Virtual power plant instruction decomposition method, system and device and computer readable medium

Abstract

The invention discloses a virtual power plant instruction decomposition method, system and device and a computer readable medium, and belongs to the technical field of power system regulation and control, and the method comprises the steps: collecting adjustable resource information of a virtual power plant; constructing a virtual power plant user excitation-adjustable margin model based on multi-modal artificial intelligence; constructing a virtual power plant instruction dynamic decomposition model fusing the adjustable margin and the economical efficiency, and issuing different control instructions to each user according to actual requirements on the adjustable margin and the scheduling cost based on the adjustable margin and the control cost of each user; the invention discloses a virtual power plant internal control instruction issuing interaction method based on credit points. According to the virtual power plant instruction decomposition method, system and device and the computer readable medium, the user participation enthusiasm and the virtual power plant resource regulation and control efficiency are improved, the execution precision and economy of instruction decomposition are improved, an efficient and transparent internal instruction interaction mechanism is established, internal friction is reduced, and the overall response speed and stability of the system are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power system control, and specifically relates to a virtual power plant instruction decomposition method, system, device and computer-readable medium. Background Art

[0002] Virtual Power Plant (VPP) contains a variety of distributed resources (such as wind power, photovoltaics, energy storage, electric vehicles, etc.), each with different characteristics and response capabilities. Decomposing global instructions into specific tasks adapted to each type of resource helps ensure that the instructions are accurately executed; after the instructions are decomposed, each distributed resource can directly execute the adapted instructions without secondary parsing, reducing response delays and improving control accuracy. After decomposing the instructions, control tasks can be flexibly allocated according to the operating status, performance constraints, remaining capacity and other conditions of different resources to maximize resource utilization; after the instructions are decomposed, the tasks of each distributed resource can be collaboratively optimized according to the overall goals (such as economy, carbon emission reduction, stability) to improve the overall efficiency of the system; avoid concentrating load or power generation control tasks on a certain resource, disperse pressure, and improve the robustness of the system; virtual power plants can comprehensively consider operating costs, carbon emissions, energy efficiency and other goals during execution to achieve multi-objective optimization. However, the existing methods have the following shortcomings: 1) Insufficient user incentive mechanism, lack of accurate user portraits based on multi-dimensional user behavior data (such as electricity consumption habits, price sensitivity, etc.), unable to effectively motivate users to participate in virtual power plant regulation; existing user incentive means are single, such as only based on electricity price fluctuations, and other user incentive factors (such as environmental awareness, policy drive) are not fully explored. 2) The adjustable margin assessment is not accurate. The current methods are mostly based on static evaluation models, and fail to track the margin changes of adjustable loads in real time and dynamically; they do not fully integrate multimodal information on the user side (such as equipment type, load attributes, and environmental impact) to improve the assessment accuracy. 3) Limitations of the virtual power plant command decomposition model. Command decomposition is often based on a single goal (such as economy or regulation demand), ignoring the balance between adjustable margin and economy; lack of decomposition mechanism considering multi-objective optimization and dynamic response, resulting in the possibility that the regulation results deviate from actual demand. 4) The internal control command interaction efficiency is low, the command issuance lacks flexibility, and is mostly centralized, with slow response speed and unable to cope with sudden demand; no user credit evaluation mechanism has been established, making it difficult to effectively control the behavior and risks of multiple participants within the virtual power plant. Therefore, it is necessary to develop a new virtual power plant instruction decomposition method, system, device and computer-readable medium to solve the existing problems. Summary of the invention

[0003] The object of the present invention is to provide a virtual power plant instruction decomposition method, system, device and computer-readable medium to solve the existing problems.

[0004] To achieve the above object, the present invention provides the following technical solution: a virtual power plant instruction decomposition method, comprising:

[0005] Collect information on adjustable resources of virtual power plants;

[0006] Construct a virtual power plant user incentive-adjustable margin model based on multimodal artificial intelligence, train the historical adjustable margin information, sustainable control time information, and incentive value information of each user of the virtual power plant, obtain the virtual power plant user incentive function based on multimodal artificial intelligence, input the user's adjustable margin, sustainable time, and incentive value into the virtual power plant user incentive function based on multimodal artificial intelligence, and obtain the incentive value required for the user to participate in the virtual power plant scheduling;

[0007] Build a dynamic decomposition model of virtual power plant instructions that integrates adjustable margin and economy. Based on the adjustable margin and control cost of each user, different control instructions are issued to each user according to the actual demand for adjustable margin and dispatch cost.

