GA-BP混合算法在变压器色谱诊断法中的应用  

Application of Hybrid Algorithm Based on GA-BP in Transformer Diagnosis Using GAS Chromatographic Method

作　　者：王少芳[1] 蔡金锭[1] 

机构地区：[1]福州大学电气系,福州350002

出　　处：《高电压技术》2003年第7期3-6,共4页High Voltage Engineering

基　　金：福建省教育厅基金(编号:JB01035)

摘　　要：将一种改进的遗传操作与人工神经网络相结合的混合算法应用于电力变压器的故障诊断,有效地解决了常规BP算法易陷入局部极小、收敛速度慢和基本遗传算法早熟等缺点。实例仿真结果表明,该算法具有较快的收敛速度和较高的计算精度,满足电力变压器故障诊断的要求。In this paper, a hybrid algorithm which combines improved genetic manipulation with artificial network is applied to fault diagnosis of power transformer. It overcomes the .shortcomings of entrapment in local optimum, slow convergence of ordinary BP algorithm and the premature convergence of basic GA effectively. Simulation diagnosis results of practical examples show that the convergence rate of the proposed method is faster and the computational result is more accurate than others, the proposed method fulfils the requirements of fault diagnosis of power transformer.

关 键 词：电力变压器 故障诊断 色谱诊断法 人工神经网络 GA-BP混合算法 遗传算法 

分 类 号：TM407[电气工程—电器]