Multi-objective optimization of experimental and analytical residual stresses in pre-stressed cutting of thin-walled ring using glowworm swarm optimization algorithm

In this study, an analytical prediction model was utilized to predict the residual stresses induced during pre-stressed cutting of thin-walled ring, and the results were validated by experiments. General multivariate regression predictive models were presented to characterize the relationship between machining parameters and key characteristic values of residual stresses. These predictive models were used to build multi-objective optimization models to select the optimal machining parameters in process planning decision. The objectives are to minimize the tensile surface residual stress and to maximize the maximum compressive residual stress and processing efficiency and be chosen as related to industrial applications. An adaptive step-size multi-objective glowworm swarm optimization (MOGSO) algorithm was employed in optimizing parameters including cutting speed, cutting feed and pre-tightening torque. Optimization results demonstrate the superiority of the improved algorithm over the traditional MOGSO. The optimum machining parameters calculated from the predicted results were represented in both objective function and decision variable spaces. Further experimental verification results verified the effectiveness of the optimization model.