Electric machine dimensioning by global optimization

Abstract

Global optimization as a genetic algorithm and simulated annealing are proposed for synchronous machine dimensioning. By using the optimization in dimensioning, it is possible to replace the heuristic optimization process of a design engineer in an automated offer calculation system or alternatively to help the engineer in design work. The role of the optimization is to vary a part of the design parameters of a design tool. Heuristic design knowledge and knowledge of the specifications of the machines are utilized. Also, the production tools can be taken into account in the optimization process. The objectives of optimization are described by fuzzy tolerances. In this paper the descriptions of the experiments are explained and the effectiveness of optimization algorithms with different parameters and encoding techniques are studied.