Computationally Efficient Optimization of a Five-Phase Flux-Switching PM Machine under Different Operating Conditions.
Hao Chen,Xiangdong Liu,Nabeel A. O. Demerdash,Ayman M. EL-Refaie,Zhen Chen,Jiangbiao He
DOI: https://doi.org/10.1109/tvt.2019.2915239
IF: 6.8
2019-01-01
IEEE Transactions on Vehicular Technology
Abstract:This paper investigates the comparative design optimizations of a five-phase outer-rotor flux-switching permanent magnet (FSPM) machine for in-wheel traction applications. To improve the comprehensive performance of the motor, two kinds of large-scale design optimizations under different operating conditions are performed and compared, including the traditional optimization performed at the rated operating point and the optimization targeting the whole driving cycles. Three driving cycles are taken into account, namely, the urban dynamometer driving schedule (UDDS), the highway fuel economy driving schedule (HWFET), and the combined UDDS/HWFET, representing the city, highway, and combined city/highway driving, respectively. Meanwhile, the computationally efficient finite-element analysis (CE-FEA) method, the cyclic representative operating points extraction technique, as well as the response surface methodology (in order to minimize the number of experiments when establishing the inverse machine model), are presented to reduce the computational effort and cost. From the results and discussion, it will be found that the optimization results against different operating conditions exhibit distinct characteristics in terms of geometry, efficiency, and energy loss distributions. For the traditional optimization performed at the rated operating point, the optimal design tends to reduce copper losses but suffer from high core losses; for UDDS, the optimal design tends to minimize both copper losses and PM eddy-current losses in the low-speed region; for HWFET, the optimal design tends to minimize core losses in the high-speed region; for the combined UDDS/HWFET, the optimal design tends to balance/compromise the loss components in both the low-speed and high-speed regions. Furthermore, the advantages of the adopted optimization methodologies versus the traditional procedure are highlighted.