Evaluation of palm kernel oil as lubricants in cylindrical turning of AISI 304 austenitic stainless steel using Taguchi-grey relational methodology

The disadvantages of the conventional cutting fluids such as mineral oil have prompted the search for eco-friendly cutting fluids. Vegetable oils have often been recommended as environmentally friendly substitutes for traditional mineral oil. The current study examined the performance of palm kernel oil (PKO) and its mineral oil during the turning of AISI 304 steel using the minimum quantity lubrication (MQL) technique. Six litres of crude PKO were extracted from palm kernel seeds through a mechanical extraction technique. Taguchi L9 (3)3 orthogonal array was considered for the experiment. The depth of cut (DC), feed rate (FR), and spindle speed (SS) are the cutting parameters while cutting temperature (CT) and surface roughness (SR) are the response characteristics. Experimental results showed that the mineral oil outperformed the PKO in terms of SR with an improvement of 48.2%. The improvement of PKO over mineral oil is approximately 0.89% in terms of cutting temperature. The highest turning temperature of mineral oil was 67.333 °C, while that of PKO was 67.8 °C. In general, the performance of PKO shows it can be a good replacement for mineral oil if produced industrially with adequate additives. The grey relational analysis (GRA) showed that the optimum DC, FR, and SS for palm kernel and mineral oils are 1.25 mm, 0.25 mm rev−1 and 870 rev min−1, and 1.25 mm, 0.10 mm rev−1, and 870 rev min−1, respectively. The results of this study demonstrated an experimental basis for the application of PKO minimal quantity lubrication turning and validated the efficacy of the integrated Taguchi-grey relational analysis (TGRA) optimization approach.


