A multi-virtual power plant and distribution network collaborative optimization scheduling method and device

Abstract

This invention discloses a method and device for collaborative optimization scheduling of multiple virtual power plants and distribution networks, comprising: S1 constructing an active load response model and an energy storage system response model, and calculating the response power respectively; S2 decoupling the centralized optimization objective function and solving it as an independent optimization objective function for each virtual power plant; S3 constructing a real-time operation model of the virtual power plants; S4 each virtual power plant transmits the interactive power and trading price to neighboring virtual power plants, and each virtual power plant iteratively solves the local decision variables, interactive power decision variables, and trading price to obtain an optimized scheduling scheme for the local distributed resources of the virtual power plant. This invention avoids the problems of computational complexity, dimensionality curse, and inability to fully explore the flexibility potential of massive and diverse distributed resources in traditional centralized optimization solutions, and can effectively improve the efficiency and flexibility of distributed resource optimization scheduling and control, while ensuring the privacy of user information.

Description

Technical Field

[0001] This invention relates to the field of distributed resource optimization scheduling technology, and in particular to a method and equipment for collaborative optimization scheduling of multiple virtual power plants and distribution networks. Background Technology

[0002] With the continuous growth of installed capacity of renewable energy generators such as wind turbines and photovoltaics, their inherent strong randomness poses a significant challenge to the safe and stable operation of existing power distribution systems. Simultaneously, diverse and massive distributed resources such as active user loads and energy storage systems are also experiencing accelerated growth. Traditional centralized optimization methods are increasingly facing numerous problems when dealing with high-dimensional, diverse, and massive adjustable distributed resources, such as complex problem-solving and lengthy computation times, large amounts of demand data, and the inability to meet the low-latency communication requirements for the regulation of diverse and massive distributed resources.

[0003] Extensive research has been conducted on the optimal scheduling of massive distributed resources. Some scholars have employed robust optimization methods to address the strong randomness of renewable energy output from sources such as wind turbines and solar power, while others have used stochastic programming methods to simulate various distributed resource scheduling scenarios. However, existing research and methods still have the following shortcomings: the optimized scheduling schemes obtained using robust optimization methods cannot fully explore the flexibility potential of diverse massive distributed resources; and stochastic programming methods cannot cope with the strong randomness of distributed resources in extreme scenarios. Summary of the Invention

[0004] The purpose of this invention is to solve the problem of improving the efficiency of optimizing and scheduling massive distributed resources, and to provide a method and equipment for collaborative optimization and scheduling of multiple virtual power plants and distribution networks.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A method for coordinated and optimized scheduling of multiple virtual power plants and distribution networks includes the following steps: S1. Construct an active load response model and an energy storage system response model for the demand response regulation of the virtual power plant, and calculate the response power of the active load and the energy storage system respectively. S2. Decouple the centralized optimization objective function and decompose it into an independent optimization objective function for each virtual power plant and solve it. S3. Construct a real-time operation model of a virtual power plant, and make local decision variables and obtain interactive power decision variables and transaction prices with neighboring virtual power plants to maintain the local power dynamic balance within the virtual power plant. S4. Each virtual power plant transmits the interactive power and transaction price to neighboring virtual power plants. Each virtual power plant iteratively solves the local decision variables, the interactive power decision variables, and the transaction price to obtain an optimized scheduling scheme for the local distributed resources of the virtual power plant.

[0006] In some embodiments of the present invention, step S2 further includes: S21, introducing Lagrange dual multiplier variables, and writing the optimization objective function under centralized solution into an augmented Lagrange function form of the centralized optimization objective function; S22, subsequently decomposing the augmented Lagrange function form of the centralized optimization objective function to obtain the optimization objective of a single virtual power plant.

[0007] In some embodiments of the present invention, the centralized optimization objective function expression is as follows: ; in The first Local decision variables and decision variables interacting with neighboring VPPs for each virtual power plant. These are the local distributed resource scheduling and operation costs of the virtual power plant and the cost of power interaction with neighboring virtual power plants, respectively. To ensure that the power distribution system meets the interactive power conservation constraint during operation. The penalty function term; For the first The selection of the virtual power plant and the first Similarly, the power decision variables for interactions between virtual power plants are also... For the first The selection of the virtual power plant and the first The power decision variables of the two virtual power plants interact, and both satisfy the power conservation constraint. Indicates the first The virtual power plant belongs to the first The set of virtual power plants adjacent to each other; the augmented Lagrangian form of the centralized optimization objective function is expressed as follows: ; The objective function expression for the single virtual power plant is as follows: ; in The physical meaning of the introduced Lagrange dual multiplier variable is the first... The selection of the virtual power plant and the first The trading price set by virtual power plants for power exchange. Then, it is the introduced slack variable, which is the first... The estimated first virtual power plant Power decision variables that a virtual power plant interacts with.

