Intelligent electricity price matching and verification method based on graph neural network and related device

Abstract

The present application relates to the technical field of power management, in particular to a smart electricity price matching and verification method based on a graph neural network and related devices, by acquiring all user electricity data and performing data processing on all user electricity data to obtain graph structure data; according to the graph structure data, a graph neural network model is established, which can effectively model the complex relationship of user electricity data, capture the similarity and difference between users, not only can quickly and accurately identify electricity price abnormal users, reduce the risk of false positives and false negatives, but also can adapt to new data distribution through continuous learning and optimization as the number of users increases and the electricity behavior changes, maintain the accuracy of verification, further improve the intelligent level of electricity price verification, and meet the challenge of user demand. The problem of poor matching of actual power consumption and recorded electricity price in the system archives in the prior art is solved.

Description

Technical Field

[0001] This invention relates to the field of power management technology, specifically to a smart electricity price matching and verification method and related apparatus based on graph neural networks. Background Technology

[0002] Electricity price implementation and electricity price verification are two crucial aspects of electricity market operation, jointly ensuring the fair, just, and efficient operation of the electricity market. The accuracy of electricity price implementation directly impacts the economic benefits of power companies and the rights of electricity users. Electricity price verification refers to the process of comprehensively checking and verifying electricity price implementation, aiming to ensure the accurate implementation of electricity price policies and fair competition in the electricity market. Electricity price implementation and electricity price verification are interdependent and mutually reinforcing. Electricity price implementation is the prerequisite and foundation for electricity price verification; only by ensuring the accuracy and compliance of electricity price implementation can effective electricity price verification be conducted. Conversely, electricity price verification serves as the supervision and guarantee of electricity price implementation. Through verification, problems and risks in electricity price implementation can be identified, allowing for timely rectification and handling, thus ensuring the accurate implementation of electricity price policies and fair competition in the electricity market.

[0003] Currently, the main method of manually verifying electricity price implementation on a household-by-household basis is not only inefficient and inaccurate, but also unable to dynamically adjust and respond to changes in user electricity consumption behavior, making it difficult to match the dynamic changes and diverse needs of users. Therefore, to improve the efficiency and accuracy of electricity price verification, intelligent methods have become an inevitable choice. The accurate implementation of electricity prices not only directly affects the economic benefits of power companies, but also relates to market fairness and user satisfaction. Therefore, improving the accuracy of electricity price implementation and effectively verifying it has become one of the key tasks of the State Grid Marketing Department.

[0004] However, in practice, due to the dynamic changes in users' electricity consumption over time, or due to unreasonable electricity pricing schemes selected during the application process, ordinary matching algorithms have significant limitations in handling complex electricity user data. They often fail to accurately capture key information within the data, affecting the accuracy and reliability of the verification results. This frequently leads to discrepancies between actual electricity consumption and the electricity prices recorded in the system archives. If such problems are not detected and addressed promptly, they may negatively impact the company's financial management and the scientific nature of electricity pricing management. Summary of the Invention

[0005] To address the problem of poor matching between actual electricity consumption and electricity prices recorded in the system archives in existing technologies, this invention provides a smart electricity price matching and verification method and related apparatus based on graph neural networks.

[0006] To achieve the above objectives, the present invention employs the following technical solution:

[0007] This invention provides a smart electricity price matching and verification method based on graph neural networks, comprising:

[0008] Obtain all users' electricity consumption data;

[0009] Data processing is performed on all users' electricity consumption data to obtain graph structure data;

[0010] Based on the graph structure data, establish a graph neural network model;

[0011] The graph neural network model is used to check the electricity price execution status of users in the current system archives, and abnormal users are marked to complete the electricity price matching and verification.

[0012] Optionally, the electricity consumption data for all users includes the electricity consumption category, electricity consumption amount, and market electricity price for all users.

[0013] Furthermore, the method for processing all user electricity consumption data to obtain graph structure data is as follows:

[0014] Classify all user electricity consumption data and obtain the electricity consumption dataset for each user category;

[0015] Normalize each feature in the electricity consumption dataset for each type of user to obtain a normalized user feature matrix;

[0016] Graph structure data is obtained based on the normalized user feature matrix.

