A virtual power plant system

Abstract

The present invention discloses a virtual power plant system, which includes an information center terminal, distributed energy resources, a dispatching center terminal, and a virtual power plant. Among them, the information center terminal is used to announce the virtual power plant's electricity purchase and sale information; the distributed energy resources dynamically aggregate based on the virtual power plant's electricity purchase and sale information announced by the information center terminal, combined with their own power supply and demand characteristics and historical information, and select their preferred virtual power plant; the dispatching center terminal is used to announce dispatching requirements; the virtual power plant is used to optimize the dispatching of the distributed energy resources aggregated to itself in combination with the dispatching requirements announced by the dispatching center terminal to obtain the operation plans of each distributed energy resource. Each distributed energy resource is uniformly dispatched via the virtual power plant, which can relieve the computational pressure on the power grid to directly dispatch a large number of distributed energy resources. The dispatching center terminal only needs to send dispatching requirements to the virtual power plant, and then the virtual power plant can achieve demand response through its own optimized dispatching, thereby ensuring the efficient operation of the power system.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power, and particularly to a virtual power plant system. Background Art

[0002] A virtual power plant is not a power plant in the traditional sense, but a regional concentration of sources, grids, loads, and storages. It can serve as a unified agent for various distributed energy sources to participate in the power market, and can effectively utilize advanced communication and aggregation technologies to combine the multi-energy complementarity effect to achieve multi-source integrated scheduling.

[0003] However, there is not much record and introduction about virtual power plants in the prior art. Therefore, how to construct a virtual power plant system is a technical problem that the prior art urgently needs to solve. Summary of the Invention

[0004] The purpose of the present invention is to overcome the above-mentioned deficiencies of the prior art and provide a virtual power plant system.

[0005] To achieve the above purpose, the technical solution of the present invention is as follows:

[0006] A virtual power plant system includes an information center terminal, distributed energy sources, a dispatching center terminal, and a virtual power plant; wherein,

[0007] The information center terminal is used to announce the power purchase and sale information of the virtual power plant;

[0008] The distributed energy sources dynamically aggregate according to the power purchase and sale information of the virtual power plant announced by the information center terminal, combined with their own power supply and demand characteristics and historical information, and select their preferred virtual power plants;

[0009] The dispatching center terminal is used to announce dispatching requirements;

[0010] The virtual power plant is used to optimize the dispatching of the distributed energy sources aggregated to itself according to the dispatching requirements announced by the dispatching center terminal to obtain the operation plans of each distributed energy source.

[0011] Further, the distributed energy sources include six types of distributed energy sources: thermal power units, gas turbine units, wind and light units, energy storage power stations, and demand response resources. Each distributed energy source can only select one virtual power plant for aggregation at a certain moment.

[0012] Further, the decision-making models of the thermal power units and gas turbines participating in the virtual power plant are similar and are constructed as follows:

[0013]

[0014]

[0015] In this model, the objective function f xRepresents the net profit of unit x; m represents the set of units; α i,x is a 0-1 decision variable, 0 represents not participating in virtual power plant i; P x (t) represents the output of unit x at time t; ρ vpp is the selling electricity price provided by the virtual power plant; a x , b x and c x are respectively the consumption characteristic coefficients of the xth conventional unit, and linear fitting is often used; γ x is the operation and maintenance cost per unit capacity of the unit; P cap,x is the capacity of the unit;

[0016] Constraint 1 is the upper and lower limit constraint of unit output, P x,max and P x,min represent the upper and lower limits of the active power output of unit x; Constraint 2 is the unit ramp constraint, ΔP x,max and ΔP x,min represent the upper and lower limits of unit ramp; Constraints 3 and 4 are the minimum continuous startup and shutdown constraints of the unit, and represent the minimum continuous startup and shutdown times of unit x; u x,t represents the startup and shutdown state of unit x at time t, which is a 0-1 variable, 0 represents shutdown; β x,t Whether unit x starts at time t, which is a 0-1 variable, 1 represents startup; Constraint 5 is the maximum startup times constraint of the unit; Constraint 6 is the relationship constraint between the startup times and the continuous operation time.

