Research On Thermal Performance Test And Robust Modeling Technology Of NC Machine Tools

With the progress of social science and technology,the industry is facing the development of intelligent manufacturing.The society has higher and higher requirements for the accuracy of CNC machine tools.A large number of studies have shown that the thermal error of machine tools accounts for 40 % to 70 % of the total error.The higher the precision of machine tools,the greater the impact of thermal error.Therefore,the thermal performance test and error reduction and control of CNC machine tools are important links to improve machining accuracy.At present,the thermal error model is mainly used to predict and compensate the thermal error of the machine tool,so as to improve the machining accuracy of the machine tool.However,most of the current modeling techniques cannot be effective for a long time and thus affect their engineering applications.Therefore,the research on machine tool thermal error robustness modeling technology is of great significance.In this paper,the Vcenter-55 three-axis CNC machining center is taken as the research object.Combined with its structural characteristics,the research and development of the thermal performance test system,the design of the thermal performance test experiment scheme,the monitoring of the temperature rise change of the key heat source,the variable optimization of the temperature measurement point participating in the modeling,the robust modeling technology and the error compensation technology are studied in depth.The main research work is as follows :1)According to the international standard IS0 230-3: 2020 and the structural characteristics of the machine tool,a thermal performance test experiment was designed.A total of 13 batches of thermal performance test experiments and data analysis were carried out.2)Optimize the selection of temperature measuring points,reduce the collinearity interference between temperature data,reduce the cost of use,and ensure the strong correlation between the measuring point variables and the thermal error.In this paper,KMEANS clustering set grey correlation method and fuzzy clustering combined with grey correlation method are used to select temperature sensitive points.3)The temperature sensitive points of each batch of experiments were calculated by using the above KWG method and FWG respectively.The multiple linear regression model(MLR)model and ridge regression model(RR)were established by using the selected temperature sensitive points as variables,and the experimental data on the time span of nearly half a year were predicted.The root mean square error(S)was used as the evaluation standard of the model,and the data preprocessing method of ’ symmetric mapping ’ was proposed to further improve the prediction accuracy and robustness of the model.In this method,the prediction accuracy of the MLR model is increased by about 46.6 %,and the prediction accuracy of the RR model is increased by about 43.1 %.4)CNC machine tool thermal error through the mechanical origin offset technology error compensation control,experimental verification,the machine tool thermal error is reduced by more than 35 % and the accuracy is maintained for at least 3 months.