Virtual power plant intelligent scheduling method and system based on space-time network model

Abstract

The invention provides a virtual power plant intelligent scheduling method and system based on a space-time network model, and belongs to the technical field of intelligent power plants. The method comprises the following steps: collecting real-time data from a distributed power station and a meteorological station; the collected data is preprocessed; constructing a space-time network model based on an encoder-decoder by using a space-time neural network unit, importing the preprocessed data, and predicting the expected working power of each device of the virtual power plant; and according to a prediction result of the space-time network model, in combination with a real-time state and an actual constraint condition of the virtual power plant, carrying out intelligent scheduling decision making. Through the end-to-end design of the space-time network model, electrical data, meteorological data, date data, equipment parameter data, market operation data and geographical location information are comprehensively considered, and detailed scheduling strategies and schemes can be directly obtained, so that the reliability and economy of power supply are improved, and the scheduling efficiency is improved. And optimization of an energy structure is facilitated, and energy conservation, emission reduction and sustainable development are promoted.

Description

Technical Field

[0001] The present invention belongs to the technical field of smart power plants, and more specifically, relates to a virtual power plant intelligent scheduling method and system based on a spatiotemporal network model. Background Art

[0002] One of the defining characteristics of virtual power plants (VPPs) is that their power sources, loads, and energy storage devices are often geographically distributed. This spatial distribution significantly impacts the formulation and implementation of scheduling strategies. While time series data primarily reflects how data changes over time, spatial distribution data reveals the distribution characteristics of data across different geographic locations. Existing VPP scheduling strategies often focus on analyzing time series data in terms of data collection and processing, ignoring the spatial distribution of the data.

[0003] The spatial distribution of data allows for forward-looking forecasts of power output, loads, and energy storage devices. For example, severe temperature drops typically occur from north to south at a consistent rate. Wind data also exhibits significant variations over time along the wind's path. Failure to fully consider the spatial distribution of distributed power sources, loads, and energy storage devices can lead to power shortages or surpluses in certain areas, even if the overall system is functioning properly, impacting the overall efficiency and stability of the power system. Summary of the Invention

[0004] The main purpose of this application is to provide a virtual power plant intelligent scheduling method and system based on a space-time network model, which can combine the spatial distribution properties of data to improve the accuracy and rationality of the scheduling strategy.

[0005] To achieve the above objectives, in a first aspect, the present application proposes a virtual power plant intelligent scheduling method based on a spatiotemporal network model, comprising: Collect real-time data from sensors, smart meters, and weather stations at distributed power stations. This includes power supply equipment data, energy storage equipment data, controllable load data, weather data, date data, market and price data, manual control data, and geographic location information. Preprocessing the collected real-time data; wherein the preprocessing includes data cleaning, format conversion, anomaly detection, and data normalization; Using the spatiotemporal network basic unit, a spatiotemporal network model based on an encoder-decoder is constructed; based on the preprocessed data, the spatiotemporal network model is used to predict the expected operating power and expected load of each device in the virtual power plant; Based on the prediction results of the spatiotemporal network model, combined with the real-time status and actual constraints of the virtual power plant, intelligent scheduling decisions are made; wherein the real-time status of the virtual power plant includes real-time information on equipment operation, real-time information on demand, and business strategies.

[0006] Furthermore, the spatiotemporal network model includes a convolutional neural network and a recursive neural network; wherein, the convolutional neural network is used to transmit the hidden state value at the previous time point to the next node through spatial convolution processing, and the recursive neural network is used to calculate the output value and the new hidden state value based on the input data and the hidden state value at the previous time point.

[0007] Furthermore, the preprocessing of the collected real-time data also includes: drawing a plane diagram of the power supply equipment, energy storage equipment, and controllable loads in the virtual power plant based on the geographic location information, dividing the plane diagram into blocks according to the geographic coordinates to form a rectangular diagram composed of several small squares, and assigning a numerical value to each square to represent the value of a certain input feature in the geographic area, so as to form multiple features at the same time into a three-dimensional tensor.

