Comprehensive Modeling Of Geometric-Thermal Error Of Machine Tool Linear Axis Based On Experimental Method

With the continuous improvement of China’s industrialization level,there is a higher demand for the accuracy of components in fields such as aviation,aerospace,automotive,and optics,which in turn puts forward higher requirements for the machining accuracy of machine tools.Therefore,in the field of precision machining,machine tool error compensation has become an important research direction,hoping to economically and effectively improve the machining accuracy of machine tools through error compensation.The foundation of error compensation implementation is to establish a high-precision and strong robustness error prediction model.This article takes the three-axis CNC experimental platform as the research object,explores the causes of thermal errors in CNC machine tools and the mechanism of thermal deformation of machine tools.It models the comprehensive positioning errors of CNC machine tools under different temperature changes,aiming to provide an error prediction model with high prediction accuracy,strong robustness,and more suitable for variable working conditions,in order to improve the efficiency and stability of error compensation.The main work of this article is as follows:The experimental modeling method is used to model the comprehensive positioning error of the machine tool.The geometric positioning error term is fitted by polynomial,and the thermal error term is modeled by data-driven model.The focus of thermal error modeling lies in the selection of temperature sensitive points and the establishment of error models.For the selection of temperature sensitive points,sensors are arranged in advance at the main heat source of the machine tool to collect temperature data based on experience,and group search method is used to select them.This article uses correlation coefficient method and maximum tree fuzzy clustering method to cluster temperature data.Multiple regression significance analysis and sampleAnalyze the thermal error sources of CNC machine tools and explore their impact on machine tool errors,namely the mechanism of thermal deformation of the machine tool.Obtain comprehensive positioning error data of the machine tool through experiments.From the shape and slope of the comprehensive positioning error curve,it can be concluded that the shape and slope of the positioning error curve are caused by different factors.The shape of the positioning error curve is mainly determined by geometric positioning errors,while the slope of thermal deformation is caused by thermal deformation.Therefore,the error separation method is used to separate the comprehensive positioning error into geometric error term and thermal error term,model them respectively,and then superimpose them to get the comprehensive positioning error model.The feasibility of this modeling method is verified through experiments.determination coefficient are proposed as effectiveness evaluation indicators for evaluating clustering results,and the effectiveness of the clustering algorithm is verified through experiments.For the establishment of the thermal error model,the traditional multiple regression model has the defect of multiple collinearity,which affects the prediction accuracy and robustness of the model.To solve this problem,this paper proposes to use the principal component regression modeling,which eliminates the multiple collinearity among the temperature variables through the feature that the principal components are not related to each other,and improves the prediction accuracy and robustness of the error model.By comparing with multiple regression models and BP neural network models,it is proven that the principal component regression model has higher performance and is more suitable for establishing thermal error models.