关键词: 进化神经网络, 遗传算法, 刀具寿命, 切削参数优化

Abstract: Artificial neural network was introduced to predict cutting tool life, and neural networkbased prediction model of cutting tool life as well as the optimizing selection of machining parameters were proposed. In the prediction process, there were some disadvantages in Back Propagation (BP) algorithm, such as low convergence speed, easily falling into local minimum point and weak global search capability. To settle these problems, a genetic algorithm was used to train BP neural network to replace classical learning algorithms. An evolutionary neural network learning algorithm was developed. Results of simulations and experiments showed that the evolutionary neural network based on genetic algorithm could effectively overcome the shortcoming of falling into local minimum point. This method could obtain higher prediction accuracy. As a result, it provided theoretical basis for the establishment of cutting tool requirements planning, the account of its cost and the selection of machining parameters, as well as reduced the cost in Manufacturing Execution System (MES).

Key words: evolutionary neural network, genetic algorithm, cutting tool life, machining parameters optimization