[0008] Build a credit point model for virtual power plant users, and calculate the credit points given and deducted from users based on whether the users participate in scheduling; build an electricity redemption model based on credit points, and users use credit points to redeem free electricity; build a credit point recharge model, and users choose to recharge to increase their credit points.

[0009] Preferably, the collected virtual power plant adjustable resource information includes: adjustable margin information of each user, sustainable adjustment time information, virtual total incentive cost information, adjustable margin threshold of each user of the virtual power plant, sustainable adjustment time threshold of the virtual power plant, and dispatch capacity information of the virtual power plant received from the upper level; as shown in formula (1):

[0010]

[0011] Where: X0(t) is the virtual power plant dispatch information collected at time t, ΔP i (t) is the adjustable margin information of user i collected at time t, ΔT i (t) is the sustainable adjustment time information of user i collected at time t, is the virtual total incentive cost information collected at time t, ΔP min is the adjustable margin threshold of each user in the virtual power plant, ΔT min is the sustainable adjustment time threshold of the virtual power plant, ΔP G (t) is the dispatch capacity information of the virtual power plant received from the superior at time t, ΔT G (t) is the sustainable scheduling time of the virtual power plant collected at time t and issued by the superior.

[0012] Preferably, the construction of a virtual power plant user incentive-adjustable margin model based on multimodal artificial intelligence includes:

[0013] The information of different modes within the virtual power plant is combined to form a unified feature vector, as shown in formula (2), which represents all the input data related to the incentive cost of each user of the virtual power plant at the current time t;

[0014]

[0015] Where: X i '(t) is the multi-modal feature vector of virtual power plant user i after fusion, ΔP i '(t) is the adjustable margin information vector of virtual power plant user i after One-Hot Encoding; ΔT i '(t) is the sustainable adjustment time information vector of virtual power plant user i after One-Hot Encoding; is the standardized incentive information of user i;

[0016] The multimodal feature vector X of virtual power plant user i after fusion i '(t) is input into the multi-layer perceptron and the model is trained through the fully connected layer; the user's incentive mean square error (MSE) is minimized, as shown in formula (3):

[0017]

[0018] Where: δ(t) is the mean square error of the excitation value of the multimodal artificial intelligence algorithm, N is the number of samples, is the historical incentive value of user i;

[0019] After training, the user incentive function of the virtual power plant based on multimodal artificial intelligence is obtained, and historical information is input The incentive value of virtual power plant user i at the next moment can be predicted As shown in formula (4),

[0020]

[0021] Where: is the incentive value of user i at time t+1 given by the multimodal artificial intelligence model, f DMT () is the multimodal artificial intelligence virtual power plant user incentive-adjustable margin function, are the historical adjustable margin, sustainable control time and incentive value of user i respectively.

[0022] Preferably, the virtual power plant instruction dynamic decomposition model integrating adjustable margin and economy includes:

[0023] The objective function of the virtual power plant command dynamic decomposition model that integrates adjustable margin and economy is to maximize the adjustable margin of the virtual power plant and minimize the dispatch cost. The objective function of the model is dynamically adjusted through the margin coefficient and economic coefficient, as shown in formula (5):

[0024]

[0025] Among them, k P is the margin coefficient, is the economic coefficient, ΔP X is the adjustable margin of the virtual power plant, is the dispatch cost of the virtual power plant;

[0026] The calculation is performed based on the adjustable margin of each user of the virtual power plant. The adjustable margin of each user is the control instruction decomposed to each user. Formula (6) is the calculation model of the total dispatch margin of the virtual power plant:

[0027]

[0028] Where: ΔP i X (t) is the adjustable margin of user i, and m is the number of users responding to dispatch within the virtual power plant;

[0029] It is calculated based on the actual incentive cost of each user within the virtual power plant. The incentive cost is a nonlinear function. Formula (7) is the incentive cost calculation model of the virtual power plant.

[0030]

[0031] in, is the actual incentive cost of user i;

[0032] Formula (8) is the adjustable margin constraint of the virtual power plant, that is, the adjustable margin of the total response of the virtual power plant must be greater than the margin sent by the upper level received by the virtual power plant;

[0033] ΔP X ≥ΔP G (t) Formula (8)

[0034] ΔP G (t) is the margin sent by the superior to the virtual power plant;

[0035] Formula (9) is the virtual power plant incentive cost constraint, that is, the total cost of the virtual power plant to regulate users shall not exceed the total incentive value issued by the virtual power plant.

