Virtual power plant hierarchical optimization scheduling method and system considering multiple distributed resources
By constructing multiple distributed resource models within a virtual power plant and employing the alternating direction multiplier method to solve the augmented Lagrangian function, the problem of difficult distributed renewable energy consumption was solved, resource coordination and profit distribution among virtual power plants were realized, and the power grid's supply and demand balance and economic efficiency were improved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- STATE GRID JIANGSU ELECTRIC POWER CO LTD NANJING POWER SUPPLY COMPANY
- Filing Date
- 2026-02-04
- Publication Date
- 2026-05-29
AI Technical Summary
In new power systems, the mismatch between the generation of distributed renewable energy and the demand for electricity leads to difficulties in the absorption of renewable energy in some areas. The existing research on hierarchical optimization and dispatch of virtual power plants is insufficient, making it difficult to achieve grid supply and demand balance and increase the proportion of clean energy absorption.
Multiple distributed resource models are constructed within the virtual power plant. The augmented Lagrangian function is solved using the alternating direction multiplier method. An upper-level optimal scheduling model is established to maximize the revenue of the virtual power plant. Resource coordination and profit allocation are carried out through a lower-level cooperative game model, forming a hierarchical optimal scheduling system.
It enables resource collaboration and fair profit distribution among virtual power plants, increases the proportion of clean energy consumption, reduces operating costs, and improves system flexibility and economic efficiency.
Smart Images
Figure CN122118945A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of virtual power plant optimization scheduling technology, and relates to a hierarchical optimization scheduling method and system for virtual power plants that considers multiple distributed resources. Background Technology
[0002] Against the backdrop of new power system construction, the integration of massive distributed power sources into the grid has become a major characteristic of the power system. However, the mismatch between distributed renewable energy generation and electricity load demand leads to difficulties in the absorption of distributed resources in some areas. The participation of virtual power plants in regulating distributed resources is of great significance for achieving grid supply and demand balance.
[0003] Currently, research on key technologies for hierarchical optimal scheduling of virtual power plants is still insufficient. Since the load demand of multiple virtual power plants differs at the same time, coordinated hierarchical control can incentivize each virtual power plant to actively explore its load-side regulation capacity. This not only helps increase the proportion of clean energy consumption and effectively responds to the call for energy conservation and emission reduction, but also improves operational profitability. In this context, virtual power plants within the same distribution network area can form multi-virtual power plant alliances through cooperative game theory, promoting internal resource synergy and the local consumption of distributed resources, enhancing system flexibility, reducing the risks of individual participation in market transactions, and improving the efficiency of each virtual power plant. Although some projects have begun to study joint trading mechanisms for multiple virtual power plants, improving their economic efficiency, research on hierarchical optimal scheduling frameworks among multiple virtual power plants remains insufficient. Summary of the Invention
[0004] To address the shortcomings of existing technologies, this invention provides a hierarchical optimization scheduling method and system for virtual power plants that considers multiple distributed resources. This method can better provide optimal strategy solutions, facilitate the multi-level coordinated operation of virtual power plant aggregation units, complete the comprehensive regulation of power generation and consumption resources, and improve overall energy utilization efficiency and economic benefits.
[0005] The present invention adopts the following technical solution.
[0006] The first aspect of this invention proposes a hierarchical optimization scheduling method for virtual power plants that considers multiple distributed resources, comprising: We construct multiple distributed resource models within a virtual power plant and, based on these models, build an upper-level optimization scheduling model with the goal of maximizing the revenue of the virtual power plant; we also construct a lower-level optimization scheduling model that considers the cooperative game between virtual power plants. Augmented Lagrangian functions are constructed for the upper-level and lower-level optimal scheduling models, respectively. The augmented Lagrangian function is solved using the alternating direction multiplier method to obtain a hierarchical optimization scheduling strategy for the virtual power plant.
[0007] Preferably, the various distributed resources include electric vehicles, energy storage systems, gas turbines, central air conditioning, and flexible loads.
[0008] Preferably, the construction of the upper-level optimization scheduling model with the goal of maximizing the revenue of the virtual power plant is as follows: (1) In the formula, This represents the profit generated by the user's energy consumption; Indicates the first i The operating costs of electric vehicles within a virtual power plant; Indicates the first i The cost of power transmission loss when sharing power among virtual power plants; Indicates the first i The operating cost of energy storage within a virtual power plant; Indicates the first i The cost of a central air conditioning system within a virtual power plant; Indicates the first i The cost of transferring transferable loads within a virtual power plant; Indicates the first i The cost of cutting off loads within a virtual power plant; Indicates the first i The operating cost of the gas turbine in a virtual power plant.
[0009] Preferably, the Lagrange augmented function for constructing the upper-level optimization scheduling model is as follows: (46) In the formula, Let be the Lagrange multiplier for energy sharing between the i-th and j-th virtual power plants at time t; For penalty parameters; Let be the variable for the interaction between virtual power plant i and virtual power plant j. Let be the variable that allows virtual power plant j to interact with virtual power plant i.
[0010] Preferably, the construction of the lower-level optimization scheduling model considering the cooperative game among virtual power plants is as follows: (47) In the formula, For the first i The cost of a virtual power plant participating in energy sharing; For the first i The cost of a virtual power plant participating in power sharing; For the first i The cost of power transmission losses when sharing power among virtual power plants. For the number of time periods, This represents the number of virtual power plants.