[0008] Furthermore, it also includes: in response to user operations, displaying the results of the spatiotemporal network model prediction, scheduling decisions and real-time data on a visual interface to support user monitoring and operation.

[0009] In a second aspect, the present invention further provides a virtual power plant intelligent dispatching system based on a spatiotemporal network model, comprising a main control center, a plurality of distributed data acquisition and equipment control modules, and a network module, wherein the main control center and the distributed data acquisition and equipment control modules are connected via the network module; The distributed data acquisition and equipment control module is used to collect real-time data from sensors, smart meters, and weather stations in distributed power stations; the collected real-time data includes power supply equipment data, energy storage equipment data, controllable load data, weather data, date data, market and price data, manual control data, and geographic location information; Preprocess the collected real-time data and send the data to the main control center through the network module; wherein the preprocessing includes data cleaning, format conversion, anomaly detection, and data normalization; The main control center uses the spatiotemporal network basic unit to build an encoder-decoder based spatiotemporal network model; based on the preprocessed data, the spatiotemporal network model is used to predict the expected working power and expected load of each device in the virtual power plant; Based on the prediction results of the spatiotemporal network model, combined with the real-time status and actual constraints of the virtual power plant, intelligent scheduling decisions are made; wherein the real-time status of the virtual power plant includes real-time information on equipment operation, real-time information on demand, and business strategies; The distributed data acquisition and equipment control module is also used to convert the instructions of the main control center for intelligent scheduling decisions into actual actions for operating equipment.

[0010] Furthermore, the spatiotemporal network model includes a convolutional neural network and a recursive neural network; wherein, the convolutional neural network is used to transmit the hidden state value at the previous time point to the next node through spatial convolution processing, and the recursive neural network is used to calculate the output value and the new hidden state value based on the input data and the hidden state value at the previous time point.

[0011] Furthermore, the preprocessing of the collected real-time data also includes: drawing a plane diagram of the power supply equipment, energy storage equipment, and controllable loads in the virtual power plant based on the geographic location information, dividing the plane diagram into blocks according to the geographic coordinates to form a rectangular diagram composed of several small squares, and assigning a numerical value to each square to represent the value of a certain input feature in the geographic area, so as to form multiple features at the same time into a three-dimensional tensor.

[0012] Furthermore, it also includes an input module and an output module. The output module is used to respond to user operations and display the results, scheduling decisions and real-time data predicted by the spatiotemporal network model on a visual interface. The input module is used to support user monitoring and operation.

[0013] Compared with the prior art, the present invention has the following beneficial effects: (1) Through the end-to-end design of the spatiotemporal network model, comprehensive consideration of power supply equipment data, energy storage equipment data, controllable load data, meteorological data (such as rainfall, humidity, temperature and wind speed), date data, market and price data, manual control data and geographic location information can directly obtain detailed dispatching strategies and plans, reducing the operating costs and maintenance costs in the power dispatch process.

[0014] (2) It optimizes the working efficiency of the output equipment and energy storage equipment of distributed power stations, improves the satisfaction of user needs, realizes differentiated management of various equipment in different regions, improves the reliability of important equipment, and contributes to the peak-valley regulation and safe and stable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A schematic diagram of a flow chart of a virtual power plant intelligent scheduling method based on a spatiotemporal network model provided by an embodiment of the present invention; Figure 2 A schematic diagram of the structure of a spatiotemporal network model provided by an embodiment of the present invention; Figure 3 A logical diagram of a CRNN unit provided in an embodiment of the present invention; Figure 4 A schematic diagram of a three-dimensional tensor provided by an embodiment of the present invention; Figure 5 A schematic diagram of the workflow of the encoder and decoder provided in an embodiment of the present invention; Figure 6 A schematic structural diagram of a virtual power plant intelligent dispatching system based on a space-time network model provided in an embodiment of the present invention.

