基于QGA算法优化支持向量机回归的数控机床主轴热误差建模研究

Research on Thermal Error Modeling of NC Machine Spindle Based on Support Vector Regression Optimized by QGA Algorithm optimization

针对精密数控机床主轴的热误差的实时监测反馈问题,提出了一种用于主轴热误差的建模方法。该方法利用QGA(Quantum Genetic Algorithm)寻优算法和支持向量机回归方法的复合建模方式,建立了机床主轴热误差回归模型。并通过搭建主轴热特性测试平台,采用聚类方法筛选主轴的温度敏感点,将采集到的热特性数据用于构建的热误差建模中。实验结果表明,该方法在机床主轴热误差预测中,残差值小于1.5μm,能够为主轴热误差的闭环控制过程提供准确反馈。

Aiming at the problem of real-time monitoring and feedback of thermal error of precision NC machine tool spindle,a modeling method for spindle thermal error is proposed.This method utilizes a composite modeling method of QGA(quantum genetic algorithm)optimization algorithm and support vector machine regression method to establish a regression model for the thermal error of machine tool spindle.By building a spindle thermal characteristic testing platform and using clustering methods to screen temperature sensitive points of the spindle,the collected thermal characteristic data is used in the thermal error modeling of the construction.The experimental results show that the residual value of the method in predicting the thermal error of machine tool spindles with residual value than 1.5μm could provide accurate feedback for the closed-loop control process of spindle thermal error.