[0008] In some embodiments of the present invention, the active load response model expression in step S1 is as follows:

[0009] in A collection of time periods throughout the 24 hours of a day. Indexes for each time period; For the first The active load aggregation power within each virtual power plant, of which For fixed load, The first A virtual power plant at any time The increased and decreased load power variables made are constrained. Indicates the first A virtual power plant at any time The increases and decreases in load power that are made have certain upper and lower limits. Then guarantee the first day The total increase and decrease in load power made by each virtual power plant are equal, thereby ensuring the electricity needs of users. This indicates that the virtual power plants will not simultaneously make decisions to increase or decrease load power; the energy storage system response model expression is as follows:

[0010] in For the first The response power of the energy storage system equipped in each virtual power plant. These represent the charging and discharging power of the energy storage system, respectively, with constraints. and To ensure that the charging and discharging power of the energy storage system remains within a limited range and that charging and discharging decisions are not made simultaneously at any given moment, among which... To characterize the energy storage system at time... The choice between charging and discharging is a 0-1 variable; the state-of-charge constraint expression for the energy storage system is as follows: ; in These refer to the charging efficiency and discharging efficiency of the energy storage system, respectively. For the capacity of the energy storage system, The time interval is specified, and the value of the state of charge must satisfy upper and lower limit constraints. ,in These are the minimum and maximum values ​​of the state of charge, respectively, and the constraints. To ensure that the energy storage system completes one cycle of charging and discharging in a day, These are the initial and final values ​​of the state of charge within a day, respectively.

[0011] In some embodiments of the present invention, with the goal of minimizing their own operating costs, the interactive power and transaction price of neighboring virtual power plants are estimated; the real-time operation model of the virtual power plant in step S3 is as follows: The constraint expressions for each virtual power plant during operation are as follows: ; ; ; in These represent the power purchased and sold between each virtual power plant and the upper-level power grid. These represent the upper limits of the power purchase and sale between each virtual power plant and the upper-level power grid. This is a 0-1 variable representing the decision of each virtual power plant to purchase or sell electricity to the upper-level power grid; These represent the power purchased and sold between each virtual power plant and its neighboring virtual power plants. These represent the upper limits of power purchase and sale between each virtual power plant and its neighboring virtual power plants. This is a 0-1 variable representing the decision of each virtual power plant to purchase or sell electricity to neighboring virtual power plants; These represent the power values ​​of the photovoltaic and wind turbine generator sets equipped within each virtual power plant.

[0012] In some embodiments of the present invention, in order to improve the local consumption rate of distributed renewable energy, the trading price during the real-time operation of the power system is subject to the following constraint expression: ; in The first The virtual power plant's purchase price from the upper-level power grid and the price it sells to the upper-level power grid at any given time. The first The price at which a virtual power plant purchases electricity from neighboring virtual power plants and the price at which it sells electricity to neighboring virtual power plants.

[0013] In some embodiments of the present invention, the optimized scheduling scheme in step S4 is as follows: The objective function expression for the operating cost of the virtual power plant is as follows:

[0014] in For the first The total operating cost of a virtual power plant; For the first Demand response scheduling cost of a virtual power plant This is the unit demand response dispatch cost, used to compensate users for changing their own electricity usage plans; For the first The energy storage charging and discharging loss cost of a virtual power plant Unit cost of energy storage charging and discharging losses; For the first The cost of power interaction between a virtual power plant and the upper-level power grid; For the first The cost of power interaction between a virtual power plant and neighboring virtual power plants.

[0015] In some embodiments of the present invention, each virtual power plant optimizes its own operating cost objective function using the distributed augmented Lagrange method; wherein the local decision variable expression is as follows: ; The expression for the interaction power decision variables between each virtual power plant and its neighboring virtual power plants is as follows: ; The electricity price expressions for each virtual power plant and its neighboring virtual power plants are as follows: .

[0016] The present invention also provides a multi-virtual power plant and distribution network collaborative optimization scheduling device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the multi-virtual power plant and distribution network collaborative optimization scheduling method as described above.

[0017] The present invention also provides a storage medium comprising a stored computer program, wherein, when the computer program is executed, it controls the device where the storage medium is located to perform the multi-virtual power plant and distribution network collaborative optimization scheduling method as described above.

[0018] The beneficial effects of this invention compared to the prior art include: The method of this invention decouples the centralized optimization objective function into independent optimization objective functions for each virtual power plant, significantly reducing the computational complexity of a single optimization objective function and solving the curse of dimensionality problem faced by centralized problems, while ensuring user data privacy and security. Simultaneously, each virtual power plant can perform parallel computation, iteratively converging to the global optimum, significantly improving computational efficiency and thus enhancing system decision-making, scheduling efficiency, and resource utilization efficiency. The method of this invention only requires optimization decisions based on local operating constraints, interactive power decision strategies of neighboring virtual power plants, and trading prices. Each virtual power plant only needs to interact with its neighboring users regarding interactive power decision variables and trading price information, greatly improving data privacy and security. Furthermore, since only local and neighboring regional communication systems need to be built, communication costs and burdens are significantly reduced. When some users experience communication failures, other users can still make optimization scheduling decisions, improving the robustness of the communication system in responding to risks.

