基于Adaboost算法的输电线路舞动预警方法  

An early warning method of transmission line galloping based on Adaboost algorithm

作　　者：李哲[1] 王建[2] 梁允[1] 熊小伏[2] 翁世杰[2] 

机构地区：[1]国网河南省电力公司电力科学研究院,郑州450052 [2]重庆大学输配电装备及系统安全与新技术国家重点实验室,重庆400044

出　　处：《重庆大学学报（自然科学版）》2016年第1期32-38,97,共8页Journal of Chongqing University

基　　金：重庆市科技攻关(应用重点)项目(cstc2012gg-yyjsB90003);国家电网公司重大基础前瞻科技项目(SG20141187)~~

摘　　要：输电线路舞动是目前尚未被全面准确认识的世界性难题,已严重威胁输电系统的安全稳定运行。文章分析影响舞动的外界气象环境因素,并在此基础上提出一种基于Adaboost集成学习算法的输电线舞动预警方法。采用基于Gini指标的决策桩作为弱分类器,通过对多个弱分类器的训练及加权求和,输出舞动预测结果及其置信度,可为电网运维人员提供决策支撑。最后,使用历史数据进行验证性实验,结果证明了所提方法的有效性。Transmission line galloping is a worldwide problem which has not been fully understood,and it has seriously threatened the safe and stable operation of a transmission system.We investigated the factors of meteorological environment that influence galloping,and proposed an early warning method of transmission line galloping based on the Adaboost ensemble learning algorithm.In this method,the decision stump based on the Gini index is used as the weak classifier.The prediction result and its confidence are obtained by training and weighted summing of multiple weak classifiers,which are helpful information for the decision making of operators and dispatchers of power grids.The effectiveness of the proposed method is proved by the verification experiment with historical data.

关 键 词：输电线 舞动 预警 ADABOOST算法 决策桩 

分 类 号：TM732[电气工程—电力系统及自动化]