基于小波神经网络的主轴热误差预测研究

Spindle Thermal Error Prediction Based on Wavelet Neural Network

以TX1600G镗铣加工中心镗削系统主轴部件为研究对象,针对其热误差问题,提出一种基于小波神经网络的预测方法。首先根据镗铣加工中心主轴部件的结构特点建立其有限元模型,基于该模型进行热-结构耦合分析,进而选取热关键点并获取其样本数据;然后利用小波神经网络建立主轴热误差预测模型,并与BP神经网络预测结果相对比;最后结果表明小波神经网络预测精度高,为该加工中心的主轴热误差预测提供了理论依据,该方法同样适用于其它主轴热误差的前期预测。

Taking the boring spindle system of TX1600G boring-milling machining center as the research object, a wavelet neural network-based prediction method is proposed to solve the thermal error problem. Firstly a finite element model of the spindle is established according to the structural characteristics of the boring-milling machining center, thus the thermal key points are selected and the sample data are obtained after the thermal-structure coupling analysis is processed based on the model above; secondly, with the method of wavelet neural network, the prediction model of spindle thermal error is built up, which compared with the prediction results of BP neural network;finally, the results indicate that the prediction based on wavelet neural network is of higher precision, which provides a theory evidence for the thermal error prediction of the machining center spindle and this method is also applicable to what predicts the spindle error of other types.