A two-layer optimization scheduling method and system for promoting source-storage-load coordination among multiple virtual grids

Abstract

This invention discloses a two-layer optimization scheduling method and system for promoting source-storage-load coordination among multiple virtual grids. The method includes: acquiring parameter information of distributed energy resources; establishing the range of transaction price decisions through a balanced power bidding mechanism based on the parameter information of distributed energy resources; setting the overall operating revenue maximization as the optimization objective in the upper-layer model; conducting non-cooperative game theory among virtual grids to make decisions on transaction plans and power generation plans among virtual grids, obtaining the interactive power between virtual grids; using the interactive power between virtual grids as a parameter, setting the minimum power generation cost of virtual grids as the optimization objective in the lower-layer model; making decisions on the output of each power generation unit within the virtual grid at different times, obtaining the source-storage-load optimization results within the virtual grid; feeding the source-storage-load optimization results back to the upper-layer model until the set iteration conditions are met, obtaining the output of each power generation unit and the interactive power with the main grid.

Description

Technical Field

[0001] This invention relates to the field of microgrid optimization scheduling technology, specifically to a two-layer optimization scheduling method and system that promotes source-storage-load coordination among multiple virtual grids. Background Technology

[0002] With the increasing penetration of distributed renewable energy resources, energy and environmental issues have become prominent. Distributed energy resources (DERs), which generate electricity from resources such as wind, solar, and hydropower, have become a major direction for future global energy development due to their high cleanliness, low generation cost, and renewability. However, distributed energy is affected by factors such as weather conditions and geographical location, resulting in relatively small generation capacity and problems such as intermittency, randomness, and difficulty in centralization due to dispersed locations. This can lead to overcapacity, reduced efficiency, and increased costs. Direct grid connection can impact grid stability and impose certain restrictions on participation in the electricity market. Aggregating DERs into a virtual grid, however, facilitates regulation.

[0003] Current research on virtual grid optimization scheduling can be divided into internal optimization scheduling and external optimization scheduling. Internal optimization scheduling mainly refers to the allocation or capacity optimization of power generation plans for distributed power sources, gas turbines, and energy storage within the virtual grid, based on the virtual grid control center. External optimization scheduling focuses on the optimized operation when the virtual grid is treated as a whole for centralized power system scheduling. The process of external optimization of the virtual grid is similar to that of internal optimization, but it needs to consider the system line power flow constraints. Therefore, a two-layer optimization scheduling model can be established for solution. For example, Chinese invention patent application CN115149578A discloses a distributed optimization scheduling method for interconnected distribution networks that takes demand response into account. It constructs a two-layer distributed optimization scheduling model; determines the objective functions and constraints of the two layers in the two-layer optimization scheduling model; generates the constraints of each layer based on decision variables; converts the constraints of the lower layer into new constraints; and uses linearization methods to process the new constraints, thereby transforming the single-layer nonlinear optimization scheduling model into a single-layer mixed-integer linear optimization model for solution. Chinese invention patent application document with publication number CN115149586A discloses a method and system for coordinated optimization of distributed energy aggregation control and autonomous control. The implementation steps include: (1) proposing a regional power grid distributed energy control system model; (2) constructing a joint optimization problem of distributed photovoltaic total output cost and power supply reliability; (3) a two-stage coordinated optimization of load aggregation control and distributed photovoltaic autonomous control based on consensus algorithm enhanced federated deep reinforcement learning. Summary of the Invention

[0004] The technical problem to be solved by this invention is how to improve the economic benefits of virtual grids and ensure efficient use of resources.

[0005] The present invention solves the above-mentioned technical problems through the following technical means:

[0006] This invention proposes a two-layer optimization scheduling method to promote source-storage-load coordination among multiple virtual grids, the method comprising:

[0007] Obtain parameter information of distributed energy sources, including multiple virtual grid photovoltaics, load forecast power, and control parameters of power generation units;

[0008] Based on the parameter information of the distributed energy, the range of the transaction electricity price decision is established through the balanced power bidding mechanism. In the upper-level model, the optimization objective is to maximize the overall operating revenue. Non-cooperative game is conducted between the virtual grids to make decisions on the transaction plan and power generation plan between the virtual grids, and the interactive electricity between the virtual grids is obtained.

[0009] Using the power exchange between the virtual grids as a parameter, the lower-level model takes the lowest power generation cost of the virtual grid as the optimization objective, and makes decisions on the output of each power generation unit in the virtual grid at different times to obtain the source-storage-load optimization results within the virtual grid.

[0010] The optimization results of the source and storage load within the virtual grid are fed back to the upper-level model, so that the upper-level model can find the optimal solution under its own constraints and transmit it to the lower-level model. This process continues until the set iteration conditions are met, and then the output of each power generation unit and the interaction power with the main power grid are obtained.

[0011] This invention addresses the coordination and optimization among multiple virtual grids based on a balanced power bidding strategy and non-cooperative game theory, aiming to maximize overall operational benefits. Within each virtual grid, the interactive power generation between grids is used as a parameter. In the lower-level model, minimizing the power generation cost of each virtual grid is the optimization objective. The output of each power generation unit within the virtual grid is decided at different times to achieve internal power allocation. Through a multi-virtual grid bi-layer optimization scheduling strategy that balances inter-grid coordination and intra-grid optimization, the optimal solution for the power generation plan is found, with the goals of maximizing overall operational benefits and minimizing power generation costs. This achieves source-storage-load coordination, improves the economic efficiency of the virtual grid, and ensures efficient resource utilization.

[0012] Furthermore, in the upper-level model, the objective function is:

[0013]

[0014] In the formula: The gain of the i-th virtual grid at time t; Let be the revenue obtained from selling electricity in the i-th virtual grid at time t; Let be the revenue generated by the interaction of the i-th virtual grid with the main power grid at time t; Let $t$ be the revenue that the $i$-th virtual grid receives from purchasing electricity from the other virtual grids at time $t$. Let t be the controllable power output cost of the i-th virtual grid in a transaction with other virtual grids;

[0015] The constraints are:

[0016] λ i,t ≥λ i,ke

[0017] In the formula: λ i,t To compete for electricity prices; λ i,ke The unit power generation cost of a controllable generating unit.

[0018] Furthermore, before using the power exchange between the virtual grids as a parameter, and taking the minimum power generation cost of the virtual grid as the optimization objective in the lower-level model to make decisions on the output of each power generation unit within the virtual grid at different time periods, and obtaining the source-storage-load optimization result within the virtual grid, the method further includes:

[0019] The power generation units within the virtual grid are each constructed as aggregate unit mathematical models, wherein the power generation units include at least two of the following: photovoltaic modules, gas turbines, and battery energy storage.

[0020] The aggregation unit data model corresponding to the photovoltaic module is as follows:

[0021]

[0022] In the formula: P pv P represents the output power of the photovoltaic module. mpp，STC γ is the output power of the photovoltaic module under MPPT control; γ is the power temperature coefficient; S is the actual irradiance; S STC T represents the reference value for light intensity; T represents the photovoltaic operating temperature; T STC For reference temperature;

[0023] The mathematical model of the aggregation unit corresponding to the gas turbine is:

[0024] P MT =A MT η MT H MT

[0025] In the formula: P MT A represents the output power of the gas turbine. MT The amount of fuel consumed by the gas turbine; η MT For power generation efficiency; H MTThe calorific value of a micro gas turbine;

[0026] The mathematical model for the aggregation unit corresponding to the battery energy storage is:

[0027]

[0028] In the formula: P B,t P is the amount of electricity stored inside the battery at time t; α is the battery's self-discharge rate; B,t-Δt η represents the amount of electricity stored inside the battery at time t-Δt; in η out Improve battery energy storage charging and discharging efficiency; P B,in P B,out Δt represents the charging and discharging power of the battery energy storage; 'a' represents the charging and discharging state of the battery energy storage, 0 for discharging and 1 for charging; Δt represents the time period.

