An artificial intelligence-based virtual power plant collaborative control and scheduling method and system

Abstract

The application discloses an artificial intelligence-based virtual power plant cooperative control and scheduling method and system, relates to the technical field of power dispatching, predicts the power consumption of a power supply area through an artificial intelligence model, determines the predicted power consumption corresponding to the power supply area, provides a reliable basis for subsequent scheduling, encodes the output of each distributed energy in the virtual power plant, takes meeting the predicted power consumption as an optimization target, can convert a complex scheduling problem into an optimization problem, reduces the generation difficulty of a scheduling scheme, improves the response capability and control precision of the entire virtual power plant to grid instructions and market signals, and finally schedules based on the target optimal output parameter coding obtained through optimization, thereby realizing efficient virtual power plant cooperative control and scheduling.

Description

Technical Field

[0001] This application relates to the field of power dispatching technology, and more specifically, to a method and system for collaborative control and dispatching of virtual power plants based on artificial intelligence. Background Technology

[0002] With the large-scale integration of renewable energy sources (such as wind power and photovoltaics) and the widespread adoption of distributed energy sources (such as energy storage, electric vehicles, and controllable loads), the operating characteristics of power systems have undergone profound changes. Virtual Power Plants (VPPs), as an innovative model for aggregating, coordinating, and managing distributed energy resources, can integrate these dispersed resources as a whole to participate in electricity market transactions or provide ancillary services, improving energy efficiency and enhancing grid stability and flexibility. However, VPPs contain various heterogeneous resources with significant differences in operating characteristics, response speed, and controllability. Traditional VPP dispatching methods are often based on fixed optimization models or simple rules, making it difficult to achieve optimal coordinated control and economic dispatch under complex and ever-changing grid environments and market rules. Summary of the Invention

[0003] This application aims to provide a virtual power plant collaborative control and scheduling method and system based on artificial intelligence, which is intended to solve the problem that existing technologies based on fixed optimization models or simple rules are difficult to achieve optimal collaborative control and economic scheduling under complex and ever-changing power grid environments and market rules.

[0004] The first aspect of this application provides a virtual power plant collaborative control and scheduling method based on artificial intelligence, comprising:

[0005] For the energy supply area corresponding to the virtual power plant, the power consumption of the power supply area is predicted by an artificial intelligence model to determine the predicted power consumption of the energy supply area.

[0006] Based on the output of each distributed energy source in the virtual power plant, parameter encoding is performed to determine multiple different output parameter codes.

[0007] With the goal of meeting the predicted electricity consumption of the energy supply area, the output parameter coding is optimized to obtain the target optimal output parameter coding;

[0008] The optimal output parameters of the target are decoded to obtain the output scheduling schemes for each distributed energy source in the virtual power plant, and the virtual power plant is coordinated and controlled according to the output scheduling schemes.

[0009] In one possible implementation, for the energy supply area corresponding to the virtual power plant, the electricity consumption of the power supply area is predicted using an artificial intelligence model to determine the predicted electricity consumption corresponding to the energy supply area, including:

[0010] Collect multi-dimensional electricity consumption impact data corresponding to the energy supply area of ​​the virtual power plant;

[0011] Retrieve the training labels of the artificial intelligence model; wherein the training labels are either trained labels or untrained labels;

[0012] If the training label of the artificial intelligence model is a trained label, then the artificial intelligence model is used to analyze and process the multi-dimensional electricity consumption impact data to determine the predicted electricity consumption corresponding to the energy supply area.

[0013] If the training label of the artificial intelligence model is an untrained label, the artificial intelligence model is trained to obtain a trained artificial intelligence model. The trained artificial intelligence model is then used to analyze and process the multi-dimensional electricity consumption impact data to determine the predicted electricity consumption corresponding to the energy supply area. The training label is then set as a trained label for subsequent analysis.

[0014] In one possible implementation, the collection of multi-dimensional electricity consumption impact data corresponding to the energy supply area of ​​the virtual power plant includes:

[0015] Collect seasonal data, weather data, temperature data, date type data, and historical electricity consumption data of the previous N days corresponding to the energy supply area of ​​the virtual power plant to obtain the multi-dimensional electricity consumption impact data;

[0016] The seasonal data includes spring, summer, autumn, or winter codes; the weather data includes rain, sunny, cloudy, or snowy codes; and the date type data includes weekday or non-working day codes.

