基于灰色理论预处理的神经网络机床热误差建模

Modeling for Machine Tool Thermal Error Based on Grey Model Preprocessing Neural Network

为最大限度减少热误差对数控机床加工精度的影响,尝试结合灰色理论和人工神经网络各自对数据处理的优点,提出一种基于灰色理论预处理的神经网络机床热误差补偿模型。在一台处于实际加工状态的数控车床上进行试验,采用数字式温度传感器测量经过优化选取的对热误差有关键影响的机床构件和加工环境的温度数据,采用非接触式位移传感器获得机床加工热误差数据,在不断调整灰色模型数据序列长度及神经网络权值、阈值的基础上,最终建立热误差补偿模型。通过与传统灰色模型和神经网络进行对比分析及试验论证表明,该补偿模型具有对原始温度和热误差数据要求低、计算简便、预测精度高、鲁棒性强等优点,可用于各种复杂实际加工场合中的数控机床热误差实时补偿。

In order to eliminate the influence of thermal error on machining precision,a novel method combining the data processing merit of grey model（GM） with that of artificial neural network（ANN） is proposed for thermal error modeling in machine tools.bining gtion Relevant temperature and thermal error data of a turning machine are measured respectively through digital temperature sensors and non-contact displacement sensors.A series of experimental data are used to establish the thermal error compensation model by means of adjusting the length of data sequence in GM and the weight and threshold value in ANN.The results show that the new model performs better than the traditional GM and ANN model.It is less-requiring for the original data,convenient in calculation,good in fitting and precise in prediction under a variety of working conditions.So the new model is more suitable for the complex industrial applications.