[0017] Furthermore, the decision-making model of the wind-solar unit participating in the virtual power plant is similar and constructed as follows:

[0018]

[0019]

[0020] In this model, W represents the set of wind turbines, S represents the set of photovoltaic generators; λ represents the penalty coefficient for wind and light abandonment; p pre,x (t) represents the predicted output of unit x at time t; Constraint 1 is the upper and lower limit constraint of unit output, and w represents the minimum predicted ratio coefficient of the actual output of the unit.

[0021] Furthermore, the decision-making model of the energy storage power station participating in the virtual power plant is constructed as follows:

[0022]

[0023]

[0024] In this model, ES represents the set of energy storage power stations; p ch,x (t), pdis,x (t) corresponds to the charging and discharging power of the energy storage at time t; δ x is the charging and discharging battery loss cost of the energy storage power station.

[0025] Constraint 1 is the constraint of the stored power and the charging and discharging relationship, E x (t) is the stored power of the energy storage power station x at time t, γ loss,x is the battery loss rate of the energy storage power station x; Constraint 2 is the upper and lower limit constraint of the stored power of the energy storage power station, α i,x is the decision variable of the corresponding energy storage power station, E x，max (t) and E x，min (t) represent the upper and lower limits of the stored power of the energy storage power station; Constraints 3 and 4 are the upper and lower limit constraints of the charging and discharging of the energy storage power station, α i,ch,x (t), α i,dis,x (t) is the charging and discharging decision variable at time t, and only one situation is allowed at each moment, p chmax,x , p dismax,x is the maximum charging and discharging power limit; Constraints 5 and 6 are the constraint of the relationship between the charging and discharging decision variable of the energy storage power station and the decision variable participating in the virtual power plant.

[0026] Furthermore, the decision-making model for the demand response resources to participate in the virtual power plant is constructed as follows:

[0027]

[0028]

[0029] In this model, DR represents the set of demand response resources; p x + (t), p x - (t) correspond to the positive power and negative power of the demand response resources at time t respectively; ε x is the per-degree response cost of the demand response resources.

[0030] Constraints 1 and 2 are the upper and lower limit constraints of the demand response capacity, where α i,x + (t), α i,x - (t) are the 0-1 decision variables of the positive and negative powers at time t, P max,s + , P max,s - are the upper and lower limits of the positive and negative powers; Constraints 3 and 4 are the relationship between the positive and negative power decision variables and the virtual power plant decision variables, and the virtual power plant decision variable constraints; Constraints 5 and 6 are the minimum start-up and shutdown time constraints of the demand response resources, where TO x , TC xrespectively represent the minimum start-up time and the minimum shut-down time of the demand response resource x; Constraint 7 is the constraint on the call times of the demand response resource; Constraint 8 is the ramp rate constraint; Constraint 9 is the constraint on the characteristics of the demand response resource, and different types of demand response resources correspond to different characteristic functions.

[0031] Furthermore, the product terms of the unknowns and the 0-1 decision variables in the model are linearized by introducing a third variable and an infinite large number M, and then the model becomes a mixed-integer linear programming model. The specific treatment is as follows:

[0032]

[0033] Furthermore, the utility functions of each distributed energy source are constructed as follows:

[0034]

[0035] This is the utility function of the distributed energy source y, which can evaluate the advantages, disadvantages and preferences of the strategies of the players; a y is the strategy set of the distributed energy source y, and a -y is the strategy set of the participants other than the distributed energy source y, and f y is the objective function of the distributed energy source y, σ is the penalty factor, and P load is the predicted value of the virtual power plant load.

