Resource scheduling method, device and equipment for virtual power plant and storage medium

Abstract

The invention provides a resource scheduling method and device for a virtual power plant, equipment and a storage medium, and the method comprises the steps: collecting power grid data of a source network load storage scene, and carrying out the preprocessing of the power grid data; performing operation prediction based on the preprocessed power grid data through a generative AI large model to obtain a prediction result of a source grid load storage scene in a future preset time period; generating a plurality of candidate scheduling strategies based on the prediction result through a multi-objective optimization model; evaluating the plurality of candidate scheduling strategies to obtain corresponding evaluation results; and determining a target scheduling strategy based on the evaluation result, and executing a resource scheduling task based on the target scheduling strategy. According to the resource scheduling scheme of the virtual power plant provided by the invention, the energy utilization efficiency can be improved, the power resource distribution can be optimized, and meanwhile, the stability and economy of a power system are ensured, so that the resource scheduling scheme adapts to a complex energy system, deals with the uncertainty of a market environment and meets the requirement of high real-time performance.

Description

Technical Field

[0001] The present application relates to the field of energy scheduling technology, and in particular to a resource scheduling method, device, equipment and storage medium for a virtual power plant. Background Art

[0002] With the transformation of the global energy structure and the development of smart grid technology, Virtual Power Plant (VPP), as a new energy management and dispatching model, has become a key technology to improve energy efficiency and cope with the volatility of renewable energy. Virtual power plants integrate and optimize distributed energy resources, such as solar energy, wind energy, and energy storage equipment, to achieve flexible dispatch and economic operation of power systems.

[0003] However, with the increasing complexity of the energy structure and the increasing uncertainty of the market environment, traditional virtual power plant resource scheduling technology has the following defects: the existing technology mainly relies on rules and experience, lacks the ability to adapt to the dynamic changes of complex energy systems, and it is difficult to achieve efficient coordination of multiple energy resources; the fluctuation of electricity market prices and the intermittent nature of new energy output increase the uncertainty of scheduling, and the existing technology is difficult to effectively respond to the rapid changes in this market and technical environment; in addition, virtual power plants need to respond quickly to the scheduling instructions of the power grid, which puts higher requirements on the real-time performance of the system, and the existing technology often faces the problems of large computational complexity and low efficiency when dealing with the scheduling problems of large-scale and multi-type energy resources.

[0004] The preceding description is intended to provide general background information and does not necessarily constitute prior art. Summary of the invention

[0005] In response to the above technical problems, the present application provides a resource scheduling method, device, electronic device and storage medium for a virtual power plant to solve the problems of low energy utilization efficiency and insufficient adaptability of the existing technology to complex energy systems and uncertain market environments.

[0006] In order to solve the above technical problems, the present application provides a resource scheduling method for a virtual power plant, comprising the following steps:

[0007] Collecting grid data of source-grid-load-storage scenarios and preprocessing the grid data;

[0008] Through the generative AI big model, operation forecasting is performed based on pre-processed power grid data to obtain the forecast results of source-grid-load-storage scenarios in the future preset time period;

[0009] Generate multiple candidate scheduling strategies based on the prediction results through a multi-objective optimization model;

[0010] Evaluate the multiple candidate scheduling strategies to obtain corresponding evaluation results;

[0011] A target scheduling strategy is determined based on the evaluation result, and a resource scheduling task is executed based on the target scheduling strategy.

[0012] Furthermore, in some embodiments of the present application, the collecting of grid data of the source-grid-load-storage scenario and preprocessing of the grid data include:

[0013] Collect real-time power data, historical data and weather data for source-grid-load-storage scenarios;

[0014] The real-time power data, the historical data and the weather data are subjected to timestamp alignment processing and / or data preprocessing.

[0015] Furthermore, in some embodiments of the present application, the generative AI big model is used to perform operation forecasting based on preprocessed power grid data to obtain forecast results of source-grid-load-storage scenarios in a preset time period in the future, including:

[0016] Performing standardization processing on the preprocessed power grid data;

[0017] The standardized power grid data is input into the pre-built generative AI large model for operation prediction, and the output is the load forecast results, power generation forecast results and electricity price forecast results of the source-grid-load-storage scenario in the future preset time period.

[0018] Furthermore, in some embodiments of the present application, the method of constructing the generative AI large model includes:

[0019] Integrating and processing the collected power grid data to obtain integrated power grid data;

[0020] According to the analysis result of the time series diagram, the integrated power grid data is segmented into a data set to obtain a training data set and a test data set;

[0021] Build an initial generative AI model using machine learning and deep learning technologies;

[0022] Performing model training on the initial generative AI big model based on the training data set and the machine learning framework and the deep learning framework to obtain a trained generative AI big model for modeling;

[0023] The trained generative AI big model is adjusted using an optimization parameter adjustment technology to obtain an adjusted generative AI big model;

[0024] The prediction accuracy of the adjusted generative AI big model is evaluated based on the test data set.

[0025] Furthermore, in some embodiments of the present application, the generating of multiple candidate scheduling strategies based on the prediction results by the multi-objective optimization model includes:

[0026] Defining optimization objectives and constraints of the multi-objective optimization model;

[0027] Analyze resource scheduling tasks and obtain resource scheduling goals;

[0028] Analyzing the adjustment parameters corresponding to the resource scheduling objectives according to the prediction results, the optimization objectives and the constraint conditions through the multi-objective optimization model;

[0029] A plurality of candidate scheduling strategies are generated based on the adjustment parameters.

[0030] Furthermore, in some embodiments of the present application, the multi-objective optimization model is constructed by:

[0031] Obtain and annotate the pre-balanced energy data and post-balanced energy data corresponding to the energy system to construct training data;

[0032] The initial energy system balance model is constructed by using a Transformer-like architecture, and the key information and multi-dimensional relationships corresponding to the energy data before the balance are extracted and compressed by an encoder, and the key information extracted by the encoder is converted into a balanced energy system of a predetermined balance target by a decoder;

[0033] Pre-training the initial energy system balance model according to the training data using a self-supervised learning method to obtain deep features of the energy system;

[0034] Use the training data of a specific project as labels for supervised learning to obtain an energy system balance model;

[0035] The energy system balance model is evaluated, and the energy system balance model is adjusted according to the evaluation result.

[0036] Further, in some embodiments of the present application, the evaluating the multiple candidate scheduling strategies to obtain corresponding evaluation results includes:

[0037] Using the simulation environment, each candidate scheduling strategy is evaluated by a machine learning evaluation index and a multi-dimensional benefit index to obtain a machine learning evaluation result and a multi-dimensional benefit evaluation result;

[0038] Based on the machine learning evaluation results and the multi-dimensional benefit evaluation results, a comprehensive evaluation result of each candidate scheduling strategy is determined.