[0019] Other beneficial effects of the embodiments of the present invention will be further described below. Attached Figure Description

[0020] Figure 1 This is a flowchart of the optimal scheduling method for multi-virtual power plants and distribution networks in the implementation of this invention; Figure 2 This is a schematic diagram of the improved IEEE 33-node power distribution system in the implementation of this invention; Figure 3 This is a typical output power and load power curve of wind turbines and photovoltaics in the system of this invention. Figure 4 This is a convergence graph showing the difference in decision results between the method proposed in this invention and the traditional centralized optimization method. Figure 5 This is a flowchart of another embodiment of the present invention regarding the optimal scheduling method for coordinated operation of multiple virtual power plants and distribution networks. Detailed Implementation

[0021] The present invention will be further described below with reference to the accompanying drawings and preferred embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0022] It should be noted that the directional terms such as left, right, up, down, top, and bottom used in this embodiment are only relative concepts or are based on the normal use of the product, and should not be considered as restrictive.

[0023] Since most of the diverse and massive distributed resources belong to the user side, and considering the privacy and security issues of individual users, power distribution system operators have difficulty obtaining user privacy information such as the status parameters of user-side equipment required for centralized optimization solutions, resulting in the inability to solve centralized problems.

[0024] Robust optimization methods select the worst-case operating cost as the objective function. Although considering the worst-case scenario can account for the fluctuations in the output of renewable energy sources such as wind turbines and photovoltaics in extreme scenarios, the optimized scheduling scheme obtained by robust optimization methods is often too conservative and cannot fully explore the flexibility potential of diverse and massive distributed resources. Although stochastic programming methods take into account multiple possible operating scenarios of distributed resources, the scheduling scheme obtained is less robust because the objective function is the expectation of multiple scenarios. It cannot cope with the strong randomness of distributed resources in extreme scenarios.

[0025] In existing research methods, each virtual power plant only considers power interaction with the upper-level power grid, without considering power interaction between multiple virtual power plants, thus failing to fully leverage the complementary and synergistic effects between multiple regions.

[0026] To address the aforementioned issues, this invention proposes an optimal scheduling method for multiple virtual power plants based on the distributed augmented Lagrange method.

[0027] Example 1 The following embodiments of the present invention propose a method for coordinated and optimized scheduling of multiple virtual power plants and distribution networks, such as... Figure 1 As shown, the process includes the following steps: S1, constructing an active load response model and an energy storage system response model for participating in the demand response regulation of the virtual power plant, and calculating the response power of the active load and the energy storage system respectively; S2, decoupling and decomposing the centralized optimization objective function into an independent optimization objective function for each virtual power plant and solving it; S3, constructing a real-time operation model of the virtual power plant, making local decision variables and deriving interactive power decision variables and trading prices with neighboring virtual power plants to maintain the local power dynamic balance of the virtual power plant; S4. Each virtual power plant transmits the interactive power and transaction price to neighboring virtual power plants. Each virtual power plant iteratively solves the local decision variables, the interactive power decision variables, and the transaction price to obtain an optimized scheduling scheme for the local distributed resources of the virtual power plant.

[0028] In this embodiment of the invention, the power system can be decomposed into multiple virtual power plants, each of which is a sub-region of the power system. The local decision variable is the scheduling strategy of the schedulable distributed resources in each sub-region (virtual power plant). By combining the scheduling strategies of all schedulable resources within the sub-region, an optimized scheduling scheme for the local distributed resources can be obtained.

[0029] In this embodiment of the invention, step S2 further includes: S21, introducing Lagrange dual multiplier variables, and writing the optimization objective function under centralized solution into an augmented Lagrange function form of the centralized optimization objective function; S22, subsequently decomposing the augmented Lagrange function form of the centralized optimization objective function to obtain the optimization objective of a single virtual power plant.

[0030] In this embodiment of the invention, the expression for the centralized optimization objective function is as follows: ; in The first Local decision variables and decision variables interacting with neighboring VPPs for each virtual power plant. These are the local distributed resource scheduling and operation costs of the virtual power plant and the cost of power interaction with neighboring virtual power plants, respectively. To ensure that the power distribution system meets the interactive power conservation constraint during operation. The penalty function term; For the first The selection of the virtual power plant and the first Similarly, the power decision variables for interactions between virtual power plants are also... For the first The selection of the virtual power plant and the first The power decision variables of the two virtual power plants interact, and both satisfy the power conservation constraint. Indicates the first The virtual power plant belongs to the first The set of virtual power plants adjacent to each other; the augmented Lagrangian form of the centralized optimization objective function is expressed as follows: ; The objective function expression for the single virtual power plant is as follows: ; in The physical meaning of the introduced Lagrange dual multiplier variable is the first... The selection of the virtual power plant and the first The trading price set by virtual power plants for power exchange. Then, it is the introduced slack variable, which is the first... The estimated first virtual power plant Power decision variables that a virtual power plant interacts with.

