Adaptive Sliding Mode Observer-Based Input Offset Voltage Estimation for PMSM

Abstract The input offset voltage of a permanent magnet synchronous motor (PMSM) carries critical information related to system faults. Accurate estimation of this voltage is crucial for precise fault diagnosis. However, conventional observers for voltage estimation are highly susceptible to variations in operating conditions and external disturbances. This paper proposes an offset voltage estimation method based on an adaptive sliding mode observer (SMO) to acquire input offset voltage signals under diverse fault scenarios. To address both open-circuit faults in power electronic switching devices and non-open-circuit faults occurring during servo motor operation, a state observer with dynamic gain adjustment is designed. This observer overcomes the limitation of conventional observers?their difficulty in balancing convergence speed and anti-disturbance capability under harsh operating conditions?by effectively distinguishing voltage deviations induced by noise from those caused by faults. Lyapunov stability theory is employed to conduct a rigorous convergence analysis of the proposed method, ensuring its stable operation and robustness under a wide range of operating conditions. Simulations conducted on the MATLAB/Simulink platform validate that the adaptive observer retains excellent estimation accuracy and rapid response performance, even in high-noise environments.