基于综合智能计算的数控系统误差补偿

摘要

提出一种基于综合智能计算的数控系统位置控制方法,其基本思想是,在现有的数控系统内部嵌入一个神经网络的小型系统(采用插件形式嵌入),利用综合智能计算,得出系统补偿误差,实现对数控系统误差补偿的智能控制.综合智能计算通过应用粗集理论对决策表进行属性约简,利用主元分析进行数据压缩及特征提取,以降低BP网络的输入维数,缩短网络的训练时间,并利用蚁群算法对BP网络参数进行全局搜索,实现网络快速收敛到全局最优值.试验仿真表明,此方法可有效提高系统补偿精度和实时性.%This paper is a study of error compensation for machine tools based on integrated intelligent computation. The primary focus was on integrated intelligent computation approach to develop a BP neural network embed in a digital control system of a machine tool. Data compression and feature extraction was realized by way of application of RST and principal component analysis, which reduced BP network input dimension and reduced network training time. Taking advantages of ant colony algorithm, a global optimum of error compensation for machine tools was obtained by a BP network convergence. Positioning error compensation capabilities were tested using industrial standard equipment and procedures. The results obtained show that this approach can improve compensation precision and real time of error compensation for machine tools effectively.