Skip to Main Content
This paper presents a new method for noninvasive diagnosis of static, dynamic, and mixed eccentricity faults in switched reluctance motors (SRMs). This method makes it possible to precisely determine the eccentricity fault features. The proposed signature in this algorithm is based on the analysis of produced current with a particular variation pattern. It is theoretically shown that the occurrence and growth of the fault level cause an increase in amplitude of the produced current signature (PCS), which can be employed to diagnose the fault and extract its details. In this method, first, occurrence of the fault and its location are detected by utilizing eccentricity level detection pattern. Afterward, a new index to precisely determine the fault level is introduced. Then, a novel strategy based on the proposed pattern and index is offered to detect the type of eccentricity as well as the exact location of the faulty phase. For this purpose, an SRM under eccentricity fault is modeled using three dimensional transient finite element method (TFEM). This precise model considers all complex motor geometry, nonlinear characteristics of the motor, end effects and axial fringing effects. A test bed in the laboratory is established to perform various measurements on the eccentric motor with different fault levels. The experimental results validate the numerical analysis outcomes. Fault detection and analysis in SRM demonstrate that the algorithm presented can assure the reliability of detection as well as the sensitivity of eccentricity fault in a SRM.