[0017] In one possible implementation, before using an artificial intelligence model to analyze and process the multi-dimensional electricity consumption impact data, the method further includes: normalizing the multi-dimensional electricity consumption impact data.

[0018] In one possible implementation, training the artificial intelligence model to obtain the trained artificial intelligence model includes:

[0019] The artificial intelligence model is trained using a particle swarm optimization algorithm, a genetic algorithm, or a combination of particle swarm optimization and genetic algorithms to obtain the trained artificial intelligence model.

[0020] In one possible implementation, the step of parameter encoding based on the output of each distributed energy source in the virtual power plant to determine multiple different output parameter codes includes:

[0021] For any distributed energy source in the virtual power plant, based on the lower limit of the output and the upper limit of the processing capacity of the distributed energy source, the output of the distributed energy source is randomly initialized to obtain the output parameters of the distributed energy source.

[0022] The output parameters of all distributed energy sources in the virtual power plant are encoded into vectors to obtain the output parameter encoding;

[0023] Multiple initialization and encoding operations are performed to obtain multiple different output parameter codes.

[0024] In one possible implementation, with the goal of satisfying the predicted electricity consumption corresponding to the energy supply area, the output parameter encoding is optimized to obtain the target optimal output parameter encoding, including:

[0025] To meet the predicted electricity consumption of the energy supply area, a target optimization function is constructed.

[0026] For any output parameter code, determine the target optimization function value corresponding to the output parameter code, and determine the current optimal output parameter code based on the target optimization function value; wherein, the smaller the target optimization function, the better;

[0027] Based on the current optimal output parameter code, the output parameter code is globally guided to obtain the output parameter code after global guidance;

[0028] For the output parameter encoding after global guidance, local guidance is performed on the output parameter encoding to determine the output parameter encoding after local guidance;

[0029] For the output parameter encoding after local guidance, a global mutation search is performed on the output parameter encoding to determine the output parameter encoding after the global mutation search;

[0030] Determine if the training termination condition is met. If so, determine the target optimal output parameter code based on the output parameter code after global mutation search. Otherwise, return to the step of determining the current optimal output parameter code.

[0031] In one possible implementation, the step of constructing a target optimization function with the objective of satisfying the predicted electricity consumption corresponding to the energy supply area includes:

[0032] Construct the loss term between the output parameter code and the predicted electricity consumption, as well as the cost function corresponding to the output parameter code;

[0033] The target optimization function is obtained by weighted summation of the loss term and the cost function.

[0034] In one possible implementation, determining whether the training termination condition is met includes:

[0035] Determine whether the current optimization is greater than or equal to the preset maximum number of optimizations. If so, the training termination condition is met; otherwise, the training termination condition is not met.

[0036] A second aspect of this application provides a virtual power plant collaborative control and scheduling system based on artificial intelligence, comprising:

[0037] The electricity consumption prediction module is used to predict the electricity consumption of the power supply area corresponding to the virtual power plant through an artificial intelligence model, and determine the predicted electricity consumption of the power supply area.

[0038] The scheduling coding module is used to perform parameter coding based on the output of each distributed energy source in the virtual power plant, and to determine multiple different output parameter codes.

[0039] The coding optimization module is used to optimize the output parameter coding to meet the predicted electricity consumption of the energy supply area and obtain the target optimal output parameter coding.

[0040] The collaborative control and scheduling module is used to decode the target optimal output parameter encoding to obtain the output scheduling scheme corresponding to each distributed energy source in the virtual power plant, and to perform collaborative control and scheduling of the virtual power plant according to the output scheduling scheme.

[0041] Beneficial effects:

[0042] This application provides a method and system for collaborative control and scheduling of virtual power plants based on artificial intelligence. It predicts the electricity consumption of a power supply area using an artificial intelligence model, determining the predicted electricity consumption for that area and providing a reliable basis for subsequent scheduling. The output of each distributed energy source in the virtual power plant is parameterized, and meeting the predicted electricity consumption is used as the optimization objective. This transforms a complex scheduling problem into an optimization problem, reducing the difficulty of generating scheduling schemes and improving the overall responsiveness and control accuracy of the virtual power plant to grid commands and market signals. Finally, scheduling is performed based on the optimized target output parameter codes, achieving efficient collaborative control and scheduling of virtual power plants. Attached Figure Description

[0043] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 This is a flowchart illustrating a collaborative control and scheduling method for virtual power plants based on artificial intelligence, as proposed in one embodiment of this application.