Introduction
AISI 304 stainless steel is one of the most frequently used materials in various applications due to its exceptional qualities.One of the key features of this steel is its high corrosion resistance [1][2][3].This material is capable of withstanding exposure to a wide range of corrosive media, including acids, alkalis, and chloride solutions.Its resistance to pitting and crevice corrosion is also excellent, making it ideal for use in applications that involve exposure to corrosive liquids, such as in the chemical and petrochemical industries [4,5].AISI 304 austenitic stainless steel also possesses excellent weldability, which makes it an ideal material for fabrication and welding applications.It is easy to form and shape, allowing it to be fabricated into various shapes and sizes, including sheets, plates, bars, and tubes.It also has a high resistance to deformation, making it ideal for use in applications that involve high temperatures and pressures Jinlong and Hongyun [6].AISI 304 stainless steel is commonly compared.Canola oil significantly improved performance due to its superior lubricating qualities.Moreover, Chang et al [31] studied the turning capabilities of titanium alloy with ceramic inserts using emulsified and raw vegetable oil.As cutting oils, coconut, and palm oils were used.The crude palm oil performed more effectively in terms of CRC, tool wear, and CT.When compared with emulsified coconut oil, crude coconut oil performed better.The effectiveness of environmentally friendly vegetable oils during the precision machining of nickelbased superalloy Inconel 718 was examined by Saleem and Mehmood [32].The performance of castor and sunflower oils was assessed and contrasted using a Taguchi L 9 (3 3 ) orthogonal array in the dry-cutting method.Combining sunflower oil with the slowest cutting speed (30 m min −1 ) and FR (0.168 mm rev −1 ), the least amount of tool wear was produced.Ross et al [33] demonstrated the cutting efficacy of MQL and cryogenic CO 2 with canola oil in high-performance Monel-400 machining.According to the results of the planned trials, CO 2 +MQL (CMQL) was discovered to be a superior cooling method by generating the best surface and reducing friction.When contrasted with MQL, CO 2 , and dry conditions, the flank wear decrease under CMQL was determined to be 37%-47%, 26%-33%, and 51%-55%, respectively.Different non-edible vegetable oils were used by Susmitha et al [34] to drill a mild steel material.In terms of SR, chip formation, and cutting force, neem oil, Karanja oil, and blends of these two oils were compared with mineral oil as cutting fluids.During the machining procedures, vegetable oil produced satisfactory outcomes.Burton et al [35] explored how emulsified vegetable oil in water performed as a cutting fluid during the process of milling.When compared to a traditional cutting fluid, emulsified vegetable oil showed lower chip thickness, cutting forces, and burr amount.According to the findings from experiments, an ultrasonic-atomized vegetable oil-in-water emulsion has a lot of potential for use as a cutting fluid.Additionally, Shukla et al [36] compared the performance of milling aluminum 6061 using flood, dry, and MQL techniques based on canola oil.The average SR for each experiment was analyzed, and cutting parameters were optimized for an improved surface finish.It concluded that, out of the three cooling processes, MQL offers the most desirable surface finish.Further, it was discovered that tool wear was far lower in the MQL conditions than it was in the flood and dry strategies.Ikumapayi et al [37] investigated the relative merits of three scenarios, namely dry cutting, mineral oil, and African star oil during ASTM A36 steel milling with respect to SR and CT.African star oil substantially decreased the cutting region temperature, which allowed for acceptable tool-chip contact, and improved MQL machining performance compared to mineral oil and dry cutting.In another study, Ikumapayi et al [38] machined 6061 aluminum alloy at different FRs, SSs, and cut depths while using African star oil as a means of cooling.Through its performance, African star oil demonstrated that it is significantly more efficient than dry machining at cutting AA6061 and produced the same effect as the mineral oil.The CT revealed efficient cooling actions and enhanced surface finish when African star oil was used as the cutting fluid.Besides, Kazeem et al [39] evaluated the cutting fluid effect of jatropha oil when turning AISI 1525 steel alloy under flood cooling conditions.In most machining conditions, jatropha oil outperformed mineral oil.Abutu et al [40] used the L 16 (2) 4 full factorial approach to develop and characterize indigenous PKO as a cutting fluid.Using Behnken's design L 15 (3) 3 design and GRA, the CT, MRR, and SR of mineral oil and PKO were examined.The ideal turning conditions for FR, DC, and cutting speed for mineral oil and PKO are, respectively, (0.2 mm rev −1 , 1.5 mm, and 800 rev min −1 ), and (0.3 mm rev −1 , 1.5 mm, and 600 rev min −1 ).Katna et al [41] developed emulsified cutting fluids based on Karanja and castor oils.In comparison to the traditional mineral oil, the cutting fluid developed from Karanja oil reduced SR by 23.57%, cutting force by 11.12%, and tool wear by 22.77%.Due to its significantly higher viscosity, castor oil-based cutting fluid performed worse than Karanja oil-based cutting fluid.According to an overview of the available research, using vegetable oil as cutting fluid offers a potential means of reducing CT, SR, and cutting forces while also creating an appropriate working environment.The machining characteristics of AISI 304 stainless steel have been discovered to be inadequately investigated with vegetable oil.As a result, this study investigates the effect of MQL with palm kernel on CT and SR while turning AISI 304 stainless steel with a commercial external threading tool at various feeds, SSs, and DCs arrangements.For the results of process parameters, Taguchi-based GRA is used.In addition, the process continues with the generation of response surface and contour plots based on the experiment data.

Extraction of vegetable oil
In this research, the vegetable oil considered as machining oil is PKO.Over 25 kg of dried palm kernel seeds were procured from Ojee Market, Ibadan, Nigeria.Oil was extracted mechanically from palm kernel seeds without prior milling of the seeds.The advantages of pure mechanical extraction include lower initial investment costs and a safer and less complex operation [42,43].The extracted oil through mechanical extraction method is referred to as crude PKO.The quantity of the extracted oil was 6 litres.

