Genetic Algorithm Approach for Improved Design of a Variable Speed Axial-Flux Permanent-Magnet Synchronous Generator

Optimal design for an axial-flux permanent magnet synchronous generator (AFPMSG) using genetic algorithm (GA) is presented, and the condition for minimum active material cost is deduced. A computer-aided design procedure based on the results of the GA is proposed. In the design procedure the practical and performance characteristics are taken into account as some restrictions for the object function used in the optimization algorithm. A 30-kW AFPMSG with two parallel connected stators and rotor-surface PMs is designed using the developed program, and then 3-D finite-element analyses are carried out to validate the design procedure.