[0045] Figure 2 This is a schematic diagram of the structure of a virtual power plant collaborative control and scheduling system based on artificial intelligence, as proposed in one embodiment of this application.

[0046] Among them, 201-Electricity consumption prediction module, 202-Schedule coding module, 203-Coding optimization module, and 204-Cooperative control and scheduling module. Detailed Implementation

[0047] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0048] like Figure 1 As shown in the figure, this application provides a virtual power plant collaborative control and scheduling method based on artificial intelligence, including:

[0049] S101. For the energy supply area corresponding to the virtual power plant, predict the electricity consumption of the power supply area through an artificial intelligence model, and determine the predicted electricity consumption corresponding to the energy supply area.

[0050] By using artificial intelligence models to predict electricity consumption in energy supply areas, compared with traditional statistics or simple models, it is possible to more accurately capture the dynamic trend of electricity consumption (such as taking into account weather, time, historical data patterns, etc.), providing a reliable basis for subsequent scheduling and reducing scheduling errors caused by prediction deviations.

[0051] S102. Based on the output of each distributed energy source in the virtual power plant, perform parameter encoding to determine multiple different output parameter codes.

[0052] Encoding the output of each distributed energy source in the virtual power plant with parameters makes it easier to use optimization algorithms (although the specific algorithm is not explicitly stated, the optimization process is implicit) to find the optimal solution in the huge parameter space. Theoretically, this can yield a better scheduling scheme than human experience or simple rules, thereby meeting energy demand more economically and efficiently.

[0053] S103. To meet the predicted electricity consumption of the energy supply area, the output parameter coding is optimized to obtain the target optimal output parameter coding.

[0054] By taking the predicted electricity demand as the optimization objective, the complex scheduling problem can be transformed into an optimization problem, generating a coordinated and optimal power output combination scheme, thereby improving the responsiveness and control accuracy of the entire virtual power plant to grid commands and market signals.

[0055] S104. Decode the target optimal output parameter encoding to obtain the output scheduling scheme corresponding to each distributed energy source in the virtual power plant, and perform collaborative control and scheduling of the virtual power plant according to the output scheduling scheme.

[0056] Optimization based on predicted electricity consumption means prioritizing the use of renewable energy sources such as wind and solar power while meeting load requirements. Flexible resources like energy storage can be used to mitigate the volatility of renewable energy or fill gaps, thereby improving the utilization rate and absorption capacity of renewable energy. The entire process, from prediction to optimization to execution, heavily relies on artificial intelligence models and optimization algorithms, reducing human intervention and achieving automated and intelligent dispatching. This improves the speed and consistency of dispatching decisions, enabling rapid responses to grid changes or market signals.

[0057] In one possible implementation, for the energy supply area corresponding to the virtual power plant, the electricity consumption of the power supply area is predicted using an artificial intelligence model to determine the predicted electricity consumption corresponding to the energy supply area, including:

[0058] Collect multi-dimensional electricity consumption impact data corresponding to the energy supply area of ​​the virtual power plant;

[0059] Retrieve the training labels of the artificial intelligence model; wherein the training labels are either trained labels or untrained labels;

[0060] If the training label of the artificial intelligence model is a trained label, then the artificial intelligence model is used to analyze and process the multi-dimensional electricity consumption impact data to determine the predicted electricity consumption corresponding to the energy supply area.

[0061] If the training label of the artificial intelligence model is an untrained label, the artificial intelligence model is trained to obtain a trained artificial intelligence model. The trained artificial intelligence model is then used to analyze and process the multi-dimensional electricity consumption impact data to determine the predicted electricity consumption corresponding to the energy supply area. The training label is then set as a trained label for subsequent analysis.

[0062] Optionally, the artificial intelligence model can be set as a BP neural network or a convolutional neural network. By detecting the training labels of the artificial intelligence model, it can be ensured that the artificial intelligence model can effectively identify data.

[0063] If the training label of the artificial intelligence model is untrained, then historical multi-dimensional electricity consumption impact data and the actual electricity consumption corresponding to the multi-dimensional electricity consumption impact data should be collected. Then, the artificial intelligence model should be trained using the historical multi-dimensional electricity consumption impact data as input and the actual electricity consumption corresponding to the multi-dimensional electricity consumption impact data as the expected output.

