Elman神经网络在高压电力计量系统故障诊断中的应用  

Application of Elman Neural Network in the Fault Diagnosis of High Voltage Electric Power Metering System

作　　者：赵建军[1] 张月阳[1] 梁威[1] 

机构地区：[1]河南大学先进控制与智能信息处理研究所,河南开封475004

出　　处：《电测与仪表》2014年第13期1-4,75,共5页Electrical Measurement & Instrumentation

基　　金：国家自然科学基金资助项目(61174112)

摘　　要：利用Elman神经网络自适应性强的优势,将其应用于高压电力计量系统故障诊断中。对计量系统的8个运行参数提取故障信息作为网络的输入值,以4位二进制数表示正常情况和9种主要故障类型,用采集得到的样本对网络进行训练,并选取4种故障类型进行网络测试,结果表明,Elman神经网络模型对高压电力计量系统故障有很好的辨识和诊断效果。Elman neural network will be able to adopted in the fault diagnosis of high voltage electric power metering system according to the advantage of the great adaptivity. Information in 8 operation parameters of metering system will be extracted to be taken as the input values of the neural network,and a 4 - bit binary number will be adopted to re-present the normal circumstance and 9 main fault types. Then the collected samples are used to train Elman network and 4 fault types will be selected to test the network. The results show that the Elman network model can effectively diagnose the fault of high voltage electric power metering system.

关 键 词：高压电力计量系统 故障诊断 ELMAN神经网络 采集样本 

分 类 号：TM93[电气工程—电力电子与电力传动]