Soft Sensing Modeling of Magnetic Suspension Rotor Displacements Based on Continuous Hidden Markov Model

Aiming at the problem that rotor displacements in magnetic bearing system are difficult to be accurately detected, a soft sensing method based on continuous hidden Markov model (CHMM) is proposed. First, the structure and operation principle of a 3-degree-of-freedom hybrid magnetic bearing (3-DOF-HMB) is described, and its nonlinear mathematical model is derived by using equivalent magnetic circuit method. Then, combining the well prediction ability of CHMM, a position prediction model is built by collecting representative current-displacement data. The accuracy of established model is promised by comparing the predicted results with the rest collected data. Finally, a 3-DOF-HMB system based on the CHMM is established for the simulation, and the mean squared error is taken as model evaluation index to compare predictive ability of the proposed model and other self-sensing models, showing that the proposed model has higher accuracy. The experiment results verify the well performance of anti-interference.