Prediction of Pareto Optimal Solutions for Multiple Responses in Cylindrical Grinding

Owing to the complexity of grinding operation it is very difficult to predict the optimal machining conditions which results in smooth surface finish, accurate geometric measurements and higher production rate. Optimization of external cylindrical grinding operation will be a solution for selecting the suitable grinding parameters in a cost-effective way. In this work, experiments are carried out according to Taguchi L27 orthogonal array and statistical models based on regression analysis are developed for various responses. These models associate the grinding parameters (work speed, feedrate and grinding depth) with machining performances (metal removal rate, and surface roughness). Using these models, Pareto optimal solutions are found out based on Multi-Objective Genetic Algorithm (MOGA) which is a popular optimization technique. Finally, the results of the proposed technique are experimentally verified.