摘要
针对低压配电网拓扑结构人工校验成本高、准确性不足的问题,提出了一种基于离群点检测的低压配电网拓扑结构校验方法。首先从用电信息采集系统提取待校验台区用户最近一段时间的智能电表电压序列数据,然后计算电压序列数据之间相关系数来度量不同用户电压曲线的相似性,最后通过基于电压曲线相似性的离群点检测识别出台区标识错误的用户,可快速、有效地验证电网GIS系统中用户与台区变压器拓扑连接关系的正确性。
Aiming at the high cost, low accuracy of low-voltage transformer area topology artificial verification, this paper proposes a new verification method based on outlier detection. For a verified transformer area, firstly the smart meter voltage sequence data of all users are extracted from power consumption information collection system： secondly the correlation coefficients among the smart meter voltage sequences are calculated to measure different users voltage curve similarity; finally, the questionable users are distinguished to use the outlier detection based on the voltage curve similarity. It can effectively verify the connectivity between electric power users and the correspondent transformer in GIS.
出处
《电力信息与通信技术》
2017年第5期61-65,共5页
Electric Power Information and Communication Technology
基金
国家电网公司科技项目"基于营配信息贯通的业务融合与数据共享服务技术"
国网河南省电力公司科技项目"支撑‘三集五大’的大数据关键技术研究与示范应用"
关键词
电压序列数据
相关系数
电压曲线相似性
离群点检测
voltage sequence data
correlation coefficient
voltage curve similarity
outlier detection