[0016] Reference numerals in the figure: 1. Main control center; 2. Distributed data acquisition and equipment control module; 3. Network module; 4. Storage and computing module; 5. Input module; 6. Output module; 7. Data collector; 8. Equipment control actuator. DETAILED DESCRIPTION

[0017] The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

[0018] The following describes the embodiments of the present disclosure through specific concrete examples, and those skilled in the art can easily understand other advantages and effects of the present disclosure from the contents disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all of the embodiments. The present disclosure can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present disclosure. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present disclosure.

[0019] Example 1 Existing technical literature on virtual power plant scheduling strategies primarily focuses on how to achieve centralized management, optimized scheduling, and resource allocation of distributed power sources through advanced algorithms and technologies. These technologies cover the entire process from data collection, processing, and analysis to scheduling decisions, aiming to improve the reliability, stability, and economic efficiency of power supply. Some technical literature proposes machine learning-based predictive models to forecast the output of distributed power sources and user demand, thereby improving scheduling decisions.

[0020] One of the defining characteristics of virtual power plants (VPPs) is that their power sources, loads, and energy storage devices are often geographically dispersed. This spatial distribution significantly impacts the formulation and implementation of dispatch strategies. Existing VPP dispatch strategies often focus on analyzing time series data for data collection and processing, while ignoring the spatial distribution of this data.

[0021] The spatial distribution of data allows for forward-looking forecasts of power output, loads, and energy storage devices. For example, severe temperature drops typically occur from north to south at a consistent rate. Wind data also exhibits significant time-varying variations along the wind's path. Failure to fully consider the spatial distribution of distributed power sources, loads, and energy storage devices can lead to power shortages or surpluses in certain areas, even if the overall system is functioning properly, impacting the overall efficiency and stability of the power system.

[0022] Based on this, see Figure 1 In the first aspect, this embodiment provides a virtual power plant intelligent scheduling method based on a spatiotemporal network model, including five parts: data collection, data preprocessing, spatiotemporal network model, scheduling decision-making, and user interaction. The specific steps are as follows: S100 to S500: S100. Collect real-time data from sensors, smart meters, and weather stations in distributed power stations. The collected real-time data includes power supply equipment data, energy storage equipment data, controllable load data, weather data, date data, market and price data, manual control data, and geographic location information. Weather data includes rainfall, humidity, temperature, and wind speed.

[0023] S200, pre-processing the collected real-time data and sending the data to the main control center through the network module; wherein the pre-processing includes data cleaning, format conversion, anomaly detection, and data normalization.

[0024] S300. Utilize the spatiotemporal network basic unit to construct a spatiotemporal network model based on an encoder-decoder; based on the preprocessed data, utilize the spatiotemporal network model to predict the expected operating power and expected load of each device in the virtual power plant.

[0025] S400. Make intelligent scheduling decisions based on the prediction results of the spatiotemporal network model and the real-time status and actual constraints of the virtual power plant; wherein the real-time status of the virtual power plant includes real-time information on equipment operation, real-time information on demand, and business strategies.

[0026] S500 , in response to user operations, displaying the prediction results, scheduling decisions and real-time data of the spatiotemporal network model on a visual interface to support user monitoring and operations.

[0027] The spatiotemporal network model of the method of this embodiment is used to predict the expected working power and expected load of each device in the virtual power plant, and obtain the operation plan (or operation diagram) of the device. The spatiotemporal network model takes into account historical electrical data, historical meteorological data (such as rainfall, humidity, temperature and wind speed), date data, device parameter data, market operation data and geographical location information, and can directly obtain the operation diagram of the device. The data at each moment is input into the network model to predict the future The operation diagram at each moment; the values of and are determined by comprehensively considering the calculation cost and control effect, and are continuously iterated over time. In case of emergency or special needs, manual intervention can also be carried out.