[0017] Furthermore, the method for normalizing each feature in each type of user electricity consumption dataset to obtain the normalized user feature matrix is ​​as follows:

[0018] (1)

[0019] The normalized user feature matrix is:

[0020] (2)

[0021] in, Indicates user In the Features in each dimension; The number of dimensions of the feature; Normalized user characteristics; N The number of users in a certain user category; This is the normalized user feature matrix.

[0022] Furthermore, based on the normalized user feature matrix, the method for obtaining graph structure data is as follows:

[0023] Based on the normalized user feature matrix, define the user node set and the electricity price node set;

[0024] Obtain the edge weights based on the user node set and the electricity price node set;

[0025] Based on the edge weights, obtain the user-electricity price graph and the adjacency matrix of the user-electricity price graph to complete the acquisition of graph structure data.

[0026] Furthermore, based on graph structure data, the method for establishing a graph neural network model is as follows:

[0027] Using graph-structured data as input, the feature vectors of each node in the neural network model are updated to obtain feature vectors of user electricity consumption behavior and electricity price adaptability.

[0028] Define an electricity price matching function, and match the feature vectors of user electricity consumption behavior and electricity price fitness with predefined electricity price categories to obtain the minimum loss function of the electricity price matching function, thus completing the establishment of the graph neural network model.

[0029] Furthermore, the method for using a graph neural network model to verify the electricity price execution status of users in the current system archives and to mark abnormal users, thereby completing the electricity price matching and verification, is as follows:

[0030] Using graph neural network models to obtain predicted electricity prices;

[0031] During the verification process, the predicted electricity price is compared with the actual electricity price to determine whether there are any abnormalities in the electricity price.

[0032] If there is an anomaly in the electricity price, obtain the predicted electricity price for the user with the anomaly and the predicted electricity price for the neighboring users of the user with the anomaly.

[0033] Anomaly scores are obtained based on the predicted electricity prices of users with abnormal electricity prices and the predicted electricity prices of their neighboring users, and it is then determined whether there are any abnormalities in the electricity price execution of users with abnormal electricity prices.

[0034] This invention also provides a smart electricity price matching and verification system based on graph neural networks, comprising:

[0035] Electricity data acquisition module: used to acquire electricity data of all users;

[0036] Graph structure data acquisition module: used to process all users' electricity consumption data and acquire graph structure data;

[0037] Graph Neural Network Model Building Module: Used to build graph neural network models based on graph structure data;

[0038] Electricity price matching and verification module: This module uses a graph neural network model to verify the electricity price execution status of users in the current system archives, and marks abnormal users to complete the electricity price matching and verification.

[0039] A terminal device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the steps of the method described above.

[0040] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method described above.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] This invention provides a smart electricity price matching and verification method based on graph neural networks. This method acquires electricity consumption data from all users and processes this data to obtain graph structure data. Graph structure data can intuitively reflect the relationships between users' electricity consumption behaviors, providing a solid foundation for subsequent model building. By establishing a graph neural network model based on the graph structure data and analyzing the data, the advantages of Graph Neural Networks (GNNs) in processing graph structure data can be utilized. GNNs can effectively model the complex relationships between users' electricity consumption data, capturing the similarities and differences between users. This not only enables the rapid and accurate identification of users with abnormal electricity prices, reducing the risk of false alarms and missed alarms, but also allows the method to adapt to new data distributions through continuous learning and optimization as the number of users increases and electricity consumption behaviors change, maintaining the accuracy of verification and further improving the intelligence level of electricity price verification to meet the challenges of user needs.

[0043] The electricity consumption data of all users includes the electricity consumption categories and electricity consumption of all users. By representing information such as users, electricity consumption behavior and market electricity prices as a graph structure, GNNs can comprehensively consider multiple factors and improve the accuracy and efficiency of electricity price matching.

