Parameter Optimization and Machining Performance of Inconel 625 with Nanoparticles Dispersed in Biolubricant

Productivity and cost-eectiveness are essential components of any long-term manufacturing system. While quantity and quality are linked to productivity, the economy focuses on energy-ecient processes that produce a high output-to-input ratio. Hard-tocut materials have always been dicult to machine because of more signicant tool wear and power losses. Inconel 625 is a hard material used in aerospace and underwater applications and is milled using biolubricants with nanoparticles. Palm oil is considered a biolubricant, and titanium dioxide (TiO2) and copper oxide (CuO) are selected as nanoparticles. When the combination of biolubricants and nanoparticles is added to the workpiece’s surface, it enhanced some properties while machining. Experiments involving four factors with four levels were carried out using the Taguchi design of experiments (DoE). e feed, depth of cut, speed, and coolant with nanoparticle additives were all factors. e responses were surface roughness, spindle vibration along X, Y, and Z axes, and material removal rate. Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) was used to alter the multiresponse optimization problem to a single-response optimization problem. e S/N of TOPSIS closeness coecients was calculated, and the optimal machining conditions were determined. Surface roughness, material removal rate, and spindle vibration were reduced by 3.10%, 6.14%, 7.54% (Vx), and 6.78% (Vz), respectively, due to the TOPSIS optimization.


Introduction
Inconel 625's enhanced mechanical properties, outstanding weldability, and high oxidation and corrosion resistance, nickel-based aerospace alloys have gained popularity. Inconel 625 is widely used in manufacturing, especially for aircraft structures, springs, turbine blades, submarine bellows, steam power plants, and oceanographic devices [1]. Due to its meager thermal conductivity, the formation of built-up edges, and a greater sticking or welding propensity to cutting edges, Inconel 625's machinability is considered poor [2] and classi ed as a di cult-to-machine material. Furthermore, due to Inconel 625's low heat transfer rate, a large portion of the cutting energy is converted into heat during machining, which remains in the tool-workpiece interface for a longer time. High localized temperatures in the machining region result from heat generation, causing tool material softening and rapid tool wear, decreasing tool life, and compromising machined surface integrity. e use of cutting uids is required to solve these issues. Most traditional machining uids contain hazardous chemical constituents that can pollute the environment, cause biological problems for workers, contaminate soil, and pollute water during disposal [3,4]. Furthermore, cutting uids account for roughly 17% of machining costs, while tooling costs account for only 8% [5,6]. Many attempts have been made to reduce cutting uids to make material removal processes more environmentally friendly [7]. e growing interest in tracking all elements of the material removal process has resulted from the metal-based industry's main challenge of increasing the quality and productivity of machined parts [8].
Dry cutting, or machining without the need for any cutting fluid, is one of the best machining options for achieving green manufacturing. However, when dry-machining Inconel, the work material bonds firmly to the tool surface, resulting in early tool failure and poor surface quality. Furthermore, Inconel's high mechanical strength and poor thermal conductivity result in unfavorable residual stresses, surface irregularities, and burning/overheating in the cutting zone when machining without coolant [9]. e surface roughness of the machined product can affect various areas of its operation, including gentle friction, heat generation, the ability to distribute and hold a lubricant, wear, and a material's ability to withstand fatigue [10]. Dry cutting also necessitates using unique cutting tool materials such as ceramic, PCD, PCBN, and careful tool geometry and specific coatings. erefore, to facilitate heat transfer from the tool-chip interface, these tacky alloys are typically machined under wet cutting conditions, which results in high manufacturing costs, worker health risks, and severe environmental problems [11]. Due to the specific inherent properties and their capacity to biodegrade, vegetable oils are seen as alternatives to mineral oils in lubricant formulations.
Vegetable oils have a high flash point, viscosity index, lubricity, and lower evaporative loss than mineral oils [12]. Plant oils are extracted by applying pressure to the pertinent part of a plant and squeezing the oil out [13]. Plant oils (edible and nonedible) can also be extracted by dissolving plant parts in water, distilling the oil, or infusing plant parts with a base oil. Various studies have demonstrated the value of edible vegetable oils such as coconut oil [14], palm oil [15], soya bean oil [16], and canola oil [17] as an environmentally friendly lubricant for machining. e novelty was premised on the fact that using cooling/lubrication circumstances and depth of cut as input variables improved the manufacturing system's sustainability and efficiency. is study is based on the idea that productivity results from quality, utilization, and efficiency working together. e utilization of nanoparticles in various base fluids has received a lot of interest in the last decade [18]. e nanoparticles dispersed in water, for instance, can improve the thermal conductivity, and it is a suitable heat transfer fluid, especially for solar collectors [19]. Nano coolants (the dispersal of nanoparticles in water or ethylene glycol) have been studied for real-world problems since the early 2000s [20]. Nanoparticles added to the lubricant are thought to provide antifriction and antiwear properties. On the other hand, the improved characteristics are entirely determined by nanoparticle characteristics like shape, size, and concentration. Moreover, it has been reported that adding a suitable amount of nanoparticles to lubricating oil improves antifriction and antiwear characteristics [21]. ese terms cover the manufacturing process, including surface roughness, material removal rate, and spindle vibration. ese responses are optimized by combining constructive process parameters like feed speed and cutting depth. Taguchi design of experiment (DoE) is used to optimize the input parameters collectively, as they would alternatively behave differently in different responses. e goal is to optimize the input parameters based on the responses that are both sustainable and constructive at the same time.
Taguchi's DoE is an excellent tool for optimization because it is simple and efficient. Taguchi assists in selecting a control variable combination that significantly reduces the impact of noise. Minimizing tool costs is necessary to be cost-effective in manufacturing [22]. Variations recommended that machining irregularity could be reduced if appropriate values and requirements were used [23]. Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) selects the alternative closest to the ideal solution and farthest from the ideal negative alternative. It is helpful in cases where there are a lot of requirements and substitutes [24]. It is based on the theory of an optimum moving solution from which it tries to negotiate the result that is the closest. e smallest distance from the positive ideal solution (PIS) and the greatest distance from the negative ideal solution (NIS) are used to rank the options. TOPSIS explores the ranges of both PIS and NIS, ranking candidates based on their relative proximity and combining the two distance measurements [25]. Taguchi's DoE was combined with the TOPSIS to identify the processing parameters for milling the Inconel 625 alloy. Surface roughness, spindle vibration, and MRR were all taken into account. ey discovered that, after optimization, machining performance improved [26]. e novelty of this research lies in the combination of biolubricants and nanoparticles that are used. Inconel 625 has not been machined with the current choice of biolubricants and nanoparticles. e responses such as surface roughness (Ra), material removal rate (MRR), and vibration have not been recorded for this particular combination.

