基于DA-SVR的2M继电保护通道在线监测和故障诊断  

On-line Monitoring and Fault Diagnosis of 2M Relay Protection Channel Based on DA-SVR

作　　者：方明 王春 李松有 FANG Ming;WANG Chun;LI Songyou(Guiyang Bureau,EHV Transmission Company,China Southern Power Grid Co.,Ltd.,Guiyang 550081,China)

机构地区：[1]中国南方电网有限责任公司超高压输电公司贵阳局

出　　处：《机械与电子》2020年第2期63-66,共4页Machinery & Electronics

基　　金：中国南方电网有限责任公司超高压输电公司(CGYKJXM20160040)

摘　　要：为了提高2 M继电保护通道在线监测故障诊断的准确率,针对SVR模型性能受惩罚参数C和核函数参数g的影响,将2 M继电保护通道在线监测技术和DA-SVR结合起来,提出一种基于DA-SVR的2M继电保护通道在线监测和故障诊断算法。研究结果表明,从各个故障类别的诊断结果还是总体正确率来看,DA-SVR均要高于GA-SVR,SVR和BP神经网络,有效提高了2M继电保护通道在线监测数据故障诊断的正确率。In order to improve the accuracy of 2M relay protection channel online monitoring fault diagnosis,the two of technology protection channel online monitoring and DA-SVR were combined with the performance of the SVR model affected by two factors of parameter g which is kernel function parameter andparameter C.and provided the 2M relay protection channel which is based on DA-SVR and support to monitor online and the ability of fault diagnosis algorithm.The results show that DA-SVR is higher than GA-SVR,SVR and BP neural network in the diagnosis of each fault category and the fault diagnosis rate of 2M relay protection channel is improved effectively.

关 键 词：保护机制 神经网络 DA-SVR 在线监测 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]