Application Research Of Power Consumption Data Anomaly Detection Based On Load Cluster Analysis

Electricity marketing is the core business and direct income source of power supply enterprises,which is the key work of power grid enterprises and also the window of direct communication with users.Abnormal power consumption detection mainly refers to the detection of electricity stealing and illegal use of electricity by users,so as to recover economic losses and maintain the order of normal electricity price,which is an important part of power marketing.In recent years,with the development of smart grid information construction promoted by power supply companies,a large number of smart meters and smart devices are connected to the power grid,which brings great convenience to the power consumption information collection of power grid.The huge amount of user electricity data can make power enterprises quickly grasp the power consumption status of users,which provides conditions for enterprises to improve the detection of abnormal electricity consumption through big data and data mining algorithm.However,with the increase of data volume,the data dimension is also growing,which brings challenges to big data applications.Therefore,according to the characteristics of data mining under the background of big data,this paper uses the fusion of principal component analysis and Kmeans clustering method to carry out data mining and feature extraction of user power consumption data,so as to improve the accuracy of power consumption abnormal data detection.The main contributions of this paper are as follows:(1)In view of the difficulty of data feature mining caused by the large amount and high dimension of original power consumption information collected by intelligent terminal,the k-means algorithm and principal component analysis method are combined in this paper to reduce the dimension of data and improve the calculation efficiency of load clustering analysis under the background of big data.(2)After the dimensionality reduction of user load data is analyzed by clustering,the users with the same load characteristics are divided into the same category by using kmeans clustering algorithm,and the typical user load characteristic curve of this category is generated.Finally,the characteristics of different kinds of loads and their influence on the operation of distribution network is analyzed.(3)Considering that the abnormal electricity consumption will be reflected in the user electricity data collected by the power grid,this paper uses the load clustering method to detect the abnormal power consumption of the regular user load curve,and puts forward the abnormal power consumption detection model based on the load clustering.The model combines the user’s typical load characteristic curve and historical load curve for comprehensive judgment,which improves the accuracy of abnormal power consumption detection and electricity stealing behavior detection.