Combined FE and Particle Swarm algorithm for optimization of high speed PM synchronous machine

Abstract

Purpose – The purpose of this paper is to optimize the stator slot geometry of a high-speed electrical machine, which is used as an assist for a turbocharger. Meanwhile, the suitability of the Particle Swarm algorithm for such a problem is to be tested. Design/methodology/approach – The starting point of the optimization is an existing design, for which the Particle Swarm algorithm is applied in conjunction with the transient time-stepping 2D finite element method. Findings – It is found that regardless of its stochastic nature, the Particle Swarm work well for the optimization of electrical machines. The optimized design resulted in an increase of the slot area and increase of the iron loss, which was compensated by a dramatic decrease in the Joule losses. Research limitations/implications – The optimization was concentrated on the stator design whereas the rotor dimensioning was carried out withing the compressor and turbine design. Originality/value – A turbocharger with electric assist is designed optimized and manufactured. The Particle Swarm algorithm is shown to be very stable.