[0028] The method of this embodiment, through the end-to-end design of a spatiotemporal network model, comprehensively considers data from power supply equipment, energy storage equipment, controllable loads, meteorological data (such as rainfall, humidity, temperature, and wind speed), date data, market and price data, manual control data, and geographic location information. This allows for direct generation of detailed dispatch strategies and plans, reducing operating and maintenance costs during power dispatch. This optimizes the efficiency of distributed power station output equipment and energy storage equipment, improves user satisfaction, enables differentiated management of various equipment in different regions, enhances the reliability of key equipment, and contributes to peak and valley regulation and safe and stable operation of the power grid.

[0029] For further information, see Figure 2 , Figure 2 The structure of the spatiotemporal network model is shown. The spatiotemporal network model includes a convolutional neural network and a recurrent neural network, and comprehensively considers the spatial distribution and temporal distribution characteristics of the features. Figure 2 Each directed line segment in represents the signal transmitted to the next node after spatial convolution processing.

[0030] The circular patterns represent inputs and outputs. represents the input data of the network, Represents the output data of the network. These results are multi-dimensional vectors, also known as tensors.

[0031] The rectangular pattern represents a spatiotemporal network basic unit (CRNN unit for short), which is the core part of the spatiotemporal network model. It is used to calculate the output value and the new hidden state value based on the input data and the hidden state value at the previous time point.

[0032] and 、 、... Represents the hidden state value of the CRNN unit, 0, 1, 2...t. represents the time sequence, and the subscript init represents the initial value.

[0033] 、 、... Represents the input value of the spatiotemporal network.

[0034] The hidden state value at the previous time point 、 and the input value at the current time point Together they determine the output of the current state , new hidden state and .

[0035] See also Figure 3 , Figure 3 It is the logical diagram of CRNN unit. Figure 3 The circular box in the figure indicates that the multiplication or addition operation is performed between the tensors according to the position. The rectangular box indicates that the activation operation is performed on each element in the tensor. The merging of the line segments indicates that two tensors are spliced. The output bifurcation indicates that the values of multiple branches are equal, that is, the tensor is copied. 、 In addition, the figure also marks 、 、 The calculation method of each variable is as follows: Among them, * represents convolution operation. The sign indicates the corresponding multiplication of the components, σ indicates the sigmoid activation function, and tanh indicates the hyperbolic tangent activation function. is the convolution kernel, the first letter of its subscript represents the source data, and the second letter represents the result data. Indicates that the source data To the result data The convolution kernel. Denotes the bias. Both the convolution kernel and the bias are parameters of the spatiotemporal network and need to be trained using input data.

[0036] Furthermore, sufficient high-quality data is crucial for the training of neural networks. In spatiotemporal networks, the data that can be used include power supply equipment data, energy storage equipment data, controllable load data, meteorological data, date data, market and price data, manual control data, and geographic location information. Meteorological data includes sample values of temperature, humidity, wind speed, wind direction, rainfall, etc. in time and space dimensions; Calendar data includes the current date's month, day of each month, day of the week, whether it is a holiday, etc. Power equipment data includes the output ratio of various new energy sources, maximum processing power, optimal output power, and output ratio of conventional units; Energy storage equipment data includes the charging and discharging power of the energy storage device; Controllable load data includes controllable load power, demand response strategy, etc. Market and price data include electricity market prices, carbon emission rights prices, etc.; Manual control of data includes adding data as needed, such as system constraints, when managers deem it necessary; The geographic location information data includes the address location information of each power supply device, energy storage device, controllable load, etc.

[0037] The collected data are preprocessed, including removing outliers, filling missing values and data normalization, to ensure the accuracy and consistency of the data.

[0038] Draw a plane graph of the power supply equipment, energy storage equipment, and controllable load in the virtual power plant based on the geographic location information, and divide the plane graph into blocks to form a rectangular graph composed of several small squares. Each square is assigned a value to represent the value of an input feature in the geographic area, so that multiple features at the same time can be formed into a three-dimensional tensor, such as Figure 4 shown.

