Decentralized Dispatch Methods Of Energy Systems With Multiple Agents

With the increasing scale of power system and the development of power market,multiarea power system has become a trend.Traditional centralized optimization methods can not meet the requirements of the operation calculation of multi-area power systems,because of the data privacy,the communication congestion and the independent dispatch of regional power system.In addition,with the development of social economy and technology,different energy supply systems have been interacted and coupled.There are multiple agents in multi-area power systems and integrated energy systems.Each agent has its own control center.Because of the control independence and information privacy,it is not realistic for all agents to jointly carry out centralized optimal dispatch.Therefore,the paper mainly studies the decentralized optimization of energy system with multiple agents.The research object expands from single power system to integrated electrical and heating system.The decentralized optimization algorithm proposed in the paper can achieve the same solution as the centralized method while guaranteeing the control independence and information privacy of each agent.It has been successfully applied in the following two aspects:(1)Reactive power optimization is essentially finding the optimal power flow for optimizing the voltage profile and power flow in the steady state.Considering that the operation of devices is restricted by their service lifetimes and regulations,dynamic reactive power optimization(DRPO)is formulated as a mixed-integer nonlinear programming problem considering the practical regulation constraints of devices.To preserve the control independence and information privacy of the distributed subnetworks,a three-stage programming approach is proposed to achieve a fully decentralized solution to the DRPO problem of the multiarea power system.In addition to incorporating optimality condition decomposition(OCD)algorithm,the decentralized DRPO method includes three stages for addressing the difficulty in handling the discrete variables,and it can achieve a fully distributed solution for the multiarea DRPO problem,with only minor boundary information to be exchanged.A forward-backward-pass dynamic programming approach with computational complexity of polynomial order is utilized to solve the stepwise fitting problem.The simulation results for three test systems demonstrate that the proposed decentralized method can obtain a high-quality solution in a decentralized manner with promising computation performance considering the practical regulation constraints.(2)The strong linkage of electric power and heat supplies can be decoupled to reduce wind power curtailment by exploiting the energy storage and regulation capabilities of the district heating network(DHN),electric boilers and heat storage tanks.In this paper,the coordinated dispatch of integrated electrical and heating systems(CDIEHS),considering the energy storage of both pipelines in the DHN and heat storage tanks,is formulated as a convex quadratic program.Since the electric power system(EPS)and the district heating system(DHS)are controlled separately by different operation organizations,CDIEHS is solved in a decentralized manner using OCD algorithm.The OCD algorithm,with guaranteed convergence for convex programs,can achieve a fully distributed solution for the CDIEHS and requires only minor boundary information exchange between the EPS and the DHS.The simulation results are discussed for two test systems,including a real power system,to demonstrate the effectiveness of the proposed decentralized method,which achieves the same solution as the centralized method in a moderate number of iterations.