[0039] Accordingly, the present application provides a resource scheduling device for a virtual power plant, comprising:

[0040] A data acquisition module is used to collect power grid data of source-grid-load-storage scenarios and pre-process the power grid data;

[0041] The operation prediction module is used to perform operation prediction based on preprocessed power grid data through a generative AI big model to obtain the prediction results of the source-grid-load-storage scenario in the future preset time period;

[0042] A strategy generation module, used to generate multiple candidate scheduling strategies based on the prediction results through a multi-objective optimization model;

[0043] A strategy evaluation module, used to evaluate the multiple candidate scheduling strategies and obtain corresponding evaluation results;

[0044] A policy execution module is used to determine a target scheduling policy based on the evaluation result, and execute a resource scheduling task based on the target scheduling policy.

[0045] The present application also provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the resource scheduling method for the virtual power plant as described above are implemented.

[0046] The present application also provides a storage medium storing a computer program that can be loaded by a processor and execute the resource scheduling method of the virtual power plant as described above.

[0047] Implementing the embodiments of the present application has the following beneficial effects:

[0048] As described above, the present application provides a resource scheduling method, device, electronic device and storage medium for a virtual power plant, and the resource scheduling method of the virtual power plant includes: collecting grid data of source-grid-load-storage scenarios and preprocessing the grid data; performing operation prediction based on the preprocessed grid data through a generative AI big model to obtain prediction results of source-grid-load-storage scenarios in a preset time period in the future; generating multiple candidate scheduling strategies based on the prediction results through a multi-objective optimization model; evaluating multiple candidate scheduling strategies to obtain corresponding evaluation results; determining a target scheduling strategy based on the evaluation results, and performing resource scheduling tasks based on the target scheduling strategy. The resource scheduling scheme for a virtual power plant provided in the present application accurately predicts the supply and demand of future preset source-grid-load-storage scenarios, and provides an optimized scheduling strategy to comprehensively coordinate and schedule complex energy systems, thereby effectively improving energy utilization efficiency and resource allocation flexibility, reducing grid load fluctuations, and enhancing grid stability and reliability; in addition, the model is continuously updated through adaptive learning, effectively improving the real-time and flexibility of the scheduling strategy, effectively adapting to complex energy systems and coping with the uncertainty of the market environment, and meeting high real-time requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present application, and together with the specification are used to explain the principles of the present application. In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments are briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor.

[0050] Figure 1 It is a schematic diagram of an application scenario of a resource scheduling method for a virtual power plant provided in an embodiment of the present application;

[0051] Figure 2 It is a flow chart of a resource scheduling method for a virtual power plant provided in an embodiment of the present application;

[0052] Figure 3 It is a structural schematic diagram of a resource scheduling device for a virtual power plant provided in an embodiment of the present application;

[0053] Figure 4 It is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0054] The realization of the purpose, functional features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. The above-mentioned drawings have shown clear embodiments of this application, which will be described in more detail later. These drawings and textual descriptions are not intended to limit the scope of the concept of this application in any way, but to illustrate the concept of this application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0055] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0056] It should be noted that, in this article, the terms "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element. In addition, components, features, and elements with the same name in different embodiments of the present application may have the same meaning or different meanings, and their specific meanings need to be determined by their explanation in the specific embodiment or further combined with the context of the specific embodiment.

[0057] It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0058] In the subsequent description, the suffixes such as "module", "component" or "unit" used to represent elements are only used to facilitate the description of the present application, and have no specific meanings. Therefore, "module", "component" or "unit" can be used in a mixed manner.

[0059] With the widespread access to distributed energy, the energy structure of virtual power plants has become more complex, bringing new challenges to scheduling. In addition, the price fluctuations in the electricity market are large, and the output of new energy is intermittent, which increases the uncertainty of scheduling. Virtual power plants also need to respond quickly to the dispatch instructions of the power grid, which places higher requirements on the real-time performance of the system. However, traditional virtual power plant resource scheduling methods are mostly based on rules and experience, which are difficult to adapt to complex energy systems and rapidly changing market environments. In addition, traditional methods often face challenges such as large computational complexity and low efficiency when dealing with large-scale, multi-type energy resource scheduling problems.

[0060] In order to solve the above technical problems, the present application provides a resource scheduling method, device, electronic device and storage medium for a virtual power plant.

[0061] Among them, the resource scheduling device of the virtual power plant can be specifically integrated in an electronic device, and the electronic device can be a smart phone, a tablet computer, a laptop or a desktop computer, but is not limited to this. The electronic device can be directly or indirectly connected to the server through wired or wireless communication. The server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers. It can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. This application does not limit this.

[0062] See also Figure 1 , Figure 1 FIG. 1 is an application environment diagram of a resource scheduling method for a virtual power plant in an embodiment. Figure 1 , the resource scheduling method of the virtual power plant can be applied to the resource scheduling system of the virtual power plant. Among them, the resource scheduling system of the virtual power plant may include a terminal 110 and a server 120. The terminal 110 and the server 120 are connected through a network, and the terminal 110 can be a desktop terminal or a mobile terminal, and the mobile terminal can be at least one of a mobile phone, a tablet computer, a laptop computer, etc. The server 120 can be implemented with an independent server or a server cluster composed of multiple servers. The terminal 110 is used to collect power grid data of the source-grid-load-storage scenario and pre-process the power grid data; the generative AI big model is used to perform operation prediction based on the pre-processed power grid data to obtain the prediction results of the source-grid-load-storage scenario in the future preset time period; based on the prediction results, a plurality of candidate scheduling strategies are generated by a multi-objective optimization model; multiple candidate scheduling strategies are evaluated to obtain corresponding evaluation results; the target scheduling strategy is determined based on the evaluation results, and the resource scheduling task is performed based on the target scheduling strategy.

[0063] It should be noted that the order of description of the following embodiments is not intended to limit the priority order of the embodiments.

[0064] See also Figure 2 , Figure 2 1 is a flow chart of a resource scheduling method for a virtual power plant provided in an embodiment of the present application. A resource scheduling method for a virtual power plant provided in this embodiment may specifically include the following steps:

[0065] S1. Collect grid data of source-grid-load-storage scenarios and pre-process the grid data;

[0066] Specifically, for step S1, the power grid data of the source-grid-load-storage scenario is collected, including but not limited to real-time power data, historical data, and weather data. These data are crucial for understanding the current state of the power grid and predicting future trends. The collected power grid data is preprocessed, including timestamp alignment processing and data preprocessing, to ensure data consistency and availability. The preprocessing step is key to ensuring that subsequent prediction and optimization models can receive high-quality and accurate data. In addition to basic data collection, this step can also include data integration and fusion technologies to integrate data from different sources and formats to improve the comprehensiveness and accuracy of the data.