[0039] The output of the neural network is similar to the input, also a multidimensional tensor, and the features of each square correspond one-to-one to the squares of the corresponding input tensor.

[0040] Furthermore, in addition to the input module, output module, and CRNN unit, the spatiotemporal network model also includes essential foundational modules such as loss functions and optimizers. Depending on the needs of the actual case, batch normalization layers (a regularization method that accelerates network convergence and improves generalization) and dropout layers (to enhance generalization) can also be added. To address the problem of input and output spanning multiple time instants, the present invention also utilizes an encoder-decoder architecture.

[0041] The spatiotemporal network model uses the mean squared error (MSE) as a loss function, using the data at the next moment as the target value for the current data prediction. The closer the two are, the smaller the MSE, indicating better network performance.

[0042] The output layer also includes an optimizer, which updates the network parameters through the error back propagation algorithm. A variety of optimizers including Adam can be used to optimize the network.

[0043] When training the spatiotemporal network model, the data at the current moment is used as input and the data at the next moment is used as output. The system optimizes the network parameters through the back propagation algorithm of the error.

[0044] Furthermore, in actual use, spatiotemporal network models generally input 、 ... common data at a moment in time and predict the future Output at each moment 、 ... , called the encoder-decoder structure, which means The information of the data at each moment is encoded and stored in the hidden state 、 and use it to predict the future The output at each moment is Figure 6 The intelligent scheduling system will perform scheduling based on the output results of the spatiotemporal network.

[0045] See also Figure 5 , Figure 5 Schematic diagram of the encoder and decoder workflow. Figure 5 The encoder on the left encodes the input data, and the decoder on the right produces the actual output. The encoder and decoder are the same neural network, both built on the basis of CRNN units. The encoder input is the actual data, while the decoder is a prediction of the future. There is no actual data at the input, but the output data of the network at the previous moment is used as the input data of the next moment. The encoder also produces output , , but only the output value of the last moment is used as the first input to the decoder and all other output values are discarded.

[0046] Figure 5 The subscripts 0, 1, and 2 in the example are used only as examples of time series and do not imply that only three time steps can be processed. In practice, the number of input and prediction sequences is independent and can be freely set. To achieve optimal results, the number of input data steps can be larger and the number of output steps can be smaller. This is determined based on the actual data, prediction results, and requirements during training and evaluation of the spatiotemporal network.

[0047] Through the above implementation, the virtual power plant intelligent scheduling method based on the spatiotemporal network model of this embodiment can achieve precise control of the operating power of each device in a distributed power station and formulate an optimal intelligent scheduling plan accordingly. This not only improves the reliability and economic efficiency of power supply, but also helps optimize the energy structure, promote energy conservation and emission reduction, and promote sustainable development.

[0048] Example 2 Based on the above implementation, please refer to Figure 6Secondly, this embodiment also provides a virtual power plant intelligent dispatching system based on a spatiotemporal network model, comprising a main control center 1, a distributed data acquisition and equipment control module 2, and a network module 3. The main control center 1 and the distributed data acquisition and equipment control module 2 are connected via the network module 3. The main control center 1 includes a storage and computing module 4, an input module 5, and an output module 6.

[0049] The storage and computing module 4 is the brain of the system, responsible for running various required software, the most important of which is the intelligent scheduling method based on the spatiotemporal network model. Other software also includes processing user input and output, network communication, etc. The input module 5 is used to receive user operations, such as system settings and manual operations. The output module 6 is used to display system status and results. The network module 3 is responsible for communicating with the various distributed data acquisition and device control modules 2.

