) and a work piece with de-ionized water as the dielectric medium. This subsequently enforces the erosion of the work piece to produce complex two and three dimensional shapes according to a numerically controlled (NC) path. Considering the beneficial aspects of the process, the components of defense related products such as cartridge, magazines appear to be effectively machined with the employment of WEDM for those materials. WEDM is expected to be a feasible process her nickel. Owing to the inadequacy in the literature on reports pertaining to WEDM for Nickel using RSM, this study attempts to address this application oriented concern. Some of the prominent literatures directly contributing to this research study are presented briefly.

Khan et al. (2013) investigated the suitability of wire electric discharge machining of high strength and low alloy (HSLA) steels. HSLA steel is considered in their study as it provides better mechanical properties and greater resistance to atmospheric corrosion. The optimum machining parameter range was obtained by using the analysis of signal-to-noise (S/N) ratio, analysis of means and analysis of variance (ANOVA). Further, the level of importance of the machining parameters on micro-hardness was identified using ANOVA. Esme et al. (2009) provided the evidence on the fact that the surface finish of the machined surface mainly depends on the pulse duration, open voltage, wire speed and dielectric flushing pressure. The mathematical relation between the work piece, surface roughness and WEDM cutting parameters were established by regression analysis method. This mathematical model may be used in estimating the surface roughness without performing any experiments. Amitesh and Jatinder (2012), presented information on the influence of machining parameters on cutting speed and material removal rate for machining of Nimonic 80A with brass wire as tool electrode during wire electrical discharge machining process. Investigation indicated that cutting speed (CS) and material removal rate (MRR), both increases with increase in pulse-on-time (Ton) and peak current(IP), while decreases with increase in pulse-off-time (Toff) and spark gap set voltage (SV). Sreenivasa Rao and Venkaiah (2013) review the effects of various WEDM process parameters such as pulse on time (Ton), pulse off time (Toff), servo voltage (SV), peak current (IP), dielectric flow rate (DFR), wire speed (WS), wire tension (WT) on different process response parameters such as material removal rate (MRR), surface roughness (Ra), wire wear ratio (WWR) and surface integrity factors. Besides various optimization methods adapted by the research in various literature repeat and constructive suggestions to overcome few intricacies posed by the process are presented, while parallelly extrapolating few ideas for feature trends. Hassan et al. (2009) provided an insight through experimental investigation on the parameters such as high tensile residual stresses, high surface roughness, presence of micro-cracks and micro-voids that proven to be a detrimental factors. These properties vary with different levels of the main machining parameters. This paper presents an experimental work that has been done in order to quantify the effect of some of the main WEDM parameters on the surface texture of AISI 4140 steel. 2D surface measurements are taken on all WEDM samples and 2D surface characterization is being carried out in order to calculate the different surface texture parameters.

1.1. Response Surface Methodology (RSM)TOP

The RSM is an empirical modeling approach is adopted her determining the relationship between various process parameters and responses with the various preferred criteria and penetrating the significance of these process parameters on the coupled responses (Myers and Montgomery, 1995). RSM is used to get just around the corner of the response around an existing working point, which depends on few main effects only. Design of experiments that allow estimating interaction and quadratic effects, which result in information about the shape of the response surface. It is a sequential testing approach for building and optimizing the experimental model. Therefore, RSM is a collection of mathematical and statistical procedures that are useful for the modeling and analysis of problems in which response of demand is affected by several variables and the objective is to optimize this response (Ozcelik and Erzurmlu, 2005). The second order mathematical models have been developed to predict the better MRR and SR condition values. The polynomial equation are considered in the present case is (Myers et al., 2009).

Where, y is the response; where weld strength; x_i represents pressure, weld time and amplitude; ε₀, ε_i, ε_ij and ε_ii represent the constant, linear, quadratic and interaction terms respectively.

Then response surface y contains the linear terms, squared terms and cross product terms. Using this quadratic model of y in this study is not only to investigate over the entire factor space, but also to locate region of desired target where the response approaches its optimum or near optimal value.

2. EXPERIMENTATIONTOP

Nickel being a stipulate material for military and aerospace systems with electroplating & corrosive resistances is used for this study. The chemical composition of the material (expressed at wt.%) is C: 0.02, Si: 0.3, Mn: 0.3, S: 0.01, Ni: min 99.00, Cu: 4.60, and Fe: 0.4, and the physical properties are melting point 1455 °C, boiling point 2730 °C, density 8.908 gr.m^–3, heat of fusion 17.48 KJ.mol⁻¹, heat of vaporization 379 KJ.mol⁻¹. The materials dimension 5×5×24 mm respectively.

