无刷直流电机滑模观测器参数优化设计方法

Brushless DC Motor Sliding Mode Observer Parameters Design Method

滑模反电势观测器的增益参数会影响观测器的收敛速度以及动态响应性能,常见的设计方法是基于观测器稳定性理论进行设计。提出一种利用遗传算法在稳定域内搜索观测误差最小的增益参数的新方法,既减少了参数设计时间,同时能够降低观测器的误差。为了改进传统遗传算法易陷入局部收敛的问题,采用多种群遗传算法求解最优增益参数。实验结果表明提出的方法能够快速求解最优反馈参数,并且观测器误差较小。

SMO gain parameters will affect the convergence rate and dynamic response speed of the observer.The common method designs gain parameters based on lyapunov theory. Proposed a new method to determine the gainparameters of SMO,which applies genetic algorithm(GA)to search the smallest observation error gain parameters instable domain. The proposed method reduced the manual debugging time and the error of the observer. In order toimprove local convergence of GA,proposed a multi-population genetic algorithm method to obtain optimal gainparameters. The experiment results have shown that the optimal gain parameters can be solved fast by the proposedmethod,and the observer error is low.