关键词: 钛合金, 插铣, 机器学习, 多目标优化

Abstract: Based on the experimental work and optimization research of plunge milling of titanium alloys, a hybrid method integrated with machine learning and multi-objective optimization algorithm was proposed herein to optimize the processing parameters considering material removal rate and cutting force as objectives. Spindle speed, radial cutting width, cutting step and feed per tooth were used as test variables. Taguchi method was used to reduce the number of test variable groups. The performance of the model given by machine learning was compared with traditional first-order and second-order regression models. Results show that machine learning has a better performance in prediction accuracies and the distributions of solutions. Four algorithms(MOEA/D, NSGA-Ⅱ, SPEA2, and NSPSO) were used and compared in solving the problem, and NSGA-Ⅱ shows a better comprehensive performance. Finally, the optimal results were compared with the initial reference. Results show that the optimal results may improve the material removal rate and reduce the cutting forces, which may help to achieve the goal of efficient and stable machining.

Key words: titanium alloy, plunge milling, machine learning, multi-objective optimization