Soft computing for sustainable drilling of AISI 316L stainless steel under formulated neem oil minimum quantity lubrication condition

Abstract This paper discusses about process input optimization to obtain desired output characteristics such as Surface Roughness (microns), Thrust Force (N) and Torque (N-m) during drilling of AISI 316 L Stainless Steel under minimum quantity formulated neem oil lubrication condition based on Taguchi Design of Experiments (TDOE), Response Surface Methodology (RSM) and Desirability Functional Analysis (DFA) by varying flow rate (ml.min−1), stand-off distance (mm), flow pressure (Bar) and nozzle exit diameter (mm). Formulated Neem Oil possesses natural lubricating properties that reduce friction and heat generation, thus prolonging the tool life and improving surface finish. Additionally, it is biodegradable and environmentally friendly, making it a sustainable choice for machining operations. From the experimental investigation using TDOE, it was observed that there was considerable improvement in thrust force, surface roughness and torque with modified neem oil as a lubricant. Further, plot for main effects and Analysis of Variance (ANOVA) are successfully used to identify the optimum process input parameters and their percentage of contribution (P%) on output parameters. RSM is successfully used to generate a second order mathematical model which can be effectively used to analyze the process parameters. Further, from Desirability Functional Analysis (DFA), minimum surface roughness (0.34 microns), thrust force (1292.37 N) and torque (14.71 N-m) value were predicted. Finally, Back Propagation Artificial Neural Network (BPANN) analysis has been adopted to predict the surface roughness, thrust force and torque with a minimal error of 1.46%, 0.017% & 0.17%, respectively. The adoption of Neem oil formulations has been successful in improving machining characteristics. Its versatility across an array of machining processes and materials, in tandem with the global momentum toward greener manufacturing paradigms, positions it as a promising lubricant for various machining practices.