[0031] In this embodiment of the invention, the expression for the active load response model in step S1 is as follows:

[0032] in A collection of time periods throughout the 24 hours of a day. Indexes for each time period; For the first The active load aggregation power within each virtual power plant, of which For fixed load, The first A virtual power plant at any time The increased and decreased load power variables made are constrained. Indicates the first A virtual power plant at any time The increases and decreases in load power that are made have certain upper and lower limits. Then guarantee the first day The total increase and decrease in load power made by each virtual power plant are equal, thereby ensuring the electricity needs of users. This indicates that the virtual power plants will not simultaneously make decisions to increase or decrease load power; the energy storage system response model expression is as follows:

[0033] in For the first The response power of the energy storage system equipped in each virtual power plant. These represent the charging and discharging power of the energy storage system, respectively, with constraints. and To ensure that the charging and discharging power of the energy storage system remains within a limited range and that charging and discharging decisions are not made simultaneously at any given moment, among which... To characterize the energy storage system at time... The choice between charging and discharging is a 0-1 variable; the state-of-charge constraint expression for the energy storage system is as follows: ; in These refer to the charging efficiency and discharging efficiency of the energy storage system, respectively. For the capacity of the energy storage system, The time interval is specified, and the value of the state of charge must satisfy upper and lower limit constraints. ,in These are the minimum and maximum values ​​of the state of charge, respectively, and the constraints. To ensure that the energy storage system completes one cycle of charging and discharging in a day, These are the initial and final values ​​of the state of charge within a day, respectively.

[0034] In this embodiment of the invention, each virtual power plant aims to minimize its own operating costs and estimates the interaction power and transaction price of neighboring virtual power plants.

[0035] In this embodiment of the invention, the real-time operation model of the virtual power plant in step S3 is as follows: The constraint expressions for each virtual power plant during operation are as follows: ; ; ; in These represent the power purchased and sold between each virtual power plant and the upper-level power grid. These represent the upper limits of the power purchase and sale between each virtual power plant and the upper-level power grid. This is a 0-1 variable representing the decision of each virtual power plant to purchase or sell electricity to the upper-level power grid; These represent the power purchased and sold between each virtual power plant and its neighboring virtual power plants. These represent the upper limits of power purchase and sale between each virtual power plant and its neighboring virtual power plants. This is a 0-1 variable representing the decision of each virtual power plant to purchase or sell electricity to neighboring virtual power plants; These represent the power values ​​of the photovoltaic and wind turbine generator sets equipped within each virtual power plant.

[0036] In this embodiment of the invention, in order to improve the local consumption rate of distributed renewable energy, the transaction price of the overall power system, including all virtual power plants, during real-time operation is subject to the following constraint expression: ; in The first The virtual power plant's purchase price from the upper-level power grid and the price it sells to the upper-level power grid at any given time. The first The price at which a virtual power plant purchases electricity from neighboring virtual power plants and the price at which it sells electricity to neighboring virtual power plants.

[0037] In this embodiment of the invention, the optimized scheduling scheme in step S4 is as follows: The objective function expression for the operating cost of the virtual power plant is as follows:

[0038] in For the first The total operating cost of a virtual power plant; For the first Demand response scheduling cost of a virtual power plant This is the unit demand response dispatch cost, used to compensate users for changing their own electricity usage plans; For the first The energy storage charging and discharging loss cost of a virtual power plant Unit cost of energy storage charging and discharging losses; For the first The cost of power interaction between a virtual power plant and the upper-level power grid; For the first The cost of power interaction between a virtual power plant and neighboring virtual power plants.

[0039] In this embodiment of the invention, each virtual power plant optimizes its own operating cost objective function using the distributed augmented Lagrange method; the local decision variable expression is as follows: ; The expression for the interaction power decision variables between each virtual power plant and its neighboring virtual power plants is as follows: ; The electricity price expressions for each virtual power plant and its neighboring virtual power plants are as follows: .

[0040] This invention proposes a distributed optimization method. In this method, each entity in the system makes local optimization decisions based on partial information, namely local operational decision constraints and decision information from neighboring entities (including transaction quantities and prices). Unlike centralized optimization methods, which require global information from all entities in the system, distributed optimization methods only need local information to achieve optimization decisions for each entity. The optimization decision-making process of a single entity can be executed independently without waiting for the collection of information from all entities, thus achieving parallel computing. This improves the efficiency of the optimization decision-making process while ensuring user privacy and security (users only need to interact with their neighboring entities and do not need to broadcast their information to the entire system).

[0041] Steps S1-S3 are all parallel computing processes. Parallel computing means that each virtual power plant (sub-region) can independently make optimization decisions based on partial information (local operating constraints and decision information from neighboring regions) without waiting to collect information from all other regions. Parallel computing effectively avoids the problem in centralized optimization methods that require information from all regions in the system to make optimization decisions. If some virtual power plants respond slower than others, centralized optimization methods must wait for the slowest region to report its own information before making optimization decisions. The distributed optimization method proposed in this embodiment fundamentally avoids this problem. Each virtual power plant can make independent optimization decisions based on partial information (local operating constraints and decision information from neighboring regions). In step S4, decision information (including energy trading quantity and price) is transmitted to the regions adjacent to it. Subsequently, the decision information (including energy trading quantity and price) of both parties is iteratively optimized and corrected in a closed loop until convergence, reducing the solution complexity of the original centralized optimization problem and improving the problem-solving efficiency.

[0042] Example 2 Another embodiment of the present invention proposes a multi-virtual power plant optimal scheduling method based on the distributed augmented Lagrange method, which consists of three steps: Step 1: Model the active load and energy storage system, including, for example Figure 5 In the flowchart shown, A1 and A2 model the active load and energy storage system participating in the demand response regulation of the virtual power plant, and calculate the response power of the active load and energy storage system respectively.

