AI-Enabled Prediction of the Thermal Expansion for High-Speed Motorized Spindles in Real World Scenarios

Up to 75% of the geometric errors in milling processes can be traced back to the thermal expansions in high-speed motorized spindles due to changing operation conditions. In this study, we investigate various machine learning approaches to predict the thermal error and hereby improve the milling accuracy. First, we introduce an industrial dataset of milling processes with varying operation conditions. Second, we propose a hybrid model to combine physical domain knowledge with deep learning to reduce the amount of data required for successful prediction of the thermal error. Last, we evaluate the performance of the proposed models and benchmark our hybrid approach against established machine learning models. On unseen data, the best model of our study achieves a R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> score of 0.96.