Nonlinear Internal Model Control for Bearingless Induction Motor Based on Neural Network Inversion

The bearingless induction motor is a nonlinear, multi-variable and strong-coupled system. For this system, a novel internal model control strategy based on neural network αth-order inverse system theory is proposed to realize the decoupling control. By cascading the αth-order inverse model approximated by the dynamic neural network with the original system, the nonlinear bearingless induction motor system is decoupled into four independent pseudo-linear subsystems, that is, two radial displacement subsystems, a speed subsystem and a rotor flux subsystem. Then, the internal model control method is introduced to the four pseudo-linear subsystems to ensure the robustness and anti-jamming ability of the closed-loop system. The effectiveness and superiority of the proposed strategy are demonstrated by simulation and experiment.