[0029] Furthermore, in the lower-level model, the constraints are power generation unit security constraints, tie-line security constraints, and power balance constraints. The objective function for the power generation cost of the virtual grid at each time step is:

[0030]

[0031] In the formula: The cost of generating electricity at time t for the i-th virtual grid; Let $\frac{i}{i}$ be the output cost of the photovoltaic modules in the $i$-th virtual grid. Let $\frac{i}{i}$ be the output cost of battery energy storage in the $i$-th virtual grid. Let $\frac{i}{i}$ be the output cost of the micro gas turbine in the $i$-th virtual grid. Let be the interaction cost of the electricity exchanged between the i-th virtual grid and the main power grid.

[0032] Furthermore, the output cost of the photovoltaic module in the i-th virtual grid is:

[0033]

[0034] In the formula: For the cost of photovoltaic power generation, Let a be the output power of the photovoltaic system in the i-th virtual grid at time t. pv b pv c pv This refers to the power generation cost coefficient of the photovoltaic power generation system.

[0035] The output cost of battery energy storage in the i-th virtual grid is:

[0036]

[0037] In the formula: The output cost of battery energy storage; M Li Total cost of energy storage battery maintenance; Let t be the output power of the battery energy storage; N be the total cycle life of the battery; U be the AC side voltage of the battery; and C be the rated capacity of the battery.

[0038] The output cost of the micro gas turbine in the i-th virtual grid is:

[0039]

[0040] In the formula: For the output cost of the gas turbine, For the fuel cost of operating micro gas turbines, This refers to the start-up and shutdown costs of the gas turbine.

[0041] Furthermore, the constraints set in the lower-level model also include:

[0042]

[0043] Where: ∑P il,t The power participating in the bidding within the virtual grid; The maximum output power of the gas turbine within the virtual grid. This represents the maximum output power of photovoltaic power generation within the virtual grid. This represents the maximum output power of energy stored within the virtual grid.

[0044] Furthermore, based on the parameter information of the distributed energy source, the range of the electricity trading price decision is established through a balanced power bidding mechanism. In the upper-level model, the optimization objective is to maximize the overall operating revenue. The virtual grids engage in non-cooperative game theory to make decisions on the trading and power generation plans among them. This can yield the Nash equilibrium solution with a high convergence speed, obtaining the interactive electricity volume between the virtual grids, including:

[0045] Step 1): Input parameters, mainly including real-time electricity price, photovoltaic output, real-time load, etc.

[0046] Step 2): Establish a non-cooperative game model of multiple virtual power plants based on the benefit function, and generate the strategy space;

[0047] Step 3): Select the initial equilibrium value J of the trading electricity price and trading volume in the policy space corresponding to each virtual grid. i0 ={λ i0 P i0}, λ i0 Let P be the initial value of the electricity trading price. i0 This is the initial value for the traded electricity volume;

[0048] Step 4): Initialize the position and velocity of the particle swarm, with the virtual grid as the game participant;

[0049] Step 5): In the k-th iteration, take the result of the previous iteration as the initial value, and make independent decisions through the adaptive mutation particle swarm algorithm. If an equilibrium point is found, proceed to step 6) and output the result; otherwise, return to step 4) to make optimization decisions until the iteration ends.

[0050] Step 6): Output the Nash equilibrium solution, which is the strategy combination under the equilibrium state.

[0051] Among all strategy combinations, if a strategy set exists... satisfy This strategy is the Nash equilibrium solution, which can be expressed as:

[0052]

[0053] In the formula, Let J1, J2, ..., J be the Nash equilibrium strategy of the game participants, representing the optimal scheduling strategy for each virtual grid when the opponent chooses their optimal strategy; that is, the highest payoff for each virtual grid under this strategy combination in the sense of Nash equilibrium. n This is a non-optimal strategy; argmaxE i (·) represents the set of variables that maximizes the objective function. This represents the virtual grid strategy except for the i-th virtual grid.

[0054] The strategy space is as follows:

[0055]

[0056] In the formula: Adjustable capacity within the virtual grid. Let be the output power of the photovoltaic module in the i-th virtual grid at time t. This represents the maximum output power of the gas turbine within the virtual grid. Load demand within the virtual grid.

[0057] Furthermore, the step of using the interactive power between the virtual grids as a parameter, and optimizing the lower-level model with the lowest power generation cost of the virtual grid as the optimization objective, makes decisions on the output of each power generation unit within the virtual grid at different times, obtaining the source-storage-load optimization results within the virtual grid. This rationally utilizes the spatiotemporal complementarity of each power generation unit within the virtual grid to achieve power coordination within a single virtual grid, thereby reducing the intermittency of distributed energy sources such as photovoltaics, improving the utilization rate of distributed energy sources, and enhancing the power supply reliability of the grid within the aggregation area. This includes:

[0058] Based on the state, load demand, and real-time electricity price of the power generation units within the virtual grid, and under the premise of satisfying the optimal operation constraints, the output of each power generation unit and its interaction power with the main power grid under the optimal objective function state are determined. The problem to be solved can be expressed as:

[0059]

[0060] In the formula, Let be the power generation cost at time t for the i-th virtual grid. Let be the output power of the battery energy storage in the i-th virtual grid at time t. Let be the output power of the micro gas turbine in the i-th virtual grid at time t.

[0061] The specific solution steps are as follows:

[0062] Step 1): Input the data of the power generation units in the virtual grid before optimization, the demand and output information, and the electricity price;

[0063] Step 2): Establish the particle swarm dimension and initialize the particle swarm; each particle represents a combination of the energy storage and power generation unit and the micro gas turbine output within the virtual grid, representing each set of scheduling data, that is, the particle swarm dimension is equal to the number of energy storage and power generation units and micro gas turbines; set the number of particles, initialize the position of each particle, select a set of data within the power allowable range of the power generation unit, that is, initialize the output power of the energy storage and gas turbines; at the same time, initialize the particle flight speed.

[0064] Step 3): Calculate the fitness value: Calculate the fitness of each particle according to the optimization objective function. This value is the gain of the virtual grid. Take the current best result as the local optimum and find the global optimum particle to complete the initialization of local and global extrema.

[0065] Step 4): Iterative optimization: Update particle position and velocity, recalculate fitness value, and determine whether to update local and global optima;

[0066] Step 5): Determine if the iteration result has converged: When the maximum number of iterations set or the global best position has not changed for a number of consecutive iterations, output the global optimal solution, that is, the optimal output power of the energy storage power generation unit and the micro gas turbine; otherwise, return to step 4) for re-looping.

[0067] Furthermore, this invention proposes a two-layer optimized scheduling system to promote source-storage-load coordination among multiple virtual grids, the system comprising:

[0068] The acquisition module is used to acquire parameter information of distributed energy sources, including multiple virtual grid photovoltaics, load forecast power, and control parameters of power generation units;

[0069] The upper-level optimization module is used to determine the range of electricity trading price decisions based on the parameter information of the distributed energy source through a balanced power bidding mechanism. In the upper-level model, the optimization objective is to maximize the overall operating revenue. The virtual grids conduct non-cooperative game theory to make decisions on the trading plan and power generation plan between the virtual grids, and obtain the interactive electricity between the virtual grids.

[0070] The lower-level optimization module is used to take the power exchange between the virtual grids as a parameter, and take the lowest power generation cost of the virtual grids as the optimization objective in the lower-level model to make decisions on the output of each power generation unit in the virtual grids at different times, so as to obtain the source-storage-load optimization results within the virtual grids.

[0071] The feedback module is used to feed back the source-storage-load optimization results inside the virtual grid to the upper-level model, so that the upper-level model can find the optimal solution under its own constraints and transmit it to the lower-level model until the set iteration conditions are met, and then obtain the output of each power generation unit and the interaction power with the large power grid.

