Application of Machine Learning in Predicting the Force of Radial Passive Magnetic Bearing

This paper investigates the application of machine learning methods for predicting the interaction force between ring-shaped permanent magnets in radial passive magnetic bearings. A semi-numerical approach based on the application of fictitious magnetic charges is used to produce datasets for training three different models: Random Forest, Polynomial regression, and Kernel Ridge regression. The findings highlight the potential of machine learning as an effective alternative to traditional force calculation methods.