Research On The Identification Algorithm Of Abnormal Electricity Consumption Behavior In The Public Transformer Station Area

The public transformer station area is the medium and low voltage connection part of the electricity consumption side of the distribution network,which mainly serves residential users and small enterprise users.Each transformer generally supplies hundreds of household scale electricity,and the number of station areas is very large.Therefore,timely sensing abnormal electricity consumption behavior and dealing with it is of great significance to ensure the safe operation of the electricity grid system and the order,safety and stability of users’ electricity consumption.In view of the identification requirements of users’ abnormal electricity consumption behavior in the public station area,this thesis uses the method of data characteristics to describe the electricity consumption behavior for intelligent analysis.The main work is as follows:(1)There are a large number of users in the public transformer area,and the daily electricity consumption information acquisition system only freezes the electricity data,which is not enough to describe the working condition of the electric energy meter.This thesis studies the monitoring scheme of the user’s electric energy meter based on the site master station,designs the data acquisition system,realizes the real-time acquisition of the working condition parameters of the electric energy meter,and ensures the time consistency of the data such as voltage,phase current,zero line current,power and power factor.(2)The working condition parameters measured by the user’s electric energy meter are composed of multiple items,and the parameters have different degrees of correlation with the electricity consumption behavior.This thesis designs an electric behavior feature extraction method based on principal component analysis,presents the correlation between the feature items by using the correlation coefficient matrix,and reflects the user’s electric behavior characteristics through the principal component expression.On this basis,the corresponding relationship between abnormal characteristic items and abnormal electricity consumption behavior is constructed,and the abnormal electricity consumption behavior of single-phase meter and three-phase meter is divided.(3)Aiming at the problem of intelligent matching between abnormal characteristics and working condition parameters,this thesis designs an abnormal electricity consumption behavior recognition algorithm based on support vector machine,which realizes the classification and recognition of abnormal behaviors such as undervoltage,overvoltage,current imbalance,low electricity factor and so on.In order to further improve the classification accuracy,the thesis introduces the idea of particle swarm optimization,designs an abnormal electricity consumption behavior recognition algorithm based on improved support vector machine,and optimizes the calculation of penalty factor and kernel function.The simulation results show that the recognition accuracy of the improved algorithm has increased from91% to 96%.