基于IA-PSO-BP模型的电主轴热误差预测方法

针对电主轴在运作时因为温升而产生热误差的问题，提出一种基于免疫粒子群优化BP神经网络（IA-PSO-BP）的电主轴热误差预测模型。通过测量电主轴在工作过程中的温升以及热位移,获取建立预测模型所需的数据，使用IA-PSO-BP模型在MATLAB中建立热误差预测模型，并与未经过优化的BP神经网络所建立的模型进行测试对比。结果显示，经过优化的BP神经网络对热误差的补偿能力高达98.4%，和当前工程常用的BP神经网络相比，平均预测误差下降了62.6%，预测误差的均方差下降了66.4%，可见其预测精度得到了显著提升。

A Thermal Error Prediction Method of Electric Spindle Based on IA-PSO-BP Model

A thermal error prediction model of electric spindle, based on BP neural network optimized by immune particle swarm optimization (IA-PSO-BP), was proposed to ddress the thermal error caused by temperature rise in the operation. The data to establish the prediction model was obtained through measuring the temperature rise and thermal displacement of the electric spindle in the working process, and the IA-PSO-BP model was used to establish the thermal error prediction model in MATLAB. Compared with the model established by BP neural network without optimization, the results show that the optimized BP neural network’s thermal error compensation ability is as high as 98.4%, compared with the commonly used BP neural network, the average prediction error is down about 62.6%, mean square error of prediction error has fallen by 66.4%, obviously the prediction accuracy has significantly improved.