Thermal Error Prediction of Ball Screws Based on Partial Least Square Method with Consideration of Friction Works

In this paper, thermal error based on Partial Least Square (PLS) method is proposed. This algorithm has the capability of feature extraction, and also has robust prediction accuracy for various working conditions. In order to improve the accuracy of the model prediction, the work dissipated by friction is integrated to the model which aims to provide operation information during manufacturing. Therefore, the model framework has to be modeled in segments which could simulate the theoretical modeling approach. In addition, the prediction accuracy is also improved by adding the thermal displacement of the rear end of screw as another new feature using the eddy current sensor. Considering the actual application in the real world, the working conditions are designed for the actual machining situation. The results show that the prediction accuracy is improved by adding the new features and the model could achieve higher accuracy compared to other models.