Workpiece, cutting tool, machine tool, and MQL set-up
The schematic of the MQL experimental set-up is shown in figure 1.In this study, some turning trials using AISI 304 stainless steel were carried out using different cutting variables and lubrication environments.Assessments were made on the relationship between the cutting parameters (DC, FR, and SS) and the CT and Sr The Taguchi design of the experiment was used for this research.An L 9 (3) 3 orthogonal array was chosen using the Taguchi design method, resulting in a total of nine experiments for each lubrication condition.The factors and levels of the orthogonal experiment are shown in table 1.
The workpiece's dimensions were 160 mm in length and 80 mm in diameter.Before the trials, a lightturning operation was used to get rid of the oxide layer and the heat-affected layer from the workpiece's surface [18].Tables 2 and 3 respectively display the chemical makeup and mechanical features of AISI 304 stainless steel.To conduct the series of experiments for this study, a three-chuck conventional lathe machine with a geared head (GH-1640ZX) was used (see figure 2).The main technical parameters are the maximum SS of 1800 rpm, rated power of 7.5 hp, and a voltage of 440 V.The cutting tools used for turning were external threading tools shown in figure 3 with the code DIN282 and a dimension of 20 by 20 mm, based on their long life and remark by the workpiece makers.The MQL system comprises an air compressor single-stage pressure machine, filter, oil tank, oil-gas mixing tube, and a lubrication gun.Compressed air at a pressure of 5 bar is mixed with a certain amount of cutting oil and finally injected into the cutting region.The flow rate of cutting oil was considered to be 2.4 ml hr −1 [24].
From the chemical characteristics table, manganese is the most abundant element in AISI 304 austenitic stainless steel.A vital component of steel is manganese.In actuality, this chemical component, which is present in all commercially available steels, contributes less than carbon to the steel's hardness and strength.The element with the second-highest abundance in AISI 304 steel is chromium.Chromium is necessary for the development of stainless steel [44].Chromium makes up approximately eighteen percent of stainless steel, giving it toughness, hardness, and greater corrosion resistance, especially at elevated temperatures.

Measurement of the turning characteristics 2.3.1. Cutting temperature measurement
This is an important parameter that affects the surface integrity and service life of a workpiece; lower CTs show better turning performance.The CT was measured using an infrared thermometer.A lens system is used in an  infrared thermometer to focus radiation onto an infrared detector, which then transforms the energy that it absorbs into an electrical signal [45].The source's emissivity is considered when adjusting the temperature calculated from the electrical impulse.The thermometer was placed carefully, about 5 cm from the toolworkpiece interface.The measurements were collected from three different points for each sample, and the mean result was used.

Surface roughness measurement
The SR of the component is a crucial metric in assessing cutting characteristics, and a lower value of SR indicates that the component will perform better in service.As a result, the lower the SR, the better.The SR of the workpiece was measured using a portable SR tester SRT − 6100 under different lubricating oil conditions in this study.The study used the mean of three values taken at random along the length of the machined surface of the workpiece sample.

Signal-to-noise ratio analysis
The 'lower is better' rule was applied to the S/N (dB) ratios to optimize the SR and CT values that had been determined.The S/N ratio for smaller-is-better responses is calculated using equation (1).
( ) Where n = number of repetitions in a trial, S/N = signal-to-noise ratio; and y = the measured quality characteristic for the i th repetition.

Evaluation of surface and contour plots
Surface or contour diagrams, which typically take the form of surfaces or cubes and represent and enable comprehension of the connections existing between the outputs and the factors examined, are the most effective means of understanding the data collected from experimental planning.However, the interaction between cutting parameters and the two responses i.e., SR and CT were studied and interpreted using response surface methodology and contour plots.The 3D surface and 2D contour plots in this study are plotted using MATLAB software.A program was written for each of the surface and contour plots before plotting on MATLAB.

Grey relational analysis
In the context of the grey systems approach, the influencing procedure known as 'GRA' examines the ambiguous relationships that exist between one major factor and all of the other variables in a system.The grey analysis aids in making up for errors in statistical regression when analyses are complex or when the methodology cannot be carried out precisely GRA measures the absolute value of the data difference between the sequences when used to assess the estimated correlation between sequences [46].

Pre-processing of data
The main sequence is transformed into an equivalent sequence using data pre-processing.Depending on a data set's characteristics, different data pre-processing techniques are available for GRA.The initial sequence displays the 'higher is better' property when the ideal value is infinite.The original sequence can be normalized using equation (2): For the lower the better characteristics, the original sequence is given as shown in equation (3) m is the number of experimental data items and n is the number of parameters.The y i * (k) is the sequence after the data pre-processing, y i °(k) denotes the original sequence, min y i 0 (k) is the smallest value of y i 0 (k), max y i 0 (k) is the largest value of y i 0 (k), and y 0 is the desired value.