Introduction
Work environment safety without compromising the production costs is the fundamental objective of many industries.Modern machining techniques tend to lean towards minimum-quantity lubrication (MQL) during machining operations.A crucial role is played by Minimum Quantity Lubrication while drilling of AISI 316 L Stainless Steels.The MQL approach involves misting or atomizing a very small amount of lubricant in a pressurized air flow that is aimed towards the cutting zone, usually kept between a flow range of (50-500 ml/hour) (Autret & Liang, 2003).Through one or more nozzles, the lubricant is sprayed using an external delivery system.In comparison to the typical amount of coolant utilized in flood cooling, the amount of coolant used in MQL is around 3-4 times lesser.MQL is also referred to as "Spatter Lubrication," "Microlubrication," and "Near-dry machining" (Aronson, 1995;Attanasio et al., 2006).Cutting fluids have largely been chosen for flood cooling applications based on their cutting performance.However, in MQL, lubricant's secondary properties, such as safety, environmental impact, biodegradability, oxidation, and storage stability, are significant (Klocke & Eisenblatter, 1997).It should be noted that in an MQL application, cooling is accomplished by the compressed air that reaches the cutting surface while lubrication is obtained by the lubricant (Wakabayashi et al., 2006).MQL systems can be divided into two categories: externally sprayed and through-the-tool lubrication.In mass production manufacturing, the spindle and tool are typically used to deliver the mist rather than an external nozzle (Filipovic & Stephenson, 2006).In MQL, the lubrication system does not recirculate the fluid.At the application site, practically all of the fluid evaporates.The primary goal was also to reduce cutting fluid consumption and the management costs associated with it.Based on a survey carried out by the Automobile Industry of Europe, lubricant expenses constitute nearly 20% of the overall manufacturing cost (Heinemann et al., 2006).Comparatively, the cutting tool cost accounts for only 7.5% of the total expenses, highlighting the significant disparity in cooling costs (Lung et al., 1995).As per the reports of Occupational Safety and Administration of Health of America (Nouari et al., 2003), and the National Institute for Occupational Safety and Health America (Tasdelen et al., 2008), the acceptable levels for cutting fluid concentration are between (5 mg/m 3 and 0.5 mg/m 3 ), respectively.In U.S. automotive parts manufacturing facilities where conventional flood coolant is utilized, the oil mist level has been estimated to range from 20 to 90 mg/m 3 (U.S.Department of Health and Human Services, 1998).Human exposure to said quantities of cutting fluid has become the leading cause of various health and safety concerns, such as dermatitis, toxicity, respiratory disorders, along with cancer.In recent years, researchers have investigated various lubrication methods for machining stainless steel, including flood cooling, MQL, nanofluid-based MQL, and vegetable oil-based MQL.For instance, a researcher (Deng et al., 2021) found that flood cooling provides better cooling and lubrication, leading to improved surface roughness and tool wear.However, it generates a lot of waste, leading to environmental concerns.On the other hand, MQL uses a small amount of lubricant, reducing the environmental impact and cost of the process (Kamat et al., 2021).Moreover, MQL can provide comparable or better results than flood cooling in terms of surface finish and tool life (Cheng et al., 2021).Vegetable oil-based MQL has also gained attention in recent years due to its eco-friendliness (AlBazi & Zitoune, 2021).In addition to traditional lubrication methods, nanofluids have emerged as a promising lubricant due to their unique properties, such as high thermal conductivity and stability (Diop et al., 2021).Nanofluids can improve the lubrication performance and reduce the cutting forces during machining of stainless steel (Alves et al., 2020).Moreover, researchers have explored the use of artificial intelligence techniques, such as neural