[0050] The distributed data acquisition and device control module 2, typically used together with the virtual power plant's power supply, energy storage, and controllable load devices, is used to collect data and receive dispatch instructions from the main control center 1. A single main control center 1 can be connected to multiple distributed data acquisition and device control modules 2. The data acquisition and device control module 2 also has a storage and computation module 4, an input module 5, and an output module 6, whose functions are similar to those of the corresponding modules in the main control center 1. In addition, it also has a data acquisition unit 7 and a device control actuator 8. The data acquisition unit 7 can connect to various data acquisition devices and transmit them to the main control center 1 via the network module 3. The device control actuator 8 converts the instructions from the main control center 1 into actions that actually operate the equipment.

[0051] In this embodiment, the virtual power plant intelligent dispatching system based on the space-time network model runs the virtual power plant intelligent dispatching method based on the space-time network model of any of the above embodiments. Specifically, The distributed data acquisition and equipment control module 2 is used to collect real-time data from sensors, smart meters and weather stations in distributed power stations. The collected real-time data includes power supply equipment data, energy storage equipment data, controllable load data, meteorological data (such as rainfall, humidity, temperature and wind speed), date data, market and price data, manual control data and geographical location information; Preprocess the collected data and send the data to the main control center 1 through the network module 3; wherein the preprocessing includes data cleaning, format conversion, anomaly detection, data normalization, etc.; The main control center 1 uses the spatiotemporal network basic unit to build an encoder-decoder based spatiotemporal network model; based on the preprocessed data, the spatiotemporal network model is used to predict the expected working power and expected load of each device in the virtual power plant; Based on the prediction results of the spatiotemporal network model, combined with the real-time status and actual constraints of the virtual power plant, intelligent scheduling decisions are made; wherein the real-time status of the virtual power plant includes real-time information on equipment operation, real-time information on demand, and business strategies; The distributed data acquisition and equipment control module 2 is also used to convert the instructions of the main control center 1 for intelligent scheduling decisions into actual actions for operating equipment.

[0052] Furthermore, the spatiotemporal network model includes a convolutional neural network and a recursive neural network; wherein, the convolutional neural network is used to transmit the hidden state value at the previous time point to the next node through spatial convolution processing, and the recursive neural network module is used to calculate the output value and the new hidden state value based on the input data and the hidden state value at the previous time point.

[0053] Furthermore, the preprocessing of the collected data also includes: drawing a plane diagram of the power supply equipment, energy storage equipment, and controllable loads in the virtual power plant based on the geographic location information, dividing the plane diagram into blocks based on the geographic coordinates to form a rectangular diagram consisting of several small squares, and assigning a numerical value to each square to represent the value of a certain input feature in the geographic area, so as to form multiple features at the same time into a three-dimensional tensor.

[0054] Furthermore, it also includes an input module 5 and an output module 6. The output module 6 is used to respond to user operations and display the results, scheduling decisions and real-time data predicted by the spatiotemporal network model on a visual interface. The input module 5 is used to support user monitoring and operation.

[0055] The virtual power plant intelligent dispatching system based on the space-time network model of this embodiment reduces the operating costs and maintenance costs in the power dispatching process through intelligent dispatching and real-time monitoring.

[0056] The above is only for explaining the embodiments of the present invention and is not intended to limit the present invention. For those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present invention without creative work should be included in the scope of protection of the present invention.

Claims

1. A virtual power plant intelligent scheduling method based on a spatiotemporal network model, characterized in that: include: Collect real-time data from sensors, smart meters, and weather stations at distributed power stations. This includes power supply equipment data, energy storage equipment data, controllable load data, weather data, date data, market and price data, manual control data, and geographic location information. Preprocessing the collected real-time data; wherein the preprocessing includes data cleaning, format conversion, anomaly detection, and data normalization; Using the spatiotemporal network basic unit, a spatiotemporal network model based on an encoder-decoder is constructed; based on the preprocessed data, the spatiotemporal network model is used to predict the expected operating power and expected load of each device in the virtual power plant; Based on the prediction results of the spatiotemporal network model, combined with the real-time status and actual constraints of the virtual power plant, intelligent scheduling decisions are made; wherein the real-time status of the virtual power plant includes real-time information on equipment operation, real-time information on demand, and business strategies.