Oils and Nanoparticles.
After carefully considering the literature review and availability of the materials, palm oil has been chosen. e nanoparticles chosen are TiO 2 and CuO. e nanoparticles are biocompatible with oils and do not cause adverse effects on the lubricant. e properties of palm oil are given in Table 1. 2.2. SEM Images. SEM image was used to examine the surface morphology of the nanoparticles. SEM images of the nanoparticles TiO 2 and CuO have been observed and shown in Figure 1. It ensures that the nanoparticles are in the nanometer size range and that the size is marked on the image. Figure 1(a) shows that CuO nanoparticles are in the range of 124 to 215 nm. According to the SEM image, as 2 Advances in Materials Science and Engineering shown in Figure 1(b), the TiO 2 nanoparticles are 60 to 100 nm in size, and the TiO 2 nanoparticles have a homogeneous spherical morphology.

EDX of CuO and TiO 2 .
e energy dispersive X-ray (EDX) analysis is used to characterize the elemental composition and chemical composition of a specimen with an atomic number.
Elemental mapping is a technique for obtaining high-resolution imaging by accumulating detailed elemental composition data across a sample area. Every pixel in the image is examined to preserve the rudimentary spectrum. e EDX spectrum of CuO Nanoparticles is shown in Figure 2. e spectrum depicts the chemical components of the sample. e dissemination of Cu (red dots) and O (green dots) components, which make up the whole body of the   Advances in Materials Science and Engineering processed sample, is homogeneous in Figure 2. For each element, the corresponding ndings are shown separately. e presence of oxygen and copper is demonstrated by the change in the distribution of both elements. Figure 2 indicates that 81.1% of Cu and 18.9% of O were presented in the sample. e elemental mapping shows that elements are correctly dispersed in aggregated Cu and O nanoparticles [27]. Figure 2 shows no CuO nanoparticle impurities, and only Cu and O elements are present [28,29]. Figure 3 shows EDX analysis of TiO 2 nanoparticles. Figure 3 reveals that 63.2% of Ti and 36.8% of O were presented in the sample. e distribution of Ti (red dots) and O (green dots) elements, which make up the entire body of the processed samples, is homogeneous [30,31]. e results of EDS revealed that no other impurities were present in the nanoparticles. From Figure 3, it was observed that there is no impurity in the TiO 2 nanoparticle, and only the elements Ti and O were present [32,33].

