Energy intelligent management method, device and equipment, and storage medium

Abstract

The application provides an energy intelligent management method and device, equipment and a storage medium, wherein the method comprises: collecting historical energy use data by using a multi-source acquisition method, preprocessing the historical energy use data to obtain a standard data set; constructing a regional change space-time graph of energy use, wherein a node in the regional change space-time graph represents a sub-region, the node features include features in the standard data set, an edge represents an energy transmission relationship between sub-regions, and a space-time graph neural network is used to output energy consumption patterns of each sub-region and energy output prediction values; obtaining historical energy demand and environmental data as external data, obtaining energy-related feedback information text through a crawler, inputting the external data and the energy-related feedback information text into a multi-modal time series model, and outputting energy demand prediction values; calculating the difference between the energy output prediction values and the energy demand prediction values of each sub-region as a management adjustment amount, and formulating an energy transmission mode between sub-regions based on the management adjustment amount.

Description

Technical Field

[0001] The present invention relates to the field of big data management technology, and in particular to an intelligent energy management method, device, equipment and storage medium. Background Art

[0002] Intelligent energy governance refers to the use of advanced intelligent control technologies to efficiently manage and optimize energy production, transmission, distribution and consumption, with the goal of improving the efficiency and reliability of energy systems.

[0003] Among the related technologies, current intelligent energy forecasting mainly relies on statistical methods and machine learning algorithms to output energy output forecasts or energy demand forecasts. Although statistical methods such as time series models and regression analysis can capture the basic trends and seasonal changes in energy changes, they are unable to handle complex nonlinear energy forecasts. Existing machine learning algorithms, especially for energy demand forecasting tasks, do not consider subjective dynamic factors such as the influence of feedback information, which affects the model's data processing and analysis capabilities. For example, discussions on social media can reflect the public's attitudes towards energy prices, supply stability, and environmental protection issues, thereby affecting their energy use behavior. In addition, there are no technical means in the field that integrate a wide range of data sources, efficient and standard data processing and integration processes, and intelligent prediction and governance solutions, which need to be continuously improved and optimized.

[0004] Based on the above analysis of the development status of this technology field, the existing technology lacks a solution that can intelligently predict energy management based on both energy output forecasts and energy demand forecasts, and consider the characteristics of feedback information in energy demand forecasts. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent energy management method, device, equipment and storage medium, aiming to solve the above-mentioned problems in the prior art.

[0006] According to a first aspect of an embodiment of the present invention, there is provided a method for intelligent energy management, comprising:

[0007] Use multi-source acquisition to collect historical energy usage data, pre-process the historical energy usage data to obtain a standard data set;

[0008] Construct a spatiotemporal graph of regional energy usage changes, where nodes represent subregions, node features include features from a standard dataset, and edges represent energy transfer relationships between subregions. Use a spatiotemporal graph neural network to output energy consumption patterns and energy output forecasts for each subregion.

[0009] Obtain historical energy demand and environmental data as external data, obtain energy-related feedback information text through crawlers, input external data and energy-related feedback information text into the multimodal time series model, and output energy demand forecast;

[0010] The difference between the predicted energy output and energy demand of each sub-region is calculated as the governance adjustment amount, and the energy transmission method between sub-regions is formulated based on the governance adjustment amount.

[0011] According to a second aspect of an embodiment of the present invention, there is provided an intelligent energy management device, comprising:

[0012] The collection and preprocessing module is used to collect historical energy usage data using a multi-source acquisition method, preprocess the historical energy usage data, and obtain a standard data set;

[0013] The energy usage prediction module is used to construct a spatiotemporal graph of regional energy usage changes. The nodes in the spatiotemporal graph represent subregions, and node features include features from standard datasets. Edges represent energy transfer relationships between subregions. A spatiotemporal graph neural network is used to output energy consumption patterns and energy output forecasts for each subregion.

[0014] The energy demand forecasting module is used to obtain historical energy demand and environmental data as external data, obtain energy-related feedback information text through crawlers, input the external data and energy-related feedback information text into the multimodal time series model, and output the energy demand forecast;

[0015] The governance adjustment module is used to calculate the difference between the predicted energy output and energy demand of each sub-region as the governance adjustment amount, and formulate the energy transmission method between sub-regions based on the governance adjustment amount.