The selection of optimum cutting parameters plays a significant role for producing high material removal rate (MRR) and greater surface roughness. Parameter selection also gets altered during cutting time, wire breakage; wire and work-piece short-circuit. In order to accommodate all these parameter while parallelly limiting the number of experimental trails, DOE is being adopted for the experimental study. The parameter levels are presented in Table 1.

The work is carried out with “Specification of CNC Ultracut Machines”, 5 axial cnc, table size 650×440 mm, best surfaces finish 0.4 Ra, maximum cutting speed 210 mm.min⁻¹. The two different wires used one for brass wire and zinc coated brass wire. In machining bras is typically preferred owing to its low cost, great conductive, high tensile strength and flush ability. Zinc coated wire is expected to posses higher conductivity, which will eventually improve the machining speed. For pure brass wire tensile strength ranging from 70,000 to 130,000 psi and a diameter of 0.025 is used. Zinc Coated Brass wires produce exceptional surface finishes when cutting tungsten carbide and are often utilized for cutting PCD and graphite (Kern, 2007).

Sparks are initiated while machining removes small amount of materials depending on the pulse on time and generated spark sizes. During pulse on time, the sparks sizes are small and the material removal rate is less which exhibits improved surfaces roughness. Bigger spark yield higher MRR and exhibit more roughness. Pulse on time produces a high temperature at the range of 1240 °C–1450 °C. During pulse off time, the dielectric fluid passes through high pressure prevailing between the wires and work pieces that removes the debris and readies the fluid to the next ionization process.

Brass wires were opted for the WEDM due to its high tensile properties. Brass wires are blend of copper and zinc. They alloying is in varying rate of 35–37% zinc and 63–65% copper. The rigidity and rating of vapor pressure of brass wire increases in addition of zinc. Besides, the rate of zinc contains in brass wire decreases its dissolving point. The cutting execution and velocity enhance in machining with brass wire because of stable release amid machining. Amidst the procedure for cutting, the zinc in the brass wire really boils off, or vaporizes, which cools the wire and conveys more usable vitality to the work zone. Expanding the zinc content in brass wire can improve the machining speed (more than 40%) nevertheless framing a wire with rate of zinc that is a weak stage in the compound will lead to intricacies. Properties of wire material are shown in Table 2.

The zinc coated wire was delivered to determine the past wire’s defects and to increment generously the cutting speed and cutting accuracy. Additionally, it was said that creating the wire with 40% zinc substance is troublesome in light of the fact that wire’s grain structure changes and it makes fragile. There are a limited number of coating materials accessible for wires in WEDM. Zinc combination is the most widely recognized material that is connected as the wire coated layer. The zinc combination was chosen because it has a low vaporization temperature compared with the core material. The wire covering is effortlessly vaporizes when a pulse is applied and core is secured because of cooling impact of coated material. Likewise, zinc must be coated on the core of metallic wire in view of low melting point. Subsequently, brass or copper ordinarily are utilized as core of these sorts of wires. An amazing combination of low cast, better flush capacity, high mechanical quality and great electrical conductivity are the benefits of zinc coated wires as opposed to brass or copper wires. Furthermore, the coated layer diminishes the danger of breaking wires since it ensures the core of the wire from thermal shock of electrical release. All specified preferences for zinc coated wires upgrade the execution of WEDM, expanding in cutting velocity and accuracy.

In order to reduce the number of experiment trials and avoiding material wastage, Tagugi L18 full factorial experimental design is being implemented. Table 3 and Table 4 respectively provides the details of the various trials conducted and the corresponding MRR and surface roughness measured. Higher machining rate is achieved with tap water as the dielectric fluid with zero electrode wear (Tariq Jilani and Pandey, 1984). Wire EDM is produced a stress free cutting.

3. RESULTS AND DISCUSSIONTOP

The material removed from the raw work piece by WEDM and the corresponding CNC display is compiled into a conclave and is furnished in Fig. 1 shows the machine diagram and its results with specimen and CNC display. The specimen dimension is 5×5×24 mm with 18 samples are shown in the picture. This trial are conducted to analyze the influence of interdisciplinary process parameters on the material removal rate and optimize using ANOVA.