[0036]

[0037] is the total incentive value issued by the virtual power plant;

[0038] Formula (10) is the adjustable margin constraint of virtual power plant users, that is, each user's adjustable margin must be greater than the margin threshold before they can participate in the virtual power plant dispatch;

[0039] ΔP i X (t)≥ΔP min Formula (10)

[0040] Among them, ΔP min The margin threshold for users to participate in virtual power plant dispatch;

[0041] Formula (11) is the duration constraint of virtual power plant users, that is, each user’s duration must be greater than the time threshold before they can participate in virtual power plant scheduling.

[0042] ΔT i (t)≥ΔT min Formula (11)

[0043] Where: ΔT min The duration threshold for users to participate in virtual power plant scheduling.

[0044] Preferably, the credit score model of the virtual power plant user includes:

[0045]

[0046] Among them, F i is the credit score of virtual power plant user i, ΔP i X (t) is the adjustable margin of user i, ΔT min is the duration threshold for users to participate in virtual power plant scheduling, λ is the participation signal fed back by users to the virtual power plant, a value of 1 indicates participation in the virtual power plant control, and a value of 0 indicates non-participation in the virtual power plant control, α F is the credit integral coefficient of the user-adjustable margin, β F is the credit score coefficient for the duration of the user participating in the control, η F is the credit integral coefficient of the user margin that does not participate in the control, μ F To not control the credit points coefficient of the user's duration;

[0047] If the internal user of the virtual power plant participates in the regulation of the virtual power plant after receiving the instruction issued by the virtual power plant, then λ=1, and the user's points are calculated according to the adjustable margin instruction and sustainable regulation time received by the user;

[0048] If the internal users of the virtual power plant do not participate in the regulation of the virtual power plant after receiving the instructions issued by the virtual power plant, then λ=0, and the credit points that need to be deducted from the user are calculated according to the issued adjustable margin instructions and the sustainable regulation time.

[0049] Preferably, the construction of a credit points-based electricity exchange model includes: as shown in formula (13):

[0050] Q dian =F i ·ω d Formula (13)

[0051] Among them, Q dian The amount of electricity that can be redeemed for free by users of virtual power plants, d is the free electricity exchange coefficient. Preferably, the construction of the credit points recharge model includes: as shown in formula (14),

[0052] Q qian =F i ·ω q Formula (14)

[0053] Where: Q qian Recharge the credit factor amount for the user, ω q Recharge the user's credit factor.

[0054] The present invention further provides a virtual power plant instruction decomposition system, the system comprising:

[0055] A collection module is used to collect the adjustable resource information of the virtual power plant;

[0056] A virtual power plant user incentive-adjustable margin model construction module based on multimodal artificial intelligence, which is used to construct a virtual power plant user incentive-adjustable margin model based on multimodal artificial intelligence;

[0057] A module for constructing a dynamic decomposition model of virtual power plant instructions integrating adjustable margin and economy, which is used to construct a dynamic decomposition model of virtual power plant instructions integrating adjustable margin and economy;

[0058] A credit score model building module for virtual power plant users, used to build a credit score model for virtual power plant users;

[0059] A module for constructing a model for electricity exchange based on credit points, which is used to construct an electricity exchange model based on credit points;

[0060] The credit points recharge model construction module is used to construct a credit points recharge model.

[0061] The present invention further provides a virtual power plant instruction decomposition device, comprising:

[0062] a memory for storing non-transitory computer-readable instructions; and

[0063] A processor is used to run the computer-readable instructions so that when the computer-readable instructions are executed by the processor, any one of the virtual power plant instruction decomposition methods is implemented.

[0064] The present invention further provides a computer-readable storage medium for storing non-temporary computer-readable instructions. When the non-temporary computer-readable instructions are executed by a computer, the computer executes any one of the virtual power plant instruction decomposition methods.