networks, to optimize the lubrication parameters for machining (Tumuluru et al., 2020).Citation numbers will remain in their original positions.Conventional machining processes have long relied on mineral oil-based cutting fluids; however, their replacement has now been mandated.A remarkable solution to this challenge has emerged in the form of vegetable oils, which have gained significant traction as eco-friendly green cutting fluids (GCF).Unlike their petroleum or chemical-based counterparts that contribute to harmful pollution, vegetable oils exhibit exceptional biodegradability and low toxicity, making them a promising alternative for safeguarding both the environment and the well-being of workers.In contrast to mineral, synthetic, and semi-synthetic oils, known for their detrimental effects on the environment and health, vegetable oils present a viable option (Shashidhara & Jayaram, 2010;Zheng et al., 2021).These oils possess several advantageous properties, including high flash point, desirable viscosity, excellent lubrication capacity, low volatility, and cost-effectiveness, leading to ecologic and economic benefits for the manufacturing industry.Their ability to form a robust and thick film during machining also proves advantageous, as it effectively cools the machining zone while reducing friction coefficient, cutting forces, tool wear, and heat generation (Agrawal et al., 2014;Debnath et al., 2014;Khan & Dhar, 2006).The realm of research exploring vegetable oils for machining applications has been extensive, encompassing various oils with distinct physicochemical properties such as cottonseed (Agrawal et al., 2014;Duan et al., 2020Duan et al., , 2021)), coconut (Chinchanikar et al., 2014;Ghatge et al., 2018;Ghuge & Mahalle, 2016;Gunjal & Patil, 2018;Lawal et al., 2012;Padmini et al., 2015Padmini et al., , 2019;;Xavior & Adithan, 2009), soybean (Ghuge & Mahalle, 2016, 2017;Gunjal & Patil, 2018;Raj et al., 2016), canola (Belluco & De Chiffre, 2004;Elmunafi et al., 2015;Gunjal & Patil, 2018;Padmini et al., 2019), castor (Elmunafi et al., 2015;Khunt et al., 2020;Pereira et al., 2017;Suresh et al., 2017), palm (Abd Rahim & Sasahara, 2011;Mahadi et al., 2017;Rahim & Sasahara, 2017;Suresh et al., 2017), groundnut oil (Ghuge & Mahalle, 2016;Suresh et al., 2017), and sunflower (Ghuge & Mahalle, 2017;Khunt et al., 2020;Pereira et al., 2017).Embracing vegetable oils and reaping their associated benefits has triggered novel manufacturing approaches, propelling the concept of green manufacturing forward.Further, a study by researchers (Padhan & Dash, 2020) has investigated power efficient machining of hardened steel using Graphene nano-particle enriched coolant with MQL and have ascertained the sustainability and low power consumption of Minimum Quantity Lubrication.
The literature regarding the application of natural lubricating oils during stainless steel machining to enhance sustainability and performance characteristics exhibits certain noteworthy gaps.One such gap pertains to the scarcity of comprehensive studies that exclusively investigate the utilization of natural lubricants, such as Neem oil or vegetable-based oils, in the context of MQL machining of stainless steel.Additionally, there are only a few researches which address the precise environmental advantages linked to these lubricants, including reductions in waste production and energy consumption.Moreover, opportunities remain for more in-depth explorations into the optimization of oil formulations and application parameters, aiming not only to meet sustainability objectives but also to enhance various aspects of machining performance, tool longevity, surface finish quality, and chip management.The systematic examination of these research voids would serve to propel sustainable machining practices within the stainless steel domain.Hence, this paper discusses optimization of surface roughness, thrust force and torque during drilling of AISI 316 L under minimum quantity formulated neem oil lubrication condition based on TDOE, RSM, DFA and ANN by varying flow rate (ml.min −1 ), stand-off distance (mm), flow pressure (Bar) and nozzle exit diameter (mm).