[0036] Furthermore, the potential function of the optimal dispatching of the virtual power plant is constructed as follows:

[0037]

[0038] Furthermore, the said dynamic aggregation includes:

[0039] Establish a dynamic evolution payment matrix of n populations and m strategies:

[0040]

[0041]

[0042] and

[0043] where, Z nm represents the payment set matrix of n populations and m strategies; f xy represents the payment matrix set when the distributed energy source x selects the virtual power plant y, and its specific representation is as shown in Equation (13), where f k1,k2,kn,,,, represents the payment function set when the distributed energy sources 1 to n select the corresponding strategy combinations; a ijDenote the payment function when distributed energy source i selects virtual power plant j, which is characterized by the revenue function of the distributed energy source under the corresponding policy combination, that is, the value of the revenue function when the decision variable of distributed energy source i with respect to virtual power plant j is 1; the set of policy coefficients is ε i , which is a 1×m matrix used to characterize the policy state of distributed energy source i.

[0044] The replicator dynamic equation of the system is characterized as follows:

[0045]

[0046]

[0047] Among them, f pxy Denote the expected revenue when distributed energy source x selects virtual power plant y; P xy Denote the probability that distributed energy source x selects virtual power plant y; Then it is the replicator dynamic equation corresponding to policy xy; Denote the average expected revenue of distributed energy source x. The replicator dynamic equation can effectively characterize the relationship between the current policy revenue and the average revenue of all policies. When the equation is greater than zero, it means that the current policy is dominant, and distributed energy source x tends to maintain the current policy and aggregate to virtual power plant y at the next moment. If it is less than zero, distributed energy source x may learn the policies of other distributed energy sources and perform imitation mutations.

[0048] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0049] Each distributed energy source is uniformly dispatched by the virtual power plant, which can relieve the computational pressure on the power grid to directly dispatch a large number of distributed energy sources. The dispatching center terminal only needs to send the dispatching requirements to the virtual power plant, and then the virtual power plant can achieve demand response through its own optimal dispatching, so as to ensure the efficient operation of the power system; the virtual power plant can also provide an efficient grid connection path for small-capacity distributed energy sources, promoting the consumption of clean power resources; the virtual power plant can also make full use of the multi-energy complementarity effect, coordinate the output of various energy sources, and efficiently utilize various distributed energy sources to provide services such as peak shaving, frequency modulation, and demand response to the power grid, improving the safety and stability of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a schematic diagram of the composition of the virtual power plant system provided by the embodiment of the present invention;

[0051] Figure 2 It is a sectionalized diagram of the dynamic aggregation and evolution trajectory of multiple virtual power plants and multiple distributed energy sources by region. DETAILED DESCRIPTION OF THE INVENTION

[0052] Embodiment:

[0053] The technical solution of the present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0054] Refer to Figure 1 As shown, the virtual power plant system provided in this embodiment mainly includes an information center terminal, distributed energy sources, a dispatching center terminal, and several virtual power plants (VPPs).

[0055] Among them, the information center terminal is used to announce the electricity purchase and sale information of the virtual power plant; at the initial moment of each cycle, the distributed energy sources dynamically aggregate according to the electricity purchase and sale information of the virtual power plant announced by the information center terminal, combined with their own power supply and demand characteristics and historical information, and select their preferred virtual power plants; the dispatching center terminal is used to announce the dispatching requirements; the virtual power plant is used to optimize the dispatching of the distributed energy sources aggregated to itself in combination with the dispatching requirements announced by the dispatching center terminal to obtain the operation plans of each distributed energy source. After the start of the next cycle, each distributed energy source re-selects a virtual power plant for aggregation in combination with historical cooperation information and the information announced by the information center, and so on.

[0056] In this way, each distributed energy source can be uniformly dispatched via the virtual power plant, which can relieve the calculation pressure on the power grid to directly dispatch a large number of distributed energy sources. The dispatching center terminal only needs to send the dispatching requirements to the virtual power plant, and then the virtual power plant can achieve demand response through its own optimized dispatching to ensure the efficient operation of the power system; the virtual power plant can also provide an efficient grid connection path for small-capacity distributed energy sources to promote the consumption of clean power resources; the virtual power plant can also make full use of the multi-energy complementarity effect, coordinate the output of various energy sources, and efficiently use various distributed energy sources to provide services such as peak shaving, frequency modulation, and demand response to the power grid, improving the safety and stability of the power system.