Methodology.
e owchart of the methodology for this research is given in Figure 4.

Experimental Setup.
Milling operations were completed on a high-rigidity Computer Numerical Control (CNC) BMV35 T12 with a machine with speci cations: maximum spindle rpm 8000, spindle power 5.5 kW, and  maximum traverse distance in x-y-z axis are 450-350-350 mm, respectively. Commercially available Inconel 625 block (150 × 50 × 50 mm) was used as the workpiece material for machining. End milling operation was selected as the machining process. e cutting tool used for machining Inconel 625 was PVD-coated carbide (Grade: VP15TF; designated as SEMT13T3AGSN-JM).

DoE.
Feed rate, spindle speed, cut depth, and palm oil are process parameters. At the same time, the surface roughness, spindle vibration, and material removal rate are considered responses. An orthogonal array (OA) matrix helps the machine operator decide the best parameters with the fewest possible experiments. e four-parameter system has a total of 15 degrees of freedom. An OA's Degree of Freedom (DoF) should be equal to or larger than the total DoF. As a result, L 16 OA was used in this study because it has a DoF of 15 and allows fewer experiments to identify the best milling parameters. e selected parameters and their levels are shown in Table 2. ere are sixteen experiments in total. Experiments are carried out after the OA has been defined, and the S/N for every experiment is calculated [34].

Preparation of Coolant.
A beaker of 200 ml was taken, and 99.5 ml of palm oil was poured into it. 0.5 g of nanoparticles was measured using a highly sensitive electronic  Advances in Materials Science and Engineering balance.
e electronic balance was air-tight to ensure minimal error. e nanoparticle was poured into the beaker, and constant stirring was done for thirty seconds using a spatula.
e stirring ensures that the oil has 0.5 wt% of nanoparticles uniformly [35]. e nanoparticles and preparation of nanolubricant are shown in Figures 5 and 6.

Milling Procedure.
e CNC machine BMV35 T12 was used for all milling operations on Inconel 625, as shown in Figure 7. e vibration sensor MPU 6050 has been soldered to the Arduino UNO board using jumper cables. e Arduino UNO board acts as an interface between the sensor and the system and is connected to a laptop using a USB-A cable. e vibration sensor MPU 6050 has been attached to the spindle using double-sided tape. Using the Arduino IDE, vibration in the spindle's x-, y-, and z-axes during the milling process has been recorded. e workpiece is cleaned with a neat cloth before fitting inside the CNC machine. Facing the workpiece has been done to 0.1 mm. After facing the material, the tool holder is removed and replaced with milling inserts. e mixture of oils and nanoparticles has been poured uniformly over the material using a dropper (10 ml). e end milling operation has been carried out according to the DoE design matrix. After machining, the surface roughness was measured using a surface roughness testing instrument (Make-Carl Zeiss. Model-E-35B). ree surface roughness values were recorded, and the average surface roughness was noted. e material removal rate (MRR) was  Advances in Materials Science and Engineering computed using the weight-loss method, and the weight of the material was recorded before and after each pass. e formula for the weight-loss method has been given as follows: Loss of material weight(g) original weight − measured weight, Volume of workpiece mass(Loss of material) density ,

MRR
Volume of workpiece machining time .
(1) e density of Inconel 625 is 8.4 g/cc. e same procedure is repeated for all experiments.

Process Parameter Optimization
2.5.1. Taguchi's S/N. Using Taguchi signal-to-noise (S/N), the optimal factors were analyzed. Larger is better (LB) and Smaller is better (SB) are the two characteristics available for optimization. LR characteristics were applied for MRR and SR characteristics for Ra and Vx, Vy, and Vz. Using the formula, S/N values for di erent responses were recorded [36][37][38]. (2) e S/N ratio was used to determine the best conditions for each response. e ideal situation is the level at which the maximum S/N is reached.

