Robust Optimization of PMLSM Based on a New Filled Function Algorithm With a Sigma Level Stability Convergence Criterion
Fei Dong, Jiwen Zhao, Jing Zhao, Juncai Song, Jie Chen, Zhilei Zheng
Abstract
Traditional design optimization methods for permanent magnet linear synchronous motor (PMLSM) always pursuit the optimal result without considering the influence of uncertainties (manufacturing errors, use wear, etc.), which will result in big fluctuations for the stability of motor thrust performance. This article proposes a new filled function algorithm (NFFA) to improve the stability of PMLSM. It searches for the structure parameters that not only lay in the smooth area of thrust ripple distribution but also satisfy the thrust ripple constraints under the effect of uncertainties. First, an analytical model of thrust ripple is established for the next-step optimization. Second, a sigma level stability criterion is proposed, which will be used to judge the robustness of optimal thrust ripple under uncertainties, and a new global robust algorithm NFFA is studied to realize the stability of thrust ripple by managing the sigma level stability criterion as the convergence criterion of traditional filled function optimization algorithm (TFFA). Third, the NFFA is used to optimize the thrust ripple of PMLSM based on the analytical model. The optimization results are proven to be stable and superior compared with other methods, such as TFFA, Taguchi robust optimization method, and design for six sigma method. Finally, the experiments confirm the feasibility and validity of the proposed method.