[0057] In a specific embodiment, the above-mentioned distributed energy sources include six types of distributed energy sources: small thermal power units, micro gas turbine units, wind-solar units, energy storage power stations, and demand response resources. Each distributed energy source can only select one virtual power plant for aggregation at a certain moment. The decision-making goal of each distributed energy source is to maximize its own net income. The characteristic modeling of the distributed energy sources is as follows:

[0058] 1) Modeling of small thermal power units and micro gas turbines

[0059] The decision-making models of small thermal power units and micro gas turbines participating in the virtual power plant are similar and can be constructed as follows:

[0060]

[0061]

[0062] In this model, the objective function f x represents the net income of unit x. m represents the set of units; αi,x is a 0-1 decision variable, where 0 represents not participating in virtual power plant i; P x (t) is the output of unit x at time t; ρ vpp is the selling electricity price provided by the virtual power plant; a x , b x and c x are respectively the consumption characteristic coefficients of the xth conventional unit, and linear fitting is often used; γ x is the operation and maintenance cost per unit capacity of the unit; P cap,x is the capacity of the unit.

[0063] Constraint 1 is the upper and lower limits of unit output constraint, P x,max and P x,min represent the upper and lower limits of the active power output of unit x; Constraint 2 is the unit ramp rate constraint, ΔP x,max and ΔP x,min represent the upper and lower limits of the unit ramp rate; Constraints 3 and 4 are the minimum continuous startup and shutdown constraints of the unit, and represent the minimum continuous startup and shutdown time of unit x; u x,t represents the startup and shutdown state of unit x at time t, which is a 0-1 variable, and 0 represents shutdown; β x,t Whether unit x starts at time t, which is a 0-1 variable, and 1 represents startup; Constraint 5 is the maximum startup times constraint of the unit; Constraint 6 is the relationship constraint between the startup times and the continuous operation time.

[0064] 2) Modeling of wind and photovoltaic units

[0065] The decision-making models of wind and photovoltaic units participating in the virtual power plant are similar and can be constructed as follows:

[0066]

[0067]

[0068] In this model, W represents the set of wind turbines, S represents the set of photovoltaic units; λ represents the penalty coefficient for wind and light abandonment; p pre,x (t) represents the predicted output of unit x at time t; Constraint 1 is the upper and lower limits of unit output constraint, and w represents the minimum predicted ratio coefficient of the actual output of the unit.

[0069] 3) Modeling of energy storage power station

[0070] The decision-making model of the energy storage power station participating in the virtual power plant can be constructed as follows:

[0071]

[0072]

[0073] In this model, ES represents the set of energy storage power stations; p ch,x (t), p dis,x (t) correspond to the charging and discharging power of the energy storage at time t respectively; δ x is the charging and discharging battery loss cost of the energy storage power station.

[0074] Constraint 1 is the constraint on the relationship between stored power and charging and discharging. E x (t) is the stored power of energy storage power station x at time t, and γ loss,x is the battery loss rate of energy storage power station x; Constraint 2 is the upper and lower limit constraint of the stored power of the energy storage power station. α i,x is the decision variable of the corresponding energy storage power station. E x，max (t) and E x，min (t) represent the upper and lower limits of the stored power of the energy storage power station; Constraints 3 and 4 are the upper and lower limit constraints of the charging and discharging of the energy storage power station. α i,ch,x (t), α i,dis,x (t) are the charging and discharging decision variables at time t. Only one situation is allowed at each moment. p chmax,x , p dismax,x is the maximum charging and discharging power limit; Constraints 5 and 6 are the constraint on the relationship between the charging and discharging decision variables of the energy storage power station and the decision variables participating in the virtual power plant.

[0075] 4) Modeling of demand response resources

[0076] The decision model for demand response resources participating in the virtual power plant can be constructed as follows:

[0077]

[0078]

[0079] In this model, DR represents the set of demand response resources; p x + (t), p x - (t) correspond to the positive and negative power of the demand response resources at time t respectively; ε x is the cost per unit power response of the demand response resources.

