摘要
客户侧窃电行为不仅造成电能资源大量流失,同时造成线路负荷过载引发火灾等重大安全事故。针对当前客户侧窃电行为的多样性与隐蔽性特征,以约束客户侧窃电行为为目的,设计了客户侧窃电态势感知及智能预警关键技术。考虑客户侧窃电行为的多样性与隐蔽性特性,选取额定电压偏离度、电压不平衡率与电流不平衡率等6个客户侧窃电态势感知指标,利用RBF神经网络构建客户侧窃电态势感知模型,将所选取的6个指标与相关数据作为模型输入,通过动态K均值聚类算法优化模型,模型输出结果即为客户侧窃电态势感知结果。基于感知结果,通过声光报警装置与智能设备实现智能预警,实验结果显示,该技术能够有效抑制客户侧窃电行为。
The customer side electricity stealing behavior not only causes the massive loss of power resources,but also causes the overload of line load,leading to fire and other major safety accidents.Aiming at the diversity and concealment characteristics of the current electricity stealing behavior in the side toilets,the key technologies of situation awareness and intelligent early warning of electricity stealing on the customer side are studied for the purpose of restraining the electricity stealing behavior on the customer side.Considering the diversity and concealment of customer side power stealing behavior,six customer side power stealing situation awareness indicators are selected,including rated voltage deviation,voltage imbalance rate and current imbalance rate,etc.The RBF neural network is used to build the customer side power stealing situation awareness model.The selected six indicators and related data are used as the model inputs,and the dynamic K-means clustering algorithm is used to optimize the model.The output of the model is the customer side power stealing situation awareness result.Based on the sensing results,intelligent early warning is realized by sound light alarm device and intelligent device.The experimental results show that the technology can effectively suppress the customer side electricity stealing behavior.
作者
陈文瑛
龙跃
傅宏
杨芾藜
周川
Chen Wenying;Long Yue;Fu Hong;Yang Fuli;Zhou Chuan(State Grid Chongqing Electric Power Company,Chongqing 400010,China;State Grid Chongqing Electric Power Company Marketing Service Center,Chongqing 400010,China)
出处
《电子技术应用》
2021年第12期69-73,共5页
Application of Electronic Technique
关键词
客户侧
窃电
态势感知
智能预警
感知指标
RBF神经网络
customer side
electricity theft
situation awareness
intelligent early warning
perception index
RBF neural network