Research On Energy Management Of Integrated Energy System Based On Multi-Agent Consistency

To solve the contradiction between the increasingly prominent energy demand growth and energy shortages,energy usage and environment protection in human society,it is great significance for the energy internet to have integrated planning,design and operation optimization on multi-energy systems,instead of the traditional pattern in which each energy supply system is designed individually and operated separately.This thesis focuses on the modeling and distributed optimization of the multi-energy system based on the idea of promoting renewable energy consumption in energy internet.The entire context takes full account of distributed renewable energy,and comprehensively considers the three major energy systems of electricity,gas and heat to realize distributed coupling among multiple energy sources and minimize the operation cost of integrated energy systems.For solving the day-ahead energy management,this thesis presents a novel distributed double-consensus algorithm,meanwhile,an event-triggered communication strategy is embedded into the execution of the above algorithm,which can dramatically reduce the communication costs,and improve the method to get more practical significance.The specific research contents are as follows:1.The overall framework of integrated energy system is analyzed.Also,it is built output model and operation cost model for source-network-load-storage structures,and established the physical constraints of distributed generation equipment,energy conversion equipment and energy storage equipment as well as the electricity-gas-heat transmission networks.It is convenient to embed models into algorithms for simulation analysis.2.The introduction of We-Energy definitions further embodies the optimization framework of distributed combined-heat-power systems.Aiming at the established a convex optimization of objective function model with constraints,a distributed double-consensus algorithm is proposed,which can be solved in a fully distributed fashion via modifying the feedback item of the supply-demand mismatch.Not only does theoretically prove the optimality and convergence of the proposed algorithm,but also through the comparison with the centralized algorithm,and in the case of time-varying load and plug-and-play,the simulation analysis of the actual examples are carried out,further proof the effectiveness and scalability of the proposed algorithm.3.With the increase of the proportion of renewable energy,its volatility and intermittent impact on the operation of the electric-gas-heat subsystems are becoming more and more significant.This thesis presents a day-ahead energy management strategy of integrated energy system based on event-triggered communications to decrease the communication burden,which is more in line with actual engineering needs.The example shows the day-ahead optimal scheduling of production devices and controllable loads in each area 24 hours and verifies the effectiveness and convergence of the proposed energy management strategy.4.Considering the valve point effect of the generation units,this thesis solves the energy management problem of day-ahead non-convex optimization by combining the multi-agent system with particle swarm optimization.Specifically,each particle is regarded as an agent,and the agent competes and cooperates with its neighbors only.In addition,this chapter sets load satisfaction as an evaluation indicator of load demand response,which takes fully consideration in distributed renewable energy output,balancing energy structure in the energy supply side and energy utilization level in the energy consumption side,and provided a brand new perspective for the optimization of multi-energy system.