Detection of Interturn Short-Circuit Fault and Demagnetization Fault in IPMSM by 1-D Convolutional Neural Network

In this paper, we propose a method to diagnose an interturn short-circuit fault (ISCF) and a demagnetization fault (DF) in an interior permanent magnet synchronous machine (IPMSM) using 1-D convolutional neural network (1-D CNN). In addition, the proposed method detects even when two faults occur at the same time. The 1-D CNN is used to automatically extract features from raw signals and use them to detect and classify the faults. The proposed method uses only three-phase current signals, so does not require additional sensors, which is cost efficient. Experimental results from a IPMSM with faults demonstrate that the proposed approach diagnoses the ISCF, DF, and even the fault in which both faults occurred simultaneously.