[0065] The technical effects and advantages of the present invention are as follows: The virtual power plant instruction decomposition method, system, device and computer-readable medium can improve user participation enthusiasm and virtual power plant resource regulation efficiency, enhance the execution accuracy and economy of instruction decomposition, establish an efficient and transparent internal instruction interaction mechanism, reduce internal friction, and improve the overall response speed and stability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0067] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0068] The present invention provides Figure 1A virtual power plant instruction decomposition method is shown in the figure. First, a virtual power plant adjustable resource information collection technology is proposed to collect information from each user. Then, a virtual power plant user incentive-adjustable margin model based on multimodal artificial intelligence is constructed. The incentive value of each user in the virtual power plant is trained based on the multimodal artificial intelligence algorithm. The virtual power plant user incentive-adjustable margin model is output to accurately give the incentive value required for the control user to participate in the virtual power plant scheduling. Secondly, a virtual power plant instruction dynamic decomposition model integrating adjustable margin and economy is proposed. The adjustable margin and control cost of each user are comprehensively considered. Different control instructions are issued to each user according to the actual demand for adjustable margin and scheduling cost, so as to realize dynamic control of each internal user. Finally, a virtual power plant internal control instruction issuance interactive method based on credit points is proposed, and a credit point model for virtual power plant users is constructed. According to whether the user participates in the scheduling, the credit points given to the user and the credit points deducted are calculated respectively. A power exchange model based on credit points is constructed, and the user can use the credit points to exchange for a certain amount of free power. A credit point recharge model is constructed, and the user can choose to recharge to increase his own credit points.

[0069] The specific steps include:

[0070] Step (1) proposes a virtual power plant adjustable resource information collection technology to provide basic information for decomposing virtual power plant dispatch instructions. The collected information includes adjustable margin information of each user, sustainable adjustment time information, virtual total incentive cost information, adjustable margin threshold of each user of the virtual power plant, sustainable adjustment time threshold of the virtual power plant, and dispatch capacity information issued by the virtual power plant.

[0071] Step (2) constructs a virtual power plant user incentive-adjustable margin model based on multimodal artificial intelligence. Multimodal artificial intelligence algorithms can train data of different modes to improve the effectiveness and accuracy of the model. The present invention proposes to use a multimodal artificial intelligence algorithm to train the incentive values ​​of each user in the virtual power plant. The algorithm first inputs the historical adjustable margin information, sustainable control time information, and incentive value information of each user of the virtual power plant, trains the above information based on the multimodal artificial intelligence algorithm, and outputs the virtual power plant user incentive-adjustable margin model. According to the user's adjustable margin and sustainable time, the incentive value required to control the user's participation in the virtual power plant scheduling is accurately given.

[0072] Step (3) proposes a dynamic decomposition model of virtual power plant instructions that integrates adjustable margin and economy, comprehensively considers the adjustable margin and control cost of each user, and issues different control instructions to each user according to the actual demand for adjustable margin and dispatch cost, so as to realize dynamic control of each internal user. The model takes the maximum adjustable margin of the virtual power plant and the minimum dispatch cost as the objective function, and can dynamically adjust the objective function of the model through the margin coefficient and the economic coefficient to meet the actual dispatch demand. Among them, the constructed virtual power plant total dispatch margin calculation model is mainly calculated based on the adjustable margin of each user of the virtual power plant, and the adjustable margin of each user is the control instruction decomposed to each user. The constructed virtual power plant incentive cost calculation model is mainly calculated based on the actual incentive cost of each user inside the virtual power plant. The incentive cost is a nonlinear function. The present invention mainly solves the function according to the method of step (2). The constraint conditions are as follows: virtual power plant adjustable margin constraint, that is, the adjustable margin of the total response of the virtual power plant must be greater than the margin issued by the upper level received by the virtual power plant. Virtual power plant incentive cost constraint, that is, the total cost of virtual power plant regulation users shall not exceed the total incentive value issued by the superior user. Virtual power plant user adjustable margin constraint, that is, each user's adjustable margin must be greater than a certain margin threshold before they can participate in virtual power plant scheduling. Virtual power plant user duration constraint, that is, each user's duration must be greater than a certain time threshold before they can participate in virtual power plant scheduling.

[0073] Step (4) proposes an interactive method for issuing internal control instructions of a virtual power plant based on credit points. Users promptly provide feedback on whether to participate in the regulation. If they participate, they provide feedback of 1, and if they do not, they provide feedback of 0. A credit point model for users of virtual power plants is constructed. If users within the virtual power plant participate in the regulation of the virtual power plant after receiving the instructions issued by the virtual power plant, the user's points are calculated according to the adjustable margin instructions and sustainable regulation time received by the user. If users within the virtual power plant do not participate in the regulation of the virtual power plant after receiving the instructions issued by the virtual power plant, the credit points that need to be deducted from the user are calculated according to the issued adjustable margin instructions and sustainable regulation time. A power exchange model based on credit points is constructed. When a user has a large number of credit points, the user can use the credit points to exchange for a certain amount of free power. A credit point recharge model is constructed. When a user has a small number of credit points, the user can choose to recharge to increase his or her credit points.