[0067] S2. Use the generative AI big model to perform operational forecasting based on pre-processed grid data, and obtain forecast results for source-grid-load-storage scenarios in a preset time period in the future;

[0068] Specifically, for step S2, the pre-built generative AI big model is used to perform operational forecasts based on the pre-processed power grid data to obtain forecast results for the source-grid-load-storage scenario for a period of time in the future, including forecasts for the load, power generation, and electricity price of the power grid, providing decision support for resource scheduling. In addition to the generative AI big model, other advanced forecasting technologies, such as deep learning and reinforcement learning, can be further used to improve the accuracy and robustness of the forecast.

[0069] S3. Generate multiple candidate scheduling strategies based on the prediction results through a multi-objective optimization model;

[0070] Specifically, for step S3, the optimization objectives and constraints are defined based on the prediction results through the multi-objective optimization model, and the multi-objective optimization technology is used to find the best balance point between multiple objectives to generate multiple candidate scheduling strategies. In the multi-objective optimization model, new optimization algorithms, such as evolutionary algorithms and mixed integer programming, can be explored and applied to improve the efficiency and effectiveness of strategy generation.

[0071] S4. Evaluate multiple candidate scheduling strategies and obtain corresponding evaluation results;

[0072] Specifically, for step S4, the generated multiple candidate scheduling strategies are evaluated to obtain the evaluation results corresponding to each candidate scheduling strategy. The evaluation process includes machine learning evaluation index evaluation and multi-dimensional benefit index evaluation to determine the performance and benefits of each strategy.

[0073] S5. Determine the target scheduling strategy based on the evaluation results, and execute the resource scheduling task based on the target scheduling strategy;

[0074] Specifically, for step S5, according to the evaluation results of each candidate scheduling strategy, the best candidate scheduling strategy in the evaluation results is selected as the final target scheduling strategy, and the resource scheduling task is performed based on the target scheduling strategy to ensure that the selected strategy can achieve the best economic benefits and system stability in actual power grid operation. This embodiment can also enable the system to dynamically adjust the scheduling strategy according to real-time data and feedback through a real-time feedback mechanism and an adaptive adjustment strategy, thereby improving the response speed and adaptability of the system.

[0075] It can be seen that this embodiment can more effectively utilize grid resources, reduce waste, and improve energy utilization efficiency through accurate prediction and optimized scheduling; the optimized scheduling strategy helps to balance supply and demand, reduce grid load fluctuations, and enhance the stability and reliability of the grid; through prediction and optimization, unnecessary energy purchases and waste can be reduced, grid operating costs can be reduced, and the economic benefits of the grid can be improved.

[0076] Further, in some embodiments, step S1 “collecting grid data of source-grid-load-storage scenario and preprocessing the grid data” may specifically include:

[0077] S11. Collect real-time power data, historical data and weather data of source-grid-load-storage scenarios;

[0078] S12. Perform timestamp alignment processing and / or data preprocessing on real-time power data, historical data, and weather data.

[0079] Specifically, step S1 mainly includes data collection and data preprocessing. For data collection, it includes collecting real-time power data, historical data and weather data of source-grid-load-storage scenarios. Collect real-time power data of source-grid-load-storage scenarios, including but not limited to load, power generation, electricity price, etc. First, track the input and output loads of the power system in real time through intelligent gateway meters, secondary anti-backflow meters and other devices to ensure the real-time and accuracy of the data; collect real-time data of distributed energy facilities, such as the output power of photovoltaic panels, the speed and power output of wind turbines, the charging status and capacity of energy storage batteries, etc.; use sensors and monitoring systems to monitor the health status and efficiency of equipment in real time to ensure the accuracy and reliability of data; access third-party meteorological services to obtain key meteorological parameters such as light, wind speed, and temperature in real time. For data preprocessing, align the timestamps of the collected energy system data for subsequent analysis and model training. Preprocess the load data, such as denoising, interpolation and smoothing, to improve data quality. Patch the meteorological data according to the timestamp of the load data for better integration with the model.

[0080] In addition, this embodiment can also integrate data from different sources and different time scales through advanced data fusion technology to improve data consistency and availability. In the data preprocessing stage, an automated anomaly detection algorithm is integrated to identify and process abnormal data points to further improve data quality. For some data that is difficult to obtain, such as data under extreme weather conditions, data enhancement technology can be used to supplement the lack of actual data by simulating and generating additional data.

[0081] This embodiment improves data accuracy through real-time data collection and precise timestamp alignment, providing a reliable data basis for subsequent prediction and scheduling; by preprocessing and completing load data and meteorological data, data integrity is enhanced, enabling the model to more comprehensively understand and predict the operating status of the energy system; high-quality preprocessed data can improve the performance of generative AI large models, making prediction results more accurate, thereby improving the effectiveness of scheduling strategies; through automated anomaly detection and data enhancement technology, the system can better adapt to different operating conditions and environmental changes, improving the robustness of the system; accurate data collection and preprocessing make resource scheduling strategies more accurate, which helps to achieve optimal energy configuration and improve energy utilization efficiency.

[0082] Further, in some embodiments, step S2 "performing operation forecasts based on preprocessed power grid data through a generative AI big model to obtain forecast results of source-grid-load-storage scenarios in a future preset time period" may specifically include:

[0083] S21. Standardize the preprocessed power grid data;

[0084] S22. Input the standardized power grid data into the pre-built generative AI big model for operation prediction, and output the load forecast results, power generation forecast results and electricity price forecast results for the source-grid-load-storage scenario within a preset time period in the future.

[0085] Specifically, for step S2, the preprocessed power grid data is standardized to ensure the consistency and comparability of the data. Since different data sources may have different dimensions and distributions, standardization can make these data have the same scale before being input into the AI ​​model, thereby improving the training efficiency and prediction accuracy of the model. The standardized power grid data is input into the pre-built generative AI big model for operation prediction. The generative AI big model uses a sliding window recursive method based on historical data and current data to output the prediction results for multiple days in the future, and outputs the load forecast results, power generation forecast results and electricity price forecast results of the source-grid-load-storage scenario in the future preset time period.

