摘要
文中构造了基于蚁群优化算法(ant colony optimization algorithm,ACOA)的容错径向基函数神经网络(radial basis function neural network,RBF-NN)故障诊断模型,它具有强逼近能力,采用ACOA优化NN可进一步改善泛化性能。该文又考虑了把故障信息受随机因素干扰而产生的变异故障样本加入NN的训练样本集中,以提高NN的容错性能。将该模型用于高压输电线系统和配电网故障诊断,并作容错性能的评估。由仿真测试表明,研究模型的容错性能要优于传统的BP-NN和GA-NN诊断模型。
In this paper, the fault diagnosis model is presented based on the fault-tolerance radial basis function neural network （RBF-NN）using ant colony optimization algorithm （ACOA）. RBF-NN possesses excellent approaching ability, and its generalization ability can be further improved by ACOA. It is also considered in the paper that the basic fault pattern （BFP）can be formed into variational fault pattern （VFP） when disturbed by the stochastic factors,and the fault-tolerance performance （FTP）can be enhanced by training the NN with VFPs. The proposed model is used for fault diagnosis of power transmission and distribution systems, and the FTP is assessed in the paper. Simulation results show the the FTP of the proposed model is superior to that of the conventional diagnosis model based on BP-NN and GA-NN,which prove its feasibility.
出处
《电力系统及其自动化学报》
CSCD
北大核心
2007年第2期44-48,102,共6页
Proceedings of the CSU-EPSA
基金
该成果得到许继奖教金资助
关键词
径向基函数神经网络
蚁群优化算法
输电配电系统
故障诊断
容错性能
radial basis function neural network （RBF-NN）
ant colony optimization algorithm （ACOA）
power transmission and distribution system
fault diagnosis
fault-tolerance performance （FTP）
