Design Optimization Methods for Electrical Machines

Abstract

This chapter presents the design optimization methods for electrical machines in terms of different optimization situations, including low- and high-dimensional, single- and multi- objectives and disciplines. Firstly, the traditional design optimization methods are briefly reviewed, and the challenges presented. Then, five new types of design optimization methods are presented to improve the optimization efficiency of electrical machines, particularly those complex structured permanent magnet machines, in terms of different optimization situations. They are (a) a sequential optimization method for design optimization of low-dimensional problems of electromagnetic devices including electrical machines, (b) a multi-objective sequential optimization method for engineering multi-objective problems, (c) a multi-level design optimization method (or sequential subspace optimization method) for high dimensional problems, (d) a multi-level genetic algorithm for high dimensional optimization problems as well, and (e) the multi-disciplinary design optimization method. Design examples with detailed experimental and optimization results are illustrated for each optimization method.