[0080] Constraints 1 and 2 are the upper and lower limit constraints of the demand response capacity. Among them, α i,x + (t), α i,x - (t) are the 0-1 decision variables of the positive and negative power at time t. P max,s + , P max,s -are the upper and lower limits of positive and negative power; Constraints 3 and 4 are the relationships between the positive and negative power decision variables and the virtual power plant decision variables, and the constraints of the virtual power plant decision variables; Constraints 5 and 6 are the minimum start-up and shutdown time constraints of the demand response resources, where TO x , TC x represent the minimum start-up time and the minimum shutdown time of the demand response resource x respectively; Constraint 7 is the constraint on the number of calls of the demand response resource; Constraint 8 is the ramp rate constraint; Constraint 9 is the characteristic constraint of the demand response resource, and different types of demand response resources correspond to different characteristic functions.

[0081] Preferably, the product term of the unknown quantity and the 0-1 decision variable in the model is linearized by introducing a third variable and an infinite large number M, then the model is a mixed integer linear programming model, which can improve the calculation speed. The specific processing is as follows:

[0082]

[0083] The above-mentioned distributed energy characteristic model based on the market environment adds virtual power plant decision variables on the basis of the power supply and demand characteristic models of each distributed energy, and can effectively reflect the decision-making initiative of the aggregation of distributed energy in the market environment.

[0084] The aggregation scheduling model of the internal members of the virtual power plant can be regarded as a game process in which multiple distributed energy parties participate and interact with each other. It is a multi-objective revenue maximization problem. Based on the potential game theory, it can be transformed into a single-objective optimization problem for solution by reasonably constructing a potential function. The model constructed based on the potential game theory has the property of finite improvement, and the changes of the utility functions of each participant will be reflected in the global potential function. Therefore, the optimal solution of the potential function is the global optimal solution, that is, by solving the optimal solution of the potential function, the multi-objective optimization problem can be converted into a single-objective optimization, while ensuring the global optimality of the solution result and reducing the calculation difficulty.

[0085] The utility functions of each distributed energy are constructed as follows:

[0086]

[0087] This model is the utility function of the distributed energy y, which can evaluate the advantages and disadvantages and preferences of the strategies of the players; a y is the strategy set of the distributed energy y, a -y is the strategy set of the participants other than the distributed energy y, f y is the objective function of the distributed energy y, σ is the penalty factor, P load is the predicted value of the virtual power plant load. Further, the potential function of the virtual power plant optimal scheduling is constructed as follows:

[0088]

[0089] The virtual power plant member aggregation scheduling model with F as the potential function conforms to the characteristics of a complete potential game, can effectively solve the optimal scheduling problems of various distributed energy sources under the goal of maximizing the global interests of the virtual power plant, and reduces the difficulty of solving.

[0090] Based on the evolutionary laws of biological populations in nature, using evolutionary game theory to solve game problems has the following advantages: 1) Evolutionary game theory is based on the premise of the bounded rationality and limited information of participants. When making decisions, participants generally go through a process of imitation, learning, and subsequent strategy adjustment. 2) The most fundamental equilibrium basis of the evolutionary game model is the evolutionary stable equilibrium, which is different from the traditional Nash equilibrium solution. The Nash equilibrium solution is a fixed point and is difficult to characterize the characteristics of market dynamic changes. 3) Evolutionary game theory can view problems from the perspective of the entire market, regarding the adjustment process of the strategies of each participant as a dynamic system affected by various factors and finally reaching a stable equilibrium process, which can be solved by relatively complex mathematical methods such as system dynamics and differential equations. Using evolutionary game theory can take into account the market incomplete information and the market game characteristics of the bounded rationality of participants while solving the bilateral decision-making dynamic aggregation problem of multiple virtual power plants and multiple distributed energy sources in the market environment.