[0072] Furthermore, in the upper-level model, the objective function is:

[0073]

[0074] In the formula: The gain of the i-th virtual grid at time t; Let be the revenue obtained from selling electricity in the i-th virtual grid at time t; Let be the revenue generated by the interaction of the i-th virtual grid with the main power grid at time t; Let $t$ be the revenue that the $i$-th virtual grid receives from purchasing electricity from the other virtual grids at time $t$. Let t be the controllable power output cost of the i-th virtual grid in a transaction with other virtual grids;

[0075] The constraints are:

[0076] λ i,t ≥λ i,ke

[0077] In the formula: λ i,t To compete for electricity prices; λ i,ke The unit power generation cost of a controllable generating unit;

[0078] In the lower-level model, the constraints are generation unit security constraints, tie-line security constraints, and power balance constraints. The objective function for the generation cost of the virtual grid at each time step is:

[0079]

[0080] In the formula: The cost of generating electricity at time t for the i-th virtual grid; Let $\frac{i}{i}$ be the output cost of the photovoltaic modules in the $i$-th virtual grid. Let $\frac{i}{i}$ be the output cost of battery energy storage in the $i$-th virtual grid. Let $\frac{i}{i}$ be the output cost of the micro gas turbine in the $i$-th virtual grid. Let be the interaction cost of the electricity exchanged between the i-th virtual grid and the main power grid.

[0081] The advantages of this invention are:

[0082] (1) This invention addresses the coordination and optimization among multiple virtual grids based on a balanced power bidding strategy and non-cooperative game theory, thereby maximizing overall operational benefits. Within the virtual grid, the interactive power between virtual grids is used as a parameter. In the lower-level model, the optimization objective is to minimize the power generation cost of the virtual grid. The output of each power generation unit within the virtual grid is decided at different times to achieve internal power allocation. Through a multi-virtual grid dual-layer optimization scheduling strategy that takes into account both inter-grid coordination and intra-grid optimization, the optimal solution of the power generation plan is obtained with the goal of maximizing overall operational benefits and minimizing power generation costs. This achieves source-storage-load coordination, improves the economic efficiency of the virtual grid, and ensures efficient utilization of resources.

[0083] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0084] Figure 1 This is a flowchart illustrating the two-layer optimization scheduling method for promoting source-storage-load coordination among multiple virtual grids proposed in Embodiment 1 of the present invention.

[0085] Figure 2 This is a schematic diagram of the overall process of the two-layer optimization scheduling method for promoting source-storage-load coordination among multiple virtual grids proposed in Embodiment 1 of the present invention;

[0086] Figure 3 This is a system model diagram of the two-layer optimization scheduling method for promoting source-storage-load coordination among multiple virtual grids in Embodiment 1 of the present invention;

[0087] Figure 4 This is a graph showing the electricity price change in Embodiment 1 of the present invention;

[0088] Figure 5These are the photovoltaic output and load variation curves of each virtual grid in Embodiment 1 of the present invention, wherein (a), (b), and (c) are the photovoltaic output and load variation curves corresponding to virtual grid 1, virtual grid 2, and virtual grid 3, respectively;

[0089] Figure 6 These are the output curves of each virtual grid under the double-layer optimization in Embodiment 1 of the present invention, where (a), (b), and (c) are the outputs inside virtual grid 1, virtual grid 2, and virtual grid 3, respectively.

[0090] Figure 7 This refers to the interaction power between virtual meshes under dual-layer optimization in Embodiment 1 of the present invention.

[0091] Figure 8 This is a schematic diagram of the structure of the two-layer optimized scheduling system for promoting source-storage-load coordination among multiple virtual grids, as proposed in Embodiment 2 of the present invention. Detailed Implementation

[0092] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0093] like Figures 1 to 2 As shown, the first embodiment of the present invention proposes a two-layer optimization scheduling method to promote source-storage-load coordination among multiple virtual grids. The method includes the following steps:

[0094] S10. Obtain parameter information of distributed energy, including multiple virtual grid photovoltaics, load forecast power, and control parameters of power generation units;

[0095] It should be noted that the parameter information of distributed energy is specifically used for the calculation of the two-layer model. The output cost of photovoltaic modules in the virtual grid is calculated from the photovoltaic power of the virtual grid; the adjustable capacity inside the virtual grid is calculated from the photovoltaic and load prediction power of the virtual grid; and the output power of the gas turbine in the virtual grid is calculated from the control parameters of the power generation unit.

[0096] S20. Based on the parameter information of the distributed energy, the range of the transaction electricity price decision is established through the balanced power bidding mechanism. In the upper-level model, the optimization objective is to maximize the overall operating revenue. Non-cooperative game is conducted between the virtual grids to make decisions on the transaction plan and power generation plan between the virtual grids, and the interactive electricity between the virtual grids is obtained.

[0097] It should be noted that when trading between virtual grids, the transaction price needs to be considered. The scope of the transaction price decision is established through the balanced power bidding mechanism. The virtual grid as a whole reports the bidding strategy in a unified manner and conducts non-cooperative game to pursue its own profit maximization.

[0098] S30. Using the power exchange between the virtual grids as a parameter, and taking the lowest power generation cost of the virtual grid as the optimization objective in the lower-level model, the output of each power generation unit in the virtual grid at different times is decided to obtain the source-storage-load optimization result inside the virtual grid.

[0099] It should be noted that, based on the status, load demand, and real-time electricity price of the power generation units within the virtual grid, the spatiotemporal complementarity of each power generation unit is rationally utilized. With the goal of minimizing the power generation cost of the virtual grid, decisions are made on the output of each power generation unit within the virtual grid at different times, thereby achieving power coordination within a single virtual grid. This improves the utilization rate of distributed energy and the power supply reliability of the power grid within the aggregated area.

[0100] S40. Feed back the source-storage-load optimization results inside the virtual grid to the upper-level model, so that the upper-level model can find the optimal solution under its own constraints and transmit it to the lower-level model until the set iteration conditions are met, and then obtain the output of each power generation unit and the interaction power with the large power grid.

[0101] It should be noted that the iteration condition described in this embodiment is the number of iterations or the iteration threshold.

[0102] This embodiment addresses the interaction between virtual grids. When trading between virtual grids, the transaction price needs to be considered. A balanced power bidding mechanism establishes the scope of the transaction price decision. The virtual grid, as a whole, reports its bidding strategy uniformly and engages in non-cooperative game theory to maximize its own profit. Within the virtual grid, the interactive power volume between virtual grids is used as a parameter. In the lower-level model, the optimization objective is to minimize the power generation cost of the virtual grid. Decisions are made on the output of each power generation unit within the virtual grid at different times to achieve internal power allocation. Through a multi-virtual grid two-layer optimization scheduling strategy that considers both inter-grid coordination and intra-grid optimization, the optimal solution for the power generation plan is found with the goal of maximizing overall operating benefits and minimizing power generation costs. This achieves source-storage-load coordination, improves the economic efficiency of the virtual grid, and ensures efficient resource utilization.

[0103] Furthermore, this embodiment is applied to a multi-virtual grid scenario composed of photovoltaic, load, and energy storage. Regarding the power interaction problem between virtual grids, considering the inherent competitive characteristics between virtual grids, the non-cooperative game algorithm enables each virtual grid to compete with its own benefit as the goal. The transaction plan and power generation plan between virtual grids are solved through non-cooperative game theory to make decisions.

[0104] In one embodiment, a balanced power bidding mechanism is used to determine the scope of electricity trading prices. This mechanism not only satisfies the relevant bidding rules but also meets the economic requirements of each virtual grid. This embodiment considers factors such as generation cost, dynamic response, available power generation, and supply-demand relationship. In the upper-level model, the objective function for inter-grid scheduling optimization is established as shown below:

[0105]

[0106] In the formula: The gain of the i-th virtual grid at time t; Let be the revenue obtained from selling electricity in the i-th virtual grid at time t; Let be the revenue generated by the interaction of the i-th virtual grid with the main power grid at time t; Let $t$ be the revenue that the $i$-th virtual grid receives from purchasing electricity from the other virtual grids at time $t$. Let t be the controllable power output cost of the i-th virtual grid in a transaction with other virtual grids;

[0107] The constraints on electricity prices during game-like bidding between virtual grids are as follows:

[0108] λ i,t ≥λ i,ke

[0109] In the formula: λ i,t To compete for electricity prices; λ i,ke The unit power generation cost of a controllable generating unit.

