Optimum design of linear synchronous motor using evolution strategy combined with stochastic FEM

Abstract

The random based evolution strategy (ES) in conjunction with stochastic finite element method (SFEM) is applied to the optimum design of electromechanical devices. (1+1) ES is adopted as an optimization method. Optimum pole shape is designed to increase the levitation and propulsion forces of the linear synchronous motor (LSM) which is used in a Maglev vehicle. The SFEM analyzes performance uncertainty due to the uncertain system parameter in the devices. The degree of randomness of objective function value is expressed by its expectation and variance. The optimization process is influenced by these values. To show validity of the proposed algorithm, results of ES combined with SFEM are compared with those of the conventional ES.