The neural network inverse decoupling control of bearingless synchronous reluctance motor

A bearingless synchronous reluctance motor (BSRM) is a nonlinear, strong-coupled complicated system. The linearization and decoupling control is a key to stable operation and practicability for BSRM. Based on the research of BSRM, the accurate mathematical model of BSRM is given in this paper. Reversibility of the model is proved, using the neural network inverse system. By using static neural network and integrator to structure BSRM neural network inverse system, the system was decoupled into two independent second-order linear displacement subsystems and a first-order linear rotor subsystem, and design the regulator to control each of the subsystems, so as to be easy to achieve the dynamic decoupling. Simulation and experimental results show the good static and dynamic decoupling performance.