[0110] Furthermore, the constraint condition of the objective function in the lower-level model is that the power participating in the bidding within the virtual grid depends on the maximum output power within the virtual grid.

[0111]

[0112] In the formula, ∑P il,t The power participating in the bidding within the virtual grid. This represents the maximum output power of the gas turbine within the virtual grid. This represents the maximum output power of photovoltaic power generation within the virtual grid. This represents the maximum output power of energy storage within the virtual grid.

[0113] In one embodiment, in a system comprising multiple virtual grids, each virtual grid contains a power generation unit comprising a photovoltaic system, an energy storage system, and a micro gas turbine system. The characteristics of each power generation unit are fully utilized and aggregated and controlled in the form of a virtual grid. Based on the output characteristics of each power generation unit within the virtual grid, an aggregated unit mathematical model is constructed. The power generation unit includes at least two of the following: photovoltaic modules, gas turbines, and battery energy storage.

[0114] In order to continuously extract the maximum energy from the solar panels, the photovoltaic power generation system is usually operated at the maximum power point (MPP). The aggregation unit data model corresponding to the photovoltaic module is as follows:

[0115]

[0116] In the formula: P pv P represents the output power of the photovoltaic module. mpp，STC γ is the output power of the photovoltaic module under MPPT control; γ is the power temperature coefficient; S is the actual irradiance; S STC T represents the reference value for light intensity; T represents the photovoltaic operating temperature; T STC For reference temperature;

[0117] The output power of the micro gas turbine power generation model is proportional to the fuel consumed. The mathematical model of the aggregation unit corresponding to the gas turbine is as follows:

[0118] P MT =A MT η MT H MT

[0119] In the formula: P MT A represents the output power of the gas turbine. MT The amount of fuel consumed by the gas turbine; η MT For power generation efficiency; H MT The calorific value of a micro gas turbine;

[0120] The mathematical model of the aggregated unit corresponding to the battery energy storage represents the battery energy storage power generation model, which simulates the relationship between the stored energy at time t and the energy of charging and discharging during the time period Δt using the following formula:

[0121]

[0122] In the formula: P B,t P is the amount of electricity stored inside the battery at time t; α is the battery's self-discharge rate; B,t-Δt η represents the amount of electricity stored inside the battery at time t-Δt; in η out Improve battery energy storage charging and discharging efficiency; P B,in P B,out Δt represents the charging and discharging power of the battery energy storage; 'a' represents the charging and discharging state of the battery energy storage, 0 for discharging and 1 for charging; Δt represents the time period.

[0123] In one embodiment, in the lower-level model, the constraints are power generation unit safety constraints, tie line safety constraints, and power balance constraints. During the optimization operation, the output of the virtual grid is usually affected by the maximum power capacity of the tie line and the installed capacity of the internal power generation units. Therefore, in order to maintain the reliability and stability of the virtual grid operation, the output of the power generation units in the grid is limited to a certain range under the premise of satisfying the supply and demand balance.

[0124] Specifically, in the lower-level model, the objective function for the power generation cost of the virtual grid at each time step is:

[0125]

[0126] In the formula: The cost of generating electricity at time t for the i-th virtual grid; Let $\frac{i}{i}$ be the output cost of the photovoltaic modules in the $i$-th virtual grid. Let $\frac{i}{i}$ be the output cost of battery energy storage in the $i$-th virtual grid. Let $\frac{i}{i}$ be the output cost of the micro gas turbine in the $i$-th virtual grid. Let be the interaction cost of the electricity exchanged between the i-th virtual grid and the main power grid.

[0127] In the lower-level model, the constraints of the objective function include:

[0128] (1) Power balance constraint

[0129] At any given time, the power output within each virtual grid must be equal to the load demand. The output power includes the output of each power generation unit within the virtual grid, the power interacting with other virtual grids, and the power interacting with the main power grid.

[0130]

[0131] In the formula: For load demand within the virtual grid, Let be the output power of the photovoltaic module in the i-th virtual grid at time t. Let be the output power of the battery energy storage in the i-th virtual grid at time t. Let be the output power of the micro gas turbine in the i-th virtual grid at time t. Let be the interaction power between the i-th virtual grid and the main power grid at time t.

[0132] (2) Photovoltaic output constraints

[0133] For photovoltaic systems, the output power cannot exceed the system's maximum output power.

[0134]

[0135] In the formula, This represents the maximum output power of the photovoltaic power generation system within the i-th virtual grid.

[0136] (3) Energy storage output constraint

[0137] To extend the lifespan of energy storage batteries, the upper and lower limits of the state of charge of the energy storage system are set to 0.2 and 0.8, respectively, and its charging and discharging power must be less than the maximum output power and maximum discharge power of the energy storage system.

[0138]

[0139] In the formula, Let be the maximum charging power of the energy storage and power generation system in the i-th virtual grid. Let be the maximum discharge power of the energy storage power generation system in the i-th virtual grid.

[0140] (4) Output and ramping constraints of micro gas turbine

[0141] While meeting its own output power limitations, the micro gas turbine power generation system must also meet the power limits for uphill and downhill ramping when facing situations that require changes in output power, such as sudden load changes.

[0142]

[0143]

[0144] In the formula, Let $\frac{i}{i}$ be the minimum and maximum outputs of the micro gas turbine power generation system in the $i$-th virtual grid, respectively. These represent the maximum downward ramp rate and the maximum upward ramp rate of the micro gas turbine power generation system, respectively.

[0145] (5) Tie line power constraints

[0146] There is a certain amount of power trading between virtual grids and the main power grid, as well as between multiple virtual grids, and this power constraint is related to the power that the transmission lines can withstand.

[0147]

[0148] In the formula, Let be the interaction power between the i-th virtual grid and the main power grid at time t. The maximum allowable interactive power for the tie line between the i-th virtual grid and the main power grid.

[0149] Furthermore, the output cost of the photovoltaic module in the i-th virtual grid is:

[0150]

[0151] In the formula: For the cost of photovoltaic power generation, Let a be the output power of the photovoltaic system in the i-th virtual grid at time t. pv b pv c pv This refers to the power generation cost coefficient of the photovoltaic power generation system.

[0152] The output cost of battery energy storage in the i-th virtual grid is:

[0153]

[0154] In the formula: The output cost of battery energy storage; M Li Total cost of energy storage battery maintenance; Let t be the output power of the battery energy storage; N be the total cycle life of the battery; U be the AC side voltage of the battery; and C be the rated capacity of the battery.

[0155] The output cost of the micro gas turbine in the i-th virtual grid is:

[0156]

[0157] In the formula: For the output cost of gas turbines, For the fuel cost of operating micro gas turbines, This refers to the start-up and shutdown costs of the gas turbine.

[0158] In one embodiment, the upper-level model is a non-cooperative game theory model, and the lower-level model is a nonlinear single-objective model. A two-layer optimization scheduling model is established, with each layer having its own objective function and constraints. The lower-level model uses the upper-level model as a parameter, solves the problem, and feeds the result back to the upper-level model. The upper-level model finds the optimal solution under its own constraints and then transmits it to the lower-level model. The specific mathematical model is as follows:

[0159] upper layer:

[0160] Lower layer:

[0161] In the formula, x and y are the decision variables for the upper and lower level optimizations, F(x,y) and f(x,y) are the objective functions for the upper and lower level optimizations, and g(x,y) and h(x,y) are the constraints.

[0162] In one embodiment, the upper-level model, i.e., the non-cooperative game model, is solved using an adaptive mutation particle swarm optimization algorithm. This allows for the determination of the Nash equilibrium solution with a high convergence speed, yielding the interactive electrical quantities between virtual grids, including:

[0163] Step 1): Input parameters, mainly including real-time electricity price, photovoltaic output, real-time load, etc.