[0016] According to a third aspect of an embodiment of the present invention, an electronic device is provided, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the steps of the intelligent energy management method provided in the first aspect of the present disclosure are implemented.

[0017] According to a fourth aspect of an embodiment of the present invention, a computer-readable storage medium is provided, on which an information transmission implementation program is stored. When the program is executed by a processor, the steps of the energy intelligent management method provided in the first aspect of the present disclosure are implemented.

[0018] The technical solution provided by the embodiment of the present invention includes the following beneficial effects: determining intelligent governance solutions from both energy demand and energy output forecasting aspects simultaneously, avoiding decision-making bias caused by one-sided reliance on a single factor; taking into account energy-related feedback information texts that can reflect the public's emotions and discussions on energy prices, supply stability and environmental protection issues when forecasting demand, so as to fully consider the impact of driving-side changes on demand and improve forecast accuracy.

[0019] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate one or more embodiments of this specification or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments recorded in this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0021] Figure 1 This is a flow chart of the energy intelligent management method according to an embodiment of the present invention;

[0022] Figure 2 This is a schematic diagram of an intelligent energy management device according to an embodiment of the present invention;

[0023] Figure 3 is a schematic diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0024] In order to enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the technical solutions in one or more embodiments of this specification will be clearly and completely described below in conjunction with the drawings in one or more embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on one or more embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this document.

[0025] Method Example

[0026] According to an embodiment of the present invention, a method for intelligent energy management is provided. Figure 1 This is a flow chart of the energy intelligent management method according to an embodiment of the present invention. Figure 1 As shown, the energy intelligent management method according to an embodiment of the present invention specifically includes:

[0027] In step S110, historical energy usage data is collected using a multi-source acquisition method, and the historical energy usage data is pre-processed to obtain a standard data set, which specifically includes:

[0028] With the rapid growth and diversification of energy data, energy data comes from a wide range of sources. Therefore, historical energy usage data is collected from energy devices, sensor devices, and smart meters. The historical energy usage data is formatted after being received. Historical energy usage data refers to indicators related to energy usage in a real environment. The embodiments of the present invention do not limit specific indicators, but include various types of energy usage, transportation volume, etc.

[0029] Clean, remove duplicates, and fill missing values ​​in historical energy usage data to obtain a standard data set;

[0030] Preferably, in order to improve the security performance during the model training process, the standard data set is divided into multiple sub-data sets and distributed to multiple participants. Each participant holds part of the data, and the MPC protocol is used for joint training to ensure that the data is not leaked during the calculation process; or the standard data set is encrypted using symmetric encryption or asymmetric encryption. The data set exists in encrypted form during storage and transmission, and the key is used to decrypt the data during subsequent training. The decrypted data exists only in the memory and is not written to the disk.

[0031] In step S120, a spatiotemporal graph of regional energy usage is constructed, wherein nodes in the spatiotemporal graph represent subregions, node features include features from the standard dataset, and edges represent energy transmission relationships between subregions. The spatiotemporal graph neural network is used to output the energy consumption pattern and energy output forecast of each subregion, specifically including:

[0032] By introducing the time dimension based on the standard dataset, a spatiotemporal map of regional changes is constructed;

[0033] Obtain a spatiotemporal graph neural network including a graph convolution module GCN and a temporal convolution module TCN;

[0034] The graph convolution module is used to extract spatial features in the regional change spatiotemporal graph to capture the mutual influence between sub-regions. The temporal convolution module is used to extract temporal features in the regional change spatiotemporal graph to capture the temporal evolution of energy consumption. Temporal features include periodic and seasonal patterns. The spatial and temporal features are integrated to obtain a unified spatiotemporal representation.

[0035] Predict the energy consumption pattern and energy output of each sub-region based on the unified spatiotemporal representation;

[0036] In the embodiment of the present invention, energy consumption patterns can be divided into industrial, commercial and residential according to industry; and can be divided into electricity, fossil fuels and renewable energy according to type.

[0037] In step S130, historical energy demand and environmental data are obtained as external data, energy-related feedback information text is obtained through crawlers, the external data and energy-related feedback information text are input into the multimodal time series model, and energy demand forecast is output, which specifically includes:

[0038] Obtain meteorological data and economic data as environmental data;

[0039] Compared to energy output forecasts, energy demand forecasts require greater attention to the impact of feedback information. Social media feedback can reflect the public's acceptance of energy regulation. Strong opposition to a plan on social media may lead to policy delays or revisions, thereby affecting energy demand.

