Research On Cooperative Optimal Scheduling Of Microgrid Clusters Based On Multi-Agent Systems

In September 2020,China proposed a climate strategy of "carbon peaking and carbon neutrality",and renewable energy will develop rapidly because of its green and low carbon.As an efficient and flexible form of renewable energy utilization,microgrids have good development prospects.However,affected by the uncertainty of renewable energy generation and load,the anti-interference ability of the current microgrid is weak,and the renewable energy consumption capacity is often limited.Combined with the theory of multiple agents,this paper studies the game relationship between different intelligent subjects in the microgrid and the collaborative optimization strategy between adjacent intelligent microgrids from the multi-time scale level,and the main tasks are as follows:Firstly,considering the specific operation mode of the devices between the microgrid and the microgrid,the collaborative optimization scheduling framework of the multi-agent microgrid group is designed based on the theory of the multi-agent system,and the connection between the various levels is analyzed;for the supply-side and demand-side agents,its functions are described and the mathematical model is established,laying a foundation for follow-up research.Secondly,in view of the shortage of supply-side adjustable resources,it is difficult to maintain the balance of supply and demand of the power system,so how the intelligent subject of microgrid operators can guide scattered and diversified demandside resources to participate in the grid demand response through price incentives is particularly important,based on the negotiation mechanism between the intelligent body,a master-slave game strategy between microgrid operators and flexible loads is proposed to achieve bilateral win-win;considering the characteristics of source-load randomness,in order to make the dispatching results more robust,the introduction of interval numbers depicting random fluctuations of source loads.This paper proposes a pre-day scheduling strategy based on interval master-slave game,which can obtain a more robust pre-day scheduling strategy.Thirdly,aiming at the problem that although the interval master-slave game can depict deviations,the interval range of the results it obtains is wide,and it is difficult to meet the needs of real-time accurate scheduling within the day,an intraday rolling optimization method based on model prediction control is proposed.In the intraday stage,based on the model prediction control repeated rolling optimization,with the goal of tracking the daily grid connection line,to solve the intraday dispatch plan;combined with the characteristics of agent self-learning,a fast intraday dispatch method based on neural network is proposed,which uses data-driven thinking to train neural network parameters,which can realize the rapid solution of intraday scheduling scheme and has higher accuracy than traditional MPC.Finally,in order to improve the energy utilization rate,considering that the energy interaction between microgrids can be directly carried out through the contact line,the collaborative optimization and regulation strategy of microgrid groups is proposed based on the collaboration mechanism of multiple agents,and the scheduling strategy in each intelligent microgrid is solved by the alternate direction multiplier method,and the proposed method can promote complementarity and mutual benefit between microgrids while alleviating the centralized scheduling pressure;considering the complexity of the actual operation process of the microgrid,Combined with the negotiation and collaboration mechanism between multiple agents,a two-layer optimal dispatch model of active distribution network based on multiple agents is proposed,and the upper and lower levels negotiate with each other to achieve the minimum loss of the upper distribution network on the basis of considering the power flow constraint of the distribution network,and the interaction between the lower microgrid agents to achieve their own optimization while meeting the output of the microgrid contact line nodes;and build a simulation analysis platform based on the real-time database,control the dispatching strategy down to the bottom Simulink model,and calculate the results and scheduling strategies by comparing the model.The effectiveness of the policy is verified.