Fault Diagnosis of Five-Phase Fault-Tolerant Permanent-Magnet Motor Based on Principal Component Neural Network

In this paper, a new fault diagnosis method for a five-phase fault-tolerant permanent-magnet (FTPM) motor by using a compact method is proposed. The key is to create a neural network based on principle component analysis (PCA). For a current signal of a five-phase FTPM motor system, PCA theory is used to extract the main element from the fault sample data. It realizes optimum compressed of fault sample data and simplifies structure of neural network in fault diagnosis. Speed and precision of the fault classification are enhanced. The obtained results verify the effectiveness of the proposed method.