A Machine Learning Genetic Algorithm (ML-GA) Approach to Optimize Process Parameters of Trochoidal Toolpath During Hybrid-Electrochemical Magnetorheological (H-ECMR) Finishing

Hybrid-Electrochemical Magnetorheological (H-ECMR) Finishing process is an advanced surface finishing technique that produces surface roughness (Ra) up to a few nanometers required for the implants. In this study, duplex stainless steel (DSS), a biomaterial constituting ferrite and austenite in equal proportion, is considered as the workpiece. DSS is widely used for its high corrosion resistance compared with other grades of stainless steel biomaterials. Furthermore, optimization of the process parameters for the trochoidal, a high-speed finishing toolpath, is carried out to achieve optimum surface roughness parameters during H-ECMR finishing. The surface roughness parameters (i.e., average Surface roughness (R_a). Kurtosis (R_ku), and Skewness (R_sk)) are attained from the integration of the Design of Experiment (DOE), Response Surface Methodology (RSM), and Machine Learning Genetic Algorithm (ML-GA). DOE is utilized to plan the experiments based on Central Composite Design (CCD) using input process parameters of the H-ECMR finishing process. The regression equation is developed with the RSM, and its optimization study is performed with ML-GA. This paper aims to analyze the impact of the process parameters for the H-ECMR finishing process on different surface roughness parameters.