Grey relational coefficient and grey relational grade
The grey relational grade (GRG) is the indicator of the relevance between two sequences or two systems.It is referred to as a local grey relation assessment when only one sequence, y 0 (k), exists as the reference sequence and all other sequences act as comparison sequences.Following data pre-processing, equation (4) can be used to express the grey relation coefficient ξ i (k) for the k th performance parameters in the i th test.
where, Δ 01 is the deviation sequence of the reference sequence and the comparability sequence.Reference and comparability sequences are indicated by the symbols y 0 * (k) and y i * (k), respectively.ξ ε [0, 1] may be modified depending on the real system standard.ξ is the identification or distinguishing coefficient.The distinguished capacity is larger and the ξ value is smaller.ξ =0.5 is used frequently.It is customary to use the mean value of the grey relational coefficient (GRC) as the GRG after deriving the GRC.Equation ( 5) is a definition of the GRG: The significance of different variables to a real engineering system, however, varies.The GRG in equation ( 6) was expanded and defined as follows in the actual situation of the various factors carrying unequal weights: where w k denotes the normalized weight of factor k. Equations ( 5) and (6) are equal when given the same weight.
The level of relationship between the comparability sequence and the reference sequence is represented by γ i .If the two sequences are the same, the GRG value is 1.The GRA also shows the degree of impact a reference sequence has on the comparability sequence.As such, the GRG for a reference sequence and a comparability sequence will be higher than that for the other GRGs.

Evaluation of experimental results
The output response parameters considered in this study are the SR and CT.The results obtained from the experiment are presented in table 4 and were subsequently converted to charts (see figures 4 and 5) for better visualization of the response-process parameter relationship.The signal-to-ratio values of the response characteristics are given in table 5.

Surface roughness
Surface roughness for palm kernel oil and mineral oil is reported to be between 1.265-4.990micrometers and 0.883-2.050micrometers, respectively.Surface roughness averages and standard deviations for palm kernel oil and mineral oil are 2.745 ± 0.824 micrometers and 1.422 ± 0.454 micrometers, respectively.The bigger the deviations, the rougher the surface, and the less the surface roughness, the smoother the surface [47,48].The analysis found that mineral oil lubricant had the lowest level of surface roughness in all nine testing runs (see figure 4).Mineral oil had a stronger impact on surface roughness than palm kernel oil.However, its use in the machining industry is discouraged due to its poor biodegradability, which has a number of negative environmental consequences, including contamination of surface and ground water, soil contamination, air pollution, and, as a result, farm produce and food contamination [49,50].Palm kernel oil with a high average value is highly suggested.The poor performance of palm kernel oil could be attributed to warping caused by heating and some machining errors caused by chatter or deflection.Surface roughness of palm kernel oil can be ascribed to both raw material qualities of austenitic stainless steel and manufacturing factors used throughout the trials.Surface roughness was reduced in conventional mineral oil by 63%, 8.5%, 82%, 60.55%, 54.7%, 53.8%, 10.7%, 24.4%, and 11.54% for tests 1, 2, 3, 4, 5, 6, 7, 8, and 9, respectively.Experiment 3 produced the highest surface roughness at spindle speeds, feed rates, and depths of cut of 415 rev min −1 , 0.25 mm rev −1 , and 0.75 mm, respectively.Kazeem et al [51] revealed superior surface finish results with mineral oil during AISI 1525 steel machining.Odusote et al [49] showed a similar pattern in cutting temperature.When groundnut oil was employed as the cutting fluid, the temperature of the mild steel was very near to that of the traditional oil, which was the lowest.The mineral oil condition helps to reduce friction and remove the chip generated from the cutting zone, leading to the development of BUE [52][53][54].

