摘要
定子电阻的准确估计是改善直接转矩控制低速性能的关键技术.为了提高定子电阻的在线估计精度和速度,本文将小波分析、自组织算法和神经网络技术相结合,提出了一种自组织小波神经网络定子电阻估计器.该网络继承了小波分析优异的局部特性和神经网络的自学习能力,具有较高的估计精度.并采用自组织算法对小波元的数量进行了离线优化,大大简化了网络结构,提高了在线估计的实时性.
Exact estimation of stator resistance is the key technology to improve the low speed performance of direct torque control. To improve the accuracy and speed of on-line stator resistance estimation, a self-organization wavelet neural network estimator is proposed in this paper by combining wavelet analysis, self-organization algorithm and neural network technology together. The proposed network inherits the excellent local performance of wavelet analysis and the self-learning ability of neural network to get high estimation accuracy, and its wavelet number is optimized off-line by using self-organization algorithm to simplify the network structure and improve the on-line estimation speed.
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
2007年第3期371-373,379,共4页
Control Theory & Applications
基金
湖南省自然科学基金资助项目(06JJ50121).
关键词
直接转矩控制
小波
神经网络
自组织
direct torque control
wavelet
neural network
self-organization
