基于谱聚类的输电线覆冰过程微气象特征提取模型  

Micro-Meteorological Feature Extraction Model for Ice Covering Process of Transmission Line Based on Spectral Clustering

作　　者：刘宣廷 李鹏[1] 苗爱敏 陈勇 沈鑫 曹敏 LIU Xuan-ting;LI Peng;MIAO Ai-min;CHEN Yong;SHEN Xin;CAO Min(School of Information,Yunnan University,Kunming 650500,China;Zhongkai University of Agriculture and Engineering,Guangzhou 510225,China;Electric Powee Research Institute Yunnan Powee Grid Co.,Ltd.,Kunming 650217,China)

机构地区：[1]云南大学信息学院,云南昆明650500 [2]仲恺农业工程学院,广东广州510225 [3]云南电网有限责任公司电力科学研究院,云南昆明650217

出　　处：《测控技术》2019年第7期89-92,共4页Measurement & Control Technology

基　　金：国家自然科学基金项目(61763049);云南省应用基础研究计划重点项目(2018FA032);云南省应用基础研究面上项目(2017FB229)

摘　　要：因输电线路覆冰造成的电网瘫痪是近年来常见的电网灾害之一,进而对电力系统安全运行的可控性以及电网灾害预警的可靠性提出了更高的要求。提出了基于谱聚类的输电线路覆冰过程特征提取模型,对覆冰过程进行了定性分析,该模型通过对影响覆冰过程的微气象因素进行特征提取(聚类),较为准确地对输电线覆冰严重程度进行了分离,并对微气象数据进行统计分析,将气象倾向率作为参考量,获得了严重覆冰、覆冰脱落等过程发生的气象条件,最后通过云南电网提供的相关数据验证了此模型的有效性,为电网部门输电线路覆冰灾害的预警提供气象参考条件。Power grid paralysis caused by icing of transmission lines is one of the common power grid disastersin recent years,which puts forward higher requirements for the controllability of power system safe operation and the reliability of power grid disaster warning.A feature extraction model based on spectral clustering for icing process of transmission lines is proposed,and the icing process is qualitatively analyzed.By feature extraction(clustering)of the micro-meteorological factors that affect the icing process,the model can accurately separate the icing severity of transmission lines and integrate the micro-meteorological data.According to the analysis,the meteorological tendency rate is taken as the reference quantity,meteorological conditions such as severe icing,ice shedding and so on were obtained.The validity of the model is verified by the relevant data provided by Yunnan Power Grid,which provides meteorological reference conditions for the transmission line icing disaster warning of power grid departments.

关 键 词：谱聚类算法 输电线路覆冰 微气象 特征提取 

分 类 号：TM75[电气工程—电力系统及自动化] TP18[自动化与计算机技术—控制理论与控制工程]