Research On Coordinated Optimal Dispatch Of Virtual Power Plant Cluster Considering Uncertain Factors Power Sharing

To seek more efficient,safe and sustainable energy utilization mode,on the basis of the increasingly mature smart grid and new energy generation technologies,it is proposed to gradually replace the traditional centralized energy with distributed energy,which transform the centralized,one-way,producer-controlled energy system into a large number of distributed,less centralized and energy network with more consumer interaction.With the rapid increase of the proportion of intermittent renewable energy resources(mainly wind and photovoltaic)in the distributed energy,the volatility and uncertainty of them has brought great challenges to the safe and stable operation of the main grid once connected to the power grid.Therefore,gave birth to a new concept,i.e.,virtual power plant,it realize the polymerization and coordination of distributed energy optimization such as fuel cell,energy storage system,controllable load,electric cars and so on,via advanced information and communication technology and software system,and participate in electricity market and power grid operation as a special power plants,providing a new way for solving the problem of large-scale distributed energy resources into power network.Firstly,a coordinated optimal operation model for virtual power plant cluster formed by multiple virtual power plants is put forward,to take the virtual power plant with high penetration of renewable energy resources for the study object,considering the daily random variation of renewable energy,electric power load and electricity price,this paper proposes a two-stage(daily-ahead and real-time)optimization scheduling strategy,the virtual power plant scheduling is decomposed into day-ahead scheduling and real-time scheduling,for the purpose of reducing the influences of the uncertain factors on the operation of system.In the day-ahead scheduling stage,VPP cluster is controlled by hierarchical optimization method with the goal of minimizing the system operation cost.Mean-variance theory is used to evaluate the risk of operation cost change caused by uncertainties.In the real-time scheduling stage,rolling scheduling strategy is adopted to optimize each unit in VPP to minimize the unbalanced cost between day-ahead and real-time market.Simulation results show that this method has advantages in reducing the operation cost caused by system risk aversion and reducing the influences of uncertain factors.Secondly,the paper puts forward the power sharing mode of virtual power plant cluster,establishes the distributed transaction decision-making model of virtual power plant cluster,and achieves the goal of minimizing the total operation cost of the cluster by coordinating the exchange power between the virtual power plants within the cluster and the exchange power between the cluster and the power grid.Considering that each virtual power plant is an independent entity with independent decision-making authority,the proposed model is transformed into a distributed optimization problem with independent decision-making of each virtual power plant by using the Alternating Direction Method of Multipliers.Three virtual power plants are selected to form a cluster for simulation analysis,and the following conclusions are drawn:Inside the Clusters,the purchase price among VPPs is lower than the price that VPPs purchase power from power grid,and the price among VPPs is higher than the price that grid purchase power from VPPs,which can greatly promote the power exchange within the cluster.By comparing three modes,namely,direct transaction among virtual power plants and power grid,electric energy transaction among virtual power plants and cluster el.ectric energy sharing transaction proposed in this paper,adopting the model of power sharing can greatly save the operation cost of virtual power plants inside the cluster.In terms of algorithm,alternating direction method of multipliers used in this paper is close to centralized decision method in performance,but the ADMM has a dominant position in protecting individual privacy,and better efficiency of parallel computing,also its computational time is not affected by the number of virtual power plant,at the same time,it can alleviate the burden of communication equipment and costs of operation and maintenance.