[0011] Preferably, the augmented Lagrangian function for constructing the lower-level optimal scheduling model includes: Taking the logarithm of the lower-level optimal scheduling model transforms the problem of finding the maximum value of the lower-level optimal scheduling model into the problem of finding the minimum value, resulting in the transformed formula: (48) Based on the transformed formula, the augmented Lagrangian function of the lower-level optimization scheduling model is established as follows: (49) In the formula, Indicates the first i The bargaining factor of a virtual power plant The Lagrange multiplier representing the backup shared between virtual power plants i and j. Indicates the penalty factor. It is an L2 norm; Let be the variable for the interaction between virtual power plant i and virtual power plant j. Let be the variable that allows virtual power plant j to interact with virtual power plant i.
[0012] Preferably, the formula for calculating the bargaining factor is: (50) In the formula, Indicates the basic bargaining weight. To prevent positive numbers with a denominator of zero, Cost sensitivity coefficient For time period t Reduced load power For a moment t Power transferred from internal load.
[0013] Preferably, the augmented Lagrangian function of the upper-level optimization scheduling model is solved using the alternating direction multiplier method, specifically including: 1) Iteratively solve the augmented Lagrangian function of the upper-level optimization scheduling model and update the variables. , of which k The iteration process is as follows: (51) in, For the first k The augmented Lagrangian function of the upper-level optimization scheduling model in the next iteration; For the first k Lagrange multipliers in the next iteration; For the first k The variables that virtual power plant i interacts with virtual power plant j in each iteration. For the firstk The variables that virtual power plant j interacts with virtual power plant i during the next iteration.
[0014] 2) In the k After the solution is obtained in the next iteration, the Lagrange multipliers are updated: (52) 3) Number of update iterations k=k+ 1; 4) Determine the convergence status of the algorithm: (53) in, , These are the upper tolerance limits for the original residual and the dual residual, respectively; For penalty parameters; When the above formula is satisfied, the iteration is complete; otherwise, the iteration continues until the number of iterations exceeds a preset value.
[0015] Preferably, the augmented Lagrangian function of the lower-level optimal scheduling model is solved using the alternating direction multiplier method, specifically including: 1) Iteratively solve the augmented Lagrangian function of the lower-level optimization scheduling model and update the variables. , of which k The iteration process is as follows: (54) in, For the first k The augmented Lagrangian function of the lower-level optimization scheduling model in the next iteration; For the first k Lagrange multipliers in the next iteration; For the first k The variables that virtual power plant i interacts with virtual power plant j in each iteration. For the first k The variables that virtual power plant j interacts with virtual power plant i during the next iteration; 2) In the k After the solution is obtained in the next iteration, the Lagrange multipliers are updated: (55) 3) Number of update iterations k=k +1; 4) Determine the convergence status of the algorithm: (56) (57) When the above formula is satisfied, the iteration is complete; otherwise, the iteration continues until the number of iterations exceeds a preset value.
[0016] A second aspect of this invention proposes a hierarchical optimization scheduling system for virtual power plants that considers multiple distributed resources, comprising: The upper-level model construction module is used to construct various distributed resource models within the virtual power plant and, based on these models, construct an upper-level optimization scheduling model with the goal of maximizing the revenue of the virtual power plant. The lower-level model building module is used to build a lower-level optimization scheduling model that considers the cooperative game between virtual power plants; The model solving module is used to construct the augmented Lagrangian functions of the upper-level and lower-level optimal scheduling models, respectively; the augmented Lagrangian functions are solved using the alternating direction multiplier method to obtain the hierarchical optimal scheduling strategy of the virtual power plant.
[0017] A third aspect of the present invention provides a terminal, including a processor and a storage medium; the storage medium is used to store instructions; the processor is used to perform operations according to the instructions to execute the steps of the method.
[0018] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method.
[0019] Compared with the prior art, the beneficial effects of the present invention include at least the following: This invention constructs an upper-level optimal scheduling model aiming to maximize the revenue of virtual power plants and a lower-level optimal scheduling model considering the cooperative game among virtual power plants. Augmented Lagrangian functions are constructed for each model, collectively forming a novel and complete hierarchical optimal scheduling solution system for virtual power plants. This system achieves for the first time a specific hierarchical optimization structure of "upper-level multi-resource scheduling - lower-level multi-virtual power plant Nash bargaining," and innovatively employs the alternating direction multiplier method as a collaborative solver connecting the upper and lower levels. Specifically, the upper-level model focuses on the optimal scheduling decisions of multiple distributed resources (such as electric vehicles, energy storage, and gas turbines) within the virtual power plant to maximize its own revenue; the lower-level model focuses on the cooperative game among multiple virtual power plants, using an asymmetric Nash bargaining model to fairly distribute alliance profits, and also employs the alternating direction multiplier method for distributed solution.
[0020] This invention constructs multiple distributed resource models within a virtual power plant, comprehensively reflecting the characteristics and operating status of different resources within the virtual power plant, thereby enabling better resource allocation. Based on this, an upper-level optimal scheduling model is constructed with the goal of maximizing the revenue of the virtual power plant, improving overall economic efficiency, maximizing the utilization of renewable energy, and reducing operating costs. A lower-level optimal scheduling model considering the cooperative game among virtual power plants is constructed, enabling each virtual power plant to coordinate and share resources during the scheduling process, avoiding resource duplication and waste. Augmented Lagrangian functions for the upper-level and lower-level optimal scheduling models are constructed respectively, transforming the optimization problem into an easily solvable form, optimizing the solution process, and making the model solution more stable and efficient. The alternating direction multiplier method is used to solve the augmented Lagrangian function, which can solve complex optimization problems step by step, thereby obtaining a hierarchical optimal scheduling strategy for virtual power plants.
