System Configuration of Tri-pole Magnetic Bearings

This paper describes the whole system configuration for a tri-pole VMB (vertical magnetic bearing) especially focused on the key issue of eccentricity at high rotation speed. To simplify the complexity of the tri-pole VMB, system modeling is developed together with current reduction, power stage reduction and commutation reduction. To meet the six-region control algorithm, methods regarding the intelligence approaches are employed rather than the traditional PID compensation. The three developed controllers by intelligence approaches are introduced, i.e. 1) NFIS (neuro-fuzzy integrated system) 2) ANN (adaptive neuro network) model controller and 3) ANN optimal controller. This paper presents various intelligence controllers for the VMB control based on an entire analysis on the system modeling in the sequence of mechanical field, magnetic field, dynamic field and control. The whole approach provides a technical thinking whenever facing a challenging of a new magnetic bearing system, and is helpful for the engineers to implement it