Temperature-sensitive point selection and thermal error modeling of spindle based on synthetical temperature information

Thermal errors are the important factors that affect machining accuracy. It is effective to improve the accuracy by establishing the error model and compensating for the errors. In this paper, a temperature-sensitive point selection method based on synthetical temperature information (STI) is proposed, and the thermal error model using support vector regression (SVR) optimized by the whale optimization algorithm (WOA) is established. Firstly, the STI matrix is constructed by combining the temperature value information, the temperature shape information, and the relationship between temperatures and errors. Multiple cluster validity indexes (CVI) are used to determine the optimal number of clusters. Secondly, the STI matrix and the optimal number of clusters are used for fuzzy C-means clustering (FCM), and the correlation coefficient is used to select the temperature-sensitive points. Thirdly, the SVR model optimized by the WOA based on STI (S-WOA-SVR) is established. Finally, the S-WOA-SVR model is verified at different speeds and compared with other traditional methods, such as the SVR model optimized by the WOA based on the traditional clustering method (T-WOA-SVR) and the SVR model optimized by the GA based on STI (S-GA-SVR). The results show that the S-WOA-SVR model has higher accuracy and robustness. Compared with the T-WOA-SVR model and the S-GA-SVR model, the RMSE of the S-WOA-SVR model is reduced by 52.3% and 46.6% for the thermal elongation in Z direction δz, respectively.