Power consumption behavior feature reconstruction and extraction method based on XGBoost and CNN

Abstract

The invention discloses a power consumption behavior feature reconstruction and extraction method based on XGBoost and CNN. The method comprises the following steps of: collecting a plurality of piecse of influence factor data and load data for different target users, and carrying out the data preprocessing; respectively training the data set of each user by using an XGBoost model, and calculating the feature importance in the training process to obtain a feature importance sequence; reconstructing the features of each target user according to the importance sequence to obtain a plurality of feature matrixes; carrying out feature extraction on each type of obtained feature matrix by using a CNN (Convolutional Neural Network); finding out a time point of concept drift in each target user load curve, segmenting to obtain a plurality of time periods, and repeating the steps in different time periods; and carrying out power consumption behavior identification, power load prediction or electricity larceny detection application based on the extracted features. The power consumption behavior modes of the users are identified according to the power consumption characteristics and potential power consumption habits of different users, personalized services are provided for the users, and the power grid service quality is effectively improved.

Description

technical field [0001] The invention relates to the technical field of data mining and power system operation analysis and planning, in particular to a method for reconstructing and extracting power consumption behavior characteristics based on XGBoost and CNN. Background technique [0002] The power system has developed to this day and has become a huge network that provides energy and power for all countries in the world. Its role is to provide various users with reliable, continuous and good-quality electric energy as economically as possible to meet the requirements of various users at any time, that is, to meet the load Require. [0003] In recent years, with the development of the Internet of Things and the advancement of informatization construction, a large amount of user data has been introduced, which contains a wealth of internal laws and derivative information of users' electricity consumption behavior. For different types of users, power consumption is related ...

AI Technical Summary

Problems solved by technology

For different types of users, power consumption is related to many factors, and the relationship between the relevant factors affecting their power consumption and the load is generally different. If the traditional method is used to extract power consumption characteristics for different users according to the same rules, It will lead to the inability to dig out effective user electricity consumption laws, reduce the operating efficiency of power equipment, and affect the economical and safe operation of the power system and the refined operation of the power service department

Images