[0044] The method for processing all user electricity consumption data to obtain graph structure data involves classifying all user electricity consumption data and obtaining electricity consumption datasets for each category. The aim is to classify users according to certain criteria (such as electricity consumption behavior, region, electricity consumption category, etc.) to facilitate subsequent processing, better understand the differences between user groups, and provide a foundation for subsequent graph structure construction. By normalizing each feature in each category of user electricity consumption dataset, a normalized user feature matrix is ​​obtained. Data with different dimensions or value ranges are transformed to the same scale for comparison and analysis. In electricity consumption data, the value ranges of different features may vary greatly; normalization can eliminate this difference, making the model more stable when processing data. Finally, based on the normalized user feature matrix, graph structure data is obtained. The normalized user feature matrix provides standardized data input for subsequent graph structure construction, helping to improve the accuracy and efficiency of the model while eliminating noise and differences in user electricity consumption data, making the model more stable and accurate when processing data. The normalized user feature matrix is ​​converted into graph structure data. Graph-structured data consists of nodes and edges, where nodes can represent users and edges can represent relationships between users (such as similarity, association, etc.). Constructing graph-structured data is crucial for the application of GNNs models because it can capture complex relationships between users and provide rich information input to the model.

[0045] In the method for obtaining graph-structured data based on a normalized user feature matrix, a user node set and an electricity price node set are defined. In this step, each row (or column, depending on the specific data representation) in the normalized user feature matrix is ​​considered a user node, thus constructing the user node set. Simultaneously, an electricity price node set is defined based on the electricity price category or specific electricity price data. Then, based on the user node set and the electricity price node set, the edge weights are obtained. The edge weights reflect the strength or similarity of the relationship between user nodes and electricity price nodes. This can be obtained by calculating the similarity, distance, or other metrics between user features and electricity price features; a larger weight indicates a stronger correlation between the user and a specific electricity price. Finally, based on the edge weights, the user-electricity price graph and its adjacency matrix are obtained, completing the acquisition of graph structure data. Using the user node set, the electricity price node set, and the edge weights, a user-electricity price graph is constructed. In this graph, user nodes and electricity price nodes are connected by weighted edges. Based on the graph structure, a corresponding adjacency matrix is ​​generated. The adjacency matrix is ​​a two-dimensional array where the values ​​of the elements represent the connection relationships between nodes in the graph (i.e., the edge weights). The introduction of graph structure data allows the model to more comprehensively understand the relationship between users and electricity prices, thereby enhancing the model's expressive power. The edge weights reflect the correlation between users and electricity prices, which is crucial for predicting users' electricity price choices in electricity price matching and verification tasks. By accurately calculating the edge weights and constructing the user-electricity price graph, the model can more accurately predict users' electricity price behavior, thus improving prediction accuracy.

[0046] Based on graph structure data, the method for building a graph neural network model uses graph structure data as input to update the feature vectors of each layer of the neural network model, obtaining feature vectors of user electricity consumption behavior and electricity price adaptability. Graph structure data contains rich user information and relationship networks, which are crucial for training the graph neural network model. By utilizing this information, the model can learn a wider range of user electricity consumption behaviors and electricity price patterns. An electricity price matching function is defined, and the feature vectors of user electricity consumption behavior and electricity price adaptability are matched with predefined electricity price categories. The minimum loss function of the electricity price matching function is obtained, completing the establishment of the graph neural network model. By minimizing the loss function during the training process, the graph neural network model can continuously optimize its weights and parameters, thereby enhancing the model's generalization ability and enabling it to be applied to a wider range of electricity price matching scenarios. In GNNs models, node features are passed and updated through the edges in the graph structure. In each layer, node features are aggregated and updated based on the features of their neighboring nodes and the weights of the edges. This information transmission mechanism enables the model to capture the complex relationships in the graph structure and generate more expressive node feature representations. By updating node features layer by layer, GNNs models can gradually extract deep information from the graph structure, which helps the model to more accurately predict users' electricity pricing behavior. This improves prediction accuracy and enables rapid processing of large-scale graph structure data, further improving computational efficiency.

