Design optimisation of brushless permanent magnet synchronous motor for electric vehicles

A novel new application of optimisation algorithm “Bess Algorithm” in the design of electric machine is presented in this thesis. The optimisation has the ability to perform global and local search and can be applied on constrained, unconstrained optimisation problem with multi-objective function, which all counted when consider optimisation algorithm for the design of electric machine. The searching procedure of the optimisation algorithm has been described in detailed. Furthermore, novel instructions and recommendation were implemented to tune the optimisation parameters, particularly for the purpose electric machine design, which in turn reduced the search space, increase efficiency and ability to find optimal solution with lower computation time. The optimisation was applied to search for optimal parameters of a benchmark electric machine with multi-objective to reduce the cost and increase the power density, power-volume ratio and efficiency. Throughout the thesis, a full detailed analytical model for the design of brushless permanent magnet synchronous motor that account for electromagnetic and thermal aspects was described. The optimisation was employed to search for optimal parameters of the analytical model that satisfy the design requirements. Then, the generated optimal parameters were evaluated and verified by Finite Element Analysis, FEA. The results from the FEA show good agreement with their corresponding values in the analytical model within acceptable range. At the same operational conditions and output specifications, the results show that the power density, volume to power ratio and cost of the new optimised motor IV were all increased by 19%, 39%, 24% respectively and the efficiency reduced only by -1%. The optimisation was also compared with one of the most usable optimisation algorithm used in the design of electric machine i.e. Genetic Algorithm. The results show that bees algorithm has more ability to cover the search space with less number of recruited bees and less number of iterations and higher computation efficiency