Abstract:

 Based on theories of the model reference adaptive system (MRAS) and the Kalman filter, the online inertia identification and state estimation of permanent magnet synchronous motor (PMSM) servo system were  respectively studied for improving the dynamic performance and robustness. In the proposed algorithm, an optimal state estimator based on the Kalman filter was used to provide exact estimation for the rotor speed, rotor position and disturbance torque in a random noisy environment. Also, the MRAS was incorporated to identify the variations of inertia moment real time, and the identified inertia was used to adapt the EKF for better dynamic performance. In addition, the disturbancerejection ability to variations of the mechanical parameters was discussed, and it was verified that the system was robust to the modeling error and system noise. Simulation and experimental results showed that, compared with the M／T method, the proposed technique had better performance in speed resolution, real-time and anti-interference ability.

Key words: model reference adaptive system, inertia identification, Kalman filter, state estimation, permanent magnet synchronous motor