QUALITY AND PRODUCTIVITY IMPROVEMENT IN THE MILLING OF POLYOXYMETHYLENE CO-POLYMER (POM C) USING RSM, SVM AND GENETIC ALGORITHM.

This paper aims to conduct an economic and environmental machining procedure (dry condition) during face milling of the Polyoxymethylene Co-polymer. In this context, an experimental study based on Taguchi's design was conducted to develop empirical models using Response Surface Methodology (RSM) on one hand and Support Vector Machine (SVM) for regression on the other hand. The ANalysis of VAriance (ANOVA) is conducted to determine the contribution and the significance of each cutting parameter on the surface quality and productivity (MRR). The obtained models are being compared to determine the most efficient approach. The last part is to find the optimum cutting combination by using Genetic Algorithm (GA) optimization based on SVM models, whether to minimize surface roughness Ra, or for composite objective to improve the quality and to increase the productivity. The results show that feed per tooth (fz) is the most affecting parameter on Ra followed by depth of cut (ap) and then the cutting velocity (Vc). SVM was more robust than RSM with less deviation error. [ABSTRACT FROM AUTHOR]