Methodology
The drilling experiments were performed using an AMS Vertical Machining Centre (Model: SPARK, BANGALORE), with max feed rate (20 m.min −1 ) and a max speed (6000 rpm).Solid carbide drill bits (dia 8 mm), parabolic flute (point angle 135°), chisel edge angle (135°), a helix angle of 35°, and a relief angle of 10° were utilized under MQL condition (as shown in Figure 1).For MQL application, neem oil was mixed with Methanol in a 3:1 ratio, and a 2% KoH catalyst was added.The mixture was then heated and separated from glycerol to obtain Methyl ester of neem oil (MENO).Finally, MENO was heated again with distilled water to remove the alcohol content.This modified neem oil lubricant was then supplied to the interface zone between chip and tool using a custom MQL setup under different flow rates (ml.min−1 ), stand-off distances (mm), flow pressures (Bar), and nozzle exit diameters (mm), with constant drilling parameters such as a 10 mm drill diameter, spindle speed of 1250 m.min −1 , and feed of 100 mm.min −1 .A comparison of the physical properties of raw neem oil and formulated neem oil is presented in Figure 2. The AISI 316 L grade steel in the form of round bars has been used, and its chemical composition and mechanical properties are displayed in Figure 3. Thrust force and torque generated during drilling of AISI 316 L stainless steel were measured by a 9257 BA KISTLER Dynamometer, and surface roughness was measured by a Talysurf Surtronic 3+ surface roughness measuring equipment.Microstructural changes were analyzed using a Trinocular inverted metallurgical microscope, JEOL JSM-6380LA Analytical Scanning Electron Microscope, and OLYMPUS BX53M System microscope.Each experiment has been repeated three times, and during each drilling condition, three measurements were taken and respective averages have been recorded.
The orthogonal array (Taguchi L 27 ) has been obtained from MINITAB software-Version 15 (Jayashree et al., 2018;Shetty et al., 2021Shetty et al., , 2019)).ANOVA was employed as an analytical tool to investigate the primary design parameters that significantly influence the quality of characteristics under study.Additionally, it aimed to determine the percentage contribution of each input process parameter.The drilling test parameters and their respective levels selected for experimentation are detailed in Table 1.The experiments were conducted following the L 27 Orthogonal Array.To  gain insights into the quality characteristics beforehand, response surface methodology (RSM) was adopted.RSM facilitates understanding the process output parameters across various experimental domains by employing a second-order model (Hegde et al., 2022;Karthik et al., 2022;Shetty et al., 2022).The application of the second-order model is particularly useful when the response function is either unknown or exhibits nonlinearity.The levels and factors of RSM are shown in Table 2.The DFA is used for optimizing multi-process output variables (Ahmad et al., 2017;Candioti et al., 2014;Castillo et al., 1996;Derringer & Suich, 1980;Harrington, 1965;Kushwaha et al., 2013;Natesh et al., 2023;Ryad et al., 2018).In this study, the DFA was used to optimize the drilling process output variables.
Figure 4, illustrates the complete design of a multilayer perception (MLP) model, which employs the Back Propagation Algorithm (BPA) for training the network.The Back Propagation Artificial Neural Network (BPANN) is a supervised learning algorithm that is widely utilized for regression tasks and pattern recognition.It is a feed-forward neural network that implements a gradient descent (GD) learning algorithm to modify the network's weights and biases and to minimize the error of prediction.BPANN operates in three stages: the forward pass, the backward pass, and  weight updating.During the forward pass, the input data is propagated through the network layer by layer, and the output is computed using the activation function at the output layer.Further, during the backward pass, the cost function is used to compute the error between the predicted output and the actual output.The error is then propagated backwards through the network to update the weights and biases of the network.The chain rule is employed to calculate cost function derivative with respect to the biases and weights.In the weight updating stage, the weights and biases of the network are adjusted to minimize the error between the predicted output and the actual output.This iterative process is repeated until the error is minimized, or a predefined number of iterations is reached.BPANN has several benefits, such as its ability to learn complicated nonlinear relationships between input and output variables, and its ability to generalize well to new input data.However, it necessitates a significant amount of training data and may suffer from over fitting if the network is overly complex or the training data is noisy.In this experiment, BPANN comprises input neurons (4) that represent flow rate (ml.where AVG x -weighted sum of input of x th processing elements (Equation 2).
where o y -output of y th processing element; (x, y)-processing elements; w bi -bias weight; w xy -weight of connection; Further, three different learning algorithms LM (Levenberg-Marquardt), GD (Gradient Descent), SCG (Scaled Conjugate Gradient) are compared with each other for best results by calculating respective RMSE (Equation 3), R 2 (Equation 4) and MEP (Equation 5) (Table 3).Training and Test set for BPANN are presented in Tables 4 and 5.

