PUBLICATIONS

This paper proposes an optimal design of an efficient switched reluctance generator (SRG) to replace permanent magnet synchronous

generator (PMSG) such a wind power system. The proposed 1kW SRG is designed on the basis of a D²L method as a basic design of 

SRG. The optimal design of SRG is performed by using a genetic algorithm (GA). In order to evenly distribute design variables in the 

design range, Latin hypercube sampling (LHS) such as design of experiment (DOE) method is used for the accuracy and reliability of 

sample point. kriging approximation modeling method is afterward used to consider the exact nonlinear characteristic in this paper. In 

addition, the torque ripple reduction method of SRG is proposed in this paper. In order to reduce the torque ripple of SRG, some 

torque sharing functions (TSF) had been researched. However, these methods only were proposed at the low-speed range. Therefore, 

the torque ripple reduction method is proposed at the low and high-speed range. In order to improve the reliability of switched 

reluctance machine (SRM) driver, two methods are proposed in this paper. Firstly, two switches of each phase are operated in 

complementary manner and operation pattern is analyzed for the fault diagnosis method. Secondly, the switching signal of SRM driver 

is analyzed for the SRM operation. The fault diagnosis method is proposed by using analyzed pattern results of switching signal. The 

simulation and experiment are performed in an SRM drive system, and the results are presented to verify the validity of the proposed 

method.