Development of an expert system for optimal design of the grinding process

The physical or empirical modeling of the grinding process and the effects of its parameters on the workpiece quality is sophisticated. This is due to the extreme complexity of the process. So far, no remarkable success could be made by the proposed models to achieve a reliable and effective design and control of the process. This article introduces an expert system to enhance the design of the grinding process. A pilot system was built considering three main grinding outputs, including surface roughness, material removal rate and normal grinding force. Based on the primary experimental results, the system suggests the proper grinding and dressing parameters to obtain the desired surface roughness with the highest possible material removal rate and least normal force. The analysis is based on the regression correlation development and the “Non dominated Sorting Genetic Algorithm II” optimization method. The validation tests conducted in three different surface roughness and normal force ranges proved the reliability and effectiveness of the proposed expert system. The surface roughness was predicted with less than 7% deviations from the experiments. The predicted normal grinding forces, in most cases, were in good agreement with the experiments with higher than 75% accuracy. Although the wheel balancing issues at the highest cutting speed (v_c = 35 m/s) caused large deviations between the predicted and measured forces.