Diagnosis of Inter-Turn Short-Circuit Incipient Fault in Permanent Magnet Synchronous Motors Using Input Current on the Power Side

Past research has shown that completing diagnosis in the incipient fault stage of low-turn, high-resistance short circuits in permanent magnet motors is still a challenge. Meanwhile, the motor signals employed in the established technology route are of poor practicality for motor systems with increasing integration. For these reasons, this article proposes an interturn short-circuit fault diagnosis method utilizing power-side input current and channel attention mechanism. First, this article derives harmonic features of the power-side input current under fault conditions. Then, in order to sufficiently exploit the spatial and temporal scale features of the incipient fault data, this article develops a residual model with an improved channel attention mechanism for interturn short-circuit fault diagnosis. Finally, a 2.3 kW permanent magnet motor has been used for experimental validation. The results show that the fault information can be effectively transmitted to the power-side input current, and the proposed method can accurately diagnose the interturn short-circuit incipient fault. Furthermore, through the feature explainability analysis of the proposed model, this article also finds that the region of the input current fluctuation is favorable for diagnosis and deserves attention.