[0047] The method described above utilizes a graph neural network model to verify the electricity pricing status of users in the current system archives and marks abnormal users to complete electricity price matching and verification. By matching the feature vector of each user in the graph neural network model with predefined electricity price categories, the most suitable electricity price type is obtained. This data-driven matching method, compared to traditional manual judgment or simple rule matching, can capture more subtle user characteristics and differences in electricity consumption behavior, thereby improving the accuracy of electricity price matching. During verification, the obtained most suitable electricity price type is compared with the actual electricity price to determine if there are any anomalies. The automatic electricity price verification using the graph neural network model reduces manual intervention and improves verification efficiency. Simultaneously, by predicting electricity prices through the model and comparing them with actual prices, potential electricity price anomalies can be intelligently identified, reducing the risk of human error. When an electricity price anomaly is detected, the method not only focuses on the predicted electricity price of the abnormal user but also considers the predicted electricity prices of its neighboring users. This graph-based analysis method can capture the correlation and mutual influence between users, thereby more accurately determining whether electricity price anomalies truly exist and avoiding misjudgments caused by anomalies in data from a single user. Anomaly scores are calculated based on the predicted electricity prices of abnormal users and their neighboring users, providing a quantitative basis for determining whether there are anomalies in electricity price implementation. This quantitative approach makes anomaly identification more objective and interpretable, aiding subsequent analysis and processing. By accurately identifying and marking users with abnormal electricity prices, this method helps maintain the fairness and transparency of the electricity market. For power companies, this helps to promptly detect and address potential electricity price violations, ensuring the normal operation of the electricity market; for users, it enhances their trust in and satisfaction with electricity price policies.

[0048] This invention also provides an intelligent electricity price matching and verification system based on graph neural networks. Through the setup of an electricity consumption data acquisition module, a graph structure data acquisition module, a graph neural network model building module, and an electricity price matching and verification module, it achieves the collection and processing of a large amount of user electricity consumption data, thereby constructing graph structure data. Based on the graph structure data, a graph neural network model is built to achieve automatic electricity price matching and verification. Compared with traditional electricity price matching strategies based on static rules or simple models, this system can dynamically learn the characteristics of user electricity consumption data and adjust the electricity price matching strategy according to these characteristics. This dynamic optimization capability allows the system to respond more flexibly to market changes. By utilizing the graph neural network model, the system can accurately verify the electricity price execution status of users in the current system archives and automatically mark abnormal users. This precise and intelligent verification method not only improves the efficiency of electricity price management but also reduces the risk of human error, providing strong support for power companies' electricity price management. This system, through automation and intelligence, comprehensively optimizes electricity price management from data collection, processing, modeling to verification. It not only reduces the labor and time costs of electricity price management but also improves its accuracy and scientific rigor. This system provides new ideas for technological innovation in the power industry and strong support for the upgrading and transformation of related industries. With continuous technological advancements and the expansion of application scenarios, the intelligent electricity price matching and verification system based on graph neural networks is expected to play a significant role in more fields.

[0049] The present invention also provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, implements the steps of the above-described method. The device has a simple structure, low modification cost, and minimal resource consumption.

[0050] A computer-readable storage medium storing a computer program, characterized in that, when executed by a processor, the computer program implements the steps of the method described above. This storage medium is highly portable and versatile. Attached Figure Description

[0051] Figure 1 This is a schematic diagram of a smart electricity price matching and verification method based on graph neural networks according to the present invention.

[0052] Figure 2 This is a schematic diagram illustrating the implementation process of a smart electricity price matching and verification method based on graph neural networks according to the present invention.

[0053] Figure 3 This is a schematic diagram of the structure of a smart electricity price matching and verification system based on graph neural networks according to the present invention. Detailed Implementation

[0054] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0055] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0056] The present invention will be further described in detail below with reference to specific embodiments. These descriptions are for explanation purposes only and are not intended to limit the scope of the invention.

[0057] See Figure 1 and Figure 2 This invention provides a smart electricity price matching and verification method based on graph neural networks, comprising:

[0058] S1: Obtain electricity consumption data for all users; the electricity consumption data for all users includes the electricity consumption category and the amount of electricity consumed by each user. By representing information such as users, electricity consumption behavior and market electricity prices as a graph structure, GNNs can comprehensively consider multiple factors and improve the accuracy and efficiency of electricity price matching.