[0091] The basic framework of evolutionary game theory consists of the replicator dynamic equation, evolutionary stable strategy, and evolutionary stable equilibrium. Among them, the replicator dynamic equation is a basic dynamic evolution mechanism used to simulate the dynamic adjustment of population strategies, which is derived from the payoff matrix. Classical evolutionary games are confined to the modeling and phase trajectory drawing within 3 populations and 3 strategies, and there are no reports on the modeling of multiple populations and multiple strategies. The present invention establishes an n-population m-strategy dynamic evolution payoff matrix in combination with the dynamic aggregation problem of multiple virtual power plants and multiple distributed energy sources:

[0092]

[0093]

[0094] And

[0095] Among them, Z nm represents the payoff set matrix of n populations and m strategies (n distributed energy sources and m virtual power plants); f xy represents the payoff matrix set when the distributed energy source x selects the virtual power plant y, and its specific representation is as shown in Equation (13), where f k1,k2,kn,,,, represents the payoff function set when the distributed energy sources 1 to n select the corresponding strategy combinations; a ijIt represents the payment function when distributed energy i selects virtual power plant j, which is characterized by the revenue function of distributed energy under the corresponding strategy combination and can be obtained from the optimized scheduling results of the virtual power plant solved by the potential game model, that is, the revenue function value when the decision variable of distributed energy i with respect to virtual power plant j is 1. The set of strategy coefficients is ε i , which is a 1×m matrix used to characterize the strategy state of distributed energy i.

[0096] The replicator dynamic equation of the system is characterized as follows:

[0097]

[0098]

[0099] Among them, f pxy represents the expected revenue of distributed energy x selecting virtual power plant y; P xy represents the probability that distributed energy x selects virtual power plant y; is the replicator dynamic equation corresponding to strategy xy; represents the average expected revenue of distributed energy x. The replicator dynamic equation can effectively characterize the relationship between the current strategy revenue and the average revenue of all strategies. When the equation is greater than zero, it means that the current strategy is dominant, and distributed energy x tends to maintain the current strategy and aggregate to virtual power plant y at the next moment. If it is less than zero, distributed energy x may learn the strategies of other distributed energy and perform imitation mutation. In addition, solving the replicator dynamic equation can obtain the dynamic aggregation evolution trajectory of the system. The existing evolution trajectory can draw at most the evolution situation of three populations and two strategies at the same time because the three-dimensional graph can represent at most three populations, and the coordinate axes can only reflect the two strategies corresponding to the 0 and 1 states. Therefore, the present invention designs a partitioned graph for the evolution trajectory of multiple populations and multiple strategies. Since the evolution equilibrium states of each distributed energy are discrete 0 and 1 states, the left vertical axis coordinate of the partitioned graph is set as multiple 0-1 proportion areas to represent the proportion of each strategy; the corresponding virtual power plant numbers are on the right vertical axis of each area; the horizontal axis is the time axis. Each evolution period corresponds to the evolution from the initial state to the evolution equilibrium state. The evolution period replacement point reflects the moment when the virtual power plant strategy is changed, and the system will use this as the initial state for the next round of evolution. The curves in the graph are the evolution trajectories of each distributed energy. In the system environment, the decisions of each distributed energy will evolve along this trajectory curve to reach the evolution equilibrium state. The partitioned graph of the evolution trajectory can be specifically drawn as Figure 2 shown.

[0100] Among them, T1 - T2 is an evolution period of the system. During this period, the decisions of various populations change along the aggregation trajectory diagram. In this process, each distributed energy source searches for the optimal decision through imitation, learning, and mutation of its game opponents, and the entire system changes dynamically until it reaches an evolutionary equilibrium. At time T2, the virtual power plant changes its power purchase and sale strategy according to the demand, and then each distributed energy source conducts a second round of aggregation evolution until the system reaches a new evolutionary equilibrium state. This sub - region and sub - segment diagram can effectively represent the aggregation evolution mechanism of the market - type virtual power plant, and at the same time intuitively reflect the market evolution trajectory obtained by the present invention based on game theory, providing decision - making reference and risk warning for market participants. In addition, the evolutionary trajectories of each participant can be effectively represented in a two - dimensional graph, solving the limitation of the number of participants and the number of decisions brought by the maximum three - dimensional representation of the evolutionary phase trajectory diagram.