[0086] Additionally, feature selection and engineering are performed before data is fed into the AI ​​model to extract the most useful information and reduce the complexity of the model. This can be done through automated feature selection algorithms or guided by the knowledge of domain experts. Model ensemble techniques, such as random forests or gradient boosting machines, are used to combine the predictions of multiple generative AI models to improve the accuracy and robustness of predictions. Models can be updated in real time to adapt to rapid changes in grid conditions. This can be done through online learning or incremental learning algorithms, allowing the model to quickly adapt to new data without the need to retrain from scratch.

[0087] This embodiment improves the prediction accuracy of the model through standardized processing and feature engineering, so that the resource scheduling strategy is more based on the actual grid operation conditions; the model integration technology can enhance the generalization ability of the model, so that the prediction results can maintain a high degree of accuracy under different grid conditions; the real-time update and adaptive learning technology enables the model to quickly respond to changes in grid conditions, improving the response speed and adaptability of the system; accurate load, power generation and electricity price prediction results provide important decision support for resource scheduling, help optimize resource allocation, and improve energy efficiency.

[0088] Furthermore, in some embodiments, the construction method of the generative AI large model may specifically include:

[0089] Integrate and process the collected power grid data to obtain integrated power grid data;

[0090] According to the analysis results of the time series diagram, the integrated power grid data is segmented into a data set to obtain a training data set and a test data set;

[0091] Build an initial generative AI model using machine learning and deep learning technologies;

[0092] Based on the training data set and the machine learning framework and deep learning framework, the initial generative AI big model is trained to obtain the trained generative AI big model for modeling;

[0093] The trained generative AI big model is adjusted by using the optimization parameter adjustment technology to obtain the adjusted generative AI big model;

[0094] The prediction accuracy of the adjusted generative AI large model is evaluated based on the test dataset.

[0095] Specifically, for the generative AI big model in this embodiment, the specific process of its construction and training is as follows: the collected power grid data is integrated and processed to obtain the integrated power grid data. For example, multi-source data such as real-time power data, historical data and weather data are integrated, and the input data is efficiently processed by regularization, normalization and other methods to form a unified data set for model training. According to the volatility and periodicity of the time series diagram, the integrated power grid data is segmented to obtain a training data set and a test data set, the purpose of which is to evaluate the performance of the model on unseen data. The initial generative AI big model is constructed by machine learning technology and deep learning technology. Select a suitable model architecture, such as a recurrent neural network (RNN), a long short-term memory network (LSTM) or a Transformer, etc., to meet the needs of time series prediction. Based on the training data set, the constructed model is trained in conjunction with frameworks such as sklearn and pytorch to obtain a trained generative AI big model. For example, adjust the model parameters to minimize the prediction error. The trained generative AI big model is adjusted by using optimization and parameter adjustment technology to obtain the adjusted generative AI big model, including hyperparameter optimization, such as learning rate, batch size, etc., and optimization and parameter adjustment technologies such as grid search and random search are used to greatly improve the model's capabilities. The prediction accuracy of the adjusted generative AI big model is evaluated based on the test data set. Various evaluation indicators, such as mean absolute error (MAE), root mean square error (RMSE), and mean absolute percentage error (MAPE), are used to quantify the model's prediction performance.

[0096] In a specific embodiment, model explanation techniques such as LIME or SHAP can be introduced to provide interpretability of model predictions and help understand the decision-making process of the model; a multi-model fusion strategy is implemented to combine the prediction results of different models to improve the robustness and accuracy of the prediction.

[0097] This embodiment improves the model's prediction accuracy for future grid load, power generation and electricity price through meticulous data preprocessing and model training; enhances the model's generalization ability for new data by segmenting the data set and evaluating the model on an independent test set; optimizes parameter adjustment and model evaluation steps to ensure that the model can adapt to changes in grid data, thereby improving the model's adaptability and flexibility; and the high-performance generative AI large model provides strong decision-making support for resource scheduling, making the scheduling strategy more scientific and precise.

[0098] Furthermore, in some embodiments, step S3 of "generating multiple candidate scheduling strategies based on the prediction results by using a multi-objective optimization model" may specifically include:

[0099] S31. Define the optimization objectives and constraints of the multi-objective optimization model;

[0100] S32. Analyze the resource scheduling task to obtain the resource scheduling target;

[0101] S33. Analyze the adjustment parameters corresponding to the resource scheduling objectives according to the prediction results, optimization objectives and constraints through the multi-objective optimization model;

[0102] S34. Generate multiple candidate scheduling strategies based on the adjustment parameters.

[0103] Specifically, for step S3, the optimization objectives of the multi-objective optimization model are clarified, including maximizing economic benefits, maximizing system stability, maximizing the utilization of renewable energy, etc. At the same time, the constraints that the model needs to comply with are defined, such as equipment operation restrictions, safety standards, environmental regulations, etc. In-depth analysis is conducted on the resource scheduling tasks, and scheduling objectives are clarified, including identifying key resources, predicting resource demand, and evaluating resource availability, etc., to ensure that the scheduling strategy can meet the needs of system operation. Through the multi-objective optimization model, according to the prediction results, optimization objectives and constraints, the adjustment parameters corresponding to the resource scheduling objectives are analyzed, including the distribution of power generation, the charging and discharging strategies of energy storage equipment, load management, etc. Based on the adjustment parameters, multiple candidate scheduling strategies are generated using the multi-objective optimization model. These strategies will consider different optimization objectives and constraints to provide a variety of scheduling options.

[0104] In a specific embodiment, based on the trained model, a prediction is made for a period of time in the future, and multiple candidate scheduling strategies are generated based on the prediction results. The resource scheduling task is decomposed using the energy big model, and the goals of each subtask and the corresponding inputs and outputs are set; the core task is the balance of the energy system, and the energy system composed of the energy block data generated by the previous steps (preprocessing, prediction, etc.) is input into the energy balance system model, and then the energy balance system after the model balance is output; the parameters that need to be adjusted are disassembled from the energy balance system after the model balance, and converted into strategies that can be issued and parameters that each specific device needs to receive.

[0105] The multi-objective optimization model provided in this embodiment can comprehensively consider multiple objectives and constraints, achieve optimal resource allocation, and improve energy utilization efficiency; through the optimization model, economic benefits can be maximized, such as reducing operating costs and increasing energy sales revenue; the optimization model helps maintain the stability of the power grid and reduces the risks of power grid load fluctuations and unstable power supply through reasonable resource scheduling; through multi-scenario simulation and real-time optimization and adjustment, the scheduling strategy's adaptability to uncertainty and changes is improved, and the system's robustness is enhanced. The multi-objective optimization model can balance economic benefits, system stability, and environmental impact to achieve comprehensive optimization.

