Optimisation Multi-Objectif des Paramètres de Coupe en Tournage Dur.

Abstract

In this work, we have developed a methodology for developing a multi-objective optimization code using an NSGA-II Hybrid Genetic Algorithm, in order to choose the optimal cutting parameters in turning of 100Cr6 bearing steel with a cBN cutting tool on a graphical interface created in MATLAB. The fitness functions expressing the three components of the cutting force are given by modeling the cutting experimental data with the Multiple Linear Regression cutting. The proposed model is based on the exploitation of NSGA-II (Non-dominated Sorting Genetic Algorithm). This algorithm is based on a classification of individuals into several levels in the sense of Pareto. It uses an elitist approach which allows saving the best solutions found in previous generations. Hybrid approaches have yielded very good results. The best individuals are allowed to reproduce. The Genetic Algorithm ends if the solution satisfies the stop criterion, that is, the population will converge when more than 95% of the individuals in the population share the same value of the "fitness" evaluation function. The case study allowed us to prove the effectiveness of the proposed model. The result is a set of optimal solutions (Pareto front) which offers the user many degrees of freedom and readability for the choice of a solution even more personalized to his needs. The quality of the solutions obtained gives us a glimpse of the real possibilities of industrial applications.