基于网络特征与用户行为分析的联合窃电检测方法

Joint detection method for electricity theft based on network characteristics and user behavior analysis

随着电力行业的迅速发展,窃电手段呈现出更复杂隐蔽的特点,给反窃电带来了很大考验.基于此,提出一种基于网络特征与用户行为分析的联合窃电检测方法.一方面,从网络特征分析角度出发,根据当前的电力网络测量数据,基于标准化残差搜索法识别与估计异常参数,准确定位疑似窃电用户所在支路,实现横向窃电检测;另一方面,从用户行为分析角度出发,利用用户的历史用电数据,结合粒子群算法(particle swarm optimization,PSO)和支持向量机(support vector machine,SVM)算法,提高窃电检测分析精度,实现纵向窃电检测.仿真结果表明,利用该联合窃电检测模型能够准确确定窃电异常支路并定位该支路上的窃电用户,有效筛选出电力网络的窃电嫌疑用户.

With the rapid development of the electric power industry,the means of electricity theft present more complex and hidden characteristics,which bring great test to the anti-electricity theft.This paper proposes a joint power based on the characteristics of network and user behavior analysis method;on the one hand,from the perspective of network characteristic analysis,based on the current electric power network measurement data and the standardized residual abnormal search method to identify and estimate parameters,accurate positioning of suspected power users in the branch,to realize transverse detection;on the other hand,from the perspective of user behavior analysis,the historical power consumption data of users is utilized;and particle swarm optimization(PSO)algorithm and support vector machine(SVM)algorithm are combined to improve the analysis accuracy of electricity theft detection and realize longitudinal electricity theft detection.The simulation results show that the combined detection model can accurately determine the abnormal branch of electric larceny and locate the users on the branch of electric larceny,and effectively screen out the suspected users of electric larceny in the power network.