Model-Based Robust Fault Diagnosis of Incipient ITSC for PMSM in Elevator Traction System

Stator winding inter-turn short circuit (ITSC) faults are among the most common issues in the permanent magnet synchronous motor (PMSM). Timely diagnosis of incipient ITSC faults is crucial in order to prevent more severe failures. However, in model-based incipient fault diagnosis processes, the intrinsic uncertainties associated with the load and parameters of the elevator traction motor (ETM) inevitably introduce disturbances to the observer residuals, thus amplifying the challenge of extracting fault features. Exploring how to separate these uncertainties from the residuals and enhance sensitivity to incipient ITSC faults is an area that has not been investigated in current model-based diagnostic strategies. In this article, we propose a model-based method for diagnosing incipient ITSC faults in PMSM, specifically addressing the influence of uncertainties in load and motor parameters during elevator operation. First, we analyze the impact of parameter mismatches on the state estimation error of the traditional nonlinear unknown input observer (NUIO). To mitigate these effects, we combine the NUIO with least squares estimation (LSE) to develop an improved NUIO (I-NUIO) that effectively decouples load disturbances and compensates for parameter mismatches. Subsequently, we employ a second-order generalized integrator (SOGI) as the fault indicator, utilizing the negative-sequence component of the fault characteristic terms as the fault characteristic value. The simulations and experimental results demonstrate its capability for detection of incipient ITSC faults in PMSM, while improving robustness of the diagnosis outcomes.