[0043] Step 2: Model the real-time operation of the virtual power plant, including, for example... Figure 5 In the flowchart shown, A3 and A4, the virtual power plant collects the power interaction requirements of each local distributed resource, and formulates a preliminary local distributed resource scheduling scheme and a scheme for power interaction with neighboring virtual power plants based on the principle of dynamic power balance.

[0044] Step 3: Final optimized scheduling scheme for multiple virtual power plants, including, for example... Figure 5 The flowchart shown in Figures A5, A6, and A7 establishes an optimized scheduling model for transactions between the virtual power plant and the upper-level power grid and neighboring virtual power plants. The distributed augmented Lagrange method is then used to solve the model, resulting in the final internal distributed resource optimization scheduling scheme for the virtual power plant.

[0045] This embodiment proposes a multi-virtual power plant and distribution network collaborative optimal scheduling method based on the distributed augmented Lagrange method, which decomposes the centralized optimization objective function into the optimization objective function of each virtual power plant and solves them independently in parallel: The objective function is:

[0046] The objective function for centralized optimization is shown in equation (1), where... The first Local decision variables and decision variables interacting with neighboring VPPs for each virtual power plant (VPP). These are the operating costs of VPP local distributed resource scheduling and the costs of power interaction with neighboring VPPs, respectively. To ensure that the power distribution system meets the interactive power conservation constraint during operation. The penalty function term; For the first The VPP selection and the first Similarly, the power decision variables that interact among VPPs are also considered. For the first The VPP selection and the first The power decision variables interact among the VPPs, and both satisfy the power conservation constraint. Indicates the first The VPP belongs to the first The set of neighboring VPPs of a VPP.

[0047] The core idea of ​​the distributed augmented Lagrange method is to decompose the objective function under centralized optimization to obtain an optimization objective that can be solved independently for each VPP, thereby achieving parallel optimization of multiple subproblems and reducing the complexity of the problem. First, it introduces Lagrange duality multiplication of subvariables. The objective function under centralized solution can be written in the augmented Lagrangian form as shown in equation (2):

[0048] The equation is then decomposed to obtain the optimization objective of a single VPP, as shown in equation (3):

[0049] In the formula The physical meaning of the introduced Lagrange dual multiplier variable is the first... The VPP selection and the first The trading price set by each VPP for power interaction Then, it is the introduced slack variable, which is the first... The estimated first VPP The power decision variables that interact with each VPP are used to achieve coupling terms in the original Lagrangian function. The relaxation allows each VPP to independently formulate an optimized scheduling scheme without relying on a centralized optimization solution.

[0050] The specific process of the distributed augmenting Lagrange method is as follows: 0) Initialization: Each VPP within the power distribution system first determines the power to be exchanged and the corresponding trading price:

[0051] Update the local decision variables according to equation (4):

[0052] According to equation (5), slack variables and the interactive power decision variables received from neighboring VPPs The convex combination form updates its own interaction power decision variables, where The specified convergence step size:

[0053] Each VPP submits the updated interactive power decision variables to its neighboring VPPs, updates its own trading price according to Equation (6), and submits the updated price to its neighboring VPPs. Repeat steps 1 to 3 iteratively until the interaction power decisions made by each VPP converge to within a given threshold range, and the game between each VPP reaches equilibrium. The convergence condition is determined by equation (7), where... For the first The convergence threshold given by each VPP:

[0054] The optimal scheduling method for multiple virtual power plants based on the distributed augmented Lagrange method includes the following three steps: Step 1: Model the active load and energy storage system; Step 2: Model the real-time operation of the virtual power plant; Step 3: Final optimization scheduling scheme for multiple virtual power plants.

[0055] The specific implementation methods for each step are as follows: (1) Modeling of active loads and energy storage systems (1-1) Active load response model

[0056] As shown in equation (8), A collection of time periods throughout the 24 hours of a day. Indexes for each time period; For the first The aggregated power of the active load within each VPP, of which For fixed load, The first Each VPP at time The increased and decreased load power variables made are constrained. Indicates the first Each VPP at time The increases and decreases in load power that are made have certain upper and lower limits. This ensures that the first [time period] within a 24-hour period of a day Each VPP at time The total increase and decrease in load power are equal, thus ensuring the electricity needs of users are met. This indicates that each VPP will not simultaneously make decisions to increase or decrease load power.

[0057] (1-2) Energy storage system response model

[0058] As shown in equation (9), For the first The response power of the energy storage system (ESS) equipped in each VPP. These are the charging power and discharging power of the ESS, respectively, with constraints. and To ensure that the ESS charging and discharging power values ​​are within a limited range and that charging and discharging decisions are not made simultaneously at any given moment, among which... To characterize ESS at time The choice between charging and discharging is a 0-1 variable; the State of Charge (SOC) of the ESS needs to satisfy constraints. In the formula These are the charging efficiency and discharging efficiency of ESS, respectively. For the capacity of ESS, The time interval is specified, and the SOC value must satisfy upper and lower limit constraints. ,in These are the minimum and maximum values ​​of SOC, respectively, and the constraints. To ensure that the ESS completes one charge and discharge cycle in a day, among which These are the initial and final values ​​of SOC within a day.

