Hybrid Model Flux Observer for PMSM Sensorless Control System Considering Parameter Disturbances and Measurement Noise

The hybrid model flux observer integrates the advantages of the voltage and current models to improve its control performance. However, at low speeds, the observer is more sensitive to the noise present in the measurement signals, causing the observation results to deviate from the actual values, and the situation of parameter mismatch will pose a serious challenge to the stability of the control system. Through mathematical derivation in this paper, the influence degree of resistance and inductance parameters on the observation angle error at different rotational speeds is obtained, and it is proposed to introduce an improved extended state observer to compensate for the system parameter disturbances to enhance the system robustness. By constructing a state equation considering measurement noise and analyzing the Bode diagram in this paper, it is concluded that the improved extended state observer, compared to the traditional observer, can reduce the sensitivity to high-frequency measurement noise and improve the estimation performance of low-frequency disturbances, and its stability conditions are derived. Finally, comparison experiments of the starting process, load mutation, parameter mutation and different flux observer methods under different load conditions are designed. The experimental results show that this control strategy, in the low-speed domain, can enhance the system’s resistance to parameter disturbances, improve the measurement accuracy of the system, and effectively suppress the measurement noise existing in the system, verifying the effectiveness and feasibility of this control strategy.