A New Approach to Fault Diagnostics for Permanent Magnet Synchronous Machines Using Electromagnetic Signature Analysis

This paper proposes a novel approach to health monitoring and multifault detection in permanent magnet synchronous machines using direct flux measurement with search coils. Unlike other spectrum-based fault detection schemes, only the fundamental frequency component of the measured voltage is utilized for fault detection. Therefore, the performance of the proposed scheme is not affected by nonstationary speed or harmonics introduced by the power supply. In addition, location of interturn short circuits and direction of static eccentricity can be detected, which have never been done by any other scheme. In spite of the invasive nature of the technique, it is very suitable for mission-critical applications and emerging applications such as off-shore wind turbines, hybrid vehicle technology, and military applications, where early detection of faults is of paramount importance. 2-D simulations using finite element analysis have been presented to validate the proposed method under different operating conditions. Experimental introduction of stator interturn short circuit, demagnetization, and static eccentricity has been discussed, and the proposed scheme is experimentally implemented to examine its effectiveness.