[0059] (2) Modeling the real-time operation process of the virtual power plant The constraints for each VPP during operation include:

[0060] in These represent the power purchased and sold between each VPP and the upstream power grid. These represent the upper limits of the power purchase and sale between each VPP and the upstream power grid. This represents a 0-1 variable representing the decision of each VPP to purchase or sell electricity to the upstream power grid; These represent the power purchased and sold between each VPP and its neighboring VPPs. These represent the upper limits for the power purchased and sold between each VPP and its neighboring VPPs. This is a 0-1 variable representing the power purchase or sale decision of each VPP with its neighboring VPPs; These represent the power values ​​of the photovoltaic and wind turbine generator sets installed within each VPP.

[0061] To improve the local consumption rate of distributed renewable energy, the trading price during the real-time operation of the system is subject to the following constraints as shown in equation (13):

[0062] in The first The VPP's purchase price from the upper-level power grid and the price it sells electricity to the upper-level power grid at any given time. The first The power purchase price of a VPP from a neighboring VPP and the power sale price to a neighboring VPP are stipulated to be higher at any given time than the purchase price from a neighboring VPP, and lower at any given time than the power sale price to a neighboring VPP. This ensures that each VPP prioritizes power exchange with neighboring VPPs to maintain local power dynamic balance, fully utilizes the complementary and synergistic effects between VPPs, and reduces pressure on the upper-level power grid.

[0063] (3) Final Optimal Scheduling Scheme for Multiple Virtual Power Plants The objective function for VPP operating costs is expressed by equation (14):

[0064] In the formula For the first Total operating cost of a VPP; For the first Demand response scheduling cost per VPP This is the unit demand response dispatch cost, used to compensate users for changing their own electricity usage plans; For the first The cost of energy storage charging and discharging losses per VPP Unit cost of energy storage charging and discharging losses; For the first The cost of power interaction between a VPP and the upstream power grid; For the first The cost of power interaction between a VPP and its neighboring VPPs.

[0065] Each VPP optimizes its own operating cost objective function using the distributed augmented Lagrange method, where the local decision variables are:

[0066] The interaction power decision variables between each VPP and its neighboring VPPs are:

[0067] The trading prices for electricity between each VPP and its neighboring VPPs are as follows:

[0068] An optimized scheduling model is established for virtual power plants to trade with the upper-level power grid and neighboring virtual power plants. Each VPP iteratively solves its own local decision variables, interactive power decision variables, and trading electricity price, and finally obtains the optimized scheduling scheme of distributed resources within each VPP.

[0069] This embodiment proposes a multi-virtual power plant and distribution network collaborative optimal scheduling method based on the distributed augmented Lagrange method, belonging to the field of distributed resource optimization scheduling technology. This embodiment decomposes the centralized resource optimization scheduling problem into a distributed resource optimization scheduling problem involving multiple virtual power plants, effectively reducing the complexity of the solution. Each virtual power plant only needs to obtain partial information to determine its local distributed resource optimization scheduling scheme, achieving truly distributed control. First, a response model for active loads and a response model for the energy storage system are constructed. Second, each virtual power plant, aiming to minimize its own operating costs, estimates the interaction power and transaction price with neighboring virtual power plants. To maintain the dynamic power balance within the virtual power plant, it formulates its local distributed resource optimization scheduling scheme and derives the power and transaction price for interaction with neighboring virtual power plants. Finally, each virtual power plant transmits the interaction power and transaction price to neighboring virtual power plants. Through an iterative algorithm, the interaction power and price between the trading parties converge to the optimal value, obtaining the final optimized scheduling scheme. This embodiment avoids the problems of computational complexity, the curse of dimensionality, and the inability to fully explore the flexibility potential of massive and diverse distributed resources in traditional centralized optimization solutions. It can effectively improve the efficiency and flexibility of distributed resource optimization scheduling and control, while ensuring the privacy of user information.

[0070] Example 3 Figure 2This is Embodiment 3 of the present invention: The effectiveness and correctness of the optimal scheduling method for multi-virtual power plants and distribution networks are verified using an improved IEEE 33-node system as a case study. The solution is obtained using the commercial solver YALMIP / GUROBI in the MATLAB environment. Considering the improved IEEE 33-node distribution system as follows... Figure 2 As shown. The system has a total active load of 3.715 MWh, a total reactive load of 3.29 MVar, and a system reference voltage of 12.66 kV. Typical wind turbine and photovoltaic output power and load power curves for the system over a day are shown below. Figure 3 As shown, the horizontal axis represents time (hours). Nodes 7 and 33 have photovoltaic (PV) power generation equipment with a capacity of 350kW; nodes 4 and 10 have PV power generation equipment with a capacity of 300kW; node 24 has PV power generation equipment with a capacity of 400kW; nodes 20 and 27 have PV power generation equipment with capacities of 300kW and 200kW respectively; nodes 13 and 18 have wind turbine power generation equipment with a capacity of 600kW; and nodes 22 and 30 have wind turbine power generation equipment with a capacity of 500kW. The energy storage system has a rated capacity of 1MWh and a rated power of 200kW, with an initial daily SOC of 0.6, upper and lower limits of SOC of 0.9 and 0.4 respectively, and an operation and maintenance factor of 2.736. / MWh, with both charging and discharging efficiencies of 0.95, and the energy storage systems are connected at nodes 15, 26, and 28.