3.1. Brass wire analysisTOP

Figure 2a, shows the mean effect plot of MRR against the input parameter such as peak current (Ip), pulse on time (Ton), pulse off time (Toff), water pressure (wp), wire feed rate (Wf), wire tension (Wt), servo voltage (Sv), servo feed setting (Sfs) on cutting speed has the selected machining condition. It is clearly noted that the pulse on time, pulse off time, Water pressure and servo voltage is proportional to the cutting speed. Moreover it is observed that there is a greater dependency of cutting speed on the Ton/Toff of the pulse. Figure 2b, shows the mean effect plot of surface roughness (SR) in brass wire Vs the all input parameters Ip, Ton, Toff, Wp, Wt, Wf, Sv, Sfs. This graph demonstrate the fact that wire tension (Wt), water pressure (wp), pulse off time (Toff) and pulse on time (Ton), majorly alters governs the surfaces roughness values. As the water pressure increase, flushing is higher and removed materials are pulsed away, yielding grater Ra. Wire tension is another affecting factor for Ra, her varying wire tension values such as 7, 8, and 9 respectively. Another factor is Toff, flushing rate which increases during pulse off timing, eventually leading to attainment or improve Ra. Ton when maintained for longer time produces sparks with higher time or sustenance yielding smoother surfaces. With more material removed due to heat.

3.2. Zinc coated wire analysisTOP

Figure 3a shows plot for MRR against Vs all input parameters on cutting speed with zinc coated brass wire. According to this Fig. 3b Toff and Ton, Ip and Sv, values are the governing parameters. MRR depends mainly on these 4 parameters. Role of Ton is more effective compared with the pure brass wire, conductivity to the wire and displays higher toughness and tensile strength. Figure 3b shows the plot for Surfaces roughness (Ra) vs. all input parameters. Surfaces roughness (Ra) mainly depends on the water pressure, wire tension, pulse on time and servo feed setting. Water pressure affects the Ra value. High water pressure range gives better flushing which improves the Ra values. Ton and wire tension contribute to Ra value, while the pulse on time is long, with a spark at less energy that removes material in small volume. The removed voids enables flushing rate indirectly governing the Ra values improved. Figure 3 (a, b) gives the interaction plot of process parameters.

Figure 4 (a, b) illustrate the surface plot of all input parameters in various trials with respect to the material removal rate (MRR) for using brass wire cutting. The plot indicate that the prominently significant factor quenching the process are Ton, Toff and water pressure (Wp). As the On time (Ton) increases, there is a corresponding increases in the materials removal rate proving a propositional relationship. MRR majorly relies on the current discharge for the stipulated time duration. Conversely, With the increases in the Toff, MRR decreases as it is directly being affected by the decreases discharge which is an important factor. Another important observation was the dependency of MRR on water pressure. As the water pressure increases, there is a step hike in the MRR owing to the high flow flush. Nevertheless, if the pressure of water is decreases, the debris are re-solidified in base material. Re-solidification layer leade to a decreases in the MRR rate due to the objection needed by the debris layers. Figure 5 (a, b) shows the corresponding contour plot for material removal rate (MRR) vs. input combinations. These plots show how the response variable is related to the two factors. The best MRR is achieved at middle of the plot range experimented. This is consistent with the fact that the output term is significant.

The material removal rate (MRR) for brass wire for p-value of less than 0.05 for the model having 95% confidence level indicates that the model terms are statistically significant. The ANOVA (Table 5) result shows that the pulse on time (Ton), pulse off time (Toff) and water pressure (Wp) are the significant model terms associated with MRR brass wire. The other model terms are not significant and thus, eliminated by backward elimination process to improve model adequacy (Sharma et al., 2013). The final mathematical models for MRR brass wire, may be used for prediction within the same design space, are given as follows:

MRR B = −17.35 − 0.8667 Ton + 0.7370 Toff + 0.7667 Wp − 1.058 WF + 3.833 WT − 0.5172 SV − 0.006167 SFS + 0.02000 Ton*Ton − 0.005200 Toff*Toff − 0.02500 Wp*Wp + 0.1117 WF*WF − 0.07333 WT*WT + 0.006367 SV*SV − 0.000100 SFS*SFS + 0.007333 Toff*WF − 0.05867 Toff*WT + 0.005867 Toff*SV