Results and discussions
Machining quality assessment on process output parameters such as surface roughness, thrust force and torque during machining of AISI 316 L stainless steel under Formulated Neem Oil Minimum Quantity Lubrication (FNOMQL) has been discussed.

Surface roughness
The characteristics of surface roughness analysis on AISI 316 L SS after drilling is of significant importance in determining the performance of the component.This is because irregularities on the surface can serve as nucleation sites for cracks or corrosion, which can compromise the structural integrity of the component.Additionally, rough surfaces are more prone to wear and possess higher co-efficient of friction, which can negatively impact their functionality.Therefore, this section aims to discuss the surface roughness after drilling AISI 316 L SS under FNOMQL.Based on the experimental results presented in Figure 5, it can be concluded that an increase in the rate of flow and inlet pressure of the formulated neem oil lubricant leads to a decrease in surface roughness.This is due to the ability of the lubricant to effectively wash away the chips and prevent them from smearing the drilled area.Furthermore, the lubricant can reduce thermal distortions that occur due to the heat generated at the cutting zone during drilling.The reduction in surface roughness achieved by the lubricant can be attributed to its ability to maintain a clean cutting environment and prevent the accumulation of debris and other contaminants.Therefore, the use of FNOMQL as a lubricant during drilling of AISI 316 L SS can improve the surface quality of the component, resulting in improved performance and durability.
The presented information in Figure 6 reveals the impact of various lubrication conditions on the machined surface.It is evident from the microscopic images that the use of a higher flow rate of formulated neem oil as a cutting fluid results in a smoother surface finish with reduced irregularities compared to the other cutting fluids.Additionally, implementing Minimal Quantity Lubrication (MQL) with a reduced amount of formulated neem oil can potentially lead to changes in the microstructure of AISI 316 L stainless steel during the drilling process.It is important to note that reduced lubrication levels can result in higher cutting temperatures, which, in turn, can alter the grain structure and promote the formation of finer grains.Such alterations can enhance the material's mechanical properties and improve its resistance to wear and corrosion, which are crucial factors in many industrial applications.Therefore, the findings suggest that the use of a higher flow rate of FNOMQL with a reduced amount of lubricant can positively impact the surface finish and mechanical properties of AISI 316 L SS during the drilling process.
Based on the analysis of the % contribution of all the factors (Table 6), it can be concluded that flow rate (39.008%) contribution towards surface roughness is maximum.Hence, it is crucial to consider flow rate conditions as an important factor.Additionally, stand-off distance (P = 10.063%),flow pressure (P = 7.42%), nozzle diameter (P = 2.75%) and their respective combinations also have a significant impact on the surface roughness of the specimen under formulated neem oil MQL.
From the surface roughness main effects plot (Figure 7) it can be derived that the selection of flow rate (15 ml.min −1 ), stand-off distance (30 mm), flow pressure (15 Bar), and nozzle diameter (4 mm) have resulted in the best possible combination of parameters for achieving the lowest surface roughness value (0.3372 microns) during drilling of AISI 316 L SS with formulated neem oil MQL.Furthermore, the second-order response surface equation which is a function of processing parameters such as flow rate, stand-off distance, flow pressure, and nozzle diameter.Equation 6 can be utilized to represent the surface roughness (in microns).From the analysis carried out with a 95% confidence level (Table 7), it is evident that, the calculated F value is greater than the table value of F (F 0.05,14,14 = 13.63).The table shows that there are also 14 degrees of freedom for this source of variation, and the sum of squares (SS) is 795.5.The adjusted sum of squares (Adj SS) is also 795.53, and the adjusted mean square (Adj MS) is 56.82.The last row in the table shows the Total variation in the response variable, which is the sum of the regression and residual variation.The table shows that there are 28 degrees of freedom for the total variation, and the sum of squares (SS) is 11,636.7.In summary, this ANOVA table indicates that the regression model is significant statistically, and the independent variables have a considerable effect on the response variable.The residual error is relatively small compared to the variation explained by the regression model.
The contour and surface plots (Figure 8) illustrate the impact of flow rate and flow pressure on surface roughness when stand-off distance (30 mm) and nozzle diameter (4 mm) are kept constant.Furthermore, the results obtained for surface roughness using the desirability function approach (Figure 9) demonstrate that the optimal drilling conditions for AISI 316 L stainless steel include a flow rate of 9.25 ml.min −1 , a stand-off distance of 30.0 mm, a flow pressure of 1.025 Bar, and a nozzle diameter of 2.0 mm.The minimum surface roughness generated under these conditions was 0.34 microns during the drilling process.

Thrust force
In the field of metal machining, the analysis of the thrust force variable has been a key component in assessing machinability characteristics.In Table 8, the P% for different factors selected has been evaluated to understand their impact on the thrust force during the machining of AISI 316 L stainless steel under FNOMQL.The results reveal that flow rate conditions (ml.min −1 ) have the highest contribution of about 37.638%.Therefore, it is essential to consider the flow rate conditions during machining of AISI 316 L SS under formulated neem oil MQL.In addition to flow rate, stand-off distance (mm) (P = 9.247%), flow pressure (Bar) (P = 7.527%), nozzle diameter (mm) (P = 8.062%), and their respective  combinations also have a considerable effect on the characteristics of thrust force on AISI 316 L SS after drilling under formulated neem oil MQL.
From Figure 11 representing the main effects plot for thrust force, the selection of flow rate (5 ml.min −1 ), stand-off distance (10 mm), flow pressure (0.1 MPa) and nozzle diameter (4 mm) have  The second order response surface representing the thrust force (N) (Equation 7) is generated.
From the analysis carried out with a 95% confidence level (Table 9), it is evident that, the calculated F value is greater than the table value of F (F 0.05,14,14 = 14.29).The table shows that there are 28 degrees of freedom for the total variation, and the SS is 12,664.8.In summary, this ANOVA table indicates that the regression model is statistically significant, and the independent variables have a significant effect on the response variable.The residual error is relatively small compared to the variation explained by the regression model.
The contour and surface plots (Figure 12) present the effect of flow rate and flow pressure on thrust force while stand-off distance (20 mm) and nozzle diameter (2 mm) are kept constant.From the results obtained for thrust force (Figure 13) using DFA it is observed that optimal drilling conditions for AISI 316 L SS flow rate is set as 9.25 ml.min −1 , stand-off distance 30.0 mm, flow pressure 15 Bar and nozzle diameter 2.0 mm.Further, from Figure 13, it is evident that minimum thrust force generated using desirability function approach was 1292.37 N during drilling of AISI 316 L SS.

