摘要
对停运输电线路进行合闸操作时为避免合闸到故障线路,通常需要对停运线路的故障状态进行判断。对于一回正常运行、另一回处于热备用状态下的同塔双回输电线路,提出一种基于感应电压特征的停运线路机器学习故障识别方法。首先对停运线路感应电压有效值进行测量,取各相电压有效值、电压平均值及故障电压占比作为样本特征。采用径向基核函数支持向量机(RBF-SVM)对停运线路的故障状态进行识别;若存在故障则利用BP神经网络对故障类型进行识别。为验证该方法的故障识别效果,以河北省6条线路的实际数据为基础,在ATP-EMTP中建立500 kV同塔双回输电线路模型。结果表明,对热备用线路上故障状态识别准确率为100%;对故障类型识别时准确率达到99.7%,为调度工作中合闸操作及停运线路故障的排除提供了参考。
In order to avoid switching to the fault line, it is necessary to judge the fault state in the closing operation of the outage line. In the double-circuit transmission lines on the same tower, one transmission line operates while the other is in the hot standby state. Thus we proposed a machine learning fault identification method based on induced voltage characteristics of outage lines for the double-circuit transmission line on the same tower. First, the root mean square(RMS) of the induction voltage was measured. The RMS of each phase voltage, the average value of each phase voltage and the fault voltage ratio of each phase were taken as the sample characteristics. We used radial basis kernel function support vector machine(RBF-SVM) to judge the fault state of the outage line. If there is a fault in the line, the BP neural network will be used to identify the fault type. To verify the fault recognition effect of this method, the models of 500 kV double-circuit transmission line on the same tower were established in ATP-EMTP based on the actual data of 6 lines in Hebei Province. The result shows that the identification accuracy of this method for the fault state on the hot standby line is 100% and that for the fault type is 99.7%. This provides a reference for the closing operation and the troubleshooting of the outage line in the dispatching work.
作者
刘英培
王祥宇
王鑫明
李少博
梁华洋
李世泽
LIU Yingpei;WANG Xiangyu;WANG Xinming;LI Shaobo;LIANG Huayang;LI Shize(School of Electrical and Electronic Engineering,North China Electric Power University,Baoding 071003,China;State Grid Hebei Electric Power Co.,Ltd.Dispatching Control Center,Shijiazhuang 050021,China;State Grid Hebei Electric Power Co.,Ltd.Maintenance Branch,Shijiazhuang 050070,China)
出处
《华北电力大学学报（自然科学版）》
CAS
北大核心
2022年第4期14-22,32,共10页
Journal of North China Electric Power University：Natural Science Edition
基金
国网河北省电力有限公司科技资助项目(5204BB200028)。
关键词
同塔双回输电线路
感应电压
RBF支持向量机
BP神经网络
故障识别
double-circuit transmission lines on the same tower
induced voltage
RBF-SVM
BP neural network
fault identification
