Optimal Dispatching Of Electric And Heating Networks Based On Controllable Load Classification And Prediction Models

Accelerating the clean energy and low-carbon transformation of the power system is an important way to achieve the goal of "double carbon".On the one hand,power grid need to promote energy saving and carbon emission reduction of industrial loads,increase the terminal utilization of electric power by means of electric-to-heat heating to reduce carbon emission.On the other hand,we will vigorously develop renewable energy,increase the installed capacity of renewable energy to the grid,and reduce the power generation from traditional coal-fired units.However,the uncertainties of renewable energy output and power loads bring severe challenges to the security,stability and economic operation of power system.Therefore,the study of power system balance and stability regulation based on controllable loads has theoretical significance and application value to improve the flexibility of power system and renewable energy consumption.This dissertation adjusts the controllable power loads and constant temperature loads to regulate the power consumption behavior of users,suppresses the power fluctuation of the electric heating system,so as to improve the energy utilization efficiency and economy of the electric heating system.Aiming at key technical problems such as classification,prediction and coordination of controllable power loads and constant temperature loads,this dissertation focuses on the classification model of controllable loads based on the recurrent neural network(RNN),the prediction model of controllable loads,and the optimal scheduling of electricthermal combined system based on controllable loads.Through theoretical analysis,simulation and experimental verification,this dissertation provides theoretical and technical basis for controllable loads to improve the stability and economy of thermoelectric combined system.The research work mainly includes the following aspects.(1)A controllable load classification model based on RNN is studied.In response to the problem of power data missing,this dissertation proposes a power data detection and repair method based on SOM-LSTM(Self-Organizing Map-Long Short-Term Memory).The data are classified by SOM,and LSTM is trained according to the characteristic values of different users to complete the detection and repair of different types of missing power data.Furthermore,in order to improve the accuracy of load classification,a method of feature extraction combined with variable mode decomposition is proposed to make the load characteristics more obvious,and the load classification is realized by training the RNN model.(2)Research on forecasting methods of adjustable industrial load and constant temperature load.The operation characteristics of adjustable power load and constant temperature load are discussed.Based on historical operation data,the load autocorrelation characteristics and the correlation with other external factors are analyzed.For adjustable power load,a prediction model based on LSTM is established by historical operation data.Aiming at constant temperature load,LSTM prediction model considering correlation is proposed which considered the correlation of adjacent users’ electricity behavior in a certain area.Furthermore,by comparing with the existing prediction models,the effectiveness of the proposed method is verified and the controllable load is predicted.(3)Research on the optimal scheduling method of electric-thermal combined system based on controllable industrial load.This dissertation built a combined electric heating system model,to analyze the operating characteristics,costs and constraints of cogeneration units,thermal energy storage systems,thermal power units,wind turbines,energy storage systems.Then this dissertation considered energy and power balance constraints to establish power networks,centralized thermal networks and distributed thermal networks.At last we established the objective function of the combined electric heating system considering the lowest economic cost,and applied the method to solve it to realize the controllable load schedulable.(4)A simulation model was built to verify the effectiveness of the proposed method in multiple scenarios.The comparative analysis shows that by flexibly adjusting the controllable electric industrial load and the constant temperature load,the power demand of the electric and thermal load users can be met,and the translational load is used to increase the wind power consumption and improve the economic efficiency of the combined heat and power system,which verifies the correctness and effectiveness of the proposed method in this dissertation.