Figure 5 (a, b) gives the interaction plot of process parameters. Figure 6 (a, b) illustrate the surface plot of all input parameters in various trials with respect to the surface roughness (SR) for using brass wire cutting. The plot indicate that the prominently significant factor quenching the process are Toff, water pressure (Wp) and wire feed rate (Wf). As the Off time (Toff) increases, there is a corresponding decreases in the surface roughness (SR) proving a opposing relationship. SR majorly relies on the current discharge for the stipulated time duration. Conversely, with the decreases in the Toff, SR increases as it is directly being affected by the decreases discharge which is an important factor. Another important observation was the dependency of SR on water pressure. As the water pressure increases, there is a low in the SR owing to the high flow flush. Nevertheless, if the pressure of water is decreases, the debris are re-solidified in base material. Re-solidification layer lead to a increases in the SR rate due to the objection needed by the debris layers. Figure 7 (a, b) shows the corresponding contour plot for surface roughness (SR) vs. input combinations. These plots show how the response variable is related to the two factors. The best SR is achieved at middle of the plot range experimented. This is consistent with the fact that the output term is significant. The surface roughness (SR) for brass wire for p-value of less than 0.05 for the model having 95% confidence level indicates that the model terms are statistically significant. The ANOVA (Table 6) result shows that the pulse off time (Toff), water pressure (Wp) and wire feed rate are the significant model terms associated with SR brass wire. The other model terms are not significant and thus, eliminated by backward elimination process to improve model adequacy (Sharma et al., 2013). The final mathematical models for SR brass wire, may be used for prediction within the same design space, are given as follows:

SR B = −32.84 − 0.1675 Ton + 0.4070 Toff + 2.777 Wp + 0.7106 Wf + 4.348 Wt − 0.4268 Sv − 0.04073 Sfs + 0.004457 Ton*Ton − 0.002806 Toff*Toff − 0.1104 Wp*Wp −0.06018 Wf*Wf − 0.1897 Wt*Wt + 0.003697 Sv*Sv + 0.000088 Sfs*Sfs − 0.007937 Toff*Wf − 0.03207 Toff*Wt + 0.006042 Toff*Sv

Figure 5 (a, b) gives the interaction plot of process parameters. Figure 5 (a, b) illustrate the surface plot of all input parameters in various trials with respect to the material removal rate (MRR) for using zinc coated wire cutting. The plot indicate that the prominently significant factor quenching the process are Ton, Toff and servo voltage. As the On time (Ton) increases, there is a corresponding increases in the materials removal rate proving a propositional relationship. MRR majorly relies on the current discharge for the stipulated time duration. Conversely, with the increases in the Toff, MRR decreases as it is directly being affected by the decreases discharge which is an important factor. Another important observation was the dependency of MRR on servo voltage. As the servo voltage increases, there is a step hike in the MRR owing to the high accurate servo movement. Nevertheless, if the servo voltage is decreases, the servo mechanism operate at a slow speed, so wire movement in-betweens the material is slow. This slow speed of wire lead to a decreases in the MRR rate. Figure 5 (a, b) shows the corresponding contour plot for material removal rate (MRR) vs. input combinations. These plots show how the response variable is related to the two factors. The best MRR is achieved at middle of the plot range experimented. This is consistent with the fact that the output term is significant. The material removal rate (MRR) for zinc coated wire for p-value of less than 0.05 for the model having 95% confidence level indicates that the model terms are statistically significant. The ANOVA (Table 7) result shows that the pulse on time (Ton), pulse off time (Toff) and servo voltage are the significant model terms associated with MRR zinc coated wire. The other model terms are not significant and thus, eliminated by backward elimination process to improve model adequacy (Sharma et al., 2013). The final mathematical models for MRR zinc coated wire, may be used for prediction within the same design space, are given as follows:

MRR z = 12.87 − 0.7096 Ton + 0.3796 Toff − 0.2795 Wp − 0.2231 Wf + 3.135 Wt − 0.2844 Sv − 0.3758 Sfs + 0.01709 Ton*Ton − 0.002573 Toff*Toff + 0.01475 Wp*Wp − 0.03613 Wf*Wf − 0.08042 Wt*Wt + 0.001388 Sv*Sv + 0.001790 Sfs*Sfs + 0.009900 Toff*Wf − 0.04133 Toff*Wt + 0.004953 Toff*Sv