Torque
Precise measurement of drilling torque plays an important role in optimizing the drilling process, as it helps in designing fixtures, predicting tool wear, and understanding the friction characteristics that lead to heat generation and temperature changes between the tool and workpiece.In this section, we will discuss the torque induced during drilling of AISI 316 L SS under formulated neem oil MQL.
Experimental data (Figure 14) indicates that torque is significantly reduced at higher flow rates and lower lubrication pressure.This reduction is due to the formation of a thin film of highly viscous formulated neem oil lubricant that reduces the friction between the tool and This phenomenon leads to a decrease in torque generated during the drilling process.However, at higher pressure of the lubricant, the formation of a thin film between the tool-work interfaces has not been observed, which leads to increased friction and torque generation.
The use of formulated neem oil MQL during drilling of AISI 316 L SS reduces the torque generated, and this can be attributed to the lubricant's ability to reduce friction and form a thin film between the tool and workpiece.The results indicate that higher flow rates and lower lubrication pressure result in reduced torque generation, which can help in the optimization of the drilling process.The knowledge of the torque generated during the drilling process is crucial for understanding the process and improving the efficiency of the drilling operation.
Based on the analysis of the % contribution of all the factors (Table 10), it can be concluded that flow rate (23.25%) contribution towards surface roughness is maximum.Hence, it is crucial to consider flow rate conditions as an important factor.Additionally, stand-off distance (P = 11.25%),flow pressure (P = 9.94%), nozzle diameter (P = 5.5%) and their respective combinations also have a significant impact on the surface roughness of the specimen under formulated neem oil MQL.
where A = Flow rate (ml.min −1 ); B = Stand-off distance (mm); C = Flow Pressure (Bar); D = Nozzle Diameter (mm).From Figure 15 indicating main effects plot for torque, the selection of flow rate (5 ml.min −1 ), stand-off distance (10 mm), flow pressure (0.1 MPa) and nozzle diameter (4 mm) have resulted the best possible combination to get the lowest torque value of 14.703 N-m during drilling of AISI 316 L SS under formulated neem oil MQL.
The second order response surface representing the torque (N-m) (Equation 8) can be expressed as a function of processing parameters such as flow rate (ml.min −1 ), stand-off distance (mm), flow pressure (Bar) and nozzle diameter (mm).
From the analysis carried out with a 95% confidence level (Table 11), it is evident that, the calculated F value is greater than the table value of F (F 0.05,14,14 = 10.38).The table shows that there are 28 degrees of freedom for the total variation, and the sum of squares (SS) is 12,664.8.In summary, this ANOVA table indicates that the regression model is statistically significant, and the independent variables have a significant effect on the response variable.The residual error is relatively small compared to the variation explained by the regression model.
The contour and surface plots (Figure 16) present the effect of flow rate and flow pressure on torque while stand-off distance (20 mm) and nozzle diameter (2 mm) are kept constant.From the results obtained for torque (Figure 17) using DFA it is observed that optimal drilling conditions for AISI 316 L SS flow rate is set as 10.65 ml.min −1 , stand-off distance 10.0 mm, flow pressure 0.944 Bar and nozzle diameter 2.0 mm.Further from Figure 17, it is evident that minimum torque generated using desirability function approach was 14.703N-m during drilling of AISI 316 L SS.

