Research On Security Portrait Of Industrial Power Users Based On Improved Clustering Algorithm

Electricity is the basic energy related to national economy and people’s livelihoods.Nowadays,the stability of the power grid is of vital importance in increasingly severe situation of production.Since industrial power users are the main consumers of electricity,the research on security portrait of industrial power users is of great practical significance for accurately and efficiently identifying different user types,formulating precise power grid electrical security control strategies,improving power supply service quality,and ensuring safe and reliable power supply.The main work and innovations of this paper are as follows:1.Put forward the idea of "clustering-modeling".Focusing on the temporal characteristics of industrial power users’ electricity consumption behaviors,this paper puts forward the idea of "clustering-modeling",which builds an evaluation system for industrial power users on the basis of cluster analysis technology.The system carries out research from three dimensions,which are electrical security,safety and controllability based on daily load mode,daily freezing amount mode,and daily maximum load rate mode.2.Propose a step-by-step clustering algorithm based on wavelet transformation.This paper tests and verifies the superiority of the algorithm in convergence and complexity.In the experiment and analysis stage of the model,multiple sets of comparative experiments prove that the algorithm effectively solves the problems of the loss of key information in traditional dimensionality reduction algorithms and the high computational complexity in clustering algorithms.3.Construct a security portrait model of industrial power users based on the step-by-step clustering algorithm.Based on the actual data collected by the industrial power users of the S City Power Supply Company of China Southern Power Grid,this paper carries out case analysis,proposes differentiated electricity safety management strategy based on the profiling results and further explores the construction of a "cluster-regression" daily load forecasting model.The experimental results prove that the step-by-step clustering algorithm based on wavelet transformation performs better than the corresponding traditional clustering algorithm and dimensionality reduction + traditional clustering series algorithm in each dimension of the portrait.The research on portrait of industrial power users’ electrical security level based on step-by-step clustering takes user electricity data as the core of the model,and forms a multi-dimensional user portrait through "clustering-modeling",which fills the gap in the research on security portrait of industrial power users and puts forward the application of data mining methods.The method assists the precise management and expands the application of data mining technology in formulating electrical security strategy to ensure the safe and stable operation of the power grid.