[0021] This invention deeply couples physical operational constraints with economic bargaining factors within the underlying asymmetric Nash bargaining model and its solution framework. Specifically, it overcomes the limitation of fixed-value bargaining factors in traditional models by introducing dynamic bargaining factors. This dynamically links the physical operational characteristics of the virtual power plant with its economic bargaining power, achieving both dynamism and precision in the bargaining factors: they are no longer preset static parameters but are dynamically calculated from actual operational data. This includes the virtual power plant's operating costs, resource sharing contributions, and flexible load adjustments, enabling bargaining power to reflect its actual physical contribution in real time and accurately. Furthermore, it constructs a fair allocation mechanism with dual incentives of cost and contribution: through the product structure of cost-saving rewards (index term) and resource-sharing rewards (proportional term), transparent incentive rules are established. This ensures that profit distribution depends not only on bargaining power but also on quantifiable actual contributions, greatly improving the fairness and incentive compatibility of the distribution. In addition, it forms a closed-loop optimization of physical operation and game decision-making: the dynamic bargaining factor acts as a bridge, feeding back the physical operation results output by the upper-level scheduling model to the lower-level game model, directly affecting the profit distribution scheme in this round. The new distribution scheme, in turn, affects the scheduling strategy of each virtual power plant in the next round. This closed-loop feedback mechanism forces each virtual power plant to consider its overall physical contribution to the alliance when seeking to maximize its own interests, thereby driving the system towards the overall optimal operating state and achieving a high degree of synergy between the physical and economic layers. Attached Figure Description
[0022] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a flowchart illustrating the implementation of the method of the present invention; Figure 3 This refers to the convergence status of the residuals; Figure 4 The revenue iteration status of virtual power plant 1; Figure 5 The revenue iteration status of Virtual Power Plant 2; Figure 6 This describes the revenue iteration of Virtual Power Plant 3. Detailed Implementation
[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.
[0024] Embodiment 1 of the present invention provides a hierarchical optimization scheduling method for virtual power plants that considers multiple distributed resources, such as... Figure 1 and Figure 2 As shown, the method includes the following steps: Step 1: Construct multiple distributed resource models within the virtual power plant and, based on these models, construct an upper-level optimization scheduling model with the goal of maximizing the revenue of the virtual power plant. More preferably, a model of the internal components of the upper-level virtual power plant and a profit maximization function model are established, as follows: Establish a revenue-maximizing function model for the upper-level virtual power plant: (1) In the formula, This represents the profit generated by the user's energy consumption; Indicates the first i The operating costs of electric vehicles within a virtual power plant; Indicates the first i The cost of power transmission loss when sharing power among virtual power plants; Indicates the first i The operating cost of energy storage within a virtual power plant; Indicates the first i The cost of a central air conditioning system within a virtual power plant; Indicates the first i The cost of transferring transferable loads within a virtual power plant; Indicates the first i The cost of cutting off loads within a virtual power plant; Indicates the first i Operating costs of gas turbines within a virtual power plant; This refers to the number of time periods, specifically time period 1 to time period 24.
[0025] The internal units of the virtual power plant include electric vehicle models, energy storage system models, gas turbine models, central air conditioning models, and flexible load models; The following section will introduce the specific internal model of the virtual power plant involved in the above model and the constraints: (1) Electric vehicle charging and discharging model In this invention, it is assumed that electric vehicles travel according to certain patterns and that the charging and discharging power is fixed, then the model is as follows: (2) Maximum energy reached: (3) The amount of electricity fully charged for the electric vehicle is: (4) In the formula, The time frame represents the start time of charging the electric vehicle. The time period refers to a time interval, while time period t refers to a specific time period. This refers to the charging power of electric vehicles; It is the discharge power of the electric vehicle; This is the power of the electric vehicle during time period t; , This refers to the maximum charging and discharging power of electric vehicles. yes The battery level of electric vehicles during a given time period. The initial charge level of the electric vehicle; This refers to the battery level of an electric vehicle before it is charged. This indicates the maximum energy capacity of an electric vehicle battery; This indicates the electric vehicle's willingness to charge and discharge. When it is 0, it means the electric vehicle is only charging; when it is 1, it means the electric vehicle is discharging.
[0026] Furthermore, to correspond with actual conditions, the charging cost in this invention is set to the maximum charging cost within the electric vehicle access period. Each electric vehicle will choose to charge and discharge when its own benefit is maximized. The charging and discharging cost of the electric vehicle is: (5) In the formula, Indicates the first i The cost of using electric vehicles within a virtual power plant This represents the unit charging cost of an electric vehicle. , Indicates the charging and discharging power of electric vehicles; T Indicates the first T During the period, N Indicates the first iThere are virtual power plants N A cluster of electric vehicles.
[0027] (2) Energy storage system model Energy storage systems are primarily designed to balance electrical energy levels and can only charge or discharge at any given time. Their model is as follows: (6) In the formula, , ; , This represents the maximum charging and discharging power of the energy storage system. , For energy storage systems t The charging and discharging power at any given moment.
[0028] Energy storage charging and discharging power: (7) In the formula, Indicates electric vehicles t Power at any moment express t Charging power at any time express t Discharge power at any given time , This indicates the charging efficiency and discharging efficiency of an electric vehicle.
[0029] Energy stored: (8) in, The time difference is within 1 hour.
[0030] Energy storage constraints: (9) (10) In the formula, , Indicates the minimum and maximum capacity of energy storage; Indicates the rated capacity of energy storage; This indicates the amount of electricity stored in the last hour. This indicates the amount of electricity stored at the initial moment. Indicates the first i A virtual power plant t The amount of electricity stored at all times Indicates the first i The power of energy stored in a virtual power plant at time t.
[0031] Therefore, the cost of operating energy storage is: (11) In the formula, Indicates the cost of energy storage. , This represents the charging and discharging cost coefficient, set at 0.05-0.15 yuan / kWh. , This indicates the energy storage charging and discharging power.