Chip formation
Chip formation during drilling of AISI 316 L SS can be influenced by the use of Minimal Quantity Lubrication (MQL) with formulated Neem oil.The flow rate of the lubricant can impact the type of chips that are produced.The morphology of chips formed under different lubricant flow rates 5-15 ml.min −1 with constant pressure (15 Bar), stand-off distance (30 mm) and nozzle diameter (4 mm) is presented in Figure 18.Further, from the figure, it can be concluded that under higher lubricant flow rates helical chips were formed which indicates less wear during drilling, whereas in the case of minimal lubricant flow rate, long ribbon-like chips are formed because of increased wear between tool-work interfaces (Sultan et al., 2015).With a low flow rate of lubricant, the cutting  forces can be high, leading to the formation of long, ribbon-like chips.These chips are characterized by their length and thin cross-section, which can increase the risk of tool breakage and reduce the efficiency of the cutting process.With a higher flow rate of lubricant, the cutting forces can be reduced, leading to the formation of helical chips.chips have a coiled shape and are more tightly packed, reducing the risk of tool breakage and increasing the efficiency of the cutting process.The increased lubrication also helps to reduce friction and heat generated during the cutting process, further improving the performance and prolonging the life of the tool.Thus, the choice of lubricant flow rate for MQL during drilling of AISI 316 L SS can have a significant impact on the type of chips formed and the efficiency and performance of the cutting process.The optimal flow rate will depend on various factors, including the type of tool, cutting conditions, and desired outcome.

Validation of TDOE and RSM with BPANN
The results of the experimental drilling of AISI 316 L SS with formulated Neem oil MQL have been validated using a Back Propagation artificial neural network.This network was trained with a learning factor of 0.6 and 8000 cycles, which resulted in accurate predictions of the drilling performance.The errors obtained using the Back Propagation algorithm were compared with those obtained using the Response Surface Methodology.In terms of surface roughness, the error obtained using Back Propagation was 1.46%, while the error using Response Surface Methodology was 0.08%.For the thrust force, the error obtained using Back Propagation was 0.017%, while error obtained using Response Surface Methodology was 0.22%.Finally, for torque, the error obtained using Back Propagation was 0.17%, while the error obtained using Response Surface Methodology was 0.38%.
These results indicate that Back Propagation can be effectively applied to predict the drilling performance of AISI 316 L SS with formulated Neem oil MQL.The lower errors obtained using Back Propagation suggest that it can provide more accurate predictions than Response Surface Methodology.This can be valuable information for industries looking to optimize their drilling processes and improve their efficiency.The use of artificial neural networks, such as Back Propagation, can provide a more comprehensive understanding of the relationship between process parameters and drilling performance.

Conclusions
During process optimization of drilling of AISI 316 L SS under formulated neem oil MQL with L 27 array following conclusion are drawn: • The drilling of AISI 316 L SS under formulated neem oil MQL was optimized using an L 27 orthogonal array to determine the best process output variables for surface roughness, thrust force, and torque.
• The surface roughness of drilled holes in AISI 316 L SS decreased as the rate of flow and inlet pressure of the lubricant increased.The fluid flow rate had the highest contribution of about 39.008% to the reduction of surface roughness.
• The lowest thrust force values were derived at the highest flow pressure and rate of flow of the formulated neem oil lubricant.Flow rate conditions had the highest contribution of about 37.638% to the reduction of thrust force.
• At higher flow rates and lower lubrication pressure, torque reduced significantly due to the formation of a thin film of highly viscous formulated neem oil lubricant that reduces the friction between tool and workpiece.Flow rate conditions had the highest contribution of about 23.25% to the reduction of torque.
• The analysis of chips showed that helical chips were formed under higher lubricant flow rates, while long ribbon-like chips were formed in the case of minimal lubricant flow rate.
• The TDOE, RSM and BPANN models accurately predicted the surface roughness, thrust force, and torque with minimal error.
• The utilization of MQL Formulated Neem Oil in industrial contexts carries profound and diverse practical implications.Its application presents industries with an environmentally sustainable and ecologically responsible lubrication solution, aligning with the escalating emphasis on ecological stewardship.Neem oil's intrinsic properties exhibit remarkable capacity to diminish machining expenditures through the prolongation of tool life, enhancement of surface quality, and attenuation of thermal energy generation.This effectuates an upswing in overall productivity and financial efficiency.Additionally, its inherent antimicrobial attributes contribute to the cultivation of a more sterile and secure machining environment.

