基于神经网络的数控机床丝杠热评价模型

Thermal evaluation model of NC machine tool screw based on Neural Network

对在使用年限中的数控机床滚珠丝杠进行热评价,提出了基于支持向量机(support vector machine,SVM)与BP神经网络的数控机床滚珠丝杠热评价模型。采集同一类型、不同年限数控机床的滚珠丝杠温度变化数据,利用粒子群优化支持向量机(PSO-SVM)、BP神经网络以及分数阶粒子群优化BP(FPSO-BP)神经网络分别建立热评价模型并且仿真验证,验证结果显示PSO-SVM与FPSO-BP神经网络准确率都高于90%。模型的提出为数控机床的设计使用和故障诊断提供了思路和方法。

For the thermal evaluation of CNC machine tool ball screw in service life,a thermal evaluation model of CNC machine tool ball screw based on support vector machine(SVM)neural network and BP neural network was proposed.The ball screw temperature data of the same type and different years of CNC machine tools were collected,and the thermal evaluation models were established and simulated by using particle swarm optimization support vector machine(PSO-SVM),BP neural network and fractional order particle swarm optimization BP(FPSO-BP)neural network.The simulation results show that the accuracy of PSO-SVM and FPSO-BP neural network is higher than 90%.The model provides ideas and methods for the design,use and fault diagnosis of CNC machine tools.