[0032] (3) Gas turbine model Gas turbines incur certain fixed losses during operation, which are related to their power generation capacity. (12) In the formula, , , The correlation coefficient represents the cost of gas turbine power generation. The price is 0.02-0.05 yuan / kW²h. The price is 0.1-0.3 yuan / kWh. For 10-50, This indicates the power output of the gas turbine. Indicates the first i The operating cost of the gas turbine in a virtual power plant.
[0033] In addition, the operating constraints of gas turbines are as follows: (13) (14) In the formula, ramp This indicates the gradeability of the gas turbine. , These represent the minimum and maximum output of the gas turbine, respectively. express t The output of the gas turbine at all times.
[0034] (4) Central air conditioning model A central air conditioning system mainly consists of three circulation systems: chilled water circulation, cooling water circulation, and refrigerant circulation. It includes one or more refrigeration units and a corresponding number of chilled water pumps, cooling water pumps, and cooling towers, as well as terminal equipment such as fan coil units or fresh air units.
[0035] For a person with N 1 chiller unit N 2 chilled water pumps N 3 cooling coil fans, N 4 cooling water pumps and N The energy consumption model of an air conditioning system with 5 cooling towers can be expressed as follows: (15) In the formula, This represents the total energy consumption of the central air conditioning system. For the first u Energy consumption of the chiller unit For the first v Power of chilled water pump, P CWpump,m For the first m The power of the cooling water pump, P coil,k For the first k Each fan coil unit power, P tower,n For the first n The power of the cooling tower.
[0036] The chiller unit is the core of a central air conditioning system, mainly composed of a compressor, condenser, expansion valve, and evaporator. The compressor is the primary energy-consuming component. For multi-chiller air conditioning systems, the energy consumption of the chiller units depends on the coefficient of performance (COP), which can be represented by the following model: (16) In the formula, Q e The chilled water circulation side cooling load; the COP (Coefficient of Performance) is related to the chiller's evaporation temperature, condensation temperature, and load rate, and can be expressed as: (17) In the formula, r The chiller unit load rate can be expressed as cooling load. Q e With rated load Q nom The ratio, i.e. r = Q e / Q nom . and These are constant coefficients related to COP, with values ranging from 2.5 to 4.0 and from -0.1 to 0.1, respectively. T e The evaporation temperature. T c The condensation temperature can be calculated using the following formula: (18) (19) In the formula, T chwr and T cwsThese represent the chilled water return temperature and the cooling water supply temperature, respectively. Q c This indicates the load on the cooling water circulation side. F chw ( m chw )and F cw ( m cw ) It concerns chilled water flow rate m chw and cooling water flow rate m cw The empirical formula.
[0037] The water pumps in a central air conditioning system mainly include chilled water pumps and cooling water pumps, which provide power for the water system circulation. Currently, most building air conditioning systems operate at their rated flow rate, constantly running at full load, resulting in unnecessary waste.
[0038] This invention assumes that all water pumps are equipped with variable frequency control systems to facilitate flow control and reduce energy consumption. The power of the variable frequency water pump depends on the water flow rate, i.e.: (20) (twenty one) (twenty two) In the formula, For chilled / cooling water pump power, This refers to the chilled / cooling water flow rate. for t Time period s Chilled water pump in the scenario l and cooling water pump n Water flow ratio, This refers to the rated flow rate of chilled / cooling water. , , and For constant coefficients related to the mechanical efficiency of water pumps or fans, the values range from 0.8 to 1.2, 0.1 to 0.3, 0.05 to 0.15, and 0.01 to 0.05, respectively. For cooling towers p Mechanical efficiency k p and A p The correlation coefficients have values ranging from 0.5 to 2.0 and from 0.1 to 1, respectively.
[0039] Fan coil units are terminal devices in central air conditioning systems. The chilled water in the coils exchanges heat with the indoor air to achieve cooling. The heat exchange process can be represented as: (twenty three) In the formula, Q room.k For the corresponding area's cooling load, m sa,k and m chw,k These are the fan speed and the chilled water flow rate in the coil, respectively. , , These are the relevant constant coefficients for the air handling unit, with values ranging from 0.8 to 1.2, -4.5 to -3.5, and -0.5 to 0.5, respectively. T chws Indicates the chilled water supply temperature. T ma,k The mixed temperature of indoor and outdoor air inside the bellows can be expressed as: ,in For the room k Indoor temperature, This represents the outdoor temperature. Simultaneously, the sum of the chilled water flow rates within the coils equals the total chilled water flow rate in the system, i.e.: (twenty four) A cooling tower is a heat dissipation device in a central air conditioning system. It absorbs heat from the indoor air and releases it into the atmosphere through the circulation of cooling water. The heat dissipation process can be represented as: (25) In the formula, For cooling load demand, m ta,n and m cw,n These are the wind speed and cooling water flow rate in the cooling tower, respectively. T cwr and T wb These are the chilled water return temperature and the cooling tower wet-bulb temperature, respectively. Furthermore, the total cooling water flow rate in the air conditioning system remains constant, i.e.: (26) The energy consumption of fan coil units and cooling towers both originate from the fans. Their energy consumption model is similar to that of variable frequency water pumps, and can be expressed as: (27) (28) In the formula, for t Time period s Cooling tower in the scenep power, This refers to the airflow ratio of the cooling tower. For the mechanical efficiency of the cooling tower, The mechanical efficiency coefficient of the cooling tower fan is 0.5-2. This is the heat transfer area coefficient of the cooling tower, with a value ranging from 0.1 to 1. and m sa / ta.nom These are the wind speed and rated wind speed of the fans in the fan coil unit and cooling tower, respectively.