[0064] In one possible implementation, the collection of multi-dimensional electricity consumption impact data corresponding to the energy supply area of ​​the virtual power plant includes:

[0065] Collect seasonal data, weather data, temperature data, date type data, and historical electricity consumption data of the previous N days corresponding to the energy supply area of ​​the virtual power plant to obtain the multi-dimensional electricity consumption impact data;

[0066] The seasonal data includes spring, summer, autumn, or winter codes; the weather data includes rain, sunny, cloudy, or snowy codes; and the date type data includes weekday or non-working day codes.

[0067] Representing various data using digital encoding makes it easier to input for building artificial intelligence models. It's worth noting that, in addition to the multi-dimensional data affecting electricity consumption mentioned above, other data influencing electricity consumption can also be used to improve the accuracy of electricity consumption predictions.

[0068] In one possible implementation, before using an artificial intelligence model to analyze and process the multi-dimensional electricity consumption impact data, the method further includes: normalizing the multi-dimensional electricity consumption impact data.

[0069] Normalizing the multi-dimensional electricity consumption impact data before identification can effectively reduce data complexity and improve data identification efficiency. To ensure accurate identification of the normalized data, the input training data also needs to be normalized during the training of the artificial intelligence model.

[0070] In one possible implementation, training the artificial intelligence model to obtain the trained artificial intelligence model includes:

[0071] The artificial intelligence model is trained using a particle swarm optimization algorithm, a genetic algorithm, or a combination of particle swarm optimization and genetic algorithms to obtain the trained artificial intelligence model.

[0072] It is worth noting that other intelligent optimization algorithms can also be used to train the artificial intelligence model. For example, the gradient descent algorithm can also be used to train the artificial intelligence model.

[0073] In one possible implementation, the step of parameter encoding based on the output of each distributed energy source in the virtual power plant to determine multiple different output parameter codes includes:

[0074] For any distributed energy source in the virtual power plant, based on the lower limit of output and the upper limit of processing corresponding to the distributed energy source, the output of the distributed energy source is randomly initialized to obtain the output parameters corresponding to the distributed energy source.

[0075] The output parameters of all distributed energy sources in the virtual power plant are encoded into vectors to obtain the output parameter encoding.

[0076] Multiple initialization and encoding operations are performed to obtain multiple different output parameter codes.

[0077] Encoding the output of each distributed energy source in the virtual power plant with parameters makes it easier to use optimization algorithms (although the specific algorithm is not explicitly stated, the optimization process is implicit) to find the optimal solution in the huge parameter space. Theoretically, this can yield a better scheduling scheme than human experience or simple rules, thereby meeting energy demand more economically and efficiently.

[0078] In one possible implementation, with the goal of satisfying the predicted electricity consumption corresponding to the energy supply area, the output parameter encoding is optimized to obtain the target optimal output parameter encoding, including:

[0079] S103.1. Construct a target optimization function with the goal of meeting the predicted electricity consumption corresponding to the energy supply area.

[0080] S103.2. For any output parameter code, determine the target optimization function value corresponding to the output parameter code, and determine the current optimal output parameter code based on the target optimization function value. Wherein, the smaller the target optimization function, the better.

[0081] S103.3. Based on the current optimal output parameter encoding, the output parameter encoding is globally guided to obtain the globally guided output parameter encoding as follows:

[0082]

[0083]

[0084]

[0085]

[0086]

[0087]

[0088]

[0089] in, Indicates the first t During the second optimization process, the first j Each output parameter is encoded. Indicates the first j Output parameter encoding after global guidance j =1,2,…,M, where M represents the total number of output parameter codes. This represents the current optimal output parameter encoding. This indicates the encoding of the suboptimal output parameters. The third optimal output parameter encoding represents the minimum objective function value of the current optimal output parameter encoding, the second optimal output parameter encoding represents the second minimum objective function value, and the third optimal output parameter encoding represents the third minimum objective function value. Indicates the first intermediate code. This indicates the second intermediate code. Indicates the third intermediate code. Represents the first random number between (0,1). This represents the second random number between (0,1). This represents a third random number between (0,1). Indicates the first guiding weight. Indicates the second guiding weight. Indicates the third guiding weight. Represents the cosine function. This represents an exponential function with base e. Influence factors expressed as constant terms Represents pi (π). This indicates the fitness of the current optimal output parameter encoding. This indicates the fitness of the suboptimal output parameter encoding. This represents the fitness of the third optimal output parameter encoding, where fitness = 1 / (target optimization function value + 0.0001).