Cutting temperature
Material loss occurs during machining owing to plastic distortion in the major shear zone, which causes friction at the flank and rake face of the cutting tool [55].In the cutting region, the energy expended to cause the plastic deformation is converted into heat.The cooling/lubrication technique aids in decreasing the impact of heat created via effective heat dissipation.Figure 5 depicts the temperature of the tool-chip contact under various cutting parameters and cooling/lubricating conditions.Under MQL, the minimal tool-chip interface temperature was reached using vegetable oil.The basis of this is the rapid heat dispersion from the cutting area through the oil mist and corresponding decrease in friction, as well as the efficient dispersion of vegetable oil at the cutting area due to the generation of tiny particles with compressed air.Under experiment 7, at high spindle speeds of 870 rev min −1 , the highest tool-chip contact temperatures for mineral oil (67.8 °C) and palm kernel oil (67.333 °C) were both measured.This is due to the fact that as rotational speed rises, the temperature at the cutting tool and workpiece interface will increase.Additionally, for both lubrication situations, the lowest cutting temperature was achieved at a low spindle speed of 415 rev min −1 .This decrease is due to a shorter toolchip contact length and, consequently, a lower friction tool-chip interface.In comparison to traditional mineral oil, which has a mean and standard deviation of 47.189 ± 10.186 °C, palm kernel oil has a lower mean and standard deviation of 46.767 ± 12.7 °C.Palm kernel oil outperformed its mineral oil counterpart by 0.89%.Palm kernel oil performed marginally better than mineral oil, according to the cutting temperature analysis.This demonstrates that compared to mineral oil, palm kernel oil has superior fluidity and a quicker cooling capacity.The outcome of the present investigation is consistent with that of Abegunde et al [56], who found that palm kernel oil performed similarly to mineral oil in terms of cutting temperature.Against mineral oil, there was hardly any temperature difference-about 2%.Alaba et al [24] also found that using palm kernel oil during MQL machining of AISI 1039 steel resulted in an improvement of 12% over mineral oil.The structure of plants with various fatty acid content can be used to explain the behavior of palm kernel oil.At higher values of the examined parameters, the saturated character of palm kernel oil confers a great resistance to oxidative stability [18].6 illustrates the relationship between FR and Sr It was discovered that while PKO had the highest SR at high speed, mineral oil had the lowest SR at that speed.It was also observed that for mineral oil, SR increases as FR increases from low to moderate FR, but gradually decreases from moderate to high FR.From low to the medium FR, the SR of PKO decreases significantly, and from medium to high FR, it slightly increases.The line's slope for PKO is 1.93 and is positive.Accordingly, the SR rises by 1.93 units for every unit increase in FR.In other words, the surface gets rougher as the FR rises.The slope of the line for mineral oil is negative and equal to −2.23.This indicates that the SR decreases by 2.23 units for every unit increase in FR.In other words, the surface becomes smoother as the FR rises.The relationship between SS and SR is depicted in figure 7.For each SS setting, the average SR was assessed.The SR for PKO was found to be decreasingly correlated with SS from low to high levels.Additionally, it was discovered that for mineral oil, SR decreases as SS rises from low to moderate speeds.Additionally, the SR of sloppily increases from a moderate to a high FR. Figure 8 illustrates the variation in SR and DC for PKO and mineral oil during the turning process of AISI 304 steel.From low to high depths of cut, the SR continuously increases, but the increase from moderate to high depths of cut happens more quickly.From low to moderate DC, SR decreases sharply as the DC rises; however, from moderate to high DC, SR slightly increases.The slope of the line for PKO is a negative value of 1.58.This indicates that the SR decreases by 1.58 units for every unit increase in depth.In other words, the surface smooths out as the depth rises.The slope of the line for mineral oil is positive and equal to 1.28.This means that for every unit increase in depth, the SR increases by 1.28 units.In other words, as the depth increases, the surface becomes rougher.