[0074] For step (1), the collected virtual power plant adjustable resource information is as shown in formula (1), which mainly includes the adjustable margin information of each user, the sustainable adjustment time information, the virtual total incentive cost information, the adjustable margin threshold of each user of the virtual power plant, the sustainable adjustment time threshold of the virtual power plant, and the dispatch capacity information of the virtual power plant received from the superior.

[0075]

[0076] Where: X0(t) is the virtual power plant dispatch information collected at time t. ΔP i (t) is the adjustable margin information of user i collected at time t, ΔT i (t) is the sustainable adjustment time information of user i collected at time t, is the virtual total incentive cost information collected at time t, ΔP min is the adjustable margin threshold of each user in the virtual power plant, ΔT min is the sustainable adjustment time threshold of the virtual power plant, ΔP G (t) is the dispatch capacity information of the virtual power plant received from the superior at time t. G (t) is the sustainable scheduling time of the virtual power plant collected at time t and issued by the superior.

[0077] For step (2), the virtual power plant user incentive-adjustable margin model based on multimodal artificial intelligence is constructed as shown in formulas (2)-(4). The present invention first uses a feature fusion method to combine information of different modes within the virtual power plant to form a unified feature vector, as shown in formula (2), which represents all input data related to the incentive cost of each user of the virtual power plant at the current time t.

[0078]

[0079] Where: X i '(t) is the multi-modal feature vector of virtual power plant user i after fusion, ΔP i '(t) is the adjustable margin information vector of virtual power plant user i after One-Hot Encoding; ΔT i '(t) is the sustainable adjustment time information vector of virtual power plant user i after One-Hot Encoding; is the standardized incentive information of user i.

[0080] Then, the fused multimodal feature vector X of virtual power plant user i is i '(t) is input into a multi-layer perceptron (MLP) and the model is trained through a series of fully connected layers. The goal of this function in the present invention is to minimize the error between the trained stimulus value and the true value. The present invention minimizes the user's stimulus mean square error (MSE), as shown in formula (3).

[0081]

[0082] Where: δ(t) is the mean square error of the excitation value of the multimodal artificial intelligence algorithm, N is the number of samples, is the historical incentive value of user i.

[0083] After the above training, the user incentive function of the virtual power plant based on multimodal artificial intelligence can be obtained, as shown in formula (4), with the input historical information The incentive value of virtual power plant user i at the next moment can be predicted

[0084]

[0085] Where: is the incentive value of user i at time t+1 given by the multimodal artificial intelligence model, f DMT () is the multimodal artificial intelligence virtual power plant user incentive-adjustable margin function, are the historical adjustable margin, sustainable control time and incentive value of user i respectively.

[0086] For step (3), the dynamic decomposition model of the virtual power plant command that integrates the adjustable margin and economy is shown in formulas (5)-(11).

[0087] Formula (5) is the objective function of the dynamic decomposition model of virtual power plant instructions that integrates adjustable margin and economy, with the maximum adjustable margin of the virtual power plant and the minimum dispatch cost as the objective function. The objective function of the model can be dynamically adjusted through the margin coefficient and economic coefficient to meet the actual dispatch requirements.

[0088] Formula (6) is the calculation model of the total dispatch margin of the virtual power plant, which is mainly calculated based on the adjustable margin of each user of the virtual power plant. The adjustable margin of each user is the control instruction decomposed to each user.

[0089] Formula (7) is a virtual power plant incentive cost calculation model, which is mainly calculated based on the actual incentive cost of each user in the virtual power plant. The incentive cost is a nonlinear function, and the present invention mainly solves the function based on the method of step (2).

[0090] Formula (8) is the adjustable margin constraint of the virtual power plant, that is, the adjustable margin of the total response of the virtual power plant must be greater than the margin sent by the superior to the virtual power plant.

[0091] Formula (9) is the virtual power plant incentive cost constraint, that is, the total cost of the virtual power plant to regulate users shall not exceed the total incentive value issued by the virtual power plant.

[0092] Formula (10) is the adjustable margin constraint of virtual power plant users, that is, each user’s adjustable margin must be greater than a certain margin threshold before they can participate in the virtual power plant dispatch.

