基于模糊神经网络的双凸极永磁电机非线性建模

Nonlinear modeling for doubly salient permanent magnetic motor based on fuzzy neural network

中文关键词  双凸极永磁电机  非线性模型  自适应模糊神经网络  混合算法  遗传算法

英文关键词  doubly salient permanent magnet motor  nonlinear modeling  adaptive-network-based fuzzy inference system  hybrid algorithm  genetic algorithm

中文摘要

      双凸极永磁电机的电感、磁链等特性呈严重非线性, 常规的线性或准线性模型难以准确反映双凸极永磁电机的实际特性, 影响双凸极永磁电机的控制精度和工作性能. 为此, 本文提出采用自适应模糊神经网络建立双凸极永磁电机模型的新方法. 首先在介绍了自适应模糊神经网络结构后, 采用改进的递推最小二乘法修改网络参数, 同时采用遗传算法对遗忘因子和学习率进行了优化, 仿真计算和实测结果表明, 该模型有很快的收敛性和很高的精确度, 最后给出了利用模型实现双凸极永磁电机优化控制的方法.

英文摘要

      The doubly salient permanent magnet (DSPM) machine has high nonlinear characteristics of inductance and flux linkage, etc. The normal linear or quasi-linear modeling can not reflect the real nonlinear charactersitics of the DSPM machine, degrading control precision and operational performance. In this paper, a new modeling method is proposed for the DSPM machine to take into account its nonlinearity more accurately by using adaptive-network-based fuzzy inference system. After the structure of adaptive-network-based fuzzy inference system is introduced, the recursive least squares method is improved and applied to modify the parameters of the network. Moreover, the forgetting factor and learning rate are optimized by using genetic algorithm. Both simulation and experiment have shown that the developed modeling offers the advantages of fast convergence and high precision. Finally, the optimal controller based on the modeling for the DSPM motor is developed.