TOPSIS.
To turn a multiresponse optimization problem into a single-response optimization problem, TOPSIS is used. e steps involved in the TOPSIS approach are depicted below. A normalized matrix is utilized in TOPSIS. Calculate the PIS and NIS using normalized weighted values and Euclidean distance using formulae [39,40]. e S/N for closeness coe cient was computed from that the optimal parameter for multiresponse was identi ed.
(1) Normalization matrix r ij is calculated out using where i 1, 2, 3, . . ., m, j 1, 2, 3, . . ., n and a ij represent the i th value of the j th experimental run. r ij represents the normalized data for the corresponding test. (2) Compute the weight w ij of each response.
(3) e weighted normalized data is computed by multiplying the normalized data with its equivalent weight. e weighted normalized data V ij is computed using where i 1, 2, . . ., m, j 1, 2, . . ., n and w j represents the weight of the j th criterion n j 1 w j 1.

Single Response Optimization Using
Taguchi's S/N. e surface roughness, MRR, and spindle vibrations are tabulated in Table 3.
Taguchi's S/N values were used to find the best parameters for individual responses. e SB characteristics were used for surface roughness and vibration signals. Maximizing the MRR is a critical criterion in metal removal processes [41]. e MRR must be determined to attain excellent machinability. As a result, for MRR, the LB characteristic was used. From Table 4, it can be seen that the highest S/N value produces the best results. For palm oil, the minimum surface roughness can be obtained when the spindle speed is 3000 rpm, feed rate of 125 mm/min, depth of cut (DoC) of 0.15 mm, and palm oil, with CuO nanoparticles being used as a coolant. Minimal vibration on the x-axis during the machining operation was obtained with the spindle speed of 2000 rpm, feed rate of 175 mm/min, DoC of 0.15 mm, and palm oil without nanoparticles. Similarly, for the y-axis and z-axis speeds of 1500 rpm and 3000 rpm, a feed rate of 150 mm/min for both DoC of 0.25 mm for both 3.2. ANOVA. Analysis of Variance (ANOVA) is used to find the most significant parameter influencing the response. When using ANOVA, the method is quite beneficial for determining the level of risk and the effect of milling parameters on a specific response. It is utilized to determine each control factor's relative influence in the response evaluation to ensure that the quality of the most critical aspects of the product should be carefully monitored [22,42]. Table 5, both speed and feed rate have the same level of contribution of 31% to surface roughness, followed by DoC of 18.68% and coolant of 17.82%, respectively. From the P-value (P < 0.05), it was found that all parameters have significantly impacted machining. e average surface roughness for the speed of 1500 rpm is 0.16 μm, and when the speed increases to 2500 rpm, the surface roughness is reduced by 18.75%. Similarly, for the speed of 3000 rpm, there is a 38.75% decrease in surface roughness. As the spindle speed increases, the built-up edge advancement slows down, and heat in the shear zone rises, making it more straightforward for machining and improving surface quality [43]. With an increase in feed, the surface roughness steadily increased, generating force on the machined surface that causes vibration, which raises the roughness. An identical pattern was observed in the literature [21,44].

Surface Roughness. From
It is observed that when CuO and TiO 2 (hybrid mode) nanoparticles are mixed with palm oil, the lowest surface roughness is obtained when compared to all other combinations [45]. e mechanism could involve rolling CuO nanoparticles rather than forming a layer or repairing surfaces [46]. CuO nanoparticles act as a third body between the two mating parts, preventing metal-metal contact and thus reducing surface roughness, as evidenced by the lower coefficient of friction and roughness values observed [47]. Table 5, it was observed that feed has a significant impact on the MRR, contributing 47.36%. Moreover, the speed gives 37.06% of the contribution to the machining process. DoC and coolant have an insignificant impact of 6.07% and 7.73%. e feed has more impact than speed and as the feed increases, machining the material to the desired length takes less time, increasing the MRR [48]. When palm oil is mixed with the hybrid combination of nanoparticles at a speed of 2000 rpm, a feed of 175 mm/min, and 0.15 mm DoC, the optimal results are obtained. e hybrid mode of nanoparticles formulates a third layer between the workpiece and the tool. e surface is slippery, will be long-lasting, and is ideal for machining for extended periods [49]. Table 5, it can see that x-axis feed can have a significant impact of 42.31%, followed by a speed of 25.66%, and coolant of 18.83%. DoC also has a minimal impact of 11.65%. All parameters will have a significant impact on the machining process. For the y-axis, the DoC had a significant contribution of 49.56%, followed by coolant with 25.9% and speed with 16.48%, and feed had an insignificant impact of 6.43% [50]. For the z-axis, DoC had a significant impact of 56.23%, followed by a speed of 22.54%, and both the coolant and depth of cut had an insignificant impact of 9.6% and 9.2%, respectively. e spindle moves in the same direction as the z-axis. As a result, the contribution of the DoC is more on the z-axis [34,51].