[0040] Through crawling, we extract content related to energy prices, supply stability, and environmental protection issues from local social forums in the region, use the topic model algorithm to extract hot events in the content, and use the sentiment classification model to identify the emotions reflected in the hot events. The hot events and corresponding emotions are combined as energy-related feedback information text.

[0041] Obtain a multimodal time series model including LSTM modules and Transformer modules;

[0042] Use the LSTM module to extract the first feature corresponding to the external data, and use the Transformer module to extract the second feature corresponding to the energy-related feedback information text. Establish the cross attention between the first and second features, and use Formula 1 to calculate the cross attention:

[0043] Formula 1;

[0044] in, Indicates cross attention, represents the activation function, Represents the first feature, represents the transpose of the second feature, Represents feature dimensions; the core goal of cross-attention is not to directly adjust the weights of different features, but to explore the interactive relationship between different features;

[0045] The first feature, the second feature and the cross attention are fused to obtain a unified multimodal representation. Based on the unified multimodal representation, the energy demand forecast of each sub-region is predicted through the output layer.

[0046] It should be noted that the embodiment of the present invention uses a TGN network for extraction in step S120, and an LSTM network for extracting demand features in step S130. This is because the regional change spatiotemporal graph has a graph structure, and TGN can capture the complex relationship patterns between nodes and allow customized information transmission between nodes; while LSTM is more suitable for general time series prediction.

[0047] In step S140, the difference between the predicted energy output and the predicted energy demand of each sub-region is calculated as the governance adjustment amount. Based on the governance adjustment amount, the energy transmission mode between the sub-regions is formulated, which specifically includes:

[0048] Calculating the difference between the predicted energy output and the predicted energy demand can help us understand the energy gap.

[0049] Maintain the sub-regions with differences greater than or equal to zero and sort them in descending order of difference to obtain a positive list. Maintain the sub-regions with differences less than zero and sort them in descending order of absolute difference to obtain a negative list. The higher the position in the positive list, the greater the energy surplus at the end of the forecast period. The higher the position in the negative list, the greater the energy gap at the end of the forecast period.

[0050] Obtaining a predefined energy transmission rule table, wherein the energy transmission rule table defines that each sub-area can transmit energy to other sub-areas, because in actual applications, not every two sub-areas can transmit energy, or the types of transmission are limited;

[0051] Traverse the sub-areas in the negative list one by one, and select the sub-area that is ranked first in the positive list and is allowed to be transmitted to this area in the energy transmission rule table as the scheduling point. Adjust the positive list and negative list in real time, and obtain the energy transmission method after one traversal.

[0052] Assume that the negative list is F1, then F2; the positive list is Z1, then Z2, then Z3. First, we need to develop an energy transfer control method for F1, which has the most gap. The energy transfer rule table allows transfers between F1, Z2, and Z3. Therefore, Z1, which has the largest surplus, should transfer to F1. However, since the set rules do not allow it, Z2 starts to transfer to F1, until F1 is adjusted to positive or the traversal is completed.

[0053] Regardless of whether F1 becomes positive, the next adjustment will be F2's turn, because intelligent management is based on predicted values ​​rather than current values, and not all sub-regions will eventually have a supply exceeding demand. Energy transmission methods do not necessarily need to be implemented immediately, but can be achieved gradually during the forecast period.

[0054] The method further comprises:

[0055] In step S150, the energy transmission mode is visualized on the regional map, and the energy usage of each sub-region and the energy transmission between sub-regions are monitored in real time, specifically including:

[0056] By creating interactive dashboards or simulation software for visualization, the focus is on displaying key energy indicators and providing early warning functions, issuing alarms when a sub-area exceeds the set threshold.

