Establishment of micropit diameter prediction models based on the support vector machine optimization

Abstract

Many studies on microtextures have focused on the effect of laser processing parameters on microtextured morphology from the experimental point of view, and not on its influence on the physical field and stress field around the microtextures. This has resulted in the lack of accurate predictions on micropit diameters, thus reducing the processing efficiency. In order to address these problems, the effects of laser processing parameters on the precision machining, physical field, and stress field of micropit textures for a ball-end milling cutter were studied by experiments and simulation. The optimum laser processing parameters include laser power of 70%, scanning speed of 1700 mm/s, and scanning of seven times. Based on the optimized data of a support vector machine, the prediction models of micropit diameters for laser machining were established. The errors of the models are less than 10%, thus verifying their accuracy. This study can serve as a theoretical reference for the efficient processing of titanium alloys.