[0106] Furthermore, in some embodiments, the construction method of the multi-objective optimization model may specifically include:

[0107] Obtain and annotate the pre-balanced energy data and post-balanced energy data corresponding to the energy system to construct training data;

[0108] The initial energy system balance model is constructed using a Transformer-like architecture. The encoder extracts and compresses the key information and multi-dimensional relationships corresponding to the energy data before balancing. The decoder converts the key information extracted by the encoder into a balanced energy system with a predetermined balance target.

[0109] The self-supervised learning method is used to pre-train the initial energy system balance model based on the training data to obtain the deep characteristics of the energy system;

[0110] Use the training data of a specific project as labels for supervised learning to obtain an energy system balance model;

[0111] The energy system balance model is evaluated and adjusted according to the evaluation results.

[0112] Specifically, for the construction of the multi-objective optimization model in this embodiment, the energy data before and after the balance corresponding to the energy system are obtained, and the energy data after the balance are annotated to construct training data. This includes energy consumption data, energy storage data, new energy power generation data, etc., which will be used to train the model to understand and predict the balance state of the energy system. The initial energy system balance model is constructed using a Transformer-like architecture. The architecture extracts and compresses the key information and multi-dimensional relationships of the energy data before the balance through an encoder, and converts this information into a balanced energy system with a predetermined balance target through a decoder. The initial energy system balance model is pre-trained according to the training data using a self-supervised learning method to obtain the deep features of the energy system. This step helps the model learn the inherent laws and patterns of energy data. The training data of a specific project is used as a label for supervised learning to obtain an energy system balance model, so that the model can adapt to the special needs of energy system balance in a specific scenario. The energy system balance model is evaluated, and the model is adjusted according to the evaluation results to ensure the effectiveness and accuracy of the model in practical applications.

[0113] In a specific embodiment, the construction of the energy system balance model first starts with the processing of training data, using a large amount of energy system balance data that has been annotated in the early stage, including a large amount of energy data before balance (energy consumption, energy storage, new energy generation, etc.), and energy data after balance generated by the energy balance optimization model and further corrected by the operation and maintenance team. These data constitute the training data of the energy system balance model. The structure of the energy system balance model is based on a transformer-like architecture. The encoder part is responsible for extracting and compressing the information of the energy data before balance and the multi-dimensional relationship between each other. The decoder part is responsible for using the information of the unbalanced energy system extracted by the encoder part to convert it into a balanced energy system based on a predetermined balance target. The input and output are both energy systems composed of multi-dimensional energy blocks. The model training uses the PyTorch framework, combined with distributed training technology, to accelerate the processing of massive data and model iteration. The training stage adopts a "multi-stage training" strategy: the first stage is based on an energy system composed of multi-dimensional energy block data before and after large-scale balance, and uses a self-supervised learning method to pre-train the model to capture the deep characteristics of the energy system. The second stage uses the multi-dimensional energy block data of specific projects as labels for supervised learning, so that it can adapt to the special needs of energy system balance in specific scenarios.

[0114] This embodiment improves the prediction accuracy of the model for the balance state of the energy system by combining self-supervised learning and supervised learning; the Transformer-like architecture and ensemble learning method enhance the generalization ability of the model to different power grid environments; the efficient energy system balance model can quickly respond to power grid changes and improve the efficiency and response speed of resource scheduling; accurate energy system balance prediction helps to optimize the allocation and utilization of energy and improve energy utilization efficiency; through model evaluation and adjustment, as well as reinforcement learning technology, the robustness of the system in the face of uncertainty and change is enhanced.

[0115] Further, in some embodiments, step S4 of “evaluating multiple candidate scheduling strategies to obtain corresponding evaluation results” may specifically include:

[0116] S41. Use the simulation environment to evaluate the machine learning evaluation index and the multi-dimensional benefit index of each candidate scheduling strategy, and obtain the machine learning evaluation result and the multi-dimensional benefit evaluation result;

[0117] S42. Based on the machine learning evaluation results and the multi-dimensional benefit evaluation results, determine the comprehensive evaluation results of each candidate scheduling strategy.

[0118] Specifically, for step S4, a simulation environment is used to evaluate each candidate dispatching strategy using machine learning evaluation indicators and multi-dimensional benefit indicators. The simulation environment provides a platform that simulates actual operations, allowing the effectiveness of the strategy to be tested and evaluated without actually affecting the operation of the power grid. The performance of the candidate dispatching strategy is quantitatively evaluated using machine learning evaluation indicators such as accuracy, recall, F1 score, etc. These indicators help evaluate the effectiveness of the strategy in forecasting and scheduling tasks. Multi-dimensional benefit indicator evaluation is performed, including economic benefits, system stability, environmental impact, etc., such as evaluating the comprehensive performance of the strategy in different aspects to determine its overall benefit. Based on the machine learning evaluation results and the multi-dimensional benefit evaluation results, the comprehensive evaluation results of each candidate dispatching strategy are determined, and the performance of the strategy in each dimension is comprehensively considered to select the best dispatching strategy.

[0119] In a specific embodiment, a simulation environment is used to evaluate each candidate strategy, and its performance under different objectives is calculated, such as economic benefits, system stability, etc. Specifically, the evaluation can be carried out from the following two aspects: First, from the perspective of machine learning, the output of the energy system balance model is essentially the output of a type of continuous numerical task, which can be evaluated using the general evaluation indicators (MAE, MSE) in the field of machine learning, so that the model can grasp the basic fit of the energy industry's operational optimization capabilities. Secondly, from a business perspective, a set of evaluation systems oriented towards actual benefits is designed. The model is evaluated through multi-dimensional benefit indicators (such as improved economic benefits, enhanced safety, increased proportion of green energy, etc.) to ensure that it not only has excellent theoretical performance, but also can bring considerable benefits in actual operations. Secondly, multiple typical process flow scenarios are simulated to detect whether the model can balance economy and system stability under complex production constraints.

[0120] Through comprehensive evaluation, this embodiment improves the accuracy of selecting the best scheduling strategy and ensures the efficiency and effectiveness of resource scheduling; real-time data integration evaluation and reinforcement learning evaluation methods improve the system's adaptability to dynamic changes in the power grid; accurate evaluation helps to optimize the allocation and utilization of resources, improve energy efficiency and reduce costs; multi-dimensional benefit indicator evaluation ensures the performance of the scheduling strategy in economic benefits and helps to improve the economic benefits of the virtual power plant; by evaluating the impact of the strategy on the system stability, it can be ensured that the selected strategy will not have a negative impact on the stability and security of the power grid.