[0093] Formula (11) is the duration constraint of virtual power plant users, that is, each user's duration must be greater than a certain time threshold before they can participate in virtual power plant scheduling; Formula (6)-Formula (11) are constraint conditions;

[0094]

[0095]

[0096] ΔP X ≥ΔP G (t) (8)

[0097]

[0098] ΔP i X (t)≥ΔP min (10)

[0099] ΔT i (t)≥ΔT min (11)

[0100] Where: k P is the margin coefficient, is the economic coefficient, ΔP X is the adjustable margin of the virtual power plant, is the dispatch cost of the virtual power plant. ΔP i X (t) is the adjustable margin of user i, is the actual incentive cost of user i, and m is the number of users who respond to the dispatch within the virtual power plant. ΔP G (t) is the margin received by the virtual power plant from its superior. is the total incentive value issued by the virtual power plant. ΔP min is the margin threshold for users to participate in virtual power plant dispatch, ΔT min The duration threshold for users to participate in virtual power plant scheduling.

[0101] For step (4), the proposed interactive method for issuing internal control instructions of a virtual power plant based on credit points is shown in formulas (12)-(14).

[0102] Formula (12) is the credit score model of the virtual power plant user. If the internal user of the virtual power plant participates in the regulation of the virtual power plant after receiving the instruction issued by the virtual power plant, then λ = 1, and the user's score is calculated according to the adjustable margin instruction and sustainable regulation time received by the user. If the internal user of the virtual power plant does not participate in the regulation of the virtual power plant after receiving the instruction issued by the virtual power plant, then λ = 0, and the credit score of the user to be deducted is calculated according to the issued adjustable margin instruction and sustainable regulation time.

[0103] Formula (13) is a power exchange model based on credit points. When a user has more credit points, the user can use the credit points to exchange for a certain amount of free power.

[0104] Formula (14) is a credit score recharge model. When a user's credit score is low, the user can choose to recharge to increase his or her credit score.

[0105]

[0106] Q dian =F i ·ω d (13)

[0107] Q qian =F i ·ω q (14)

[0108] Where: F i is the credit score of virtual power plant user i, ΔP i X (t) is the adjustable margin of user i, ΔT min is the duration threshold for users to participate in virtual power plant scheduling, λ is the participation signal fed back by users to the virtual power plant, a value of 1 indicates participation in the virtual power plant control, and a value of 0 indicates non-participation in the virtual power plant control. F is the credit integral coefficient of the user-adjustable margin, β F The credit points coefficient for the duration of the user's participation control. F is the credit integral coefficient of the user margin that does not participate in the control, μ F The credit factor for the duration of the non-control user. dian The amount of electricity that can be redeemed for free by users of virtual power plants, d is the free electricity exchange coefficient. qian Recharge the credit factor amount for the user, ω q Recharge the user's credit factor.

[0109] The present invention further provides a virtual power plant instruction decomposition system, the system comprising:

[0110] A collection module is used to collect the adjustable resource information of the virtual power plant;

[0111] A virtual power plant user incentive-adjustable margin model construction module based on multimodal artificial intelligence, which is used to construct a virtual power plant user incentive-adjustable margin model based on multimodal artificial intelligence;

[0112] A module for constructing a dynamic decomposition model of virtual power plant instructions integrating adjustable margin and economy, which is used to construct a dynamic decomposition model of virtual power plant instructions integrating adjustable margin and economy;

[0113] A credit score model building module for virtual power plant users, used to build a credit score model for virtual power plant users;

[0114] A module for constructing a model for electricity exchange based on credit points, which is used to construct an electricity exchange model based on credit points;

[0115] The credit points recharge model construction module is used to construct a credit points recharge model.

[0116] The present invention further provides a virtual power plant instruction decomposition device, comprising:

[0117] a memory for storing non-transitory computer-readable instructions; and

[0118] A processor is used to run the computer-readable instructions so that when the computer-readable instructions are executed by the processor, any one of the virtual power plant instruction decomposition methods is implemented.

[0119] The present invention further provides a computer-readable storage medium for storing non-temporary computer-readable instructions. When the non-temporary computer-readable instructions are executed by a computer, the computer executes any one of the virtual power plant instruction decomposition methods.

[0120] The credit point mechanism provided by the present invention can urge users to participate in the regulation of virtual power plants and improve the effectiveness of virtual power plant control.

[0121] After the above steps, effective control instructions can be issued to each user within the virtual power plant, thereby achieving effective regulation and control of each user within the virtual power plant.

