摘要
为提高电机磁链观测器的观测性能,实现磁链的准确观测,提出一种RBF神经网络定子磁链观测器。采用RBF神经网络重构基于电压模型的带幅值和相位补偿的变截止频率定子磁链观测器,使磁链观测器的截止频率能跟随电机定子电信号频率的变化而变化。RBF神经网络磁链观测器实现了变截止频率,结构简单,自适应能力强,无直流偏移和初始相位问题,可在定子电信号频率变化和负载变化情况下实现较为精确的定子磁链观测。实验结果证明了方法的有效性。
For improving stator flux observer,a RBF neural network stator flux observer is presented.A variable cutoff frequency stator flux estimator with amplitude and phase compensation was reconstructed with RBF neural networks,and the cutoff frequency of observer can vary with the stator voltage frequency changes.With variable cutoff frequency,simple structure,and better adaptive capability,without the problem of DC offset and initial phase,the RBF neural networks observer can accurately estimate motor stator flux when the stator voltage frequency varies.Experimental results demonstrate the validity of the method.
出处
《电机与控制学报》
EI
CSCD
北大核心
2011年第8期81-87,共7页
Electric Machines and Control
基金
湖南省自然科学湘潭市联合基金重点项目(09JJ8006)
关键词
定子磁链观测器
截止频率
RBF神经网络
直接转矩控制
stator flux estimator
cutoff frequency
radial basis function network
direct torque control