[0090] This application provides a global guide for the output parameter encoding, which enables the output parameter encoding to quickly search for regions with better fitness in the high-dimensional solution space, thereby achieving rapid optimization.

[0091] S103.4. For the output parameter encoding after global guidance, perform local guidance on the output parameter encoding to determine the output parameter encoding after local guidance as follows:

[0092]

[0093]

[0094]

[0095]

[0096]

[0097] in, Indicates the first t During the second optimization process, the first m Output parameter encoding after global guidance Indicates the first m Output parameter encoding after local guidance m =1,2,…,M, Indicates the first t +1 optimization process m Optimization speed of output parameter encoding after global guidance Indicates the first t During the second optimization process, the first m Optimization speed of output parameter encoding after global guidance Indicates a local guiding factor. This represents the fourth random number between (0,1). This represents the adaptive speed control factor. This represents the local center of all output parameter codes; Indicates local center The d Dimensional parameters, d =1,2,…,D, where D represents the total dimension of the output parameter encoding. This represents the encoding of the current optimal output parameters. d Dimensional parameters, Indicates the first q The first random other output parameter encoding d The dimension parameter, Q, represents a random integer term between 1 and M / 2. Indicates the first q The fitness of a randomized encoding of other output parameters, This represents the first constant term, and is set to 2; This represents the second constant term, and is set to 1. This represents the third constant term, set to 0.3; iter represents the preset maximum number of optimizations.

[0098] The embodiments of this application provide local guidance for the output parameter encoding, which not only improves the precision of optimization but also facilitates local search by combining the information of the output parameter encoding, thereby discovering a better position in the high-dimensional solution space and improving the optimization effect.

[0099] S103.5. For the output parameter encoding after local guidance, perform a global mutation search on the output parameter encoding to determine the output parameter encoding after the global mutation search as follows:

[0100]

[0101]

[0102] in, Indicates the first t During the second optimization process, the first k The output parameter encoding after the first local guidance d Dimensional parameters, Indicates the first k Output parameter encoding after a global mutation search This represents the variation factor (which can be set to a constant between (0, 0.2) or (0, 0.5)). Indicates Cauchy mutation, Represents the Cauchy variant. This represents the fifth random number between (0, 1).

[0103] Optionally, the global mutation search process can be controlled by annealing simulation algorithm, thereby achieving strong global optimization capability in the early and middle stages of the algorithm, and improving the convergence capability in the later stages of the algorithm. This enables the algorithm to effectively find the global optimal solution in the high-dimensional solution space, obtain the target optimal output parameter encoding, and realize the solution of the output of distributed energy.

[0104] S103.6 Determine whether the training termination condition is met. If so, determine the target optimal output parameter code based on the output parameter code after global mutation search. Otherwise, return to the step of determining the current optimal output parameter code.

[0105] Optionally, after each optimization of the output parameter encoding (i.e., when the output parameter encoding changes), limit-crossing processing can be performed on the output parameter encoding. For example, if a parameter in a certain dimension of the output parameter encoding exceeds its corresponding upper limit, the parameter that exceeds the upper limit is directly set to the corresponding upper limit; if it exceeds the lower limit, the parameter that exceeds the lower limit is directly set to the corresponding lower limit.

[0106] In one possible implementation, the step of constructing a target optimization function with the objective of satisfying the predicted electricity consumption corresponding to the energy supply area includes:

[0107] Construct the loss term between the output parameter code and the predicted electricity consumption, as well as the cost function corresponding to the output parameter code;

[0108] The target optimization function is obtained by weighted summation of the loss term and the cost function.