[0059] S2: Process all user electricity consumption data to obtain graph structure data. The specific method is as follows:

[0060] Classify all user electricity consumption data and obtain the electricity consumption dataset for each category: Divide the known monthly electricity consumption dataset for users according to different electricity consumption categories. Assuming there are N users, the electricity consumption dataset for a certain category is as follows: ;in, Indicates user In the Features in several dimensions, such as monthly electricity consumption and peak-valley electricity consumption ratio; The number of dimensions of the feature;

[0061] For each feature in the electricity consumption dataset of each user category, normalization is performed to obtain the normalized user feature matrix:

[0062] (1)

[0063] The normalized user feature matrix is:

[0064] (2)

[0065] in, Indicates user In the Features in each dimension; The number of dimensions of the feature; Normalized user characteristics; N The number of users in a certain user category; This is the normalized user feature matrix.

[0066] Based on the normalized user feature matrix, obtain the graph structure data:

[0067] Based on the normalized user feature matrix, the user node set is defined as follows: ,in Indicates user Nodes; Define the set of electricity price nodes ,in Indicates the first Various electricity pricing schemes;

[0068] Based on the user node set and the electricity price node set, obtain the edge weights:

[0069] (3)

[0070] (4)

[0071] in, User eigenvectors; Electricity pricing scheme eigenvectors;

[0072] Based on the edge weights, obtain the user-electricity price graph and its adjacency matrix, thus completing the acquisition of the graph structure data:

[0073] Users can choose from methods such as cosine similarity and Euclidean distance to construct a user-electricity price graph.

[0074] (5)

[0075] in, It is a set of nodes; It is an edge set. It is the weight threshold.

[0076] Adjacency matrix of a graph Represented as:

[0077] (6)

[0078] S3: Based on the graph structure data, establish a graph neural network model, specifically as follows:

[0079] Using graph-structured data as input, the feature vectors of each node in the neural network model are updated to obtain feature vectors for user electricity consumption behavior and electricity price suitability. GNNs can model the characteristics of graph-structured data, comprehensively considering the similarity between users and the impact of different factors on electricity price matching. Information is passed through the adjacency relationships of nodes, continuously updating node features during multi-layer propagation. Each node's feature is represented as a vector, containing information such as the user's electricity consumption behavior, industry classification, and transformer capacity. The initial feature vector for each user... This can include basic user information, such as industry classification, electricity consumption, and transformer capacity. Assuming there are a total of... Each user node can be represented by a feature matrix. To represent the initial characteristics of all user nodes:

[0080] (7)

[0081] in, This represents the dimension of each node's features. This matrix contains the initial features for each user;

[0082] In each layer of a graph neural network, the features of a node are updated based on information from its neighboring nodes. Assume the... The characteristics of layer nodes are So in The feature update formula for the layer is:

[0083] (8)

[0084] in: Indicates the first Users in the layer eigenvectors; Indicates user The set of neighboring nodes; It is an adjacency matrix The elements in the text represent nodes. With nodes The strength of the relationship (similarity). It is the first The layer's weight matrix is ​​used for feature transformation; It is an activation function; the linear rectifier unit is used to increase the nonlinearity of the network.

[0085] The feature update formula in this layer represents how each node updates its own features based on information from its neighbors. This information propagation mechanism occurs layer by layer in graph neural networks, gradually capturing the complex relationships between users;

[0086] After passing information through multiple layers of graph neural networks to the features of nodes, each node will eventually learn a feature vector representing the user's electricity consumption behavior and electricity price adaptability. in, This refers to the number of layers in the network. This feature vector contains complex information about the relationship between users and electricity prices.

[0087] Define an electricity price matching function, and match the feature vectors of user electricity consumption behavior and electricity price adaptability with predefined electricity price categories. Obtain the minimum loss function of the electricity price matching function, and complete the establishment of the graph neural network model.