[0122] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0123] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0124] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0125] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0126] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0127] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

[0128] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for decomposing instructions of a virtual power plant, characterized by: include: Collect information on adjustable resources of virtual power plants; Construct a virtual power plant user incentive-adjustable margin model based on multimodal artificial intelligence, train the historical adjustable margin information, sustainable control time information, and incentive value information of each user of the virtual power plant, obtain the virtual power plant user incentive function based on multimodal artificial intelligence, input the user's adjustable margin, sustainable time, and incentive value into the virtual power plant user incentive function based on multimodal artificial intelligence, and obtain the incentive value required for the user to participate in the virtual power plant scheduling; Build a dynamic decomposition model of virtual power plant instructions that integrates adjustable margin and economy. Based on the adjustable margin and control cost of each user, different control instructions are issued to each user according to the actual demand for adjustable margin and dispatch cost. Build a credit score model for virtual power plant users, and calculate the credit points given to users and the credit points deducted according to whether the users participate in scheduling; build an electricity exchange model based on credit points, and users use credit points to exchange for free electricity; Build a credit points recharge model, and users can choose the recharge method to increase their credit points.

2. A virtual power plant instruction decomposition method according to claim 1, characterized in that: The collected virtual power plant adjustable resource information includes: adjustable margin information of each user, sustainable adjustment time information, virtual total incentive cost information, adjustable margin threshold of each user of the virtual power plant, sustainable adjustment time threshold of the virtual power plant, and dispatch capacity information of the virtual power plant received from the upper level; as shown in formula (1): Where: X0(t) is the virtual power plant dispatch information collected at time t, ΔP i (t) is the adjustable margin information of user i collected at time t, ΔT i (t) is the sustainable adjustment time information of user i collected at time t, is the virtual total incentive cost information collected at time t, ΔP min is the adjustable margin threshold of each user in the virtual power plant, ΔT min is the sustainable adjustment time threshold of the virtual power plant, ΔP G (t) is the dispatch capacity information of the virtual power plant received from the superior at time t, ΔT G (t) is the sustainable scheduling time of the virtual power plant collected at time t and issued by the superior.

3. A virtual power plant instruction decomposition method according to claim 1, characterized in that: The construction of a virtual power plant user incentive-adjustable margin model based on multimodal artificial intelligence includes: The information of different modes within the virtual power plant is combined to form a unified feature vector, as shown in formula (2), which represents all the input data related to the incentive cost of each user of the virtual power plant at the current time t; Where: X i '(t) is the multi-modal feature vector of virtual power plant user i after fusion, ΔP i '(t) is the adjustable margin information vector of virtual power plant user i after One-Hot Encoding; ΔT i '(t) is the sustainable adjustment time information vector of virtual power plant user i after One-Hot Encoding; is the standardized incentive information of user i; The multimodal feature vector X of virtual power plant user i after fusion i '(t) is input into the multi-layer perceptron and the model is trained through the fully connected layer; the user's incentive mean square error (MSE) is minimized, as shown in formula (3): Where: δ(t) is the mean square error of the excitation value of the multimodal artificial intelligence algorithm, N is the number of samples, is the historical incentive value of user i; After training, the user incentive function of the virtual power plant based on multimodal artificial intelligence is obtained, and the historical information ΔP is input i (t),ΔT i (t), The incentive value of virtual power plant user i at the next moment can be predicted As shown in formula (4), Where: is the incentive value of user i at time t+1 given by the multimodal artificial intelligence model, f DMT () is the multimodal artificial intelligence virtual power plant user incentive-adjustable margin function, ΔP i (t),ΔT i (t), are the historical adjustable margin, sustainable control time and incentive value of user i respectively.

