点击次数：

发表时间：2022-10-10

发表刊物：大连理工大学学报

所属单位：机械工程学院

期号：6

页面范围：814-818

ISSN号：1000-8608

摘要：To predict and control cutting force effectively is very important for machining process to achieve high quality and low cost. Based on the feed-forward multi-layer neural networks, trained by the error back-propagation (BP) algorithm, the predicting cutting force program by using oriented-object language Visual C++ has been built up. In order to make calculating fast and accurate, two methods of optimization, the conjugated gradient and Quasi-Newton, have been applied to the training process of the network to avoid the main problems of supersaturation and local minimum. The experimental data of milling and grinding of two kinds of materials which are difficult to machine have been applied to the program. The results indicate that the maximum relative errors obtained by the traditional empirical formula method and the artificial neural network method are 24.9% and 2.01% respectively. The approach to the prediction of cutting force with the help of artificial neural network is valuable for application.

备注：新增回溯数据