[0121] Further, in some embodiments, step S5 “determine the target scheduling strategy based on the evaluation result, and execute the resource scheduling task based on the target scheduling strategy” may specifically include:

[0122] S51. Convert the selected target scheduling strategy into corresponding control instructions;

[0123] S52. Send control instructions to each energy device to execute resource scheduling tasks.

[0124] Specifically, for step S5, after evaluating the candidate scheduling strategies in the simulation environment, comprehensive evaluation results including machine learning evaluation indicators and multi-dimensional benefit indicators are collected. These results provide an overview of the performance of each candidate strategy, including economic benefits, system stability, environmental impact, etc. Based on the comprehensive evaluation results, one or more optimal candidate scheduling strategies are selected as the target scheduling strategy. Factors considered in the selection include the overall benefit of the strategy, the risk level, the cost-effectiveness ratio, and the impact on system stability and security. After determining the target scheduling strategy, it needs to be fine-tuned to adapt to the specific requirements of actual grid operation. This includes adjusting the parameters in the strategy to better adapt to the actual operating conditions and market changes of the grid. The target scheduling strategy is converted into specific control instructions and issued to each energy device. This includes adjusting the output of power generation equipment, managing the charging and discharging of energy storage equipment, optimizing load distribution, etc. During the execution of the resource scheduling task, the grid status and equipment performance are monitored in real time to ensure that the implementation effect of the strategy meets expectations. Data during the execution process is collected to provide feedback for subsequent strategy adjustments and model optimization.

[0125] This embodiment improves the efficiency and accuracy of resource scheduling through comprehensive evaluation and optimization, ensuring that the power grid operates in the best state; the real-time monitoring and feedback mechanism enhances the system's adaptability to emergencies and market changes and improves the system's robustness; by selecting the optimal scheduling strategy, economic benefits can be maximized, such as reducing operating costs and increasing energy sales revenue; the execution of the target scheduling strategy helps maintain the stability and security of the power grid and reduce the risk of power grid failures and accidents.

[0126] In a specific embodiment, after step S5 "determine the target scheduling strategy based on the evaluation results, and execute the resource scheduling task based on the target scheduling strategy", this embodiment also provides a feedback and learning process, and designs a friendly human-computer interaction interface to facilitate manual viewing of strategy recommendation results, modification and confirmation. Manual personnel can modify and adjust the recommended strategy based on actual conditions and experience, and feedback to the system. The system uses manual feedback and actual execution results as new data to continuously train and optimize the model and improve the accuracy of decision-making. As the market environment and system status change, the system can adaptively adjust the strategy to improve the flexibility and robustness of the system.

[0127] Among them, adaptive adjustment of strategies can be achieved specifically through the following two aspects: first, industry experts and operation and maintenance personnel will initially revise the strategies issued daily, and re-enter the revised energy balance system data into the relevant model for fine-tuning, so that the model can better adapt to the laws of the current scenario; second, by using technologies such as reinforcement learning based on multi-faceted benefit feedback, the model can update parameters in a timely manner according to the changes in benefits brought about by changes in small parameters in real-time scenarios, and better allocate weights to understand the current scenario.

[0128] In summary, the resource scheduling method of the virtual power plant provided in this embodiment collects the grid data of the source-grid-load-storage scenario and pre-processes the grid data; performs operation prediction based on the pre-processed grid data through a generative AI big model to obtain the prediction results of the source-grid-load-storage scenario in the future preset time period; generates multiple candidate scheduling strategies based on the prediction results through a multi-objective optimization model; evaluates multiple candidate scheduling strategies to obtain corresponding evaluation results; determines the target scheduling strategy based on the evaluation results, and performs resource scheduling tasks based on the target scheduling strategy. A resource scheduling scheme for a virtual power plant provided in this embodiment accurately predicts the supply and demand of future preset source-grid-load-storage scenarios, and provides an optimized scheduling strategy to comprehensively coordinate and schedule complex energy systems, thereby effectively improving energy utilization efficiency and resource allocation flexibility, reducing grid load fluctuations, and enhancing grid stability and reliability; in addition, the model is continuously updated through adaptive learning, effectively improving the real-time and flexibility of the scheduling strategy, effectively adapting to complex energy systems and coping with the uncertainty of the market environment, and meeting high real-time requirements.

[0129] In order to facilitate better implementation of the resource scheduling method of the virtual power plant in the embodiment of the present application, the embodiment of the present application also provides a resource scheduling device for a virtual power plant. The meanings of the terms are the same as those in the resource scheduling method of the virtual power plant mentioned above, and the specific implementation details can refer to the description in the method embodiment.

[0130] See also Figure 3 , Figure 3 A schematic diagram of the structure of a resource scheduling device for a virtual power plant provided in an embodiment of the present application, wherein the resource scheduling device for the virtual power plant may specifically include a data acquisition module 201, an operation prediction module 202, a strategy generation module 203, a strategy evaluation module 204 and a strategy execution module 205, which may be specifically as follows:

[0131] The data acquisition module 201 is used to collect the grid data of the source-grid-load-storage scenario and pre-process the grid data;

[0132] Operation prediction module 202, used to perform operation prediction based on preprocessed power grid data through a generative AI big model, and obtain prediction results of source-grid-load-storage scenarios in a preset time period in the future;

[0133] A strategy generation module 203, used to generate multiple candidate scheduling strategies based on the prediction results through a multi-objective optimization model;

[0134] A strategy evaluation module 204 is used to evaluate multiple candidate scheduling strategies and obtain corresponding evaluation results;

[0135] The policy execution module 205 is used to determine the target scheduling policy based on the evaluation result, and execute the resource scheduling task based on the target scheduling policy.

[0136] Furthermore, in some embodiments, the data acquisition module 201 may specifically include:

[0137] Data collection unit, used to collect real-time power data, historical data and weather data of source-grid-load-storage scenarios;

[0138] The data preprocessing unit is used to perform time stamp alignment processing and / or data preprocessing on real-time power data, historical data and weather data.

[0139] Furthermore, in some embodiments, the operation prediction module 202 may specifically include:

[0140] A standardization unit, used for standardizing the pre-processed power grid data;

[0141] The prediction unit is used to input the standardized power grid data into the pre-built generative AI large model for operation prediction, and output the load prediction results, power generation prediction results and electricity price prediction results of the source-grid-load-storage scenario within a preset time period in the future.