[0109]

[0110]

[0111] in, Indicates the first weighting coefficient. Denotes the second weighting coefficient, and and The sum equals 1. The fluctuation term is represented as a constant. Indicates the first in the virtual power plant i A distributed power source during the time period The output power, which is used to characterize the amount of electricity generated. Indicates the first in the virtual power plant i A distributed power source during the time period Maximum output i =1,2,…, I , I This represents the total number of distributed power sources in the virtual power plant. =1,2,…, T , T Indicates the total number of forecast periods. This indicates the energy supply area corresponding to the virtual power plant during the time period. The predicted electricity consumption Describe the objective optimization function. show contribution The corresponding power generation cost, Indicates governance l The cost required for each type of pollutant Indicates power supply node i The corresponding number l Emission coefficients of various pollutants l =1,2,…,L, where L represents the total number of pollutants generated in the virtual power plant. =0 indicates a power supply node i There is no first l Pollutants in the middle, Indicates the first in the main power grid l Emission coefficients of pollutants in China This represents the output of the main power grid, i.e., the output of the virtual power plant during the time period. Electricity purchased from the main power grid.

[0112] When the power supply of a virtual power plant is sufficient to meet the electricity consumption of its service area, it may or may not need to purchase electricity from the main grid. Since the generation costs of each node in the virtual power plant may differ, power supply scheduling for each node is performed based on the predicted generation of the service area to minimize generation costs and environmental pollution. It is worth noting that the distributed power source referred to in this application embodiment can be wind power, solar power, small-scale thermal power, or energy from other energy storage devices.

[0113] In one possible implementation, determining whether the training termination condition is met includes:

[0114] Determine whether the current optimization is greater than or equal to the preset maximum number of optimizations. If so, the training termination condition is met; otherwise, the training termination condition is not met.

[0115] This application provides an artificial intelligence-based collaborative control and scheduling method for virtual power plants. It uses an AI model to predict electricity consumption in a power supply area, determining the predicted electricity consumption for that area and providing a reliable basis for subsequent scheduling. The method encodes the output of each distributed energy source in the virtual power plant as a parameter, using the target of meeting the predicted electricity consumption as the optimization objective. This transforms a complex scheduling problem into an optimization problem, reducing the difficulty of generating scheduling schemes and improving the overall responsiveness and control accuracy of the virtual power plant to grid commands and market signals. Finally, scheduling is performed based on the optimized target output parameter encoding, achieving efficient collaborative control and scheduling of virtual power plants.

[0116] like Figure 2 As shown in the figure, this application provides an artificial intelligence-based virtual power plant collaborative control and scheduling system, including:

[0117] The power consumption prediction module 201 is used to predict the power consumption of the power supply area corresponding to the virtual power plant through an artificial intelligence model, and determine the predicted power consumption of the power supply area.

[0118] The scheduling coding module 202 is used to perform parameter coding based on the output of each distributed energy source in the virtual power plant, and determine multiple different output parameter codes.

[0119] The coding optimization module 203 is used to optimize the output parameter coding to meet the predicted electricity consumption of the energy supply area and obtain the target optimal output parameter coding.

[0120] The collaborative control and scheduling module 204 is used to decode the target optimal output parameter encoding to obtain the output scheduling scheme corresponding to each distributed energy source in the virtual power plant, and to perform collaborative control and scheduling of the virtual power plant according to the output scheduling scheme.

[0121] The virtual power plant collaborative control and scheduling system based on artificial intelligence provided in this application embodiment can execute the above-mentioned method and technical solution. Its principle and beneficial effects are similar, and will not be repeated here.

[0122] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0123] This application describes embodiments with reference to flowchart illustrations and / or block diagrams of methods, apparatuses, electronic devices, and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0124] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0125] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