[0101] The above - mentioned embodiments are only for explaining the technical concept and characteristics of the present invention. The purpose is to enable ordinary technicians in the field to understand the content of the present invention and implement it accordingly, and it cannot be used to limit the protection scope of the present invention. Any equivalent changes or modifications made according to the essence of the content of the present invention should be covered within the protection scope of the present invention.

Claims

1. A virtual power plant system, characterized in that, It includes an information center terminal, distributed energy resources, a dispatching center terminal, and a virtual power plant; among them, the information center terminal is used to announce the electricity purchase and sale information of the virtual power plant; the distributed energy resources perform dynamic aggregation by combining the electricity purchase and sale information of the virtual power plant announced by the information center terminal, in combination with their own power supply and demand characteristics and historical information, and select their preferred virtual power plant; the dispatching center terminal is used to announce dispatching requirements; the virtual power plant is used to perform optimal dispatching on the distributed energy resources aggregated to itself by combining the dispatching requirements announced by the dispatching center terminal, and obtain the operation plans of each distributed energy resource; the dynamic aggregation includes: establishing an n-population m-strategy dynamic evolution payoff matrix: And Among them, Z nm represents the payoff set matrix of n populations with m strategies; f xy represents the set of payoff matrices when distributed energy x selects virtual power plant y, and its specific representation is as shown in Equation (13), where f k1,k2,kn,，，, represents the set of payoff functions when distributed energy 1 to n select the corresponding strategy combinations; ε ij represents the payoff function when distributed energy i selects virtual power plant j, which is characterized by the revenue function of the distributed energy under the corresponding strategy combination, that is, the revenue function value when the decision variable of distributed energy i with respect to virtual power plant j is 1; the set of strategy coefficients is ε i , which is a 1×m matrix used to characterize the strategy state of distributed energy i. the replicator dynamic equation of the system is characterized as follows: Among them, f pxy represents the expected revenue of distributed energy x choosing virtual power plant y; P xy represents the probability of distributed energy x choosing virtual power plant y; is the replicator dynamic equation corresponding to strategy xy; represents the average expected revenue of distributed energy x; the replicator dynamic equation can effectively characterize the relationship between the current strategy revenue and the average revenue of all strategies. When the equation is greater than zero, it means that the current strategy is dominant, and distributed energy x tends to maintain the current strategy and aggregate to virtual power plant y at the next moment. If it is less than zero, distributed energy x may learn the strategies of other distributed energy sources and imitate and mutate.

2. The virtual power plant system according to claim 1, wherein the distributed energy resources include thermal power units, gas turbine units, wind and light units, energy storage power stations, and demand response resources. Each distributed energy resource can only select one virtual power plant for aggregation at a certain moment.

3. The virtual power plant system according to claim 2, wherein The decision-making model of the thermal power units and gas turbine units participating in the virtual power plant is constructed as follows: In this model, the objective function f x represents the net revenue of unit x; m represents the set of units; α i,x is a 0-1 decision variable, where 0 means not participating in virtual power plant i; P x (t) represents the output of unit x at time t; ρ vpp is the selling electricity price provided by the virtual power plant; a x , b x and c x are the consumption characteristic coefficients of the x-th conventional unit respectively, and linear fitting is adopted; γ x is the operation and maintenance cost per unit capacity of the unit; P cap,x is the capacity of the unit; Constraint 1 is the upper and lower limits constraint of the unit output, P x,max and P x,min represent the upper and lower limits of the active power output of unit x; Constraint 2 is the unit ramp rate constraint, ΔP x,max and ΔP x,min represent the upper and lower limits of the unit ramp rate; Constraints 3 and 4 are the minimum continuous on - and off - line constraints of the unit, and represent the minimum continuous on - line and off - line times of unit x; u x,t represents the start - stop status of unit x at time t, which is a 0 - 1 variable, and 0 means shutdown; β x,t represents whether unit x starts at time t, which is a 0 - 1 variable, and 1 means start; Constraint 5 is the maximum start - up times constraint of the unit; Constraint 6 is the relationship constraint between the start - up times and the continuous operation time.

