基于GRNN的多故障自适应电力系统故障诊断  

Adaptive Multi-Fault Diagnosis of Power System Based on GRNN

作　　者：廖志伟[1] 叶青华[1] 王钢[1] 文福拴[1] 

机构地区：[1]华南理工大学电力学院,广东广州510640

出　　处：《华南理工大学学报（自然科学版）》2005年第9期6-9,共4页Journal of South China University of Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(50477029)

摘　　要：为了实现快速而准确的电网故障诊断,利用广义回归神经网络(GRNN)在逼近能力、分类能力和学习速度方面的优势,建立了基于GRNN的电网故障诊断模型.仿真分析表明:在输入信息因干扰而畸变的情况下,文中所构造的模型能够快速、正确地实现电网的故障诊断;在电网拓扑结构改变的情况下,该模型也具有良好的自适应能力.In order to realize fast and correct fault diagnosis in electric network, this paper presents a novel fault diagnosis model employing the abilities of GRNN （ General Regression Neural Network） in approximation, classification and learning. Simulated results demonstrate that the proposed model not only can make fast and accurate diagnoses with good fault-tolerance performance, but also possesses excellent self-adaptive ability for various topological structures of electric network.

关 键 词：电力系统 故障诊断 广义回归神经网络 自适应能力 

分 类 号：TM711[电气工程—电力系统及自动化]