[0142] Furthermore, in some embodiments, the operation prediction module 202 also includes an AI model construction unit, which is specifically used to: integrate and process the collected power grid data to obtain integrated power grid data; divide the integrated power grid data into data sets according to the timing diagram analysis results to obtain training data sets and test data sets; construct an initial generative AI big model through machine learning technology and deep learning technology; train the initial generative AI big model based on the training data set and the machine learning framework and the deep learning framework to obtain a trained generative AI big model; adjust the trained generative AI big model using the optimization parameter adjustment technology to obtain an adjusted generative AI big model; and evaluate the prediction accuracy of the adjusted generative AI big model based on the test data set.

[0143] Furthermore, in some embodiments, the strategy generation module 203 may specifically include:

[0144] A definition unit is used to define the optimization objectives and constraints of the multi-objective optimization model;

[0145] A task analysis unit is used to analyze resource scheduling tasks and obtain resource scheduling targets;

[0146] A parameter adjustment unit, used to analyze adjustment parameters corresponding to resource scheduling objectives according to prediction results, optimization objectives and constraint conditions through a multi-objective optimization model;

[0147] The strategy generating unit is used to generate a plurality of candidate scheduling strategies based on the adjustment parameters.

[0148] Furthermore, in some embodiments, the strategy generation module 203 may specifically include a multi-objective optimization model construction unit, which may be specifically used to: obtain and annotate the pre-balanced energy data and post-balanced energy data corresponding to the energy system to construct training data; use a Transformer-like architecture to construct an initial energy system balance model, extract and compress the key information and multi-dimensional relationships corresponding to the pre-balanced energy data through an encoder, and convert the key information extracted by the encoder into a balanced energy system with a predetermined balance target through a decoder; use a self-supervised learning method to pre-train the initial energy system balance model according to the training data to obtain the deep features of the energy system; use the training data of a specific project as a label for supervised learning to obtain an energy system balance model; evaluate the energy system balance model, and adjust the energy system balance model according to the evaluation results.

[0149] Furthermore, in some embodiments, the policy evaluation module 204 may specifically include:

[0150] A first evaluation unit is used to use a simulation environment to perform machine learning evaluation index evaluation and multi-dimensional benefit index evaluation on each candidate scheduling strategy, and obtain a machine learning evaluation result and a multi-dimensional benefit evaluation result;

[0151] The second evaluation unit is used to determine the comprehensive evaluation result of each candidate scheduling strategy based on the machine learning evaluation result and the multi-dimensional benefit evaluation result.

[0152] In summary, the resource scheduling device of the virtual power plant provided in this embodiment collects the grid data of the source-grid-load-storage scenario through the data acquisition module 201, and pre-processes the grid data; the operation prediction module 202 performs operation prediction based on the pre-processed grid data through the generative AI big model to obtain the prediction results of the source-grid-load-storage scenario in the future preset time period; the strategy generation module 203 generates multiple candidate scheduling strategies based on the prediction results through the multi-objective optimization model; the strategy evaluation module 204 evaluates multiple candidate scheduling strategies to obtain corresponding evaluation results; the strategy execution module 205 determines the target scheduling strategy based on the evaluation results, and executes the resource scheduling task based on the target scheduling strategy. The resource scheduling device of the virtual power plant provided in this embodiment accurately predicts the supply and demand of the future preset source-grid-load-storage scenario, and provides an optimized scheduling strategy to comprehensively coordinate the scheduling of complex energy systems, thereby effectively improving the energy utilization efficiency and the flexibility of resource allocation, reducing grid load fluctuations, and enhancing the stability and reliability of the grid; in addition, the model is continuously updated through adaptive learning, effectively improving the real-time and flexibility of the scheduling strategy, effectively adapting to the uncertainty of complex energy systems and coping with the market environment, and meeting high real-time requirements.

[0153] In addition, the present application also provides an electronic device, such as Figure 4 , which shows a schematic diagram of the structure of an electronic device involved in an embodiment of the present application. Specifically, the electronic device may include a processor 301 with one or more processing cores, a memory 302 with one or more computer-readable storage media, a power supply 303, an input unit 304 and other components. Those skilled in the art can understand that Figure 4 The electronic device structure shown in the figure does not constitute a limitation on the electronic device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0154] The processor 301 is the control center of the electronic device. It uses various interfaces and lines to connect various parts of the entire electronic device. By running or executing software programs and / or modules stored in the memory 302, and calling data stored in the memory 302, it executes various functions of the electronic device and processes data, thereby monitoring the electronic device as a whole. Optionally, the processor 301 may include one or more processing cores; preferably, the processor 301 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, and the modem processor mainly processes wireless communications. It is understandable that the above-mentioned modem processor may not be integrated into the processor 301.

[0155] The memory 302 can be used to store software programs and modules. The processor 301 executes various functional applications and resource scheduling methods of virtual power plants by running the software programs and modules stored in the memory 302. The memory 302 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the electronic device, etc. In addition, the memory 302 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory 302 may also include a memory controller to provide the processor 301 with access to the memory 302.

[0156] The electronic device also includes a power supply 303 for supplying power to each component. Preferably, the power supply 303 can be logically connected to the processor 301 through a power management system, so that the power management system can manage charging, discharging, power consumption and other functions. The power supply 303 can also include one or more DC or AC power supplies, recharging systems, power failure detection circuits, power converters or inverters, power status indicators and other arbitrary components.

[0157] The electronic device may further include an input unit 304, which may be used to receive input digital or character information and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control.

[0158] Although not shown, the electronic device may further include a display unit, etc., which will not be described in detail herein. Specifically in this embodiment, the processor 301 in the electronic device will load the executable files corresponding to the processes of one or more application programs into the memory 302 according to the following instructions, and the processor 301 will run the application programs stored in the memory 302, thereby realizing various functions, as follows:

[0159] Collect power grid data of source-grid-load-storage scenarios and pre-process the power grid data; use a generative AI big model to perform operational forecasts based on the pre-processed power grid data to obtain forecast results for source-grid-load-storage scenarios in a preset time period in the future; use a multi-objective optimization model based on the forecast results to generate multiple candidate dispatching strategies; evaluate multiple candidate dispatching strategies to obtain corresponding evaluation results; determine the target dispatching strategy based on the evaluation results, and execute resource dispatching tasks based on the target dispatching strategy.

[0160] The specific implementation of the above operations can be found in the previous embodiments, which will not be described in detail here.

[0161] The embodiments of the present application accurately predict the supply and demand of future preset source-grid-load-storage scenarios, and provide optimized scheduling strategies to comprehensively coordinate and schedule complex energy systems, thereby effectively improving energy utilization efficiency and resource allocation flexibility, reducing grid load fluctuations, and enhancing grid stability and reliability. In addition, by continuously updating the model through adaptive learning, the real-time and flexibility of the scheduling strategy are effectively improved, effectively adapting to complex energy systems and coping with the uncertainty of the market environment, and meeting high real-time requirements.

