Affiliations: [a] School of Electrical and Information Engineering, Tianjin University, Tianjin, China
Correspondence:
[*]
Corresponding author: Hong Wei Fang, School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, China. Tel.: +86 138 2103 3642; E-mail: hongwei_fang@tju.edu.cn
Abstract: The magnetic gear integrated permanent magnet synchronous generator (MG-PMSG) can reduce the acoustic noise and mechanical loss, which are caused by the mechanical gear box. It also has the merits of increasing efficiency and reducing system volume when it is used for wave energy conversion system. In this paper, an improved bat algorithm (BA) based on velocity weighting factor is proposed. The improved BA is applied for the optimization design of permanent magnet (PM) to reduce the cogging torque of MG-PMSG. The numerical model is constructed by response surface methodology (RSM). The influences of key pole shape parameters on cogging torque were investigated, including the eccentric distance, the pole-arc coefficient and the permanent magnet thickness. A global optimization design is then carried out by using the improved BA, so that the magnet dimensions corresponding to the optimal cogging torque are obtained. Finally, the performances of the MG-PMSG with the optimized permanent magnet are analyzed by finite element method. Results show that cogging torque, steady torque ripple and back electromotive force (EMF) waveform distortion of the optimized MG-PMSG are reduced.
Keywords: Magnetic gear, wave energy conversion, cogging torque, bat algorithm, response surface methodology