[0164] Step 2): Establish a non-cooperative game model of multiple virtual power plants based on the benefit function, and generate the strategy space;

[0165] Step 3): Select the initial equilibrium value J of the trading electricity price and trading volume in the policy space corresponding to each virtual grid. i0 ={λ i0 P i0}, λ i0 Let P be the initial value of the electricity trading price. i0 This is the initial value for the traded electricity volume;

[0166] Step 4): Initialize the position and velocity of the particle swarm, with the virtual grid as the game participant;

[0167] Step 5): In the k-th iteration, take the result of the previous iteration as the initial value, and make independent decisions through the adaptive mutation particle swarm algorithm. If an equilibrium point is found, proceed to step 6) and output the result; otherwise, return to step 4) to make optimization decisions until the iteration ends.

[0168] Step 6): Output the Nash equilibrium solution, which is the strategy combination under the equilibrium state.

[0169] Among all strategy combinations, if a strategy set exists... satisfy This strategy is the Nash equilibrium solution, which can be expressed as:

[0170]

[0171] In the formula, Let J1, J2, ..., J be the Nash equilibrium strategy of the game participants, representing the optimal scheduling strategy for each virtual grid when the opponent chooses their optimal strategy; that is, the highest payoff for each virtual grid under this strategy combination in the sense of Nash equilibrium. n This is a non-optimal strategy; argmaxE i (·) represents the set of variables that maximizes the objective function. This represents the virtual grid strategy except for the i-th virtual grid.

[0172] The strategy space is as follows:

[0173]

[0174] In the formula: Adjustable capacity within the virtual grid. Let be the output power of the photovoltaic module in the i-th virtual grid at time t. This represents the maximum output power of the gas turbine within the virtual grid. Load demand within the virtual grid.

[0175] Furthermore, in the lower-level model, the optimization within the virtual mesh must satisfy the following power balance constraint:

[0176]

[0177] In the formula: Let be the interaction power between the i-th virtual grid and the main power grid at time t.

[0178] This embodiment takes into account that algorithms with poor convergence performance may take a long time to find the optimal solution. Therefore, an adaptive mutated particle swarm optimization algorithm is used in both layers. In the policy space, the adaptive mutated particle swarm optimization algorithm is used to find the Nash equilibrium solution of the non-cooperative game, so as to select the optimal solution with a higher convergence speed.

[0179] In one embodiment, the lower-level optimization model is a nonlinear single-objective model, which is also solved using an adaptive mutant particle swarm optimization algorithm. According to the following formula, during real-time operation, based on the state of the power generation units within the virtual grid, load demand, and real-time electricity price, the output of the power generation units and their interaction power with the main grid under the optimal objective function state are determined under the premise of satisfying the optimization operation constraints. Therefore, the problem to be solved can be expressed as:

[0180]

[0181] In the formula, Let $i$ be the power generation cost of the $i$-th virtual grid at a certain moment. Let t be the output power of the energy storage power generation system. Let be the output power of the micro gas turbine in the i-th virtual grid at time t.

[0182] The specific solution steps are as follows:

[0183] Step 1): Input the data of the power generation units in the virtual grid before optimization, the demand and output information, and the electricity price;

[0184] Step 2): Establish the particle swarm dimension and initialize the particle swarm; each particle represents a combination of the energy storage and power generation unit and the micro gas turbine output within the virtual grid, representing each set of scheduling data, that is, the particle swarm dimension is equal to the number of energy storage and power generation units and micro gas turbines; set the number of particles, initialize the position of each particle, select a set of data within the power allowable range of the power generation unit, that is, initialize the output power of the energy storage and gas turbines; at the same time, initialize the particle flight speed.

[0185] Step 3): Calculate the fitness value: Calculate the fitness of each particle according to the optimization objective function. This value is the gain of the virtual grid. Take the current best result as the local optimum and find the global optimum particle to complete the initialization of local and global extrema.

[0186] Step 4): Iterative optimization: Update particle position and velocity, recalculate fitness value, and determine whether to update local and global optima;

[0187] Step 5): Determine if the iteration result has converged: When the maximum number of iterations set or the global best position has not changed for a number of consecutive iterations, output the global optimal solution, that is, the optimal output power of the energy storage power generation unit and the micro gas turbine; otherwise, return to step 4) for re-looping.

[0188] In this embodiment, a system comprising three virtual grids is constructed, each containing a photovoltaic system, an energy storage system, and a micro gas turbine system, establishing a system as follows: Figure 3 The simulation model diagram is shown below, and the relevant parameters are shown in Tables 1, 2, and 3. Since the objective function is related to the revenue from selling electricity in the virtual grid, the revenue from electricity exchanged between the virtual grid and the main grid, and the revenue from purchasing electricity from other virtual grids, and the revenue from selling electricity is related to the load price, different electricity prices are set for different time periods, such as... Figure 4 As shown in Table 1, the photovoltaic parameters are shown in Table 2, the energy storage parameters are shown in Table 3, the basic parameters of the particle swarm optimization are shown in Table 4, and the number of iterations for the two-layer optimization is 50.

[0189] Table 1 Photovoltaic parameters

[0190]

[0191] Table 2 Energy Storage Parameters

[0192]

[0193] Table 3 Parameters of the Micro Gas Turbine

[0194]

[0195] Table 4 Basic parameters of the adaptive mutant particle swarm optimization algorithm

[0196]

[0197] The photovoltaic output and load changes of each virtual grid are assumed, based on the characteristics of photovoltaic and load changes, as follows: Figure 5 As shown.

[0198] To address this issue, a two-layer optimization scheduling method for multiple virtual grids is proposed: a two-layer optimization control model for virtual grids is established, with the upper layer optimizing the transaction plan and power generation plan between virtual grids, and the lower layer optimizing the power output plan within the virtual grid; after optimization through game theory at the upper layer, the interactive power between virtual grids is obtained and transmitted to the lower layer for optimization of source and load storage within the grid.

[0199] First, the power output of each virtual grid before the interaction of power sources is calculated based on an adaptive mutation algorithm during standalone operation. This yields the energy storage charging and discharging power, the output of the micro gas turbine, and the interaction power with the main grid within each virtual grid. Figure 6 As shown, each virtual grid allocates power generation tasks to micro gas turbines and energy storage according to an optimization algorithm, while simultaneously purchasing electricity from the grid to compensate for power shortages. Energy storage charges when photovoltaic output exceeds load demand, discharges when photovoltaic output cannot meet load demand, and stops supplying power when the State of Charge (SOC) drops to 0.2. The power generation tasks of the micro gas turbines change with grid prices and load demand. During the periods of 1h–5h and 21h–24h, since the grid price is lower than the discharge cost of the micro gas turbines in virtual grid 2, virtual grid 2 purchases electricity from the grid, generating power when the power generation cost is lower than the grid price.

[0200] according to Figure 6 As a result, the interactive electricity generated at each moment is used as demand to compete with other virtual grids in order to obtain electricity at a price lower than the grid price or to sell electricity at a price higher than the grid price in order to reduce costs or obtain more profits. Figure 7 This illustrates the interaction power between the virtual grids after the game, with arrows indicating the direction of power flow. From Figure 7 It can be seen that there are certain transactions between different virtual grids between 4h and 22h. Through game theory strategies, the transaction volume between virtual grids and the main grid is reduced to some extent, indicating that virtual grids achieve internal supply and demand balance through transactions with other virtual grids, realizing the spatial transfer and satisfaction of load demand. However, during the periods of 1h-3h and 22h-24h, the interactive power between virtual grids does not occur because the grid price is low at these times and the bidding constraints are not met.

[0201] The effectiveness of the proposed two-level optimization method is verified from an economic perspective. In each scheduling stage, virtual grids engage in game theory and adjust their trading prices based on their own power generation costs to secure more power generation tasks and obtain more profits. Therefore, comparing the benefits of two-level optimization and intra-grid optimization, it can be seen that virtual grids that trade with each other coordinate and optimize scheduling through inter-grid game theory to secure a certain amount of power trading, making the overall optimization result better than that of intra-grid optimization alone. The benefits for the entire scheduling period are shown in Table 5.

