基于偏相关分析的数控机床温度布点优化及其热误差建模

NC Machine Tool Temperature Measuring Point Optimization and Thermal Error Modeling Based on Partial Correlation Analysis

提出了一种基于偏相关分析的数控机床温度布点优化方法。数控机床热误差建模一般采用多元线性回归方法,该方法中由于自变量之间的相互作用,各自变量与因变量之间的相互关系不再与简单相关系数所反映的情况完全吻合。使用偏相关分析对温度变量进行优化选择,实现了温度测点优化布置,并建立了数控机床热误差的多元线性回归优化模型,提高了热误差模型的精确性和鲁棒性。

This paper presented a kind of NC machine tool thermal error measurement point optimization method based on partial correlation analysis.The thermal error modeling of NC machine tool commonly uses multiple linear regression method.In multiple linear regression modeling,because of the interaction among the independent variables,the relationship among the independent variables and the dependent variables are no longer perfect matched by the simple correlation coefficient.Here the partial correlation analysis was used to select optimal temperature variables,optimal layout of the temperature measuring point was achieved,and a multiple linear regression optimization model for NC machine tool thermal errors was established.The accuracy and robustness of the error modeling are improved.