LAUSR

Implementation of fault diagnosis for power converter is critical to improve the reliability of electric vehicle (EV) with switched reluctance motor (SRM) drive. Conventional fault diagnosis methods are usually suitable for a single chopping mode or a certain control strategy. To avoid the problem, a virtual current coefficient-based diagnosis method for power transistors is proposed considering different chopping modes, control strategies, and fault cases simultaneously. The multipoint current detection method is proposed to obtain more current information, which is capable of providing remarkable fault features. The sensor current variations under various fault conditions are analyzed. The fault is first detected by identifying the error between the real sensor current and the value calculated with the mixed logic model. Three virtual current coefficients are introduced to obtain the fault location principles. The fault types and fault power transistors are recognized according to the difference between the real current coefficients and the virtual current coefficients. The applications of the fault diagnosis method are extended with the proposed scheme. The method does not require complex calculation, and thus has a good dynamic performance. Experiments on a three-phase 12/8 SRM drive validate the effectiveness of the proposed fault diagnosis method.