[0040] The constraints of a central air conditioning system mainly include the interactions between equipment and the physical constraints of variables. Specifically, the interaction between the chiller unit and the chilled / cooling water cycle can be expressed as: (29) (30) In the formula, T cwr This refers to the chilled water return temperature. T chwr and T cws These represent the chilled water return temperature and the cooling water supply temperature, respectively. The cooling water eliminates heat on the cooling water side based on the energy balance principle, including the heat generated by the compressor and the heat transferred from the chilled water circulation side to the condenser via the evaporator. Therefore, the cooling load in the cooling water circulation... Q c,i = Q e,i + That is, the cooling load demand should be the sum of the cooling load on the chilled water circulation side and the energy consumption of the chiller unit.
[0041] To ensure optimal performance of the air conditioning system, the variables in the system must be kept within an acceptable range.
[0042] (31) (32) (33) (34) (35) (36) In the formula, The formula above sets the upper and lower limits of the chilled water pump flow rate, the upper and lower limits of the chilled water and cooling water supply temperature, the upper and lower limits of the pump flow rate, and the upper and lower limits of the fan speed in the chiller unit.
[0043] Therefore, the operating cost of a central air conditioning system is: (37) In the formula, The user discomfort coefficient ranges from 0.1 to 1. A higher value indicates that the user is more sensitive to discomfort. For the first i Indoor temperature of users within a virtual power plant; Indicates the first i The most comfortable temperature for users within a virtual power plant.
[0044] (5) Flexible load model Flexible loads include transferable loads and reduceable loads, with the following costs: (38) (39) In the formula, This indicates that load costs can be reduced. Indicates the cost of transferable loads; This is the load transfer compensation coefficient; This is the load reduction compensation factor; For time period t Reduced load power For a moment t Power transferred from internal load.
[0045] Transferable and reduceable loads have the following constraints: (40) (41) (42) (43) In the formula, Indicates the shortest time the load can operate continuously. and Indicates the minimum and maximum time during which the load is allowed to be cut off continuously at one time. and This is a Boolean variable; a value of 1 indicates transfer / cutoff, and a value of 0 indicates no transfer / cutoff. for The amount of load transferred within a given time period; , Indicates the upper and lower limits of load transferability. This represents the maximum number of interruptions allowed for the load in a day, with a value ranging from 0 to 24.
[0046] (6) Profits generated from user energy consumption and energy transmission costs: (44) (45) In the formula, This represents the weighting coefficient, with a value range of (0, 1], reflecting the user's electricity consumption efficiency; Indicates the power consumed by the user; , The energy transmission loss coefficients between virtual power plants are set to 0.5 and 0.3, respectively. Indicates the first i The cost of power transmission losses when sharing power among virtual power plants.
[0047] Step 2: Construct a lower-level optimization scheduling model that considers the cooperative game between virtual power plants; Step 3: Construct the augmented Lagrangian functions for the upper-level and lower-level optimal scheduling models, respectively; More preferably, a lower-level asymmetric Nash bargaining profit distribution model based on cooperative game theory for multiple virtual power plants is established; Establish an asymmetric Nash bargaining model for multiple virtual power plants based on cooperative game theory.
[0048] First, the Lagrange augmented function of the upper-level virtual power plant cost-minimizing function model is established to solve the asymmetric Nash bargaining problem: (46) In the formula, Cost of using electric vehicles; Costs associated with electricity sharing; For energy storage costs; Cost of using the central air conditioning system; For load transfer costs; Cost of load cut-off; For gas turbine operating costs; Let be the Lagrange multiplier for energy sharing between the i-th and j-th virtual power plants at time t; The penalty parameter is the step size, and >0; For user benefit; , Let be the electrical energy shared between the i-th and j-th virtual power plants, where Let be the variable for the interaction between virtual power plant i and virtual power plant j. Let be the variable that allows virtual power plant j to interact with virtual power plant i.
[0049] Secondly, a lower-level asymmetric Nash bargaining function model is established to distribute the profits from the cooperation among multiple virtual power plants: The cost reduction resulting from cooperative game theory among virtual power plants represents the benefit achieved through virtual power plant cooperation. Based on cooperative game theory, the following asymmetric Nash bargaining model is established: (47) In the formula, For virtual power plants i The costs before participating in energy sharing; For virtual power plants i The costs of participating in electricity sharing; This refers to the cost of electricity sharing. It is understandable that this is a global definition, and the formula in (48) also satisfies the constraints here. .
[0050] Next, taking the logarithm of the above model transforms the problem of finding the maximum value into the problem of finding the minimum value. The transformed formula is shown below: (48) Then, the augmented Lagrange function is established as follows: (49) In the formula, Represents a virtual power plant i Bargaining factors The Lagrange multiplier representing the backup shared between virtual power plants i and j. Denotes the penalty factor, and >0, It is an L2 norm; , These are the shared backups between virtual power plants i and j, where... Let be the variable for the interaction between virtual power plant i and virtual power plant j. Let be the variable that allows virtual power plant j to interact with virtual power plant i.
[0051] In practice, A fixed bargaining factor value can be used, which can be set according to the actual situation. For example, the values for virtual power plant 1, virtual power plant 2, and virtual power plant 3 are 2.45, 2.68, and 1.85, respectively. Furthermore, the dynamic bargaining factor shown in formula (50) can be further: (50) in, This represents the basic bargaining weight. In this embodiment, the corresponding values for virtual power plant 1, virtual power plant 2, and virtual power plant 3 can be 2.45, 2.68, and 1.85, respectively. To prevent small positive numbers with a denominator of zero, The cost sensitivity coefficient should be a normal value. For time periodt Reduced load power For time period t Power transferred from internal load.
