Theoretical Analysis And Application Of Robust Thermal Error Compensation Model For CNC Machine Tools

As a symbol of national manufacturing and the level of comprehensive national strength,the status of high-grade CNC machine tools in modern industry is increasing.In all kinds of error sources of machine tools,the proportion of thermal error reaches 40%~70%,so controlling and reducing the thermal error is very important for improving the working accuracy of precise machine tools.From the domestic and foreign researches,the thermal error compensation technology mainly focuses on three aspects,includes temperature-sensitive points selection,thermal error modeling and the thermal error compensation,and it has formed a relatively complete theoretical system and compensation implementation technology.However,there are still some key problems which affect its wide application and need to be solved urgently,such as the effect of collinearity among the temperature variable during selecting temperature measuring points,the lack of models’ forecasting robustness and so on.Aiming at the defects of thermal error compensation technology,some theories and methods about temperature-sensitive points,thermal error modeling and compensation are put forward,besides,a large amount of experimental data are analyzed.Finally,the thermal error compensation experiment is verified by the actual machine tool.The main contents of this paper are as follows:1)One robust thermal error modeling method of CNC machine tools is put forward,it combines the temperature-sensitive points optimization method based on gray correlation degree and the modeling algorithm based on principal component regression analysis,and deals with the strategies of high weight,low coupling and less layout which should be followed during the temperature-sensitive points selection together.This method not only simplifies the complexity of measuring points selection,but also improves the model’s forecasting accuracy and robustness.2)According to the traditional temperature-sensitive points optimization algorithm named fuzzy clustering combines grey relational degree,the thermal error data of the machine tool from the multi quarters in one year are studied,the variation characteristics of temperature-sensitive points selected by this method and its influence on model’s precision are discussed.3)The models’ accuracy and robustness characteristics of multiple linear regression algorithm,distributed lag of times serious and principal component regression algorithm are studied,and the application range of the three models in thermal error compensation is summarized.In addition,a method for improving model’s accuracy based on principal component regression analysis and distributed lag is proposed as well.4)The implement method of thermal error compensation technology is introduced,includes CNC machine tools,thermal error integrated measurement system,optimal selection of temperature-sensitive points,the establishment of mathematical thermal error model,thermal error compensation and accuracy evaluation.And the results show that the compensated effect is remarkable.