[0088] To determine the most suitable electricity price matching strategy, each user's final feature vector can be matched with a predefined electricity price category. Assume there are... Define a price matching function for each electricity price category. This maps the user's final feature vector to the most suitable electricity price category. The specific formula is:

[0089] (9)

[0090] in, It is a weight matrix for electricity price matching, representing the relationship between each electricity price category and user characteristics; It is a bias term; Provided user The most suitable electricity pricing type;

[0091] Through this electricity price matching function, the system can dynamically match the most suitable electricity price type based on each user's electricity consumption behavior and characteristics;

[0092] Electricity Price Matching Loss Function Training: To optimize the electricity price matching accuracy of the graph neural network, an electricity price matching loss function is defined to measure the accuracy of the model's predicted electricity prices. A commonly used loss function is the cross-entropy loss, as shown in the following formula:

[0093] (9)

[0094] in, User Actual electricity price category (one-hot code) Indicates user Category ,otherwise ; The user predicted by the model Belongs to the electricity price category The probability, ; This refers to the number of electricity price categories.

[0095] By minimizing the loss function, graph neural networks can continuously adjust their parameters. and This improves the accuracy of electricity price matching. A graph neural network learns the similarities between users and, based on these similarities and user characteristics, accurately matches the most suitable electricity price. This process not only improves the efficiency of electricity price matching but also takes into account dynamic changes in user behavior.

[0096] S4: Utilize a graph neural network model to verify the electricity pricing status of users in the current system archives, and mark abnormal users to complete electricity price matching and verification. Specifically: for users... The actual electricity price recorded by the system is The electricity price predicted by the graph neural network is... Then the electricity price error It can be represented as:

[0097] (10)

[0098] in, User The actual electricity price recorded in the system; The predicted electricity price is obtained through a graph neural network model; User The electricity price error represents the difference between the actual electricity price and the predicted electricity price; for each user, if the electricity price error... Greater than or equal to the set threshold ,Right now If so, it is considered that the user's electricity pricing is abnormal and further testing and adjustment are required.

[0099] The next step in electricity price verification is anomaly detection, which involves identifying which users have issues with their electricity pricing. If anomalies are found, the predicted electricity prices for the users with anomalies and their neighboring users are obtained. This process is then analyzed. The relationship between a user's electricity price error and that of their neighboring users can be analyzed using the anomaly detection function of a graph neural network to calculate an anomaly score for each user. Assuming... User Given the set of neighboring nodes, then the user... The anomaly score can be calculated using the following formula:

[0100] (11)

[0101] in, User The predicted electricity price; Neighboring users The predicted electricity price; User The set of neighboring nodes; User The anomaly detection score represents the user's... The degree of difference between your electricity price and that of your neighboring users; the greater the difference, the higher the anomaly score, indicating that the user... There may be a problem with the electricity price. If there are abnormal scores... Greater than or equal to the preset threshold ,Right now Then determine the user There are irregularities in the implementation of electricity pricing.

[0102] Finally, for users judged as abnormal The system will provide adjustment suggestions based on its historical electricity consumption data and electricity price forecast results. For example, assuming the historical electricity consumption data predicts the electricity price as follows: Therefore, a weighted average of historical electricity consumption data can be used to provide users with [the following information / services]. Provide suggestions for adjusted electricity prices.

[0103] Adjusted electricity price It can be represented as:

[0104] (12)

[0105] in: This refers to the proposed adjusted electricity price; User Historical electricity consumption node count; Electricity price forecasts are based on historical electricity consumption.

[0106] In this way, the adjusted electricity price can comprehensively consider historical electricity consumption data and electricity price implementation, thereby achieving a more scientific and reasonable electricity price adjustment.

[0107] Electricity price verification and anomaly detection can not only identify abnormal users, but also provide corresponding electricity price adjustment suggestions. The feedback process includes the following steps:

[0108] Abnormal User List Generation: Generates a list containing all abnormal users and provides each user with an anomaly detection score and electricity price adjustment suggestions.

[0109] Confirmation and Implementation of Adjustments: The relevant departments (such as the electricity marketing department) confirm the electricity price adjustment recommendations and adjust the electricity prices for abnormal users.

[0110] Feedback and Optimization: Based on the adjusted electricity prices, the system performs feedback learning on the graph neural network model, continuously optimizing the accuracy of electricity price matching and anomaly detection. Through this optimization process, the system can improve the accuracy of electricity price verification and gradually enhance the level of intelligence in electricity price management.