[0202] Table 5 Analysis of Profit Optimization Results

[0203]

[0204] For example, the above comparison results show that the addition of gas turbines and energy storage not only ensures power generation within the virtual grid but also provides adjustable space for power trading between virtual grids. By adjusting the power generation plan based on the results of inter-grid game theory, load transfer within the multi-virtual grid aggregation area can be achieved to obtain additional revenue or reduce power generation costs, further reducing power interaction between the virtual grid and the power grid, and achieving power autonomy to a certain extent.

[0205] For example, consider a multi-virtual mesh comprising a photovoltaic system, an energy storage system, and a micro gas turbine. Simulation results are as follows: Figure 5 , Figure 6 and Figure 7 As shown, energy storage charges when photovoltaic output exceeds load demand, discharges when photovoltaic output cannot meet load demand, and stops supplying power when the SOC drops to 0.2. The power generation task of the micro gas turbine changes with the grid price and load demand. During the time periods of 1h-5h and 21h-24h, since the grid price is less than the discharge cost of the micro gas turbine in virtual grid 2, virtual grid 2 purchases electricity from the grid and generates power when the power generation cost is less than the grid price. Simulations also show that virtual grids that trade with each other coordinate and optimize scheduling through inter-grid game theory to secure a certain amount of power trading, resulting in a better overall optimization outcome compared to optimization within the grid only. This demonstrates the rationality and effectiveness of the two-layer optimization scheduling method disclosed in this embodiment that promotes source-storage-load coordination among multiple virtual grids. It solves a series of problems such as source-storage-load coordination, improves the economic efficiency of multiple virtual grids, and ensures efficient resource utilization.

[0206] like Figure 8 As shown, the second embodiment of the present invention also proposes a two-layer optimization scheduling system to promote source-storage-load coordination among multiple virtual grids, the system comprising:

[0207] The acquisition module 10 is used to acquire parameter information of distributed energy, including multiple virtual grid photovoltaics, load forecast power, and control parameters of power generation units;

[0208] The upper-level optimization module 20 is used to establish the range of transaction electricity price decision based on the parameter information of the distributed energy through the balanced power bidding mechanism. In the upper-level model, the optimization objective is to maximize the overall operating revenue. The virtual grids conduct non-cooperative game to make decisions on the transaction plan and power generation plan between the virtual grids and obtain the interactive electricity between the virtual grids.

[0209] The lower-level optimization module 30 is used to take the power exchange between the virtual grids as a parameter, and take the lowest power generation cost of the virtual grids as the optimization objective in the lower-level model to make decisions on the output of each power generation unit in the virtual grids at different times, so as to obtain the source-storage-load optimization results inside the virtual grids.

[0210] Feedback module 40 is used to feed back the source-storage-load optimization results inside the virtual grid to the upper-level model, so that the upper-level model can find the optimal solution under its own constraints and transmit it to the lower-level model until the set iteration conditions are met, and then obtain the output of each power generation unit and the interaction power with the large power grid.

[0211] This embodiment addresses the interaction between virtual grids. When trading between virtual grids, the transaction price needs to be considered. A balanced power bidding mechanism establishes the scope of the transaction price decision. The virtual grid, as a whole, reports its bidding strategy uniformly and engages in non-cooperative game theory to maximize its own profit. Within the virtual grid, the interactive power volume between virtual grids is used as a parameter. In the lower-level model, the optimization objective is to minimize the power generation cost of the virtual grid. Decisions are made on the output of each power generation unit within the virtual grid at different times to achieve internal power allocation. Through a multi-virtual grid two-layer optimization scheduling strategy that considers both inter-grid coordination and intra-grid optimization, the optimal solution for the power generation plan is found with the goal of maximizing overall operating benefits and minimizing power generation costs. This achieves source-storage-load coordination, improves the economic efficiency of the virtual grid, and ensures efficient resource utilization.

[0212] In one embodiment, the objective function in the upper-level model is:

[0213]

[0214] In the formula: The gain of the i-th virtual grid at time t; Let be the revenue obtained from selling electricity in the i-th virtual grid at time t; Let be the revenue generated by the interaction of the i-th virtual grid with the main power grid at time t; Let $t$ be the revenue that the $i$-th virtual grid receives from purchasing electricity from the other virtual grids at time $t$. Let t be the controllable power output cost of the i-th virtual grid in a transaction with other virtual grids;

[0215] The constraints on electricity prices during game-like bidding between virtual grids are as follows:

[0216] λ i,t ≥λ i,ke

[0217] In the formula: λ i,t To compete for electricity prices; λ i,ke The unit power generation cost of a controllable generating unit;

[0218] Constraints on the power participating in bidding within the virtual grid:

[0219]

[0220] In the formula, ∑P il,t The power participating in the bidding within the virtual grid. This represents the maximum output power of the gas turbine within the virtual grid. This represents the maximum output power of photovoltaic power generation within the virtual grid. This represents the maximum output power of energy storage within the virtual grid.

[0221] It should be noted that the bidding strategy between virtual grids is as follows: a balanced power bidding mechanism is established, and the traded electricity is treated as a transfer of load demand when processing the electricity traded between virtual grids. The balance electricity available for a virtual grid can be determined by the surplus electricity of the virtual grid. The trading price between virtual grids participating in the electricity trade is determined by their own generation costs and price ceilings. Distributed power sources are aggregated into multiple virtual grids, and a multi-virtual grid collaborative scheduling model is constructed based on non-cooperative game theory. The problem of real-time power imbalance within a single virtual grid is solved through mutual coordination between virtual grids, realizing the interaction between "source" and "load" within the aggregated area of ​​multiple virtual grids.

[0222] Furthermore, in the lower-level model, the constraints are power generation unit security constraints, tie-line security constraints, and power balance constraints. The objective function for the power generation cost of the virtual grid at each time step is:

[0223]

[0224] In the formula: The cost of generating electricity at time t for the i-th virtual grid; Let $\frac{i}{i}$ be the output cost of the photovoltaic modules in the $i$-th virtual grid. Let $\frac{i}{i}$ be the output cost of battery energy storage in the $i$-th virtual grid. Let $\frac{i}{i}$ be the output cost of the micro gas turbine in the $i$-th virtual grid. Let be the interaction cost of the electricity exchanged between the i-th virtual grid and the main power grid.

[0225] In one embodiment, in a system comprising multiple virtual grids, each virtual grid contains a power generation unit including a photovoltaic system, an energy storage system, and a micro gas turbine system. The characteristics of each power generation unit are fully utilized and aggregated and controlled in the form of a virtual grid. The power generation units within the virtual grid are respectively constructed as aggregated unit mathematical models, wherein the power generation unit includes photovoltaic modules, a gas turbine, and battery energy storage.

[0226] In order to continuously extract the maximum energy from the solar panels, the photovoltaic power generation system is usually operated at the maximum power point (MPP). The aggregation unit data model corresponding to the photovoltaic module is as follows:

[0227]

[0228] In the formula: P pv P represents the output power of the photovoltaic module. mpp，STC γ is the output power of the photovoltaic module under MPPT control; γ is the power temperature coefficient; S is the actual irradiance; S STC T represents the reference value for light intensity; T represents the photovoltaic operating temperature; T STC For reference temperature;

[0229] The output power of the micro gas turbine power generation model is proportional to the fuel consumed. The mathematical model of the aggregation unit corresponding to the gas turbine is as follows:

[0230] P MT =A MT η MT H MT

[0231] In the formula: P MT A represents the output power of the gas turbine. MT The amount of fuel consumed by the gas turbine; η MT For power generation efficiency; H MT The calorific value of a micro gas turbine;

[0232] The mathematical model of the aggregated unit corresponding to the battery energy storage represents the battery energy storage power generation model, which simulates the relationship between the stored energy at time t and the energy of charging and discharging during the time period Δt using the following formula:

[0233]

[0234] In the formula: P B,t P is the amount of electricity stored inside the battery at time t; α is the battery's self-discharge rate; B,t-Δt η represents the amount of electricity stored inside the battery at time t-Δt; in η out Improve battery energy storage charging and discharging efficiency; P B,in P B,out Δt represents the charging and discharging power of the battery energy storage; 'a' represents the charging and discharging state of the battery energy storage, 0 for discharging and 1 for charging; Δt represents the time period.

