Optimization of milling process of AISI 4340 Steel for enhanced tool life and surface quality using response surface methodology and bayesian technique

Milling is a widely used machining process in the manufacturing industry employed to remove material from a workpiece. Cutting inserts suffer wear during milling operations, lowering their useful tool life. In order to reduce friction and energy consumption, several nanoparticles (NPs) have been studied as additives for cutting fluids. Furthermore, optimizing milling parameters such as spindle rotational speed, depth of cut, and feed rate is crucial for increasing tool life and improving productivity. In this study, TiO₂ NPs were added to the cutting fluid used for milling AISI 4340 steel. Laboratory experiments in a four-ball tribotester showed that a NPs can reduce the wear scar diameter and surface roughness R_a of steel balls. A Box-Behnken design of experiments was performed to optimize the input milling parameters and NP content to enhance the response characteristics of R_a (related to the surface quality of the workpiece), insert radius (tool wear), and spindle load (related to energy consumption of the process). Response Surface Methodology (RSM) and Bayesian Optimization (BO) technique were used to optimize the milling process. RSM presented R² values of 99%, 51%, and 58% for spindle load, insert wear, and R_a, respectively. Through BO the insert radius obtained an R² of 59% considering spindle load as an input variable, so that it can be utilized in a monitoring system to predict tool wear, therefore enhancing tool life and utilization ratio. The analysis of the relationships and effects among the response variables showed that insert radius is the most important factor for the R_a, and the R² for this BO model was 98%. The Multi-Response Surface Methodology showed that values of N of 723 rpm, d of 0.008 in, f of 12.5 in/min and 0.05 wt.% NPs minimize all response variables together. It was observed that the addition of NPs may reduce response variables due to a mending and load-bearing tribological effect. Finally, this optimization study allows for improved quality of manufactured parts, reduced energy consumption, and increased productivity.