Evaluation of selected surface and contour plots
The SR is the response variable in figure 12, while FR and SS are the independent variables.Two distinct peaks demonstrate that the SS and FR are combined at, respectively, 400-450 rev min −1 and 0.19-0.21mm rev −1 .The highest SR would be produced by a SS and FR combination of 0.1 to 0.12 rev min −1 and 850 to 900 mm rev −1 , respectively.A combination of FR and SS between 0.24 and 0.25 mm rev −1 and 400 to 500 rev min −1 , respectively, results in a valley in the plot that indicates a lower Sr Between 0.1 and 0.12 mm rev −1 FR and 730 to 800 rev min −1 SS, as well as between 0.12 and 0.2 FR and 730 to 900 rev min −1 SS, the contour lines are linear.The linear contour lines demonstrate that in these regions, the change in FR and SS occurs at a constant rate between the contour lines.The contour lines are curved and more concentrically spaced than parallel, which suggests that the FR and SS are interacting to alter the SR more than either factor acting alone would have predicted.The surface plot In figure 13 has a convex shape and illustrates a diminishing return relationship  between the FR, SS, and Sr This implies that the SR increases at a decreasing rate as the FR and SS increase.In other words, as the FR and SS increase, the rate of change of the SR decreases.SR should be kept to a minimum of about 1 μm, and the ideal FR and SS for this are 0.1 mm rev −1 and 900 rev min −1 , respectively.Because the peak is off-center, the plot is not symmetric.The SR is less sensitive to changes in the FR and SS in that area, according to the relatively shallow slope.This indicates that significant changes in the SR may necessitate significant changes in the FR and SS.The independent variables in figure 14 are SS and DC and the response variable is Sr There is one peak that shows that the combination of SS and DC between 400-450 rev min −1 and 0.75-0.8mm, respectively would yield the highest Sr A valley in the plot that denotes a lower SR is from a combination of SS and DC between 830-900 rev min −1 and 0.85-1.1 mm, respectively.In the region of DC between 0.75-1.0mm, and 400-640 rev min −1 SS, the DC and SS change at a constant rate since the contour lines are linear.At lower SR  in regions between 700-900 rev min −1 SS and 0.75-1.25 mm DC, the contour lines are non-linear, and as such the independent variables are changing at a non-constant rate.The contour lines are parallel at higher SR values which suggests that there is no interaction between the SS and the DC and the effect of each on the SR is independent of the other.The contour lines at lower SR values depict the opposite as they are curved and suggest a strong interaction between SS and DC in decreasing Sr In figure 15, the peak shows that the combination of FR and DC between 0.1-0.2mm rev −1 and 1.24-1.25 mm, respectively, would yield the highest SR for mineral oil during the AISI 304 turning operation.A valley in the plot that denotes a lower SR is from a combination of FR and DC between 0.24-0.25 mm rev −1 and 0.75-0.85mm, respectively.In the region of DC between 1.1-1.25 mm and FR between 0.1-0.22 mm rev −1 , the DC and FR change at a constant rate while in regions between 0.75-1.05mm DC and 0.1-0.22 mm rev −1 FR, the contour lines are nonlinear showing that the independent variable change at a non-constant rate.The contour lines are more parallel than curved at higher SR values which suggests that there is no Interaction between the FR and the DC and the effect of each on the SR is  independent of the other.The shape of this surface plot in figure 16 (for mineral oil lubricant) is irregular, showing a diminishing return relationship between the FR, SS, and CT.This suggests that the relationship between the FR, SS, and the CT is non-linear.In this case, more complex models such as polynomial regression, spline regression, or machine learning algorithms can be used to describe the relationship.The desirable optimum minimum value for CT is about 38 0 C and the combination of values for this optimum is 0.25 mm rev −1 and 400 rev min −1 , respectively.The plot is symmetric and that suggests that the CT is equally sensitive to changes in the FR and SS in either direction.In figure 17, the independent variables are SS and DC, and the response variable is CT.The peak of the plot shows that the combination of SS and DC between 850-900 rev min −1 and 1.1-1.25 mm, respectively, would yield the highest CT.In regions between SS of 700-900 rev min −1 and DC of 0.9-1.25 mm, the contour lines are linear, and hence the SS and DC change at a constant rate in this region.The contour lines are parallel at higher CTs which suggests that there is no interaction between the SS and the DC and the effect of each on the CT is independent of the other.

