Design and multi-objective optimization method of switched reluctance machines

The non-sinusoidal phase current and high saturated magnetic field in switched reluctance machine (SRM) make its design very difficult, and traditional design methods are complex, require professional knowledge are difficult to obtain optimal scheme. In this paper, the original scheme is obtained by traditional method, and the influencing modes of geometric dimensions on performances are revealed by sensitivity analysis. Then, genetic algorithm is improved according to convergence rate and global accuracy. On this basis, the original scheme is optimized to obtain higher efficiency and smaller torque ripple, and the global optimum values of key dimensions and control parameters are obtained. The study in this paper has certain reference value for design and optimization of SRM.