Study And Implementation Of Electricity Consumption Behavior Analysis Platform Of Industrial And Commercial Users For Load Data

With the widespread application of smart grids today,power grid companies in various regions have collected massive amounts of load data.How to use big data mining technology to analyze customers’ electricity consumption behaviors,so as to grasp the power users’ demand,and further explore the implications commercial value is a key issue that has yet to be resolved by power grid companies.At the same time,general industrial and commercial electricity use involves a large range of small and medium-sized enterprises.It accounts for a relatively large proportion of the service income of power grid companies,so industrial and commercial customers have higher requirements for power supply services.However,the current power data analysis dimension is still only based on the load data.A comprehensive multi-source data analysis perspective hasn’t yet been formed.And the load data analysis is based on traditional analysis methods,which have poor applicability to massive high-dimensional load data,and do not make full use of the electricity consumption behavior characteristics of industrial and commercial users.Thus,it is not conducive to accurate insight into the electricity consumption of general industrial and commercial customers’ needs,expectations and preferences.Therefore,this article is based on the analysis scenario of electricity consumption behaviors of industrial and commercial users,and mainly proopses the research algorithms of the internal influence factors of electricity consumption and the modeling analysis algorithm of electricity consumption behaviors,to assist the power grid operation decision-making and optimize the operation utility of the smart grid.The key algorithm of this paper is to analyze the electricity consumption behavior of industrial and commercial users.Firstly,it studies the "intrinsic influencing factors of the electricity consumption behavior of industrial and commercial users" from the perspective of multi-sources data,and proposes "The feature importance evaluation algorithm based on SHAP" and "Static attribute research based on tag identification".The former is innovatively based on the model’s feature interpretation framework to perform targeted analysis of the static electricity consumption portraits of industrial and commercial users;the latter uses deep convolutional neural networks as a trainable high-dimensional load feature extractor,and then the evaluation index of the classification model ranks the influence degree of the corresponding intrinsic attributes.The experiments verified that it has a higher attribute recognition accuracy compared with other related machine learning technology selection.Then,the "Deep Variational Auto-Encoding(DVAE)-based Fine-Grained Electricity Behavior Modeling and Analysis Algorithm" is proposed,which draws on the variational auto-encoding model and uses convolutional layers,anomaly score indicators and a sliding window adaptive framework to improve electricity consumption behavioral modeling analysis.Experimental results show that the proposed algorithm performs better than other benchmark methods in identifying abnormal electricity consumption behaviors of industrial and commercial users.At the same time,it generates load patterns based on the principle of DVAE’s generation model,and has high effectiveness and superiority in the discovery of fine-grained typical electricity consumption behaviors.Furthermore,in order to provide an intelligent analysis platform for the proposed research algorithms for business use,this article proposes a load data-oriented analysis platform for electricity consumption behavior of industrial and commercial users,so that relevant personnel in the electric power field without relevant coding algorithm experience can directly use this platform that provides the visualization of selective analysis of regions and industries.The interface is used to analyze the electricity consumption behavior of industrial and commercial users,understand the ranking of relevant internal influencing factors,load patterns,and perform abnormal electricity consumption behavior detection.Therefore,based on the data analysis results provided by the platform,the focus of power analysts is to combine the relevant knowledge of the power field to assist the regulation and operation of the smart grid.Firstly,this article explains the research background and significance of the whole subject,and further introduces the current research status and related technologies at home and abroad.Then it conducts demand analysis on the research content of the analysis platform,and then puts forward key questions and introduces the proposed solution algorithm and experimental process in detail.Then it explains the overall design and implementation of the entire system,as well as the deployment and testing of the platform,and clarifies the functional interaction between the user and the platform.Finally,the work of the thesis is summarized and prospected.