4. A virtual power plant instruction decomposition method according to claim 1, characterized in that: The virtual power plant command dynamic decomposition model integrating adjustable margin and economy includes: The objective function of the virtual power plant command dynamic decomposition model that integrates adjustable margin and economy is to maximize the adjustable margin of the virtual power plant and minimize the dispatch cost. The objective function of the model is dynamically adjusted through the margin coefficient and economic coefficient, as shown in formula (5): Among them, k P is the margin coefficient, is the economic coefficient, ΔP X is the adjustable margin of the virtual power plant, is the dispatch cost of the virtual power plant; The calculation is performed based on the adjustable margin of each user of the virtual power plant. The adjustable margin of each user is the control instruction decomposed to each user. Formula (6) is the calculation model of the total dispatch margin of the virtual power plant: Where: ΔP i X (t) is the adjustable margin of user i, and m is the number of users responding to dispatch within the virtual power plant; It is calculated based on the actual incentive cost of each user within the virtual power plant. The incentive cost is a nonlinear function. Formula (7) is the incentive cost calculation model of the virtual power plant. in, is the actual incentive cost of user i; Formula (8) is the adjustable margin constraint of the virtual power plant, that is, the adjustable margin of the total response of the virtual power plant must be greater than the margin sent by the upper level received by the virtual power plant; ΔP X ≥ΔP G (t) Formula (8) ΔP G (t) is the margin sent by the superior to the virtual power plant; Formula (9) is the virtual power plant incentive cost constraint, that is, the total cost of the virtual power plant to regulate users shall not exceed the total incentive value issued by the virtual power plant. is the total incentive value issued by the virtual power plant; Formula (10) is the adjustable margin constraint of virtual power plant users, that is, each user's adjustable margin must be greater than the margin threshold before they can participate in the virtual power plant dispatch; ΔP i X (t)≥ΔP min Formula (10) Among them, ΔP min The margin threshold for users to participate in virtual power plant dispatch; Formula (11) is the duration constraint of virtual power plant users, that is, each user’s duration must be greater than the time threshold before they can participate in virtual power plant scheduling. ΔT i (t)≥ΔT min Formula (11) Where: ΔT min The duration threshold for users to participate in virtual power plant scheduling.

5. A virtual power plant instruction decomposition method according to claim 1, characterized in that: The credit score model of the virtual power plant user includes: Among them, F i is the credit score of virtual power plant user i, ΔP i X (t) is the adjustable margin of user i, ΔT min is the duration threshold for users to participate in virtual power plant scheduling, λ is the participation signal fed back by users to the virtual power plant, a value of 1 indicates participation in the virtual power plant control, and a value of 0 indicates non-participation in the virtual power plant control, α F is the credit integral coefficient of the user-adjustable margin, β F is the credit score coefficient for the duration of the user participating in the control, η F is the credit integral coefficient of the user margin that does not participate in the control, μ F To not control the credit points coefficient of the user's duration; If the internal user of the virtual power plant participates in the regulation of the virtual power plant after receiving the instruction issued by the virtual power plant, then λ=1, and the user's points are calculated according to the adjustable margin instruction and sustainable regulation time received by the user; If the internal users of the virtual power plant do not participate in the regulation of the virtual power plant after receiving the instructions issued by the virtual power plant, then λ=0, and the credit points that need to be deducted from the user are calculated according to the issued adjustable margin instructions and the sustainable regulation time.

6. A virtual power plant instruction decomposition method according to claim 1, characterized in that: The construction of the electricity exchange model based on credit points includes: as shown in formula (13): Q dian =F i ·ω d Formula (13) Among them, Q dian The amount of electricity that can be redeemed for free by users of virtual power plants, d It is the free electricity redemption coefficient.

7. A virtual power plant instruction decomposition method according to claim 1, characterized in that: The construction of the credit score recharge model includes: as shown in formula (14), Q qian =F i ·ω q Formula (14) Where: Q qian Recharge the credit factor amount for the user, ω q Recharge the user's credit factor.

8. A virtual power plant instruction decomposition system, the system comprising: A collection module is used to collect the adjustable resource information of the virtual power plant; A virtual power plant user incentive-adjustable margin model construction module based on multimodal artificial intelligence, which is used to construct a virtual power plant user incentive-adjustable margin model based on multimodal artificial intelligence; A module for constructing a dynamic decomposition model of virtual power plant instructions integrating adjustable margin and economy, which is used to construct a dynamic decomposition model of virtual power plant instructions integrating adjustable margin and economy; A credit score model building module for virtual power plant users, used to build a credit score model for virtual power plant users; A module for constructing a model for electricity exchange based on credit points, which is used to construct an electricity exchange model based on credit points; The credit points recharge model construction module is used to construct a credit points recharge model.

9. A virtual power plant instruction decomposition device, comprising: a memory for storing non-transitory computer-readable instructions; as well as A processor is used to run the computer-readable instructions so that when the computer-readable instructions are executed by the processor, the virtual power plant instruction decomposition method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium for storing non-temporary computer-readable instructions, which, when executed by a computer, enables the computer to execute the virtual power plant instruction decomposition method described in any one of claims 1 to 7.

Images