摘要
如何科学、准确地识别异常用电量对于改善电力系统用电管理至关重要。文中提出一种基于密度聚类技术的电力系统用电量异常分析算法。该算法通过基于密度的聚类技术和局部离群点要素给出异常用电波动区间的离群度,利用关联分析法构造关联规则,同时给出其关联规则支持度,并结合当前用电量综合分析获取异常用电得分。最后以异常用电百分比实现用电量信息异常情况的快速、可靠分析。仿真和实验测试结果表明该异常分析算法能够高效识别用电信息异常数据,从而提高用电量异常分析的准确率。
In the electricity information acquisition network,how to scientifically and accurately identify the abnormal power consumption is crucially important to improve the management of power system.A novel analysis algorithm for abnormal power consumption based on density-based spatial clustering of applications with noise（DBSCAN）in power system is presented.Firstly,the abnormal power range is given by DBSCAN and local outlier factor（LOF）.Secondly,the association analysis is used to get association rules and support degree.Then,the score of abnormal power consumption can be gained with combining the before analysis and current power consumption.The identification of abnormal information about power consumption can be realized quickly and reliably by using the percentage of abnormal scores.The simulation results demonstrate that the abnormal analysis algorithm can detect abnormal power consumption effectively,improving the accuracy of the abnormal power consumption analysis.
出处
《电力系统自动化》
EI
CSCD
北大核心
2017年第5期64-70,共7页
Automation of Electric Power Systems
基金
重庆市研究生科研创新项目(CYS15158)
国家电网公司科技项目(SGJSSZOOFZJS1501091)
重庆市基础与前沿研究计划(cstc2015jcyjA40007)~~
关键词
用电量异常分析
密度聚类
局部离群点要素
关联分析
abnormal power consumption analysis
density-based spatial clustering of applications with noise（DBSCAN）
local outlier factor（LOF）
association analysis