[0057] To sum up, in response to the current problems, this invention proposes an intelligent energy governance method, which determines the intelligent governance plan from the two forecast aspects of energy demand and energy output at the same time, avoiding decision-making bias caused by one-sided reliance on a single factor; when forecasting demand, it takes into account energy-related feedback information texts that can reflect the public's emotions and discussions on energy prices, supply stability and environmental protection issues, so as to fully consider the impact of driving-side change factors on demand and improve forecast accuracy; when forecasting energy demand, it takes into account the cross-attention between the first feature corresponding to the external data and the second feature corresponding to the feedback information text, and combines the complex relationship between learning features across domains, which helps to reveal potential causal relationships; in the process of formulating the energy transmission method, a round of adjustments can be used to quickly maintain the supply and demand balance to the greatest extent, thereby improving the computational efficiency of intelligent governance; the overall solution comprehensively collects data, efficiently processes, accurately analyzes and makes intelligent decisions.

[0058] Device embodiment

[0059] According to an embodiment of the present invention, an intelligent energy management device is provided. Figure 2 Schematic diagram of an intelligent energy management device according to an embodiment of the present invention. Figure 2 As shown, the device according to the embodiment of the present invention specifically includes:

[0060] The collection and preprocessing module 20 is used to collect historical energy usage data using a multi-source acquisition method, preprocess the historical energy usage data, and obtain a standard data set. Specifically, it is used to:

[0061] Collect historical energy usage data from energy devices, sensor devices, and smart meters;

[0062] The historical energy usage data are cleaned, deduplicated and missing values ​​filled to obtain a standard data set.

[0063] The energy usage prediction module 22 is used to construct a spatiotemporal graph of regional energy usage changes, where nodes in the spatiotemporal graph represent subregions, node features include standard data concentration features, and edges represent energy transmission relationships between subregions. The spatiotemporal graph neural network is used to output the energy consumption pattern and energy output forecast of each subregion. Specifically, it is used to:

[0064] Obtain a spatiotemporal graph neural network including a graph convolution module and a temporal convolution module;

[0065] The graph convolution module is used to extract spatial features in the regional change spatiotemporal graph, and the time series convolution module is used to extract temporal features in the regional change spatiotemporal graph. Temporal features include periodic and seasonal patterns. The spatial and temporal features are fused to obtain a unified spatiotemporal representation.

[0066] The energy consumption pattern and energy output forecast of each sub-region are predicted based on the unified spatiotemporal representation.

[0067] The energy demand forecasting module 24 is used to obtain historical energy demand and environmental data as external data, obtain energy-related feedback information text through a crawler, input the external data and energy-related feedback information text into a multimodal time series model, and output energy demand forecasts. Specifically, it is used to:

[0068] Obtain meteorological data and economic data as environmental data;

[0069] Through crawling, we extract content related to energy prices, supply stability, and environmental protection issues from local social forums in the region, use the topic model algorithm to extract hot events in the content, and use the sentiment classification model to identify the emotions reflected in the hot events. The hot events and corresponding emotions are combined as energy-related feedback information text.

[0070] Obtain a multimodal time series model including LSTM modules and Transformer modules;

[0071] Use the LSTM module to extract the first feature corresponding to the external data, use the Transformer module to extract the second feature corresponding to the energy-related feedback information text, and establish cross-attention between the first and second features;

[0072] The first feature, the second feature and the cross attention are fused to obtain a unified multimodal representation, and the energy demand forecast of each sub-region is predicted based on the unified multimodal representation.

[0073] The governance adjustment module 26 is used to calculate the difference between the predicted energy output and the predicted energy demand of each sub-region as the governance adjustment amount, and formulate the energy transmission method between the sub-regions based on the governance adjustment amount, specifically for:

[0074] Maintain the sub-regions whose difference is greater than or equal to zero and sort them in descending order according to the difference to obtain a positive list; maintain the sub-regions whose difference is less than zero and sort them in descending order according to the absolute value of the difference to obtain a negative list;

[0075] Obtaining a predefined energy transmission rule table, wherein the energy transmission rule table indicates a definition of whether each sub-region can transmit energy to other sub-regions;

[0076] Traverse the sub-areas in the negative list one by one, and take the sub-area that is ranked first in the positive list and is allowed to be transmitted to this place in the energy transmission rule table as the scheduling point. Adjust the positive list and negative list in real time, and get the energy transmission method after one traversal.

[0077] The device further comprises:

[0078] The visualization monitoring module 28 is used to visualize the energy transmission mode in the regional map and monitor the energy usage of each sub-region and the energy transmission between sub-regions in real time.

