数控机床温度测点优化及热误差预测方法研究

Research on optimization of temperature measuring point and thermal error prediction method of CNC machine tools

热误差补偿技术是提高机床加工精度经济有效的方法,确定最佳关键温度测点布置位置和数目将极大提高机床热误差模型的精度和鲁棒性。针对一台立式加工中心,进行了机床热误差测量试验,根据其温度场,提出了模糊聚类与信息论相结合的方法,寻找最佳温度测点布置位置。该方法根据温度变量间的相似性,对温度变量聚类分组,然后利用互信息法对组内变量单独寻优,实现温度测点优化布置,最后利用多元线性回归分析建立机床热误差预测模型。在VMC1165立式加工中心进行了试验验证,温度测点减少为4个,热误差模型的拟合最大残差降低到5μm以内,相对于其他方法进一步提高机床热误差预测精度。

Thermal error compensation technology is a cost-effective way to improve the machining ac- curacy. The key point to determine the optimum temperature measurement arrangement positions and quantity will greatly improve the accuracy and robustness of the machine tool error model. A thermal experiment is per- formed on a vertical machining center. Based on the temperature field of the machine tool, this paper presents a method combining fuzzy clustering analysis and information theory to find the best combination of temperature measuring point＇ s position. Based on the similarity between the variables temperature, the temperature meas- uring point has been classified and optimized in each group by information theory. Optimization for temperature measuring points has been arranged. Finally, thermal error prediction model is built based on optimized tem- perature variables utilizing multiple linear regression analysis. Results show that the number of temperature measuring points reduces to 4, the maximum residual can be reduced to lower than 5 μm and compared with other methods, the thermal error prediction accuracy of machine tool can be further improved, which proves the validity of the method.