[0052] When the cost of the aforementioned dynamic bargaining factor decreases slightly, the incentive growth slows down; when it decreases significantly, the incentive growth accelerates, which can more effectively incentivize members to perform in-depth optimization.
[0053] Step 4: Solve the augmented Lagrangian function using the alternating direction multiplier method to obtain the hierarchical optimization scheduling strategy for the virtual power plant.
[0054] More preferably, a solution method for the upper and lower level models based on the alternating direction multiplier method is established: First, based on the alternating direction multiplier method shown in formulas (51) to (53), the upper-level virtual power plant cost minimization function model is solved. The specific steps are as follows: 1) Solve the upper-level augmented Lagrangian function model and update the variables. ,in k Indicates the first k The next iteration.
[0055] (51) in, For the first k The Lagrange multipliers and the energy interaction variables at the next iteration, among which For the first k The variables that virtual power plant i interacts with virtual power plant j in each iteration. For the first k The variables that virtual power plant j interacts with virtual power plant i during the next iteration; Other virtual power plants also perform the same operation upon receiving the information, updating the variables. .
[0056] 2) In the k After the next iteration, the Lagrange multipliers are updated: (52) When it does not converge, the updated Substitute into formula (50) for the next iteration until convergence is determined; 3) Number of update iterations k=k+ 1; 4) Determine the convergence status of the algorithm: (53) in, , These are the upper tolerance limits for the original residual and the dual residual, respectively, both set to 0.001; The iteration is complete when both the original residual and the dual residual are less than predetermined values, and the output is given. , ;otherwise k The iteration continues by incrementing by 1. When the number of iterations exceeds a preset value, the program fails to converge and the iteration ends.
[0057] Secondly, solve the lower-level asymmetric Nash bargaining Lagrange augmented function model: 1) Solve the lower-level asymmetric Nash bargaining model and update the variables. ,in k Indicates the first k The next iteration.
[0058] (54) in, For the first k The augmented Lagrangian function of the lower-level optimization scheduling model in the next iteration (Equation (49)); For the first k The Lagrange multipliers and alternative interaction variables at the next iteration; where For the first k The variables that virtual power plant i interacts with virtual power plant j in each iteration. For the first k The variables that virtual power plant j interacts with virtual power plant i during the next iteration; Virtual power plant j Upon receiving the price information, the augmented Lagrangian function is also calculated simultaneously, and the variables are updated. .
[0059] 2) In the k After the next iteration, the Lagrange multipliers are updated: (55) 3) Number of update iterations k=k +1; 4) Determine the convergence status of the algorithm: (56) (57) The iteration is complete when both the original residual and the dual residual are less than predetermined values, and the output is given. , ;otherwise k+ 1. Continue iterating. When the number of iterations exceeds the preset value, the program does not converge and the iteration ends.
[0060] The GAMS software is used to solve the upper-level virtual power plant revenue maximization model and the lower-level multi-virtual power plant asymmetric Nash bargaining model based on cooperative game theory, and the results are output.
[0061] To verify the effectiveness of the present invention, a simulation example consisting of three virtual power plants was used to verify the effectiveness of the proposed method.
[0062] Each virtual power plant includes electric vehicles, energy storage systems, gas turbines, central air conditioning, and flexible loads.
[0063] The purchase and sale price of virtual power plants is calculated according to the electricity price promulgated by a certain province in 2021. The electricity price is divided into peak hours (8:00-11:00, 17:00-22:00), off-peak hours (0:00-8:00), and normal hours (11:00-17:00, 22:00-24:00).
[0064] According to the current distributed power surplus electricity feed-in tariff policy, the electricity purchase price remains at 0.85 yuan / kWh throughout the day. Data for the virtual power plant aggregation unit is shown in Table 1.
[0065] To compare the revenue of virtual power plant consortia under different trading models, this invention sets up the following comparative examples: Case 1: Virtual power plants do not share energy with each other, but only trade energy with the upper-level power grid; Case 2: Virtual power plants participate in energy sharing and also trade with the upper-level power grid.
[0066] Table 1. Internal Aggregation Parameters of the Virtual Power Plant
[0067] Based on the calculation examples, the operational benefits of each virtual power plant under case 1 and case 2 are shown in Table 2.
[0068] Table 2 Profit Distribution of Virtual Power Plants at the Lower Level
[0069] As shown in Table 2, after adopting the asymmetric Nash bargaining method for profit distribution, the revenue increased by RMB 369.678, RMB 406.23, and RMB 275.817 respectively, and the extent of the revenue increase for each virtual power plant mainly depended on the size of the bargaining factor. This approach allows virtual power plants that share more electricity to receive greater revenue, thus promoting fairness.
[0070] Table 3. Total Electricity Purchased by the Virtual Power Plant Consortium and its Renewable Energy Consumption Rate
[0071] As shown in Table 3, Case 2 achieved 100% full utilization of renewable energy, completely resolving the 25.2% curtailment issue in Case 1. The alliance reduced its externally purchased electricity from 1850.320 kWh to 620.924 kWh, a decrease of 66.5%, highlighting the efficiency of internal resource synergy and localized utilization. By optimizing the scheduling of flexible resources such as energy storage and electric vehicles, the system's flexibility and overall energy efficiency have been fundamentally improved.
[0072] Figure 3 The graph shows the changes in the original and dual residuals during the solution process of the asymmetric Nash bargaining model. As can be seen from the graph, the original and dual residuals are already very small after only 20 iterations, indicating that the algorithm has good convergence.