[0079] To sum up, in response to the existing problems, the energy intelligent management device invented this time determines the intelligent management plan from the two forecast aspects of energy demand and energy output, avoiding decision-making bias caused by one-sided reliance on a single factor; when forecasting demand, it takes into account the energy-related feedback information text that can reflect the public's emotions and discussions on energy prices, supply stability and environmental protection issues, so as to fully consider the impact of driving-side change factors on demand and improve forecast accuracy; when forecasting energy demand, it takes into account the cross-attention between the first feature corresponding to the external data and the second feature corresponding to the feedback information text, and combines the complex relationship between learning features across fields, which helps to reveal the potential causal relationship; in the process of formulating the energy transmission method, a round of adjustments can be used to quickly maintain the supply and demand balance to the greatest extent, thereby improving the computational efficiency of intelligent management; the overall plan comprehensively collects data, efficiently processes, accurately analyzes and makes intelligent decisions.

[0080] Electronic device embodiment

[0081] Figure 3 is a schematic diagram of an electronic device according to an embodiment of the present invention. Electronic device 300 may include at least one processor 310 and memory 320. Processor 310 can execute instructions stored in memory 320. Processor 310 is communicatively coupled to memory 320 via a data bus. In addition to memory 320, processor 310 may also be communicatively coupled to input device 330, output device 340, and communication device 350 via the data bus.

[0082] The processor 310 may be any conventional processor, such as a commercially available CPU. The processor may also include a graphics processor (GPU), a field programmable gate array (FPGA), a system on chip (SOC), an application specific integrated circuit (ASIC), or a combination thereof.

[0083] The memory 320 may be implemented by any type of volatile or non-volatile memory device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk.

[0084] In the embodiment of the present disclosure, executable instructions are stored in the memory 320, and the processor 310 can read the executable instructions from the memory 320 and execute the instructions to implement all or part of the steps of any of the energy intelligent management methods in the above exemplary embodiments.

[0085] Computer readable storage medium embodiments

[0086] In addition to the above-mentioned methods and devices, the exemplary embodiments of the present disclosure may also be a computer program product or a computer-readable storage medium storing the computer program product, wherein the computer product includes computer program instructions that can be executed by a processor to implement all or part of the steps described in any of the energy intelligent management methods in the above-mentioned exemplary embodiments.

[0087] The computer program product may be written in any combination of one or more programming languages ​​to write program code for performing the operations of the embodiments of the present application, including object-oriented programming languages ​​such as Java, C++, etc., as well as conventional procedural programming languages ​​such as "C" or similar programming languages ​​and scripting languages ​​(e.g., Python). The program code may be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0088] Computer-readable storage media may be any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may include, for example, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or components, or any combination thereof. More specific examples of readable storage media include: static random access memory (SRAM) with one or more wires electrically connected, electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disk, or any suitable combination thereof.

[0089] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent energy management method, characterized in that: include: Collecting historical energy usage data using a multi-source acquisition method, and preprocessing the historical energy usage data to obtain a standard data set; Construct a spatiotemporal graph of regional energy usage changes, where nodes in the spatiotemporal graph represent subregions, node features include the features in the standard dataset, and edges represent energy transmission relationships between subregions. Use a spatiotemporal graph neural network to output energy consumption patterns and energy output forecasts for each subregion, specifically including: Obtain a spatiotemporal graph neural network including a graph convolution module and a temporal convolution module; The graph convolution module is used to extract spatial features in the regional change spatiotemporal graph, and the temporal convolution module is used to extract temporal features in the regional change spatiotemporal graph, wherein the temporal features include periodic patterns and seasonal patterns, and the spatial features and temporal features are fused to obtain a unified spatiotemporal representation; predicting energy consumption patterns and energy output forecasts for each sub-region based on the unified spatiotemporal representation; Acquire historical energy demand and environmental data as external data, obtain energy-related feedback information text through crawlers, input the external data and the energy-related feedback information text into a multimodal time series model, and output energy demand forecast, specifically including: Acquiring meteorological data and economic data as the environmental data; Using a crawler to extract content related to energy prices, supply stability, and environmental issues from local social forums, a topic model algorithm was used to extract hot topics from the content. A sentiment classification model was used to identify the emotions reflected in these hot topics. These hot topics and corresponding emotions were combined as energy-related feedback information text. Obtain a multimodal time series model including LSTM modules and Transformer modules; Using the LSTM module to extract a first feature corresponding to the external data, using the Transformer module to extract a second feature corresponding to the energy-related feedback information text, and establishing a cross-attention between the first feature and the second feature; fusing the first feature, the second feature, and the cross attention to obtain a unified multimodal representation, and predicting the energy demand forecast of each sub-region according to the unified multimodal representation; The difference between the predicted energy output and the predicted energy demand of each sub-region is calculated as the governance adjustment amount, and the energy transmission method between the sub-regions is formulated based on the governance adjustment amount.

