基于Logistic回归模型的超高压输电线路雷击跳闸识别  

Identification for Lightning Tripping of EHV Transmission Lines Based on Logistic Regression Model

作　　者：苏浩辉 陈彦州[1] 陈岳[1] 

机构地区：[1]中国南方电网有限责任公司超高压输电公司检修试验中心,广东广州510663

出　　处：《广东电力》2017年第4期85-89,共5页Guangdong Electric Power

摘　　要：以105起超高压输电线路雷击跳闸历史事件作为训练集,对多元线性回归模型和Logistic回归模型的雷击跳闸识别能力进行对比,得出Logistic回归模型拟合优度较高的结果,进而提出基于Logistic回归模型的超高压输电线路雷击跳闸自动识别方法。以近期20起超高压输电线路故障跳闸事件作为测试集,对该雷击跳闸自动识别方法进行验证,雷击跳闸识别率达90%,测试结果说明该方法有效。Taking 105 historical eventsof lightning trippingofextrahighvoltage （EHV） transmission this paper compares identification abilities for lightning tripping of multiple linear regression modelmodel, and gains a result that the Logistic regression model has higher matching goodness. And matic identification method for lightning tripping of EHV transmission lines based on Logistic regressthis automatic identification method by taking 20 faulted tripping events of EHV transmission lines as testing sd. Lightningtripping identification rate reaches to 90 % which declares effectiveness of this method.

关 键 词：超高压输电线路 雷击跳闸 Logistics回归模型 自动识别 

分 类 号：TM732[电气工程—电力系统及自动化]