Figure 5 (a, b) gives the interaction plot of process parameters; illustrate the surface plot of all input parameters in various trials with respect to the Surface Roughness (SR) for using zinc coated wire cutting. The plot indicate that the prominently significant factor quenching the process are Ton, water pressure (Wp) and wire tension (Wt). As the On time (Ton) decreases, there is a corresponding increases in the surface roughness (SR) proving a opposing relationship. SR majorly relies on the current discharge for the stipulated time duration. Conversely, with the decreases in the Ton, SR increases as it is directly being affected by the decreases discharge which is an important factor. Another important observation was the dependency of SR on water pressure. As the water pressure increases, there is a low in the SR owing to the high flow flush. Nevertheless, if the pressure of water is decreases, the debris are re-solidified in base material. Re-solidification layer lead to a increases in the SR rate due to the objection needed by the debris layers. Figure 5 (a, b) shows the corresponding contour plot for surface roughness (SR) vs. input combinations. These plots show how the response variable is related to the two factors. The best SR is achieved at middle of the plot range experimented. This is consistent with the fact that the output term is significant. The surface roughness (SR) for zinc coated wire for p-value of less than 0.05 for the model having 95% confidence level indicates that the model terms are statistically significant.

The ANOVA (Table 8) result shows that the pulse on time (Ton), water pressure (Wp) and wire tension are the significant model terms associated with SR brass wire. The other model terms are not significant and thus, eliminated by backward elimination process to improve model adequacy (Sharma et al., 2013). The final mathematical models for SR zinc coated wire, may be used for prediction within the same design space, are given as follows:

SR z = − 91.65 − 0.07772 Ton + 0.002553 Toff + 2.257 Wp – 1.335 Wf + 1.714 Wt − 0.1759 Sv + 1.551 Sfs + 0.002830 Ton*Ton − 0.000622 Toff*Toff − 0.08874 Wp*Wp + 0.006512 Wf*Wf − 0.06008 Wt*Wt − 0.000348 Sv*Sv − 0.007838 Sfs*Sfs + 0.02888 Toff*Wf − 0.01495 Toff*Wt + 0.004329 Toff*Sv

From these analysis, it may be concluded that only Ton, Toff, Wp, Wt and Sv has considerable impact on the MRR and SR whereas the other process parameters has no or negligible impact on the MRR and SR. The plot obtained from the ANOVA – Mini Tab Software depicting the variation of MRR and SR with the governing parameter.

3.3. Scanning electron microscopy (SEM)TOP

Scanning Electron Microscopy (SEM) perspectives of the cutted surfaces were taken with a specific end goal to look at the impact of various wire types on machining of different work piece materials by examining surface integrity and attributes. The surface integrity comprises of roughness, size of pits and depth of a recast layer and heat affected zone that are examined in this area. Figure 6 (a, b) demonstrates the surface of nickel work pieces which are machined with brass wire materials.

From the examination of the surface attributes that are appeared in Fig. 6 (a, b), it is watched that fast brass wire brought about a poor surface completion. Additionally, it can be seen zinc coated wire creates less edge on the work piece surface observed in Fig. 7 (a, b). Low quality surfaces by uncoated wire results from the non-uniform thermal load on the sample. This non-uniform thermal load is because of non-uniform flash arrangement and insufficient flushing qualities. Improper flushing can also bring arc formation, which can promote harm the surface.

To be sure, the harmed layers comprise of 3 layers, in particular white layer, recast layer and heat affected zone. The recast layer is characterized as the material liquefied by electrical sparks and not removed by flushing. Same conclusions can be made when utilizing nickel work piece and different electrode materials as indicated by Fig. 7 (a, b). The heat-affected zone was observed to be restricted in its thickness and couldn’t be recognized effectively. Therefore, the depth of harmed layer reported here was measured from the extreme surface to the end of recast zone. As appeared in Fig. 7 (a, b) the thickness of harmed layer of nickel machined with zinc coated wire is lower than the other one created by brass wire. Zinc coated wire has great electrical conductivity that helps the wire exchange energy to work piece effectively. At that point, it prompts apply low heat energy to the specimens’ surface and decline the harmed layer framed. High cutting temperature, more energy conveyed to the work piece and solid grip between the electrode and the nickel lead to poor surface quality. These issues are made because of the low heat conductivity and high compound reactivity of titanium composites. The microstructures reveal a reduction in the grain size at the HAZ as compared to the base material there by increasing the hardness slightly in the region without significantly altering the brittleness. This technique is technically viable due to its higher accuracy and preciseness avoiding material loss and reduction in cost. In longer run, this technique for nickel/ nickel alloy cutting proves to be economical.

4. CONCLUSIONSTOP

ACKNOWLEDGEMENTTOP

The authors thank the management and officials in CIPET, Chennai for extending WEDM facility to carry out this research study.

REFERENCESTOP