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

[0127] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0128] The above provides a detailed description of the hierarchical relationship analysis method for influencing factors provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. An artificial intelligence-based virtual power plant collaborative control and scheduling method, characterized in that, include: For the energy supply area corresponding to the virtual power plant, the power consumption of the power supply area is predicted by an artificial intelligence model to determine the predicted power consumption of the energy supply area. Based on the output of each distributed energy source in the virtual power plant, parameter encoding is performed to determine multiple different output parameter codes. To meet the predicted electricity consumption of the energy supply area, the output parameter coding is optimized to obtain the target optimal output parameter coding, including: To meet the predicted electricity consumption of the energy supply area, a target optimization function is constructed. For any output parameter code, determine the target optimization function value corresponding to the output parameter code, and determine the current optimal output parameter code based on the target optimization function value; wherein, the smaller the target optimization function, the better; Based on the current optimal output parameter encoding, the output parameter encoding is globally guided to obtain the globally guided output parameter encoding as follows: wherein, represents the t th output parameter code in the j th optimization process, represents the output parameter code after the j th global guidance, j = 1, 2, …, M, M represents the total number of output parameter codes, represents the current optimal output parameter code, represents the suboptimal output parameter code, represents the third optimal output parameter code, the target optimization function value of the current optimal output parameter code is the smallest, the target optimization function value of the suboptimal output parameter code is the second smallest, and the target optimization function value of the third optimal output parameter code is the third smallest; represents the first intermediate code, represents the second intermediate code, represents the third intermediate code, represents the first random number between (0, 1), represents the second random number between (0, 1), represents the third random number between (0, 1), represents the first guidance weight, represents the second guidance weight, represents the third guidance weight, represents the cosine function, represents the exponential function with the natural constant e as the base, represents the influence factor which is a constant term, represents the circular constant, represents the fitness of the current optimal output parameter code, represents the fitness of the suboptimal output parameter code, represents the fitness of the third optimal output parameter code, the fitness = 1 / (target optimization function value + 0.0001). For the output parameter encoding after global guidance, local guidance is performed on the output parameter encoding to determine the output parameter encoding after local guidance as follows: in, Indicates the first t During the second optimization process, the first m Output parameter encoding after global guidance Indicates the first m Output parameter encoding after local guidance m =1,2,…,M, Indicates the first t +1 optimization process m Optimization speed of output parameter encoding after global guidance Indicates the first t During the second optimization process, the first m Optimization speed of output parameter encoding after global guidance Indicates a local guiding factor. This represents the fourth random number between (0,1). This represents the adaptive speed control factor. This represents the local center of all output parameter codes; Indicates local center The d Dimensional parameters, d =1,2,…,D, where D represents the total dimension of the output parameter encoding. This represents the encoding of the current optimal output parameters. d Dimensional parameters, Indicates the first q The first random other output parameter encoding d The dimension parameter, Q, represents a random integer term between 1 and M / 2. Indicates the first q The fitness of a randomized encoding of other output parameters, This represents the first constant term, and is set to 2; This represents the second constant term, and is set to 1. This indicates the third constant term, set to 0.3; iter indicates the preset maximum number of optimizations. For the output parameter encoding after local guidance, a global mutation search is performed on the output parameter encoding to determine the output parameter encoding after the global mutation search as follows: in, Indicates the first t During the second optimization process, the first k The output parameter encoding after the first local guidance d Dimensional parameters, Indicates the first k Output parameter encoding after a global mutation search This represents the variation factor, set to a constant between (0, 0.2); Indicates Cauchy mutation, Represents the Cauchy variant. Represents the fifth random number between (0, 1); Determine whether the training termination condition is met. If so, determine the target optimal output parameter code based on the output parameter code after global mutation search. Otherwise, return to the step of determining the current optimal output parameter code. The optimal output parameters of the target are decoded to obtain the output scheduling schemes for each distributed energy source in the virtual power plant, and the virtual power plant is coordinated and controlled according to the output scheduling schemes.

2. The artificial intelligence-based virtual power plant collaborative control and scheduling method according to claim 1, characterized in that, For the energy supply area corresponding to the virtual power plant, the power consumption of the supply area is predicted using an artificial intelligence model to determine the predicted power consumption corresponding to the energy supply area, including: Collect multi-dimensional electricity consumption impact data corresponding to the energy supply area of ​​the virtual power plant; Retrieve the training labels of the artificial intelligence model; wherein the training labels are either trained labels or untrained labels; If the training label of the artificial intelligence model is a trained label, then the artificial intelligence model is used to analyze and process the multi-dimensional electricity consumption impact data to determine the predicted electricity consumption corresponding to the energy supply area. If the training label of the artificial intelligence model is an untrained label, the artificial intelligence model is trained to obtain a trained artificial intelligence model. The trained artificial intelligence model is then used to analyze and process the multi-dimensional electricity consumption impact data to determine the predicted electricity consumption corresponding to the energy supply area. The training label is then set as a trained label for subsequent analysis.