Future scope
The adoption of Neem oil formulations has been successful in improving machining characteristics.Its versatility across an array of machining processes and materials, in tandem with the global momentum toward greener manufacturing paradigms, positions it as a promising lubricant for various machining practices.This entails further research and development studies aimed at finetuning Neem oil formulations for specific applications, rendering it an indispensable asset for industries pursuing streamlined and ecologically responsible machining methodologies.

Figure 1 .
Figure 1.Experimental set up for formulated neem oil MQL assisted drilling.

Figure 2 .
Figure 2. Properties and chemical make-up of AISI 316L SS.

Figure
Figure 3. Properties of formulated neem oil.
min −1 ), stand-off distance (mm), flow pressure (Bar), and nozzle exit diameter (mm), with output neurons (3) that represent surface roughness (microns), thrust force (N), and torque (N-m).There is one hidden layer with 27 neurons.The training was conducted with normalized input values.The class used for standardization of input data is Standard Scaler and further using train_test_split function the data is split into training and testing sets.Data with 20 sets of experimental results is used as training set while dataset with 9 sets of experimental results is used as test set to validate the BPANN model.Log' Sigmoid is used as a transfer function (Equation 1).

Figure
Figure 5.Effect of (a) surface roughness v/s flow rate with stand-off distance (30mm) and nozzle diameter (4mm); (b) surface roughness v/s stand-off distance with flow rate (15ml/ min) and nozzle diameter (4mm); (c) surface roughness v/ s nozzle diameter with flow rate (15ml/min) and stand-off distance (30mm).
The thrust force induced during drilling of AISI 316 L Stainless Steel under formulated neem oil MQL has been examined.The experimental results presented in Figure10indicate that the lowest thrust force values are achieved when the formulated neem oil lubricant is applied at the highest flow pressure and flow rate.These results are consistent with the literature and demonstrate that the formulated neem oil applied through MQL at high pressure and flow rate during drilling of AISI 316 L Stainless Steel can significantly reduce thrust force.In contrast, when lubricant flow rate and pressure are low, ploughing of the surface occurs, leading to an increase in the thrust force values.Specifically, under flow rate conditions of 15 ml.min −1 , flow pressure of 15 Bar, stand-off distance of 10 mm, and nozzle diameter of 4 mm, the thrust force values are increased due to the ploughing effect.

Figure
Figure 10.Effect of (a) thrust force v/s flow rate with standoff distance (30mm) and nozzle diameter (4mm); (b) thrust force v/s stand-off distance with flow rate (15ml/min) and nozzle diameter (4mm); (c) thrust force v/s nozzle diameter with flow rate (15ml/min) and stand-off distance (30mm).

Figure
Figure 14.Effect of (a) torque v/s flow rate with stand-off distance (30mm) and nozzle diameter (4mm) constant; (b) torque v/s stand-off distance with flow rate (15ml/min) and nozzle diameter (4mm) constant; (c) torque v/s nozzle diameter with flow rate (15ml/ min) and stand-off distance (30mm).

Figure
Figure 15.Main effects plot for torque (N).

Figure
Figure 16.Contour plot and surface plot of torque (microns).

Figure
Figure 19.Validation of TDOE experimental surface roughness value using RSM and BPANN.

Figure
Figure 21.Validation of TDOE experimental torque value using RSM and BPANN.

Table 8 . ANOVA for S/N ratios (thrust force) Source DF F P P (%)
resulted in the best combination to get the lowest thrust force value of 1292.37 N during drilling of AISI 316 L SS under formulated neem oil MQL.