摘要
线损精益化管理一直都是国家电网公司一项重点工作,近几年线损数据获取技术日趋成熟,但是分析方法依旧匮乏,严重影响了高损治理的效果。为解决电网运行中高损治理不佳的问题,我们提出了用大数据分析方法来解决传统线损管理问题--供电企业基于三层分析模型的线损异常分析及处理研究。首先,搭建了一个由Hampel抗差算法、加权皮尔逊算法和随机森林算法三种不同算法构成的三层线损异常分析模型;通过该模型,可以结合不同用户用电量大小不一的特点,自上而下的分析大电量异常用户和小电量异常用户。然后将该模型嵌入公司已有的线损监控平台中,可以实现对全省线损数据的实时监测、有效挖掘、深度分析、精准定位和工单管控,形成基于三层分析模型的线损异常分析及处理新方法。该方法对分析处理10 k V高损线路和0. 4 k V高损台区中效果显著。
Line loss lean management has always been a key task of the State Grid Corporation. In recent years,line loss data acquisition technology has become increasingly mature,but the analysis methods are still scarce,which seriously affects the effect of high-loss management. In order to solve the problem of poor high-loss management in power grid operation,a big data analysis method is proposed to solve the traditional line loss management problem-the line loss anomaly analysis and processing based on the three-layer analysis model of power supply enterprises. Firstly,a three-layer linear damage anomaly analysis model composed of Hampel resistance algorithm,weighted Pearson algorithm and random forest algorithm is built. this model can combine the characteristics of different users’ different power consumption,and analyze the users of large power anomaly and small power anomaly from top to bottom. Then the model is embedded in the existing line loss monitoring platform of the company,which can realize real-time monitoring,effective mining,depth analysis,precise positioning and work order control of the line loss data in the province,and form a new method of line loss anomaly analysis and processing based on the three-layer analysis model. This method is effective in analyzing and treating 10 k V high loss lines and 0. 4 k V high loss platform areas.
作者
邵丹
石立彬
史静远
郭晓松
秦晓丹
SHAO Dan;SHI Libin;SHI Jingyuan;GUO Xiaosong;QIN Xiaodan(State Grid Xingtai Power Supply Company,Xingtai 054001 Heibei,China)
出处
《电力大数据》
2019年第10期78-83,共6页
Power Systems and Big Data
关键词
汉佩尔抗差算法
加权皮尔逊算法
随机森林算法
三层线损分析模型
线损监控平台
Hampel robust algorithm
weighted Pearson algorithm
stochastic forest algorithm
three-layer line loss analysis model
line loss monitoring platform
