Research On Flexible Load Identification And Potential Evaluation Technology For Residential Users

Recently,energy saving,emission reduction and energy transformation have become research hotspots under the background of peaking carbon dioxide emissions and carbon neutral.The society will realize the production and lifestyle of green energy and low-carbon development.However,due to the characteristics of intermittency and volatility,it will bring challenges to the operation of the power grid when a large number of clean energy such as wind energy and solar energy flows into the power grid.Thus,to maintain real-time power balance between supply and demand,the grid must have enough flexible resources for dispatch.In recent years,the load of residential users has increased sharply,forming a double peak load of cooling in summer and heating in winter,which has a huge demand response potential.Therefore,studying the response potential of flexible load of residential users and improving grid security operation through demand response are of great significance to ensure the balance of power supply and demand.This dissertation studies the non-intrusive load identification of residents and the assessment of the potential of residential users’ demand response.And the main works are as follows:(1)Firstly,the composition of the demand response system model and the common indicators of load characteristics are introduced.Secondly,the composition and classification of residential users’ demand response resources and the subjective and objective factors that affect residential users’ participation in demand response are analyzed.Finally,the load characteristics of residential users’ power equipment are summarized.(2)A non-intrusive load identification algorithm based on improved particle swarm optimization random forest is proposed.For the three hyperparameters in the random forest algorithm: the number of decision trees,the minimum number of samples and the minimum number of samples of leaf nodes,the improved particle swarm optimization algorithm is used for global optimization to improve the accuracy of nonintrusive load identification.The simulation analysis is carried out by comparing with the default parameters and grid search two hyperparameter determination methods in the REDD public dataset.The results show that the proposed algorithm can effectively improve the identification accuracy of residential users’ electrical equipment.(3)Four metrics: peak usage,combined with frequency of usage,usage consistency consistency,and peak energy consumption are designed to assess demandresponse potential and deal with electricity consumption by a single consumer.Additionally,considering the correlation of demand response potential metrics,the independent weight method is used to set different weights for different metrics.Finally,extensive experiments on the demand response potential of users and electrical equipment in different scenarios demonstrate the effectiveness of the metrics and methods proposed in this paper.In this dissertation,the non-intrusive load identification method is used to identify the flexible load of residential users and evaluate their demand response potential.The proposed demand response potential index and evaluation calculation method can provide reference for residential users to participate in demand response.