Skip to Main Content
This paper presents a comprehensive method for eccentricity fault diagnosis in switched reluctance machine (SRM) during offline and standstill modes. In this method, the fault signature is achieved by processing the differential currents resulted from injected high-frequency diagnostic pulses to the motor windings. For feasibility assessment of the proposed method, the correlation between differential current and eccentricity occurrence is analytically demonstrated. A 3-D transient finite-element method is utilized to analyze the proposal, and its results are validated by experimental tests on a prototype machine in the laboratory. According to the obtained results by experimental as well as numerical analysis, a pattern is introduced in order to detect fault presence. In this stage, an algorithm with two methods is proposed for the detection of fault location or faulty phase. Furthermore, the direction of fault is recognized by a simple comparative technique. Next, a normalized indicator is proposed in order to estimate the eccentricity fault severity as well as its occurrence. This indicator is not dependent of the healthy motor information, which makes it reliable and applicable tool for all sizes of SRMs. Finally, a simple procedure based on achieved pattern and indicator is implemented to distinguish the static, dynamic, and mixed eccentricity in the SRM.