[0162] A person of ordinary skill in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be completed by instructions, or by controlling related hardware through instructions. The instructions may be stored in a computer-readable storage medium and loaded and executed by a processor.

[0163] To this end, an embodiment of the present application provides a storage medium in which a plurality of instructions are stored, and the instructions can be loaded by a processor to execute the steps in any one of the resource scheduling methods for a virtual power plant provided in the embodiment of the present application. For example, the instructions can execute the following steps:

[0164] Collect power grid data of source-grid-load-storage scenarios and pre-process the power grid data; use a generative AI big model to perform operational forecasts based on the pre-processed power grid data to obtain forecast results for source-grid-load-storage scenarios in a preset time period in the future; use a multi-objective optimization model based on the forecast results to generate multiple candidate dispatching strategies; evaluate multiple candidate dispatching strategies to obtain corresponding evaluation results; determine the target dispatching strategy based on the evaluation results, and execute resource dispatching tasks based on the target dispatching strategy.

[0165] The specific implementation of the above operations can be found in the previous embodiments, which will not be described in detail here.

[0166] The storage medium may include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc. Since the instructions stored in the storage medium can execute the steps in any of the resource scheduling methods for a virtual power plant provided in the embodiments of the present application, the beneficial effects that can be achieved by any of the resource scheduling methods for a virtual power plant provided in the embodiments of the present application can be achieved. For details, please refer to the previous embodiments and will not be repeated here.

[0167] The above is a detailed introduction to the resource scheduling method, device, electronic device and storage medium of a virtual power plant provided in an embodiment of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A resource scheduling method for a virtual power plant, characterized in that: The steps include: Collecting grid data of source-grid-load-storage scenarios and preprocessing the grid data; Through the generative AI big model, operation forecasting is performed based on pre-processed power grid data to obtain the forecast results of source-grid-load-storage scenarios in the future preset time period; Generate multiple candidate scheduling strategies based on the prediction results through a multi-objective optimization model; Evaluate the multiple candidate scheduling strategies to obtain corresponding evaluation results; A target scheduling strategy is determined based on the evaluation result, and a resource scheduling task is executed based on the target scheduling strategy.

2. The resource scheduling method of a virtual power plant according to claim 1, characterized in that: The collecting of grid data of the source-grid-load-storage scenario and preprocessing of the grid data include: Collect real-time power data, historical data and weather data for source-grid-load-storage scenarios; The real-time power data, the historical data and the weather data are subjected to timestamp alignment processing and / or data preprocessing.

3. The resource scheduling method of a virtual power plant according to claim 1, characterized in that: The generative AI big model performs operation forecasting based on preprocessed power grid data to obtain forecast results of source-grid-load-storage scenarios in a preset time period in the future, including: Performing standardization processing on the preprocessed power grid data; The standardized power grid data is input into the pre-built generative AI large model for operation prediction, and the output is the load forecast results, power generation forecast results and electricity price forecast results of the source-grid-load-storage scenario in the future preset time period.

4. The resource scheduling method of a virtual power plant according to claim 3, characterized in that: The method of constructing the generative AI large model includes: Integrating and processing the collected power grid data to obtain integrated power grid data; According to the analysis result of the time series diagram, the integrated power grid data is segmented into a data set to obtain a training data set and a test data set; Build an initial generative AI model using machine learning and deep learning technologies; Performing model training on the initial generative AI big model based on the training data set and the machine learning framework and the deep learning framework to obtain a trained generative AI big model for modeling; The trained generative AI big model is adjusted using an optimization parameter adjustment technology to obtain an adjusted generative AI big model; The prediction accuracy of the adjusted generative AI big model is evaluated based on the test data set.

5. The resource scheduling method of a virtual power plant according to claim 1, characterized in that: The method of generating a plurality of candidate scheduling strategies based on the prediction results by using a multi-objective optimization model includes: Defining optimization objectives and constraints of the multi-objective optimization model; Analyze resource scheduling tasks and obtain resource scheduling goals; Analyzing the adjustment parameters corresponding to the resource scheduling objectives according to the prediction results, the optimization objectives and the constraint conditions through the multi-objective optimization model; A plurality of candidate scheduling strategies are generated based on the adjustment parameters.

6. The resource scheduling method of a virtual power plant according to claim 5, characterized in that: The multi-objective optimization model is constructed in the following manner: Obtain and annotate the pre-balanced energy data and post-balanced energy data corresponding to the energy system to construct training data; The initial energy system balance model is constructed by using a Transformer-like architecture, and the key information and multi-dimensional relationships corresponding to the energy data before the balance are extracted and compressed by an encoder, and the key information extracted by the encoder is converted into a balanced energy system of a predetermined balance target by a decoder; Pre-training the initial energy system balance model according to the training data using a self-supervised learning method to obtain deep features of the energy system; Use the training data of a specific project as labels for supervised learning to obtain an energy system balance model; The energy system balance model is evaluated, and the energy system balance model is adjusted according to the evaluation result.

7. The resource scheduling method of a virtual power plant according to claim 1, characterized in that: The evaluating the multiple candidate scheduling strategies to obtain corresponding evaluation results includes: Using the simulation environment, each candidate scheduling strategy is evaluated by a machine learning evaluation index and a multi-dimensional benefit index to obtain a machine learning evaluation result and a multi-dimensional benefit evaluation result; Based on the machine learning evaluation results and the multi-dimensional benefit evaluation results, a comprehensive evaluation result of each candidate scheduling strategy is determined.

8. A resource scheduling device for a virtual power plant, characterized in that: include: A data acquisition module is used to collect power grid data of source-grid-load-storage scenarios and pre-process the power grid data; The operation prediction module is used to perform operation prediction based on preprocessed power grid data through a generative AI big model to obtain the prediction results of the source-grid-load-storage scenario in the future preset time period; A strategy generation module, used to generate multiple candidate scheduling strategies based on the prediction results through a multi-objective optimization model; A strategy evaluation module, used to evaluate the multiple candidate scheduling strategies and obtain corresponding evaluation results; A policy execution module is used to determine a target scheduling policy based on the evaluation result, and execute a resource scheduling task based on the target scheduling policy.

9. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the resource scheduling method for a virtual power plant as described in any one of claims 1 to 7 are implemented.

10. A storage medium, characterized in that: A computer program is stored which can be loaded by a processor and execute the resource scheduling method for a virtual power plant as described in any one of claims 1 to 7.

Images