摘要
为提高诊断台区线损状态的诊断率,提出基于大数据挖掘技术的线损智能诊断模型设计。采集线损数据并计算得到线损率,根据线路的特征设置标准阈值,以时间离散度分析结合多次聚类分析检测出线损的异常状态;根据采集数据的波动率,构建离散度转换方程,得到线损离群点特征的欧氏距离近似度矩阵;利用大数据挖掘技术,提取出线损的离群点,利用关联分析法,构建线损智能诊断模型。实验结果表明,设计模型不仅可以提高线损诊断率,还可以诊断出用户的窃电行为。
In order to improve the diagnosis rate of line loss in the diagnosis station area,an intelligent diagnosis model of line loss based on big data mining technology is proposed.Collect line loss data and calculate the line loss rate,set the standard threshold according to the characteristics of the line,and detect the abnormal state of line loss by time dispersion analysis combined with multiple cluster analysis;According to the volatility of the collected data,the dispersion transformation equation is constructed to obtain the Euclidean distance approximation matrix of line loss outlier characteristics.Using big data mining technology,outliers of line loss are extracted,and an intelligent diagnosis model of line loss is constructed by using association analysis method.The experimental results show that the designed model can not only improve the diagnosis rate of line loss,but also diagnose the electricity stealing behavior of users.
作者
王坚
刘畅
周阳
张爱梅
石晶
WANG Jian;LIU Chang;ZHOU Yang;ZHANG Ai-mei;SHI Jing(Guiyang Power Supply Bureau of Guizhou Power Grid Co.,Ltd.,Guiyang 550002 China;Guiyang Wudang Power Supply Bureau of Guizhou Power Grid Co.,Ltd.,Guiyang 550002 China)
出处
《自动化技术与应用》
2023年第8期141-144,178,共5页
Techniques of Automation and Applications
基金
基于大数据线损诊断智能(AI)建模核心技术研究及应用示范项目(060100KK52180041)。
关键词
大数据挖掘技术
智能台区
诊断模型
线损状态
异常状态
离群点
Big data mining technology
Intelligent station area
Diagnostic model
Line loss status
Abnormal state
Outliers