抄録
In optimal design of permanent magnet motor (PM motor), many design variables are required to consider some device properties. These variables are, for example, the shape of core and magnet, the teeth length, the number of turns, the winding radius, and so on. Moreover, many restrictions must be considered toward the practical PM motor design. These restrictions are, for example, the slot space factor, the cogging torque, and so on. However, the optimization problem, which has many design variables by using finite element method (FEM), has not been reported. In this paper, the efficiency of PM motor under considering described above is optimized by using FEM and optimization algorithm. In this problem, an objective function has many local minima and it is difficult to calculate its gradient. For these reasons, the genetic algorithm (GA) and the simulated annealing method (SA), which are stochastic method, are used for optimization method, becase of unnecessarity of the gradient and possibility of the global range search. Adding to both optimization methods, in this paper, SA combined with GA is used for one of optimization methods. As the results, the solutions optimized by these methods are reasonable from engineering point of view.
