Prediction of CNC Machine Tool Cutting Energy Consumption by BP Neural Network

10.6567/IFToMM.14TH.WC.OS20.005

S. H. Sun；T. S. Wu；C. H. Yang；Y. J. Cheng；C. H. Kuo

Green Production ； Energy Consumption Restrain ； Neural Network

Proceedings of the 14th IFToMM World Congress

14th-6（2015 / 11 / 11）

58 - 63

英文

The international machine tool producer is toward green production. The high-efficiency, low pollution, low energy consumption, energy recovery and reuse of resources are developed and machine tools are expected to be included in the energy consumption restrain. By this trend, a study of prediction of machine tool cutting energy consumption is proposed. Spindle rotation velocity, feed rate, cutting depth of machining are modeled as the input parameters and the corresponding energy consumption is modeled as the output parameter. The relationship between the input and output parameters are established based on BP neural network. From the study results, the average error between the neural network output and the measured value is 1.8% for the trained data set and 4.9% for the untrained data set. These two small errors show the great capability of the neural network on the function mapping between spindle rotation velocity, feed rate, cutting depth and energy consumption.

1. 肖彬(2016)。基于卷积神经网络的癫痫病预测模型。交通大學資訊科學與工程研究所學位論文。2016。1-32。