[0073] Figures 4 to 6 The algorithm demonstrates the revenue of each virtual power plant. With iteration, the revenue of each virtual power plant gradually increases and eventually stabilizes, indicating that the game between the virtual power plants resulted in each plant reaching its maximum profit. The algorithm also exhibits a certain degree of convergence.
[0074] Embodiment 2 of the present invention provides a hierarchical optimization scheduling system for virtual power plants that considers multiple distributed resources, comprising: The upper-level model construction module is used to construct various distributed resource models within the virtual power plant and, based on these models, construct an upper-level optimization scheduling model with the goal of maximizing the revenue of the virtual power plant. The lower-level model building module is used to build a lower-level optimization scheduling model that considers the cooperative game between virtual power plants; The model solving module is used to construct the augmented Lagrangian functions of the upper-level and lower-level optimal scheduling models, respectively; the augmented Lagrangian functions are solved using the alternating direction multiplier method to obtain the hierarchical optimal scheduling strategy of the virtual power plant.
[0075] Embodiment 3 of the present invention provides a terminal, including a processor and a storage medium; the storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the steps of the method.
[0076] Embodiment 4 of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method.
[0077] Compared with the prior art, the beneficial effects of the present invention include at least the following: This invention constructs an upper-level optimal scheduling model aiming to maximize the revenue of virtual power plants and a lower-level optimal scheduling model considering the cooperative game among virtual power plants. Augmented Lagrangian functions are constructed for each model, collectively forming a novel and complete hierarchical optimal scheduling solution system for virtual power plants. This system achieves for the first time a specific hierarchical optimization structure of "upper-level multi-resource scheduling - lower-level multi-virtual power plant Nash bargaining," and innovatively employs the alternating direction multiplier method as a collaborative solver connecting the upper and lower levels. Specifically, the upper-level model focuses on the optimal scheduling decisions of multiple distributed resources (such as electric vehicles, energy storage, and gas turbines) within the virtual power plant to maximize its own revenue; the lower-level model focuses on the cooperative game among multiple virtual power plants, using an asymmetric Nash bargaining model to fairly distribute alliance profits, and also employs the alternating direction multiplier method for distributed solution.
[0078] This invention constructs multiple distributed resource models within a virtual power plant, comprehensively reflecting the characteristics and operating status of different resources within the virtual power plant, thereby enabling better resource allocation. Based on this, an upper-level optimal scheduling model is constructed with the goal of maximizing the revenue of the virtual power plant, improving overall economic efficiency, maximizing the utilization of renewable energy, and reducing operating costs. A lower-level optimal scheduling model considering the cooperative game among virtual power plants is constructed, enabling each virtual power plant to coordinate and share resources during the scheduling process, avoiding resource duplication and waste. Augmented Lagrangian functions for the upper-level and lower-level optimal scheduling models are constructed respectively, transforming the optimization problem into an easily solvable form, optimizing the solution process, and making the model solution more stable and efficient. The alternating direction multiplier method is used to solve the augmented Lagrangian function, which can solve complex optimization problems step by step, thereby obtaining a hierarchical optimal scheduling strategy for virtual power plants.
[0079] This invention deeply couples physical operational constraints with economic bargaining factors within the underlying asymmetric Nash bargaining model and its solution framework. Specifically, it overcomes the limitation of fixed-value bargaining factors in traditional models by introducing dynamic bargaining factors. This dynamically links the physical operational characteristics of the virtual power plant with its economic bargaining power, achieving both dynamism and precision in the bargaining factors: they are no longer preset static parameters but are dynamically calculated from actual operational data. This includes the virtual power plant's operating costs, resource sharing contributions, and flexible load adjustments, enabling bargaining power to reflect its actual physical contribution in real time and accurately. Furthermore, it constructs a fair allocation mechanism with dual incentives of cost and contribution: through the product structure of cost-saving rewards (index term) and resource-sharing rewards (proportional term), transparent incentive rules are established. This ensures that profit distribution depends not only on bargaining power but also on quantifiable actual contributions, greatly improving the fairness and incentive compatibility of the distribution. In addition, it forms a closed-loop optimization of physical operation and game decision-making: the dynamic bargaining factor acts as a bridge, feeding back the physical operation results output by the upper-level scheduling model to the lower-level game model, directly affecting the profit distribution scheme in this round. The new distribution scheme, in turn, affects the scheduling strategy of each virtual power plant in the next round. This closed-loop feedback mechanism forces each virtual power plant to consider its overall physical contribution to the alliance when seeking to maximize its own interests, thereby driving the system towards the overall optimal operating state and achieving a high degree of synergy between the physical and economic layers.
[0080] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.
[0081] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.
[0082] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.
[0083] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages such as Smalltalk, C++, etc., and conventional procedural programming languages such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.
[0084] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.
Claims
1. A hierarchical optimization scheduling method for virtual power plants considering multiple distributed resources, characterized in that, include: We construct multiple distributed resource models within a virtual power plant and, based on these models, build an upper-level optimization scheduling model with the goal of maximizing the revenue of the virtual power plant; we also construct a lower-level optimization scheduling model that considers the cooperative game between virtual power plants. Augmented Lagrangian functions are constructed for the upper-level and lower-level optimal scheduling models, respectively. The augmented Lagrangian function is solved using the alternating direction multiplier method to obtain a hierarchical optimization scheduling strategy for the virtual power plant.
2. The hierarchical optimization scheduling method for virtual power plants considering multiple distributed resources according to claim 1, characterized in that: The various distributed resources include electric vehicles, energy storage systems, gas turbines, central air conditioning, and flexible loads.
