基于动态模糊神经网络的电力谐波精确测量  

Precision Meaturement of Power Harmonic Based on DFNN

作　　者：徐传忠[1] 杨冠鲁[1] 王永初[2] 

机构地区：[1]华侨大学信息学院 [2]华侨大学机电学院,福建泉州362021

出　　处：《太原理工大学学报》2011年第4期357-359,364,共4页Journal of Taiyuan University of Technology

基　　金：福建省自然科学基金项目(2009J01257);厦门市科技计划项目(3502Z2009302T)

摘　　要：利用动态模糊神经网络测量电力谐波幅度,采用误差下降率调节模糊神经网络的结构,同时用分级学习方法调节模糊神经网络参数,使神经网络结构呈现动态变化,结构紧凑,避免了过拟合及过训练现象,提高了系统的泛化能力。通过对电力谐波信号的测量仿真,表明该方法具有很高的测量精度。IFNN was used to meature power harmonic in this paper. Structure of DFNN was regulated by error drop rate and parameters of DFNN were regulated by classification learning. The I）FNN changed dynamically in compact structure form, provented. Overfitting and overtraining,and there fore enhanced generalization ability. Simulation proved the measurement accuracy of the method.

关 键 词：谐波 动态模糊神经网络 修剪技术 

分 类 号：TM930.12[电气工程—电力电子与电力传动]