The selection of key temperature measuring points for the compensation of thermal errors of CNC machining tools

Statistically, up to 40% of machining errors are given with thermal errors and the proportion is as high as 70% in precision and ultra-precision machine tools. A compensation technique with creating a compensation model of the thermal error based on the relationship between temperature fields and thermal errors of machine tools is one of the most effective methods to enhance accuracy of machine tools. The key temperature measuring points have to be selected before building the thermal error compensation model because they has a great influence on the accuracy and robustness of error compensation model. In this paper, a new method to select the key temperature measuring points is presented. This method involves two phases: firstly using stability analysis of thermal error sensitivity to select the measuring points with strong correlation to thermal error; and then employing fuzzy cluster analysis to further reduce the number of the key temperature measuring points. To evaluate the performance of this method, a thermal error compensation model is built based on BP neural network to validate the selected temperature measuring points on the milling machining centre CR5116. [Received 30 April 2016; Accepted 16 November 2016]