3. The hierarchical optimization scheduling method for virtual power plants considering multiple distributed resources according to claim 1, characterized in that: The upper-level optimization scheduling model, which aims to maximize the revenue of the virtual power plant, is constructed as follows: （1） In the formula, Indicates the first i A virtual power plant t Power consumed by users during different time periods The profits generated; Indicates the first i The operating costs of electric vehicles within a virtual power plant; Indicates the first i The cost of power transmission loss when sharing power among virtual power plants; Indicates the first i The operating cost of energy storage within a virtual power plant; Indicates the first i The cost of a central air conditioning system within a virtual power plant; Indicates the first i The cost of transferring transferable loads within a virtual power plant; Indicates the first i The cost of cutting off loads within a virtual power plant; Indicates the first i The operating cost of the gas turbine in a virtual power plant. This represents the number of time periods.
4. The hierarchical optimization scheduling method for virtual power plants considering multiple distributed resources according to claim 3, characterized in that: The Lagrange augmented function for constructing the upper-level optimization scheduling model is as follows: （46） In the formula, Let be the Lagrange multiplier for energy sharing between the i-th and j-th virtual power plants during time period t; For penalty parameters; Let be the variable for the interaction between virtual power plant i and virtual power plant j. Let be the variable for the interaction between virtual power plant j and virtual power plant i. For the number of time periods, This represents the number of virtual power plants.
5. The hierarchical optimization scheduling method for virtual power plants considering multiple distributed resources according to claim 1, characterized in that: The construction of the lower-level optimization scheduling model considering the cooperative game among virtual power plants is as follows: （47） In the formula, For the first i The cost of a virtual power plant participating in energy sharing; For the first i The cost of a virtual power plant participating in power sharing; For the first i The cost of power transmission losses when sharing power among virtual power plants. For the number of time periods, This represents the number of virtual power plants.
6. The hierarchical optimization scheduling method for virtual power plants considering multiple distributed resources according to claim 5, characterized in that: Constructing the augmented Lagrangian function for the lower-level optimal scheduling model includes: Taking the logarithm of the lower-level optimal scheduling model transforms the problem of finding the maximum value of the lower-level optimal scheduling model into the problem of finding the minimum value, resulting in the transformed formula: （48） Based on the transformed formula, the augmented Lagrangian function of the lower-level optimization scheduling model is established as follows: （49） In the formula, Indicates the first i The bargaining factor of a virtual power plant The Lagrange multiplier representing the backup shared between virtual power plants i and j. Indicates the penalty factor. It is an L2 norm; Let be the variable for the interaction between virtual power plant i and virtual power plant j. Let be the variable that allows virtual power plant j to interact with virtual power plant i.
7. A hierarchical optimization scheduling method for virtual power plants considering multiple distributed resources as described in claim 6, characterized in that: The formula for calculating the bargaining factor is as follows: （50） In the formula, Indicates the basic bargaining weight. To prevent positive numbers with a denominator of zero, Cost sensitivity coefficient For time period t Reduced load power For time period t Power transferred from internal load.
8. The hierarchical optimization scheduling method for virtual power plants considering multiple distributed resources according to claim 1, characterized in that: The augmented Lagrangian function of the upper-level optimal scheduling model is solved using the alternating direction multiplier method, specifically including: 1) Iteratively solve the augmented Lagrangian function of the upper-level optimization scheduling model and update the variables. , of which k The iteration process is as follows: （51） in, For the first k The augmented Lagrangian function of the upper-level optimization scheduling model in the next iteration; For the first k Lagrange multipliers in the next iteration; For the first k The variables that virtual power plant i interacts with virtual power plant j in each iteration. For the first k The variables that virtual power plant j interacts with virtual power plant i during the next iteration. 2) In the k After the solution is obtained in the next iteration, the Lagrange multipliers are updated: （52） 3) Number of update iterations k=k+ 1; 4) Determine the convergence status of the algorithm: （53） in, , These are the upper tolerance limits for the original residual and the dual residual, respectively; For penalty parameters; When the above formula is satisfied, the iteration is complete; otherwise, the iteration continues until the number of iterations exceeds a preset value.
9. The hierarchical optimization scheduling method for virtual power plants considering multiple distributed resources according to claim 1, characterized in that: The augmented Lagrangian function of the lower-level optimal scheduling model is solved using the alternating direction multiplier method, specifically including: 1) Iteratively solve the augmented Lagrangian function of the lower-level optimization scheduling model and update the variables. , of which k The iteration process is as follows: （54） in, For the first k The augmented Lagrangian function of the lower-level optimization scheduling model in the next iteration; For the first k Lagrange multipliers in the next iteration; For the first k The variables that virtual power plant i interacts with virtual power plant j in each iteration. For the first k The variables that virtual power plant j interacts with virtual power plant i during the next iteration; 2) In the k After the solution is obtained in the next iteration, the Lagrange multipliers are updated: （55） 3) Number of update iterations k=k +1; 4) Determine the convergence status of the algorithm: （56） （57） When the above formula is satisfied, the iteration is complete; otherwise, the iteration continues until the number of iterations exceeds a preset value.
10. A hierarchical optimization scheduling system for virtual power plants considering multiple distributed resources, operating the method described in any one of claims 1-9, characterized in that, The system includes: The upper-level model construction module is used to construct various distributed resource models within the virtual power plant and, based on these models, construct an upper-level optimization scheduling model with the goal of maximizing the revenue of the virtual power plant. The lower-level model building module is used to build a lower-level optimization scheduling model that considers the cooperative game between virtual power plants; The model solving module is used to construct the augmented Lagrangian functions of the upper-level and lower-level optimal scheduling models, respectively; the augmented Lagrangian functions are solved using the alternating direction multiplier method to obtain the hierarchical optimal scheduling strategy of the virtual power plant.
11. A terminal, comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the method according to any one of claims 1-9.
12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method according to any one of claims 1-9.