Finite element based multi-objective optimization of a brushless Doubly-Fed Induction Machine

Although the brushless Doubly-Fed Induction Machine (DFIM) has great potential as generator system in large-scale wind turbines, its complexity has so far retained a commercial breakthrough. This paper contributes by combining brushless DFIM Finite Element (FE) modelling with multi-objective optimization. A static brushless DFIM FE model is applied with the NSGA-II multi-objective optimization algorithm. The result is an accurate and fast brushless DFIM design optimization tool. This tool is then used to generate a machine design with optimized performance, that fits a fixed volume frame size D180. Optimization results of different brushless DFIM construction variations are compared. The best designs of each construction variation are then analysed in more detail using a time-stepping brushless DFIM FE model. This provides good insight in the effects of different construction variations on torque ripple and induced time-harmonics. The optimized design of the best construction variation will be selected to be manufactured as prototype machine.