2. The method according to claim 1, characterized in that The method further comprises: The energy transmission mode is visualized in a regional map, and the energy usage of each sub-region and the energy transmission between sub-regions are monitored in real time.

3. The method according to claim 1, characterized in that The method of collecting historical energy usage data using a multi-source acquisition method and preprocessing the historical energy usage data to obtain a standard data set specifically includes: Collect historical energy usage data from energy devices, sensor devices, and smart meters; The historical energy usage data is cleaned, deduplicated, and missing values ​​are filled to obtain a standard data set.

4. The method according to claim 1, wherein The energy transmission mode between sub-regions is formulated based on the governance adjustment amount, specifically including: Maintain the sub-regions whose difference values ​​are greater than or equal to zero and arrange them in descending order according to the difference values ​​to obtain a positive list; maintain the sub-regions whose difference values ​​are less than zero and arrange them in descending order according to the absolute value of the difference to obtain a negative list; Obtaining a predefined energy transmission rule table, wherein the energy transmission rule table indicates a definition of whether each sub-region can transmit energy to other sub-regions; Traverse the sub-areas in the negative list in sequence, take the sub-area that is ranked first in the positive list and is allowed to be transmitted to this location in the energy transmission rule table as the scheduling point, adjust the positive list and the negative list in real time, and obtain the energy transmission method after one traversal.

5. An intelligent energy management device, characterized in that: include: An acquisition and preprocessing module is used to acquire historical energy usage data using a multi-source acquisition method, and preprocess the historical energy usage data to obtain a standard data set; The energy usage prediction module is used to construct a spatiotemporal graph of regional energy usage changes, wherein the nodes in the spatiotemporal graph represent subregions, the node features include the features of the standard data set, and the edges represent the energy transmission relationship between subregions. The spatiotemporal graph neural network is used to output the energy consumption pattern and energy output forecast of each subregion, specifically for: Obtain a spatiotemporal graph neural network including a graph convolution module and a temporal convolution module; The graph convolution module is used to extract spatial features in the regional change spatiotemporal graph, and the temporal convolution module is used to extract temporal features in the regional change spatiotemporal graph, wherein the temporal features include periodic patterns and seasonal patterns, and the spatial features and temporal features are fused to obtain a unified spatiotemporal representation; predicting energy consumption patterns and energy output forecasts for each sub-region based on the unified spatiotemporal representation; The energy demand forecasting module is used to obtain historical energy demand and environmental data as external data, obtain energy-related feedback information text through a crawler, input the external data and the energy-related feedback information text into a multimodal time series model, and output energy demand forecasts. Specifically, it is used to: Acquiring meteorological data and economic data as the environmental data; Using a crawler to extract content related to energy prices, supply stability, and environmental issues from local social forums, a topic model algorithm was used to extract hot topics from the content. A sentiment classification model was used to identify the emotions reflected in these hot topics. These hot topics and corresponding emotions were combined as energy-related feedback information text. Obtain a multimodal time series model including LSTM modules and Transformer modules; Using the LSTM module to extract a first feature corresponding to the external data, using the Transformer module to extract a second feature corresponding to the energy-related feedback information text, and establishing a cross-attention between the first feature and the second feature; fusing the first feature, the second feature, and the cross attention to obtain a unified multimodal representation, and predicting the energy demand forecast of each sub-region according to the unified multimodal representation; The governance adjustment module is used to calculate the difference between the predicted energy output and the predicted energy demand of each sub-region as the governance adjustment amount, and formulate the energy transmission method between the sub-regions based on the governance adjustment amount.

6. 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 the computer program, when executed by the processor, implements the steps of the energy intelligent management method as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that The computer-readable storage medium stores an implementation program for information transmission, and when the program is executed by the processor, the steps of the energy intelligent management method as described in any one of claims 1 to 4 are implemented.

Images