4. The virtual power plant system according to claim 3, wherein The decision-making model of the wind and light units participating in the virtual power plant is constructed as follows: In this model, W represents the set of wind turbines, and S represents the set of photovoltaic generators; λ represents the penalty coefficient for curtailment of wind and solar power; p pre,x (t) represents the predicted output of unit x at time t; Constraint 1 is the upper and lower limits constraint of unit output, and w represents the minimum predicted proportion coefficient of the actual output of the unit.

5. The virtual power plant system according to claim 4, characterized in that, The decision-making model of the energy storage power stations participating in the virtual power plant is constructed as follows: In this model, ES represents the set of energy storage power stations; P ch，x (t), P dis，x (t) respectively correspond to the charging and discharging power of the energy storage at time t; δ x is the charging and discharging battery loss cost of the energy storage power station; Constraint 1 is the constraint on the relationship between the stored electricity and the charge-discharge. E x (t) is the stored electricity of energy storage power station x at time t, and γ loss,x is the battery loss rate of energy storage power station x; Constraint 2 is the upper and lower limit constraint of the stored electricity of the energy storage power station, and α i,x is the decision variable of the corresponding energy storage power station, E x，max and E x，min represent the upper and lower limits of the stored electricity of the energy storage power station; Constraints 3 and 4 are the upper and lower limit constraints of the charge-discharge of the energy storage power station, and p chmax,x , p dismax,x is the maximum charge-discharge electricity limit; Constraints 5 and 6 are the constraints on the relationship between the charge-discharge decision variable of the energy storage power station and the decision variable of participating in the virtual power plant, and α i,ch,x (t), α i,dis,x (t) are the charge-discharge decision variables at time t, and only one situation is allowed to occur at each moment.

6. The virtual power plant system according to claim 5, wherein The decision-making model of the demand response resources participating in the virtual power plant is constructed as follows: In this model, DR represents the set of demand response resources; P x + (t), P x - (t) respectively correspond to the positive and negative electricity quantities of the demand response resources at time t; ε x is the cost per unit electricity of the demand response resources; Constraints 1 and 2 are the upper and lower bound constraints of the demand response capacity, where α i,x + (t), α i,x - (t) are the 0-1 decision variables of the positive and negative electricity quantities at time t, P max,s + , P max,s - are the upper and lower bounds of the positive and negative electricity quantities; Constraints 3 and 4 are the relationships between the positive and negative electricity quantity decision variables and the virtual power plant decision variables, and the virtual power plant decision variable constraints; Constraints 5 and 6 are the minimum start-up and shutdown time constraints of the demand response resources, where TO x , TC x represent the minimum start-up time and the minimum shutdown time of the demand response resource x respectively; Constraint 7 is the demand response resource call frequency constraint; Constraint 8 is the ramp rate constraint; Constraint 9 is the demand response resource characteristic constraint, and different types of demand response resources correspond to different characteristic functions.

7. The virtual power plant system according to any one of claims 3-6, characterized in that The product term of the unknown quantity and the 0-1 decision variable in the model is linearly processed by introducing a third variable and an infinite large number M, and then the model becomes a mixed-integer linear programming model. The specific processing is as follows:

8. The virtual power plant system according to claim 7, wherein, The utility functions of each distributed energy resource are constructed as follows: This model is the utility function of distributed energy y, which can evaluate the advantages, disadvantages and preferences of the strategies of the players; a y is the strategy set of distributed energy y, a -y is the strategy set of the participants other than distributed energy y, f y is the objective function of distributed energy y, σ is the penalty factor, P load is the predicted value of the virtual power plant load.

9. The virtual power plant system according to claim 8, characterized in that The potential function of the optimal dispatching of the virtual power plant is constructed as follows:

Images