Publication

Abstract : Surface roughness is a key factor in machining, influencing functional performance and dimensional accuracy; to enable its predictive control, this study develops an offline Digital Twin for cryogenic precooled turning to predict the surface roughness. Spindle speed, feed rate, and depth of cut were considered as input parameters, while surface roughness (Ra), tool wear, and material removal rate (MRR) were selected as output responses. Taguchi L27 orthogonal array was used for experimental design. ANOVA was performed for each response, and non-significant terms were removed to develop accurate regression models. These models generated 990 data combinations used to train and test an ANFIS model in MATLAB Simulink. The offline Digital Twin uses tool wear and MRR to predict Ra. Validation with nine additional trials showed strong agreement between predicted and experimental Ra values. Furthermore, a Genetic Algorithm (GA) was applied to identify the optimal machining conditions, achieving a minimum Ra of 0.4578 µm at a spindle speed of 332.36 rpm, a feed rate of 0.63 mm/rev, and a depth of cut of 0.2 mm. The proposed Digital Twin offers a data-driven solution for surface quality prediction and parameter optimization, with strong potential for smart manufacturing and Industry 4.0.