Global-Simplex Optimization Algorithm Applied to FEM-Based Optimal Design of Electric Machine

Finite element analysis-based optimal design of electric machines takes up a considerable amount of time for characteristic analyses and the analysis region has many local optima. Thus, selecting the appropriate global search optimization algorithm, known as the with fast convergence characteristics is necessary in the optimal design of electric machines. In this paper, a novel global search optimization algorithm known as the global-simplex optimization (GSO) algorithm, is proposed. The GSO is an optimization algorithm that adopts the characteristics of the simplex method for searching the optimum point but utilizes its global searching capability by scattering many initial points at the beginning of the algorithm. The global searching capability and convergence speed of GSO is validated through a comparison with particle swarm optimization, which has already been proven for its global searching capability through its application in the well-known Goldstein–Price benchmark function. In addition, the GSO has been applied to the optimal design of an interior permanent magnet synchronous machine to reduce the torque ripple and total harmonic distortion in its back electromotive force.