3. The artificial intelligence-based virtual power plant collaborative control and scheduling method according to claim 2, characterized in that, The collection of multi-dimensional electricity consumption impact data corresponding to the energy supply area of ​​the virtual power plant includes: Collect seasonal data, weather data, temperature data, date type data, and historical electricity consumption data of the previous N days corresponding to the energy supply area of ​​the virtual power plant to obtain the multi-dimensional electricity consumption impact data; The seasonal data includes spring, summer, autumn, or winter codes; the weather data includes rain, sunny, cloudy, or snowy codes; and the date type data includes weekday or non-working day codes.

4. The artificial intelligence-based virtual power plant collaborative control and scheduling method according to claim 2, characterized in that, Before using an artificial intelligence model to analyze and process the multi-dimensional electricity consumption impact data, the method further includes: normalizing the multi-dimensional electricity consumption impact data.

5. The artificial intelligence-based virtual power plant collaborative control and scheduling method according to claim 2, characterized in that, The step of training the artificial intelligence model to obtain the trained artificial intelligence model includes: The artificial intelligence model is trained using a particle swarm optimization algorithm, a genetic algorithm, or a combination of particle swarm optimization and genetic algorithms to obtain the trained artificial intelligence model.

6. The artificial intelligence-based virtual power plant collaborative control and scheduling method according to claim 1, characterized in that, The process involves encoding parameters based on the output of each distributed energy source in the virtual power plant, determining multiple different output parameter codes, including: For any distributed energy source in the virtual power plant, based on the lower and upper limits of the output corresponding to the distributed energy source, the output corresponding to the distributed energy source is randomly initialized to obtain the output parameters corresponding to the distributed energy source. The output parameters of all distributed energy sources in the virtual power plant are encoded into vectors to obtain the output parameter encoding; Multiple initialization and encoding operations are performed to obtain multiple different output parameter codes.

7. The artificial intelligence-based virtual power plant collaborative control and scheduling method according to claim 1, characterized in that, The objective optimization function is constructed with the goal of satisfying the predicted electricity consumption corresponding to the energy supply area, including: Construct the loss term between the output parameter code and the predicted electricity consumption, as well as the cost function corresponding to the output parameter code; The weighted sum of the loss term and the cost function yields the objective optimization function as follows: in, Indicates the first weighting coefficient. Denotes the second weighting coefficient, and and The sum equals 1. The fluctuation term is represented as a constant. Indicates the first virtual power plant i A distributed power source during the time period The output power, which is used to characterize the amount of electricity generated. Indicates the first virtual power plant i A distributed power source during the time period Maximum output i =1,2,…, I , I This represents the total number of distributed power sources in the virtual power plant. =1,2,…, T , T Indicates the total number of forecast periods. This indicates the energy supply area corresponding to the virtual power plant during the time period. The predicted electricity consumption Denotes the objective optimization function. show contribution The corresponding power generation cost, Indicates governance l The cost required for each type of pollutant Indicates power supply node i The corresponding number l Emission coefficients of various pollutants l =1,2,…,L, where L represents the total number of pollutants generated in the virtual power plant. =0 indicates a power supply node i There is no first l Pollutants in the middle, Indicates the first in the main power grid l Emission coefficients of pollutants in China This represents the output of the main power grid, i.e., the output of the virtual power plant during the time period. Electricity purchased from the main power grid.

8. The artificial intelligence-based virtual power plant collaborative control and scheduling method according to claim 1, characterized in that, The determination of whether the training termination condition is met includes: Determine whether the current number of optimizations is greater than or equal to the preset maximum number of optimizations. If so, the training termination condition is met; otherwise, the training termination condition is not met.

9. An AI-based virtual power plant collaborative control and scheduling system, wherein the AI-based virtual power plant collaborative control and scheduling system is used to execute the AI-based virtual power plant collaborative control and scheduling method according to any one of claims 1-8, characterized in that, include: The electricity consumption prediction module is used to predict the electricity consumption of the power supply area corresponding to the virtual power plant through an artificial intelligence model, and determine the predicted electricity consumption of the power supply area. The scheduling coding module is used to perform parameter coding based on the output of each distributed energy source in the virtual power plant, and to determine multiple different output parameter codes. The coding optimization module is used to optimize the output parameter coding to meet the predicted electricity consumption of the energy supply area and obtain the target optimal output parameter coding. The collaborative control and scheduling module is used to decode the target optimal output parameter encoding to obtain the output scheduling scheme corresponding to each distributed energy source in the virtual power plant, and to perform collaborative control and scheduling of the virtual power plant according to the output scheduling scheme.

Images