Brushless DC motor design: an optimisation procedure based on genetic algorithms

In this paper, a new approach to the optimised design of a surface-mounted permanent magnet (PM) brushless DC motor is presented. The novelty lies in combining a fully analytical procedure for the design (including thermal constraints) with a genetic algorithm procedure in order to optimise an objective function, as the motor efficiency or the material cost or some other motor figures or a combination of them. The motor design procedure consists of a system of nonlinear equations which imposes PM characteristics, motor performance, magnetic stresses and thermal limits. Taking advantage of the analytical form of the design procedure, the optimisation technique based on genetic algorithms has been readily implemented. An optimised design is included in the paper, as well as a comparison between the proposed optimisation technique and that based on the classical climbing-hill direct-search method.