[0071] The optimal scheduling method of the system is solved according to the proposed method. The convergence process of the difference between the decision result obtained by the method of this embodiment and the result obtained by the traditional centralized optimization method is as follows. Figure 4 As shown, the horizontal axis represents the iteration time. The experimental results show that the method in this embodiment can still converge to the global optimum of the centralized optimization scheduling method under the condition that optimization decisions are made only based on local information. Furthermore, the solution speed and convergence speed of the method in this paper are both good. As the system scale increases, the beneficial effects that can be obtained by using the method in this embodiment are more significant than those of the traditional centralized optimization method, which proves the effectiveness and superiority of the method in this embodiment.

[0072] In view of the numerous challenges and problems faced by existing traditional centralized dispatching methods in actual power system operation and dispatching decisions, the beneficial effects achieved by the embodiments of the present invention are as follows: 1) Traditional centralized optimization methods require decision-making based on global system information. As the system scale increases, the scale of decision variables in traditional centralized optimization methods will exhibit an exponential growth trend, facing the curse of dimensionality. Directly solving the global optimization objective function will cause a huge computational burden and reduce the efficiency of problem solving, making it impossible to quickly make decisions to schedule internal system resources and reducing the efficiency of system resource utilization. The distributed optimization method proposed in this embodiment of the invention decouples the global objective function and decomposes it into local optimization objective functions for each user, which greatly reduces the computational complexity of a single optimization objective function. At the same time, each user can perform parallel computation and iterate with each other to converge to the global optimum, which significantly improves the computational efficiency, thereby improving the system decision scheduling efficiency and resource utilization efficiency. 2) Traditional centralized optimization problems require all information of every user in the system as the basis for decision-making, which is not conducive to ensuring the data privacy and security of each user. At the same time, it places high demands on the system's communication network, increasing the communication burden and communication costs. In actual operation, it is also prone to communication line congestion and failure problems, and lacks the ability to avoid risks from malicious attacks on information networks. The distributed optimization method proposed in this embodiment only needs to make optimization decisions based on local information. Each user only needs to interact with its neighboring users on a partial basis, which greatly improves the privacy and security of data. At the same time, since only local and neighboring area communication systems need to be built, communication costs and communication burdens are significantly reduced. When some users experience communication failures, other users can still make optimization scheduling decisions, which improves the robustness of the communication system in dealing with risks.

[0073] The embodiments of the present invention decompose the traditional centralized optimization problem into multiple sub-problems through decoupling, thereby avoiding the dimensionality curse problem faced by centralized problems, reducing the solution complexity of a single optimization problem, and achieving fast solution speed, while ensuring user data privacy and security.

[0074] In this embodiment of the invention, the original centralized optimization objective function is decoupled and decomposed into optimization objective functions for individual users based on the coupling constraints between users.

[0075] In a preferred embodiment, predictive analysis of the power status of distributed resources is performed to establish a set of distributed resource power status scenarios, which helps each user to better formulate local optimization scheduling schemes.

[0076] This invention also provides a multi-virtual power plant and distribution network collaborative optimization scheduling device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the multi-virtual power plant and distribution network collaborative optimization scheduling method as described above.

[0077] This invention also provides a storage medium, which includes a stored computer program, wherein the computer program, when running, controls the device where the storage medium is located to execute the multi-virtual power plant and distribution network collaborative optimization scheduling method as described above.

[0078] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, several equivalent substitutions or obvious modifications can be made without departing from the concept of the present invention, and all such modifications, achieving the same performance or purpose, should be considered within the scope of protection of the present invention.

