Research On Load Resource Management And Game Strategy Based On Demand Response

The reform of the electricity market is a key step to promote China’s energy transition and achieve the “carbon peak and neutrality” goals.The reform aims to break the monopoly,intro-duce competition,optimize resources allocation,and form an open and orderly market system.Demand Response(DR)will play an important role in promoting new energy consumption and building a power system with friendly interaction among source,grid,load,and storage.Since diversified resources appear on the demand side(such as distributed power generation,energy storage equipment,electric vehicles,intelligent equipment),tap the demand side potential to replace fossil energy peaking resources,and develop a reasonable market mechanism and load management model,will become an important subject for constructing a new type of power system.Given these backgrounds,this dissertation studies the demand-side management model and electricity purchase and sale strategy in block market under the one-retailer-multi-customer mode,as well as the double-sided non-cooperative game in hourly market and block market un-der the multi-supplier-multi-customer mode.The main innovative contributions are as follows:(1)Demand-side management model and solution approach in the block market.In this dissertation,we innovatively propose a demand-side multi-objective optimization model based on shiftable load aggregation and customer satisfaction.The concept of “deformable” rectangle is adopted,which eliminates unnecessary shape restrictions on shiftable loads and can make short load segments more flexible.The aggregation of short segments into a long load segment helps to reduce the power cost related to start-stop.The modeling of customer satisfaction reflects the customers’ experience and feedback of participating in DR.Aiming at the multi-objective optimization problem,the number of solving times and the priority of the objective problems are introduced to improve the traditional NSGA-II multi-objective genetic algorithm to solve the problem.In the simulation,we quantitatively analyze the selection strategies of different Pareto optimal solutions from the perspective of socioeconomics.(2)Operation mode and electircity purchase and sales strategy under the one-retailer-multi-customer mode in the block market.In this dissertation,we innovatively propose an operation mode and a bi-level optimization model that consider customers’ willingness to participate in DR in the day-ahead block market.This model jointly solves the problem of optimal pricing and DR resources(fixed and shiftable loads)optimal scheduling problem.Aiming at the difficulty of solving the multi-peak,large-scale and discrete DR resources scheduling problem,which is a mixed integer nonlinear problem,we propose an approximate solving algorithm that does not rely on initial values and has good stability.In the simulation,we make a comparative analysis on the DR proportion effects in terms of peak filling and valley filling,increasing the profit of power-retailing companies,and reducing customers’ electricity purchase costs.(3)Double-sided non-cooperative game under the multi-supplier multi-customer mode in the hourly market.In this dissertation,we model the bilaterally open hourly market as a double-sided non-cooperative game with multiple electricity firms and customers,based on the Supply Function Euqilibrium(SFE)model and the DR resources scheduling model,so as to realize each market participants’ utility maximization.On the supply side,four parameterization methods of the SFE model are studied with a rigorous mathematical proof of the existence and uniqueness of Nash equilibrium;on the demand side,instead of simplifying demand as a function of price as the usual way,we consider fixed loads and shiftable loads characteristics and propose a detailed model to benefit each customer and guide the electricity use in reality.In order to obtain the Nash equilibrium solution with good convergence performance,a distributed iterative algorithm based on gradient ascent method and projected gradient method is proposed.Based on real data,the simulation results show that the proposed model has good performance in peak regulation and improving market participants’ utility.(4)Market mechanism and double-sided non-cooperative game model incorporating block and hourly rules.This dissertation innovatively combines the block-based supply side with the hourly-base demand side,and establishes a block-hourly distributed market structure and a double-sided non-cooperative game model.By the block-hourly transformation of the load curve and the electricity price,this model can realize the optimal block bidding based on the supply curve and the optimal customer’s self-utility based on hourly DR.The advantage is that the model not only reflects the advantages of block rules to save electricity costs on the supply side,but also highlights the value of hourly rules that are easy to understand and respond to on the demand side.A distributed iterative algorithm based on gradient ascent and active-set is proposed to solve the double-sided optimization problem.In the simulation based on real data,the effects of the model on peak shaving,saving electricity cost,and operability in practice are evaluated from three dimensions of DR effects,algorithm performance,and strategy analysis.