Thermal Error Analysis of Machine Tool Spindle Based on BP Neural Network

Abstract

With the continuous improvement of the accuracy of machine tools, the proportion of the thermal error of machine tools in the total error is increasing. In this paper, the thermal error of horizontal machining center is analyzed by finite element simulation, a three-dimensional model is established in SolidWorks, and some details (such as bolt holes) are simplified and imported into ANSYS Workbench to determine the heat generation model, heat dissipation model, convection heat transfer coefficient and other boundary conditions. On the basis of the temperature field, the thermal-structural coupling analysis is carried out, and the temperature cloud field of the whole machine tool is obtained through the analysis, which provides a theoretical basis for the experimental design. Then, the data are divided into four categories by fuzzy cluster analysis, and finally, the thermal error model is established by BP neural network based on time series. It provides a theoretical reference for the compensation of thermal error.