Stator Interturn Fault Detection in BLDC Motors: A Signal-Processing-Based Method

Abstract

In this letter, we present a signal-processing-based method for detecting stator interturn faults in brushless direct current (BLDC) motors. Utilizing current probes as measurement sensors, the proposed approach starts by transforming the current waveforms into a synchronous rotating reference frame ($dq$-axis) using the Park transformation matrix. Faults are then identified through a combination of the Savitzky–Golay smoothing filter, a modified cumulative-sum method, and a novel ratio-based index. The proposed technique is both simple and efficient, demonstrating high adaptability to various BLDC motor conditions without requiring changes to its threshold settings. The method is evaluated using current signals collected from a laboratory BLDC motor test bench. Experimental results confirm its high speed and accuracy. In addition, a comparative analysis with other similar methods highlights the effectiveness and robustness of the proposed approach across different operating scenarios.