Spindle Vibration. From
It is observed that when using vibration on the x-axis, the values are lower when TiO 2 is mixed with palm oil [52]. Similarly, for the z-axis, the values are lower when TiO 2 is mixed with palm oil [53]. Similar research discovered that adding nanoparticles to the lubricant can decrease friction and wear, increase allowable bearing capacity, and remove heat under higher temperatures and high load conditions, reducing bearing wear and achieving vibration suppression [54].

Multiresponse Optimization Using TOPSIS
TOPSIS can be used to conduct multiresponse optimization. Table 6 displays the normalized data. (3) can be used to conduct data normalization. Table 7 shows the weighted normalization and separation measures. (4) is used to calculate the weighted normalization. Equations (8), (9), and (11) calculate the separation measures and the closeness coefficient CC i . Table 8 shows the S/N values of CC i . From Table 8, it can see that the optimal parameter can be identified. e optimal parameters are the speed of 2000 rpm, feed of 175 mm/min feed, DoC of 0.15 mm, and palm oil with 0.25 wt% of CuO and 0.25wt% of TiO 2 nanoparticles. From Table 9, it can be seen that the feed had a contribution of 43.93%, followed by a speed of 25.10%, the coolant of 23.3%, and DoC of 6.08%. Speed, feed, and coolant significantly impacted the machining process. From this, it can be observed that the multiresponse characteristics are impacted by feed speed and coolant.  Table 10 showed the confirmation test results and discovered that the surface roughness decreased by 3.10%. e MRR was increased by 6.14%. e spindle vibration in the x & z-axis decreased by 7.54% and 6.78%. On the other hand, vibration in the y-axis increased by 5.44%. e consumption of nanoparticles (0.25 to 0.5 wt%) along with palm oil is significantly less, and the overall cost is also reasonably minimum. Further, it enhances the surface roughness, MRR, and vibration features. is will significantly enhance the life span of the machine.

Conclusion
Inconel 625 was machined with SEMT-13T3AGSN-JM VP15 TF with palm oil and CuO and TiO 2 nanoparticles as additives. Taguchi's DoE was applied to design the experiments. Taguchi's DoE coupled with TOPSIS was used to optimize the process parameters. e following conclusions have been drawn from the experimentation: (i) Surface roughness was measured as a feed and speed function and depended on it. (ii) Both the speed and feed significantly impact MRR. (iii) e spindle speed vibration in the x-axis depends on the speed and feed. Similarly, the y-axis depends on the DoC and coolant, and the z-axis depends on the depth of cut and speed. (iv) Taguchi's S/N analysis was used to find the best parameters for individual responses. (v) TOPSIS was used to perform the multiresponse optimization, with the best parameters being 2000 rpm, 175 mm/min feed, 0.15 mm depth, and coolant of palm oil with 0.25 wt% of CuO and 0.25wt% of TiO 2 nanoparticles. (vi) According to ANOVA for the closeness coefficient, speed and feed have physical significance, with 25.10% and 43.93%, respectively. (vii) Surface roughness, material removal rate, and spindle speed vibration were reduced by 3.10%, 6.14 percent, 7.54% (Vx), and 6.78%(Vz) due to TOPSIS optimization. is will significantly improve the machining performance.

Data Availability
e data used to support the findings of this study are included within the article.

Conflicts of Interest
e authors declare that they have no conflicts of interest.