[0111] By using graph neural networks for electricity price verification and anomaly detection, the system can accurately detect problems in user electricity price implementation and provide reasonable electricity price adjustment suggestions. This process not only improves the efficiency of electricity price verification but also optimizes the scientific nature and accuracy of electricity price management through deep learning technology. This helps power grid companies achieve intelligent and precise electricity price management, promoting fair and orderly electricity use and increasing user satisfaction.

[0112] See Figure 3 This invention provides a smart electricity price matching and verification system based on graph neural networks, comprising:

[0113] Electricity data acquisition module: used to acquire electricity data of all users;

[0114] Graph structure data acquisition module: used to process all users' electricity consumption data and acquire graph structure data;

[0115] Graph Neural Network Model Building Module: Used to build graph neural network models based on graph structure data;

[0116] Electricity price matching and verification module: This module uses a graph neural network model to verify the electricity price execution status of users in the current system archives, and marks abnormal users to complete the electricity price matching and verification.

[0117] This system, through its modules for electricity consumption data acquisition, graph structure data acquisition, graph neural network model building, and electricity price matching and verification, collects and processes a large amount of user electricity consumption data. This data is then used to construct graph structure data, and a graph neural network model is built based on this data to achieve automatic electricity price matching and verification. Compared to traditional electricity price matching strategies based on static rules or simple models, this system can dynamically learn the characteristics of user electricity consumption data and adjust the electricity price matching strategy accordingly. This dynamic optimization capability allows the system to respond more flexibly to market changes. By utilizing the graph neural network model, the system can accurately verify the electricity price execution status of users in the current system archives and automatically mark abnormal users. This precise and intelligent verification method not only improves the efficiency of electricity price management but also reduces the risk of human error, providing strong support for power companies' electricity price management. Through automation and intelligence, from data collection, processing, modeling to verification, this system achieves comprehensive optimization of electricity price management, reducing both the labor and time costs of electricity price management and improving its accuracy and scientific rigor. This system provides new ideas for technological innovation in the power industry and strong support for the upgrading and transformation of related industries. With the continuous progress of technology and the expansion of application scenarios, the intelligent electricity price matching and verification system based on graph neural networks is expected to play an important role in more fields.

[0118] This invention provides a terminal device comprising: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps in the various method embodiments described above. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in the various device embodiments described above.

[0119] The computer program can be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention.

[0120] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0121] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0122] The memory can be used to store the computer program and / or module. The processor implements various functions of the terminal device by running or executing the computer program and / or module stored in the memory and calling the data stored in the memory.

[0123] If the modules / units integrated into the terminal device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0124] In summary, this invention provides a smart electricity price matching and verification method and related apparatus based on graph neural networks (GNNs). By utilizing GNNs to model the complex relationship between user electricity consumption data and electricity prices, it accurately captures the similarities and differences between users, thereby improving the accuracy of electricity price matching. Compared to traditional manual verification or simple algorithms, GNNs can better handle complex electricity price execution situations, avoiding errors and omissions in manual verification and improving the accuracy of electricity price execution. Through the dynamic optimization capabilities of GNNs, the electricity price matching strategy can be adjusted in real time, ensuring the scientific and timely nature of electricity price execution, automatically adapting to changes in the electricity market and fluctuations in user electricity demand, and improving the flexibility and timeliness of electricity price management. The anomaly detection function based on graph neural networks can identify abnormal situations in electricity pricing in real time. By calculating the difference between the user's electricity price and the price based on historical electricity consumption data, it can quickly locate electricity pricing problems, reducing manual intervention. For abnormal users, GNNs can accurately mark them and automatically generate adjustment suggestions, ensuring the fairness and rationality of electricity pricing. At the same time, through automated data processing and intelligent analysis, it greatly reduces the need for manual intervention, lowers labor costs, improves work efficiency, and enhances the overall operational efficiency of power companies.

[0125] The above description is merely a preferred embodiment of the present invention and is not intended to limit the technical solution of the present invention in any way. Those skilled in the art should understand that, without departing from the spirit and principles of the present invention, the technical solution can be modified and replaced in several simple ways, and these modifications and replacements are all within the scope of protection covered by the claims.

