Design of Electrically Excited Synchronous Motors for Traction Applications Using Optimisation with Genetic Algorithms

This paper presents a method for designing electrically excited synchronous motors for traction applications. The initial stages of motor design are often based on empirical values, potentially ignoring preferable options. In order to overcome this issue, this paper proposes a finite element analysis coupled optimisation using genetic algorithms that allow efficient evaluation of a large solution space. For parameters described by integers, e.g. number of turns per pole and phase, a theoretically optimal number is calculated in order to minimise the volume of the machine while achieving the desired efficiency and output power. The results are then used in a final optimisation that considers feasible winding configurations, including constraints specific to hairpin windings.