[0235] It should be noted that by establishing mathematical models of aggregated units such as photovoltaic power generation, gas turbine power generation, and battery energy storage power generation within a virtual grid, and under the premise of satisfying constraints such as unit and tie-line safety, the output of each unit within the grid is decided at different times with the goal of minimizing the system's power generation cost. Initially, photovoltaic power supplies the load, while energy storage and micro gas turbines handle the remaining load and remain in standby mode. When the load demand in the region exceeds the maximum power supply capacity of the power generation system within the virtual grid, the virtual grid aggregates and purchases the remaining electricity from the main grid.

[0236] In one embodiment, the adaptive mutant particle swarm optimization algorithm is used to solve the two-layer model. The lower-layer model takes the upper-layer model as a parameter, and after solving, the result is fed back to the upper-layer model. The upper-layer model finds the optimal solution under its own constraints and then transmits it to the lower-layer model.

[0237] It should be noted that other embodiments or implementation methods of the two-layer optimized scheduling system for promoting source-storage-load coordination among multiple virtual grids described in this invention can refer to the above-described method embodiments, and will not be repeated here.

[0238] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0239] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0240] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A two-layer optimization scheduling method to promote source-storage-load coordination among multiple virtual grids, characterized in that, The method includes: Obtain parameter information of distributed energy sources, including multiple virtual grid photovoltaics, load forecast power, and control parameters of power generation units; Based on the parameter information of the distributed energy, the range of electricity trading price decision is established through a balanced power bidding mechanism. In the upper-level model, the optimization objective is to maximize the overall operating revenue. Non-cooperative game theory is conducted among the virtual grids to make decisions on the trading and power generation plans among the virtual grids, and to obtain the interactive electricity between the virtual grids. This includes establishing a multi-virtual power plant non-cooperative game model based on the benefit function and generating a strategy space; selecting the initial values ​​of the equilibrium point of the trading price and the trading electricity in the strategy space corresponding to each virtual grid; initializing the position and velocity of the particle swarm; using the virtual grids as game participants, the Nash equilibrium solution is solved using the adaptive mutated particle swarm algorithm as the optimal scheduling strategy for each virtual grid to determine the interactive electricity between the virtual grids. Using the interactive power between the virtual grids as a parameter, the lower-level model aims to minimize the power generation cost of the virtual grids. The interactive power between the virtual grids is incorporated into power balance constraints. The output of each power generation unit within the virtual grid is determined at different times, yielding the source-storage-load optimization result within the virtual grid. This includes input data on the power generation units within the virtual grid before optimization, demand output information, and electricity price. The particle swarm dimension is established, and the particle swarm is initialized, with each particle representing a combination of the output of the energy storage power generation unit and the micro gas turbine within the virtual grid. The number of particles is set, and the position and flight speed of each particle are initialized. An adaptive mutated particle swarm optimization algorithm is used to calculate the global optimal solution, which is the optimal output power of the energy storage power generation unit and the micro gas turbine. The constraints of the lower-level model's optimization objective are power generation unit safety constraints, tie-line safety constraints, and power balance constraints. The optimization results of the source and storage load inside the virtual grid are fed back to the upper-level model, so that the upper-level model can find the optimal solution under its own constraints and transmit it to the lower-level model. After the set iteration conditions are met, the output of each power generation unit and the interaction power with the large power grid are obtained. Before using the energy exchanged between the virtual grids as a parameter, and taking the lowest power generation cost of the virtual grid as the optimization objective in the lower-level model to make decisions on the output of each power generation unit within the virtual grid at different time periods, and obtaining the source-storage-load optimization results within the virtual grid, the method further includes: The power generation units within the virtual grid are each constructed as aggregate unit mathematical models, wherein the power generation units include at least two of the following: photovoltaic modules, gas turbines, and battery energy storage. The aggregation unit data model corresponding to the photovoltaic module is as follows: In the formula: P pv The output power of the photovoltaic module; P mpp，STC The output power of the photovoltaic module under MPPT control; γ The power temperature coefficient; S This represents the actual light intensity. S STC This is a reference value for light intensity; T This refers to the operating temperature of the photovoltaic system. T STC For reference temperature; The mathematical model of the aggregation unit corresponding to the gas turbine is: In the formula: P MT This refers to the output power of the gas turbine. A MT The amount of fuel consumed by the gas turbine; For power generation efficiency; H MT The calorific value of a micro gas turbine; The mathematical model for the aggregation unit corresponding to the battery energy storage is: In the formula: yes t The amount of electricity stored inside the battery at all times; The self-discharge rate of the battery energy storage; for The amount of electricity stored inside the battery at all times; , Improve the charging and discharging efficiency of battery energy storage; , The charging and discharging power for battery energy storage; a This indicates the battery's energy storage charging and discharging state, with 0 representing discharging and 1 representing charging; For a time period.

2. The two-layer optimization scheduling method for promoting source-storage-load coordination among multiple virtual grids as described in claim 1, characterized in that, In the aforementioned upper-level model, the objective function is: In the formula: For the first i A virtual grid t Benefits at all times; for t Time of the first i Revenue generated from selling electricity in a virtual grid; for t Time of the first i The benefits of electricity exchanged between a virtual grid and the main power grid; for t Time of the first i The revenue that a virtual grid receives from purchasing electricity from other virtual grids; For time t, the first i Controllable power output cost in transactions between a virtual grid and other virtual grids; The constraints are: In the formula: To compete for competitive electricity prices; The unit power generation cost of a controllable generating unit.

3. The two-layer optimization scheduling method for promoting source-storage-load coordination among multiple virtual grids as described in claim 1, characterized in that, In the lower-level model, the objective function for the power generation cost of the virtual grid at each time step is: In the formula: For the first i A virtual grid t Cost of generating electricity at any time; For the first i The output cost of photovoltaic modules in a virtual grid; For the first i The output cost of battery energy storage in a virtual grid; For the first i The output cost of a micro gas turbine in a virtual grid; For the first i The interaction cost of electricity exchange between a virtual grid and the main power grid.

4. The two-layer optimization scheduling method for promoting source-storage-load coordination among multiple virtual grids as described in claim 3, characterized in that, The lower-level model also includes constraints: In the formula: The power participating in the bidding within the virtual grid; The maximum output power of the gas turbine within the virtual grid. This represents the maximum output power of photovoltaic power generation within the virtual grid. This represents the maximum output power of energy stored within the virtual grid.

5. The two-layer optimization scheduling method for promoting source-storage-load coordination among multiple virtual grids as described in claim 3, characterized in that, No. i The output cost of the photovoltaic modules in the virtual grid is: In the formula: For the cost of photovoltaic power generation, For the first i Photovoltaic systems in a virtual grid t Output power at any moment , , This refers to the power generation cost coefficient of the photovoltaic power generation system. No. i The output cost of battery energy storage in a virtual grid is: In the formula: The output cost of battery energy storage; M Li Total cost of energy storage battery maintenance; for t The output power of the battery energy storage at all times; N Total cycle life of the battery; U This refers to the AC side voltage of the battery. C Battery rated capacity; No. i The output cost of the micro gas turbine in each virtual grid is: In the formula: For the output cost of gas turbines, For the fuel cost of operating micro gas turbines, This refers to the start-up and shutdown costs of the gas turbine.