Analysis and discussion of grey relational technique
Both SR and CT are considered to be 'lower is better' characteristics for data pre-processing in the GRA process.The pre-processing of mineral oil and PKO are displayed in tables 6 and 7, respectively.Let the comparability sequences be the results of nine experiments y i 0 (k), i = 1-9, k = 1 [57].All the sequences after data preprocessing using equation (3) are provided in tables 8-11 and are denoted as y o * (k) and y i * (k) for the comparability sequence and the reference sequence, respectively.In tables 8 and 9, respectively, the deviation sequences for mineral oil and PKO are displayed.The distinguishing coefficient ξ can be substituted into equation (4) to produce the GRC.If all the process parameters are of equal weight, then ξ is 0.5.The results of using equations (4)-( 6) to calculate the GRC and GRG values for each trial of the orthogonal array are shown in tables 10 and 11 for minerals and PKOs, respectively.The turning parameters of experiment 7 have the highest GRG for minerals and PKOs.Therefore, out of the nine trials, experiment 7 has the best multi-performance qualities and the best machining parameter setting for achieving the lowest SR and CT at the same time.The average GRG for each level of the turning parameters was calculated using the response table for the Taguchi approach, along with the best turning parameters for CT and SR.The GRG values for each level of the turning settings were calculated using equation (6).The average GRG values for PKO and mineral oil are shown in tables 12 and 13, respectively.Given that the GRG measures the degree of relationship between the comparability sequence and the reference sequence, a higher value denotes a stronger relationship between the two.In essence, superior performance is equal to a higher GRG value regardless of the category of the performance attributes.The level of cutting parameters with the highest GRG value is therefore the best.An asterisk ( * ) next to a level value indicates that turning performance has been  enhanced.The best machining results for the SR and CT will be obtained using a combination of 870 revolutions per minute SS (level 3), a 0.25 millimeter per minute FR (level 3), and a 1.25-millimeter DC (level 3).According to table 12, the maximum and minimum values for the GRG of the turning parameters fall within the following ranges: 0.132 for SS, 0.149 for FR, and 0.046 for DC.To determine which variable has the biggest impact on performance traits, these values are compared.This evaluation will highlight the significance of controllable factors in relation to multiperformance qualities.The variations of GRG values for the L 9 experimental matrix considered in this study for the corresponding process parameter setting are shown in figures 18 and 19.These means of the GRG values are further investigated to establish the ideal process parameter setting for turning AISI 304 stainless steel material using an external threading tool.The variable that could be controlled most effectively had the highest of these values.Here, 0.149 is the highest value possible among 0.132, 0.149, and 0.046.The result shows that, among the turning parameters, the FR has the greatest impact on the multiperformance characteristics.Considering these values, on the other hand, can reveal the significance of the role that each controllable factor plays in the multi-performance characteristics.The following is the order of the significance of the controllable factors to the multi-performance characteristics in the turning process: factor FR, SS, and DC (i.e., 0.149 > 0.132 > 0.046).The most effective factor in performance was the FR.This indicates that the FR had a significant impact on turning performance.Similarly, the optimum machining performance for both SR and CT for mineral oil will be obtained with a combination of 870 revolutions per minute SS (level 3),   13).Among the other turning parameters, the DC has the greatest influence on the multiperformance characteristics.Report of Taguchi analysis (main effect plot for GRG) on each of the cutting lubricant is presented in figures 20 and 21. Figure 20 reports on PKO from which the combination of 870 rev min −1 , 1.25 mm and 0.25 mm rev −1 gave an optimal result for the lubricant.Figure 21 pinpoints a combination of 870 rev min −1 , 0.10 mm rev −1 and 1.25 mm induced optimum for mineral oil.By changing cutting variables like FR, DC, and speed, the second-degree polynomial modeling equation for CT and SR was generated.A multi-regression model with a level of certainty of 95% for both response characteristics.To evaluate the model's feasibility, the value of the determination coefficients (R 2 ) was measured.The significance of the model strengthens as the R 2 value rises or approaches 1. Equations ( 7)-( 14) provide a mathematical pattern for SR and CT based on the results of experiments.The R 2 and S values of equations ( 7)-( 14) are displayed in table 14. S depicts the discrepancy between the fitted values and the raw values that is present.The model more closely mimics the reaction the lower the value of S. The results show that, for some SR and CTs, the model prediction is correct.The range of cutting parameters over that range of SR and CT can therefore be estimated using this information.