2. The method for intelligent scheduling of virtual power plants based on a spatiotemporal network model according to claim 1 is characterized in that: The spatiotemporal network model includes a convolutional neural network and a recursive neural network; wherein the convolutional neural network is used to transmit the hidden state value at the previous time point to the next node through spatial convolution processing, and the recursive neural network is used to calculate the output value and the new hidden state value based on the input data and the hidden state value at the previous time point.

3. The method for intelligent scheduling of virtual power plants based on a spatiotemporal network model according to claim 1 is characterized in that: The preprocessing of the collected real-time data also includes: drawing a plane diagram of the power supply equipment, energy storage equipment, and controllable loads in the virtual power plant based on the geographic location information, dividing the plane diagram into blocks based on the geographic coordinates to form a rectangular diagram consisting of a number of small squares, and assigning a numerical value to each square to represent the value of a certain input feature in the geographic area, so as to form multiple features at the same time into a three-dimensional tensor.

4. The method for intelligent scheduling of virtual power plants based on a spatiotemporal network model according to claim 1, characterized in that: Also includes: In response to user operations, the prediction results, scheduling decisions and real-time data of the spatiotemporal network model are displayed on a visual interface to support user monitoring and operations.

5. A virtual power plant intelligent dispatching system based on a spatiotemporal network model, characterized in that: It includes a main control center, a plurality of distributed data acquisition and device control modules and a network module, wherein the main control center and the distributed data acquisition and device control modules are connected via the network module; The distributed data acquisition and equipment control module is used to collect real-time data from sensors, smart meters, and weather stations in distributed power stations; the collected real-time data includes power supply equipment data, energy storage equipment data, controllable load data, weather data, date data, market and price data, manual control data, and geographic location information; Preprocess the collected real-time data and send the data to the main control center through the network module; wherein the preprocessing includes data cleaning, format conversion, anomaly detection, and data normalization; The main control center uses the spatiotemporal network basic unit to build an encoder-decoder based spatiotemporal network model; based on the preprocessed data, the spatiotemporal network model is used to predict the expected working power and expected load of each device in the virtual power plant; Based on the prediction results of the spatiotemporal network model, combined with the real-time status and actual constraints of the virtual power plant, intelligent scheduling decisions are made; wherein the real-time status of the virtual power plant includes real-time information on equipment operation, real-time information on demand, and business strategies; The distributed data acquisition and equipment control module is also used to convert the instructions of the main control center for intelligent scheduling decisions into actual actions for operating equipment.

6. The virtual power plant intelligent dispatching system based on the spatiotemporal network model according to claim 5 is characterized in that: The spatiotemporal network model includes a convolutional neural network and a recursive neural network; wherein the convolutional neural network is used to transmit the hidden state value at the previous time point to the next node through spatial convolution processing, and the recursive neural network is used to calculate the output value and the new hidden state value based on the input data and the hidden state value at the previous time point.

7. The virtual power plant intelligent dispatching system based on the spatiotemporal network model according to claim 5 is characterized in that: The preprocessing of the collected real-time data also includes: drawing a plane diagram of the power supply equipment, energy storage equipment, and controllable loads in the virtual power plant based on the geographic location information, dividing the plane diagram into blocks based on the geographic coordinates to form a rectangular diagram consisting of a number of small squares, and assigning a numerical value to each square to represent the value of a certain input feature in the geographic area, so as to form multiple features at the same time into a three-dimensional tensor.

8. The virtual power plant intelligent dispatching system based on the spatiotemporal network model according to claim 5 is characterized in that: It also includes an input module and an output module. The output module is used to respond to user operations and display the results, scheduling decisions and real-time data predicted by the spatiotemporal network model on a visual interface. The input module is used to support users in monitoring and operations.

Images