6. The two-layer optimization scheduling method for promoting source-storage-load coordination among multiple virtual grids as described in claim 1, characterized in that, Based on the parameter information of the distributed energy source, the range of electricity trading price decisions is established through a balanced power bidding mechanism. In the upper-level model, the optimization objective is to maximize overall operational revenue. Non-cooperative game theory is conducted between the virtual grids to determine their trading and generation plans, resulting in the exchanged electricity volume between the virtual grids, including: Step 1): Input parameters, including real-time electricity price, photovoltaic output, and real-time load; Step 2): Establish a non-cooperative game model of multiple virtual power plants based on the benefit function, and generate the strategy space; Step 3): Select the initial equilibrium values ​​for the trading electricity price and trading volume in the policy space corresponding to each virtual grid. J i0 ={ λ i0 , P i0 }, λ i0 This is the initial value of the electricity trading price. P i0 This is the initial value for the traded electricity volume; Step 4): Initialize the position and velocity of the particle swarm, with the virtual grid as the game participant; Step 5): In the k In the next iteration, the result of the previous iteration is used as the initial value. After independent decision-making through the adaptive mutation particle swarm algorithm, if an equilibrium point is found, proceed to step 6) and output the result; otherwise, return to step 4) to make optimization decisions until the iteration ends. Step 6): Output Nash Equilibrium solution, the strategy combination under equilibrium state; Among all strategy combinations, if there exists a strategy set { , , ..., } satisfies C i,t ( , )≥C i,t ( , ), i ∈ n Then the strategy is Nash Equilibrium solution Nash The equilibrium solution can be expressed as: In the formula, , … Let Nash equilibrium strategy be the strategy of the game participants, which represents the optimal scheduling strategy of each virtual grid when the other party chooses the optimal strategy. That is, under this strategy combination, each virtual grid can achieve the highest payoff in the sense of Nash equilibrium. , … This is a non-optimal strategy; Let represent the set of variables that maximizes the objective function. Representative except the first i Virtual grid strategies other than individual virtual grids; The strategy space is: In the formula: Adjustable capacity within the virtual grid. For the first i Photovoltaic modules in a virtual grid t Output power at any moment This represents the maximum output power of the gas turbine within the virtual grid. Load demand within the virtual grid.

7. The two-layer optimization scheduling method for promoting source-storage-load coordination among multiple virtual grids as described in claim 1, characterized in that, The step involves using the interactive power between the virtual grids as a parameter, and in the lower-level model, taking the minimum power generation cost of the virtual grid as the optimization objective, to make decisions on the output of each power generation unit within the virtual grid at different time periods, thereby obtaining the source-storage-load optimization results within the virtual grid, including: Step 1): Input the data of the power generation units in the virtual grid before optimization, their demand and output information, and the electricity price; Step 2): Establish the particle swarm dimension, initialize the particle swarm, and treat each particle as a combination of energy storage and power generation unit and micro gas turbine output within the virtual grid; set the number of particles, initialize the position of each particle, and initialize the particle flight speed; Step 3): Calculate the fitness value. Calculate the fitness of each particle according to the optimization objective function. The fitness is the gain of the virtual grid. Take the current best result as the local optimum and find the global optimum particle to complete the initialization of local and global extrema. Step 4): Iterative optimization, update particle position and velocity, recalculate fitness value, and determine whether to update local and global optimal solutions; Step 5): When the maximum number of iterations is completed or the global best position remains unchanged for a number of consecutive iterations, output the global optimal solution, which is the optimal output power of the energy storage power generation unit and the micro gas turbine. Otherwise, return to step 4) for re-looping.

8. A two-layer optimized scheduling system for promoting source-storage-load coordination among multiple virtual grids, characterized in that, The system includes: The acquisition module is used to acquire parameter information of distributed energy sources, including multiple virtual grid photovoltaics, load forecast power, and control parameters of power generation units; The upper-level optimization module is used to establish the range of electricity trading price decisions based on the parameter information of the distributed energy source through a balanced power bidding mechanism. In the upper-level model, the optimization objective is to maximize the overall operating revenue. The virtual grids conduct non-cooperative game theory to make decisions on the trading and power generation plans between the virtual grids, and obtain the interactive electricity between the virtual grids. This includes establishing a multi-virtual power plant non-cooperative game model based on the benefit function and generating a strategy space; selecting the initial equilibrium values ​​of the trading price and trading volume in the strategy space corresponding to each virtual grid; initializing the position and velocity of the particle swarm; using the virtual grids as game participants, and using the adaptive mutated particle swarm algorithm to solve the Nash equilibrium solution as the optimal scheduling strategy for each virtual grid to determine the interactive electricity between the virtual grids. The lower-level optimization module uses the interactive power between the virtual grids as a parameter. In the lower-level model, the optimization objective is to minimize the power generation cost of the virtual grids. The interactive power between the virtual grids is incorporated into power balance constraints. The module makes decisions on the output of each power generation unit within the virtual grid at different times, obtaining the source-storage-load optimization result within the virtual grid. This includes inputting data on the power generation units within the virtual grid before optimization, demand output information, and electricity price. The module establishes the particle swarm dimension, initializes the particle swarm, and treats each particle as a combination of the output of the energy storage power generation unit and the micro gas turbine within the virtual grid. It sets the number of particles, initializes the position and flight speed of each particle, and uses an adaptive mutated particle swarm optimization algorithm to calculate the global optimal solution. The global optimal solution is the optimal output power of the energy storage power generation unit and the micro gas turbine. The constraints of the lower-level model's optimization objective are power generation unit safety constraints, tie-line safety constraints, and power balance constraints. The feedback module is used to feed back the source-storage-load optimization results inside the virtual grid to the upper-level model, so that the upper-level model can find the optimal solution under its own constraints and transmit it to the lower-level model until the set iteration conditions are met, and then obtain the output of each power generation unit and the interaction power with the large power grid. It also includes: constructing the power generation units within the virtual grid into aggregated unit mathematical models, wherein the power generation units include at least two of photovoltaic modules, gas turbines, and battery energy storage; The aggregation unit data model corresponding to the photovoltaic module is as follows: In the formula: P pv The output power of the photovoltaic module; P mpp，STC The output power of the photovoltaic module under MPPT control; γ The power temperature coefficient; S This represents the actual light intensity. S STC This is a reference value for light intensity; T This refers to the operating temperature of the photovoltaic system. T STC For reference temperature; The mathematical model of the aggregation unit corresponding to the gas turbine is: In the formula: P MT This refers to the output power of the gas turbine. A MT The amount of fuel consumed by the gas turbine; For power generation efficiency; H MT The calorific value of a micro gas turbine; The mathematical model for the aggregation unit corresponding to the battery energy storage is: In the formula: yes t The amount of electricity stored inside the battery at all times; The self-discharge rate of the battery energy storage; for The amount of electricity stored inside the battery at all times; , Improve the charging and discharging efficiency of battery energy storage; , The charging and discharging power for battery energy storage; a This indicates the battery's energy storage charging and discharging state, with 0 representing discharging and 1 representing charging; For a time period.

9. The two-layer optimization scheduling system for promoting source-storage-load coordination among multiple virtual grids as described in claim 8, characterized in that, In the aforementioned upper-level model, the objective function is: In the formula: For the first i A virtual grid t Benefits at all times; for t Time of the first i Revenue generated from selling electricity in a virtual grid; for t Time of the first i The benefits of electricity exchanged between a virtual grid and the main power grid; for t Time of the first i The revenue that a virtual grid receives from purchasing electricity from other virtual grids; For time t, the first i Controllable power output cost in transactions between a virtual grid and other virtual grids; The constraints are: In the formula: To compete for competitive electricity prices; The unit power generation cost of a controllable generating unit; In the lower-level model, the constraints are generation unit security constraints, tie-line security constraints, and power balance constraints. The objective function for the generation cost of the virtual grid at each time step is: In the formula: For the first i A virtual grid t Cost of generating electricity at any time; For the first i The output cost of photovoltaic modules in a virtual grid; For the first i The output cost of battery energy storage in a virtual grid; For the first i The output cost of a micro gas turbine in a virtual grid; For the first i The interaction cost of electricity exchange between a virtual grid and the main power grid.

Images