Claims

1. A smart electricity price matching and verification method based on graph neural networks, characterized in that, include: Obtain all users' electricity consumption data; Data processing is performed on all user electricity consumption data to obtain graph structure data, specifically: Based on the normalized user feature matrix, define the user node set and the electricity price node set; Obtain the edge weights based on the user node set and the electricity price node set; Based on the edge weights, obtain the user-electricity price graph and the adjacency matrix of the user-electricity price graph to complete the acquisition of graph structure data; Based on the graph structure data, a graph neural network model is established, specifically as follows: Using graph-structured data as input, the feature vectors of each node in the graph neural network model are updated to obtain feature vectors of user electricity consumption behavior and electricity price adaptability. Define an electricity price matching function, match the feature vectors of user electricity consumption behavior and electricity price adaptability with predefined electricity price categories, obtain the minimum loss function of the electricity price matching function, and complete the establishment of the graph neural network model; The graph neural network model is used to verify the electricity pricing status of users in the current system archives, and abnormal users are marked to complete the electricity pricing matching and verification. Specifically: Using graph neural network models to obtain predicted electricity prices; During the verification process, the predicted electricity price is compared with the actual electricity price to determine whether there are any abnormalities in the electricity price.

2. The smart electricity price matching and verification method based on graph neural networks according to claim 1, characterized in that, The electricity consumption data for all users includes the electricity consumption category, electricity consumption amount, and market electricity price for all users.

3. The smart electricity price matching and verification method based on graph neural networks according to claim 1, characterized in that, The method for processing all user electricity consumption data to obtain graph structure data is as follows: Classify all user electricity consumption data and obtain the electricity consumption dataset for each user category; Normalize each feature in the electricity consumption dataset for each type of user to obtain a normalized user feature matrix; Graph structure data is obtained based on the normalized user feature matrix.

4. The intelligent electricity price matching and verification method based on graph neural networks according to claim 3, characterized in that, The method for normalizing each feature in each type of user electricity consumption dataset to obtain the normalized user feature matrix is ​​as follows: (1) The normalized user feature matrix is: （2） in, Indicates user In the Features in each dimension; The number of dimensions of the feature; Normalized user characteristics; N The number of users in a certain user category; This is the normalized user feature matrix.

5. The smart electricity price matching and verification method based on graph neural networks according to claim 1, characterized in that, If there is an anomaly in the electricity price, obtain the predicted electricity price for the user with the anomaly and the predicted electricity price for the neighboring users of the user with the anomaly. Anomaly scores are obtained based on the predicted electricity prices of users with abnormal electricity prices and the predicted electricity prices of their neighboring users, and it is then determined whether there are any abnormalities in the electricity price execution of users with abnormal electricity prices.

6. A smart electricity price matching and verification system based on graph neural networks, characterized in that, include: Obtain all users' electricity consumption data; Data processing is performed on all user electricity consumption data to obtain graph structure data, specifically: Based on the normalized user feature matrix, define the user node set and the electricity price node set; Obtain the edge weights based on the user node set and the electricity price node set; Based on the edge weights, obtain the user-electricity price graph and the adjacency matrix of the user-electricity price graph to complete the acquisition of graph structure data; Based on the graph structure data, a graph neural network model is established, specifically as follows: Using graph-structured data as input, the feature vectors of each node in the graph neural network model are updated to obtain feature vectors of user electricity consumption behavior and electricity price adaptability. Define an electricity price matching function, match the feature vectors of user electricity consumption behavior and electricity price adaptability with predefined electricity price categories, obtain the minimum loss function of the electricity price matching function, and complete the establishment of the graph neural network model; The graph neural network model is used to verify the electricity pricing status of users in the current system archives, and abnormal users are marked to complete the electricity pricing matching and verification. Specifically: Using graph neural network models to obtain predicted electricity prices; During the verification process, the predicted electricity price is compared with the actual electricity price to determine whether there are any abnormalities in the electricity price.

7. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1-5.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1-5.

Images