Conclusion
In this study, palm kernel oil was mechanically extracted and used as a cutting lubricant during MQL turning of AISI 304 stainless steel using external threading tools.Using the Taguchi L 9 (3 3 ) orthogonal array experimental plan, the effects of palm kernel oil and its equivalent mineral oil on surface roughness and cutting temperature were examined.The following conclusions can be made in light of the research's findings: (i) The palm kernel oil surpassed the mineral oil in terms of surface roughness by 48.2%, according to the test findings.In the case of cutting temperature, palm kernel oil performed better than mineral oil by roughly 0.89%.
(iii) The grey relational analysis showed that, for palm kernel and mineral oils, 870 rev min −1 , 0.25 mm rev −1 and 1.25 mm, and 870 rev min −1 , 0.10 mm rev −1 and 1.25 mm, respectively, are the best spindle speed, feed rate, and depth of cut.
(iv) Feed rate had the greatest impact on the grey relational grade values for both palm kernel oil while the depth of cut had the greatest contribution on depth of cut.

Figure 3 .
Figure 3. External threading tool employed for the experiment.

Figure 4 .
Figure 4. Influence of lubrication environment on SR.

Figure 5 .
Figure 5. Influence of lubrication environment on CT.

Figures 9 -
11 show the impact of cutting parameters on the temperature at which metal cutting fluids are cut.The relationship between SS and CT when PKO and mineral oil lubricants are used is depicted in figure9.For the two machining environments, it was found that the CT increased in order from 410 rev min −1 to 870 rev min −1 .At low and medium SSs, PKO is more effective at lowering the temperature at the chip-tool interface, while mineral oil performs similarly at high SSs.In figure10, the CT decreased as the FR increased abruptly from low to medium to high levels.The slope of the line for PKO is −10.53 and is negative.This indicates that the CT drops by 10.53 degrees Celsius for every unit increase in FR.In other words, the CT drops as the FR rises.The slope of the line for mineral oil is positive and equal to 3.20.This implies that the CT rises by 3.20 degrees Celsius for every unit increase in FR.In other words, the CT rises as the FR does.The effect of various levels of cut depth on the CT was examined for the lubricant under consideration, as shown in figure11.Each lubricant showed a consistent pattern as it progressed through the DC levels; their increments ranged from 0.75 mm to 1.25 mm.It is noteworthy that at 0.75 mm and 1.25 mm of cut depth, respectively, the two had their lowest and highest CTs.When cutting at a deep depth, PKO performs somewhat better than mineral oil.

Figure 6 .
Figure 6.Influence of FR on SR.

Figure 7 .
Figure 7. Effect of SS on SR.

Figure 8 .
Figure 8. Influence of DC on SR.

Figure 9 .
Figure 9. Influence of SS on CT.

Figure 10 .
Figure 10.Influence of FR on CT.

Figure 11 .
Figure 11.Influence of DC on CT.

Figure 12 .
Figure 12.Contour plot of FR, SS, and SR for mineral oil.

Figure 13 .
Figure 13.Surface plot of SS, FR, and SR for PKO.

Figure 14 .
Figure 14.Contour plot of DC, SS, and SR for PKO.

Figure 15 .
Figure 15.Contour plot of DC, FR, and SR for mineral oil.

Figure 16 .
Figure 16.Surface plot of SS, FR, and CT for mineral oil.

Figure 17 .
Figure 17.Contour plot of DC, SS, and cutting temperature for PKO.

Figure 18 .
Figure 18.GRG for the minimum SR and cutting temp.

Figure 19 .
Figure 19.GRG for the minimum SR and cutting temp.

Table 1 .
The experimental condition.

Table 4 .
Effect of cutting parameters on response characteristics.

Table 5 .
The signal-to-noise ratio for response characteristics.
Exp. no.SS (rev/min) FR (mm/rev) DC (mm) SR (μm) CT (°C) 3.2.Influence of cutting variables on SR and CT Figures 6-8 show the impact of cutting variables (FR, SS, and DC) on SR for mineral oil and PKO.Under PKO and mineral oil, figure

Table 6 .
The sequences resulting from mineral oil data preprocessing.

Table 7 .
The sequences resulting from PKO data pre-processing.

Table 8 .
The deviation sequences for mineral oil.

Table 9 .
The deviation sequences for PKO.

Table 10 .
GRC and GRG analysis for nine comparability sequences (mineral oil).

Table 12 .
The GRG response table for PKO.

Table 13 .
The GRG response table for mineral oil.

Table 14 .
Corresponding S and R 2 values.