Claims

1. A method for coordinated and optimized scheduling of multiple virtual power plants and distribution networks, characterized in that, Includes the following steps: S1. Construct an active load response model and an energy storage system response model for the demand response regulation of the virtual power plant, and calculate the response power of the active load and the energy storage system respectively. S2. Decouple and decompose the centralized optimization objective function into an independent optimization objective function for each virtual power plant and solve it. Step S2 also includes: S21. Introduce Lagrange dual multiplier variables to rewrite the optimization objective function under centralized solution as an augmented Lagrange function of centralized optimization objective function; S22. Subsequently, the augmented Lagrangian form of the centralized optimization objective function is decomposed to obtain the optimization objective of a single virtual power plant; S3. Construct a real-time operation model for virtual power plants, and make local decision variables and obtain interactive power decision variables and transaction prices with neighboring virtual power plants to maintain the local power dynamic balance of virtual power plants; with each virtual power plant aiming to minimize its own operating costs, predict the interactive power and transaction prices of neighboring virtual power plants. The real-time operation model of the virtual power plant described in step S3 is as follows: The constraint expressions for each virtual power plant during operation are as follows: ； ； ； in These represent the power purchased and sold between each virtual power plant and the upper-level power grid. These represent the upper limits of the power purchase and sale between each virtual power plant and the upper-level power grid. This is a 0-1 variable representing the decision of each virtual power plant to purchase or sell electricity to the upper-level power grid; These represent the power purchased and sold between each virtual power plant and its neighboring virtual power plants. These represent the upper limits of power purchase and sale between each virtual power plant and its neighboring virtual power plants. This is a 0-1 variable representing the decision of each virtual power plant to purchase or sell electricity to neighboring virtual power plants; These represent the power values ​​of the photovoltaic and wind turbine generator sets equipped within each virtual power plant. S4. Each virtual power plant transmits the interactive power and the transaction price to neighboring virtual power plants. Each virtual power plant iteratively solves the local decision variables, the interactive power decision variables, and the transaction price to obtain an optimized scheduling scheme for the local distributed resources of the virtual power plant. The expression for the centralized optimization objective function is as follows: ； in The first Local decision variables and decision variables interacting with neighboring VPPs for each virtual power plant. These are the local distributed resource scheduling and operation costs of the virtual power plant and the cost of power interaction with neighboring virtual power plants, respectively. To ensure that the power distribution system meets the interactive power conservation constraint during operation. The penalty function term; For the first The selection of the virtual power plant and the first Similarly, the power decision variables for interactions between virtual power plants are also... For the first The selection of the virtual power plant and the first The power decision variables of the two virtual power plants interact, and both satisfy the power conservation constraint. Indicates the first The virtual power plant belongs to the first A collection of virtual power plants near each other; The augmented Lagrange function, which is the objective function of the centralized optimization, is expressed in the following form: ； The objective function expression for the single virtual power plant is as follows: ； in The physical meaning of the introduced Lagrange dual multiplier variable is the first... The selection of the virtual power plant and the first The trading price set by virtual power plants for power exchange. Then, it is the introduced slack variable, which is the first... The estimated first virtual power plant Power decision variables that a virtual power plant interacts with.

2. The method as described in claim 1, characterized in that, The active load response model expression described in step S1 is as follows: in A collection of time periods throughout the 24 hours of a day. Indexes for each time period; For the first The active load aggregation power within each virtual power plant, of which For fixed load, The first A virtual power plant at any time The increased and decreased load power variables made are constrained. Indicates the first A virtual power plant at any time The increases and decreases in load power that are made have certain upper and lower limits. Then guarantee the first day The total increase and decrease in load power made by each virtual power plant are equal, thereby ensuring the electricity needs of users. This indicates that the virtual power plants will not simultaneously make decisions to increase or decrease load power; The energy storage system response model expression is as follows: in For the first The response power of the energy storage system equipped in each virtual power plant. These represent the charging and discharging power of the energy storage system, respectively, with constraints. and To ensure that the charging and discharging power of the energy storage system remains within a limited range and that charging and discharging decisions are not made simultaneously at any given moment, among which... To characterize the energy storage system at time... A 0-1 variable for choosing between charging and discharging; The state-of-charge constraint expression for the energy storage system is as follows: ； in These refer to the charging efficiency and discharging efficiency of the energy storage system, respectively. For the capacity of the energy storage system, The time interval is specified, and the value of the state of charge must satisfy upper and lower limit constraints. ,in These are the minimum and maximum values ​​of the state of charge, respectively, and the constraints. To ensure that the energy storage system completes one cycle of charging and discharging in a day, These are the initial and final values ​​of the state of charge within a day, respectively.

3. The method as described in claim 1, characterized in that, To improve the local consumption rate of distributed renewable energy, the trading price during the real-time operation of the power system is subject to the following constraint expression: ； in The first The virtual power plant's purchase price from the upper-level power grid and the price it sells to the upper-level power grid at any given time. The first The price at which a virtual power plant purchases electricity from neighboring virtual power plants and the price at which it sells electricity to neighboring virtual power plants.

4. The method as described in claim 1, characterized in that, The optimized scheduling scheme described in step S4 is as follows: The objective function expression for the operating cost of the virtual power plant is as follows: in For the first The total operating cost of a virtual power plant; For the first Demand response scheduling cost of a virtual power plant This is the unit demand response dispatch cost, used to compensate users for changing their own electricity usage plans; For the first The cost of energy storage charging and discharging losses for a virtual power plant. Unit cost of energy storage charging and discharging losses; For the first The cost of power interaction between a virtual power plant and the upper-level power grid; For the first The cost of power interaction between a virtual power plant and neighboring virtual power plants.

5. The method as described in claim 4, characterized in that, Each virtual power plant optimizes its own operating cost objective function using the distributed augmented Lagrange method; the expressions for the local decision variables are as follows: ； The expression for the interaction power decision variables between each virtual power plant and its neighboring virtual power plants is as follows: ； The electricity price expressions for each virtual power plant and its neighboring virtual power plants are as follows: 。 6. A multi-virtual power plant and distribution network collaborative optimization scheduling device, characterized in that, The system includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the multi-virtual power plant and distribution network collaborative optimization scheduling method as described in any one of claims 1 to 5.

7. A storage medium, characterized in that, The storage medium includes a stored computer program, wherein, when the computer program is running, it controls the device where the storage medium is located to execute the multi-virtual power plant and distribution network collaborative optimization scheduling method as described in any one of claims 1 to 5.

Images