A power grid topology identification method, system, terminal and storage medium

Abstract

The application provides a power grid topology identification method, system, terminal and storage medium, the method comprises: acquiring power grid real-time data, the power grid real-time data includes the real-time operation data and real-time topology data of power grid, the real-time operation data includes the real-time operation information and line fault information of each node, and the real-time topology data includes the connection relationship of each node;The power grid real-time data is input to the preset topology identification model, so that the topology identification model carries out topology identification on the power grid according to the power grid real-time data, and outputs topology identification result, the topology identification result includes topology structure diagram, node state information, fault positioning information and future topology state information, effectively improve the accuracy and comprehensiveness of power grid topology identification.

Description

Technical Field

[0001] This application relates to the field of power grid topology identification technology, and in particular to a power grid topology identification method, system, terminal and storage medium. Background Technology

[0002] With the development of smart grids, the complexity and dynamic characteristics of power grids are constantly increasing, and the demand for real-time identification and monitoring of power grid topology is becoming more and more urgent. Graph Convolutional Networks (GCNs) have gradually become a research hotspot for power grid topology identification due to their superiority in graph structure data processing, and can effectively capture the relationships and characteristics between power grid nodes.

[0003] Existing technical solutions mainly include topology identification based on traditional machine learning methods, such as support vector machines (SVM) and decision trees. These methods often rely on manual feature extraction and tend to overlook the interrelationships between nodes when dealing with complex power grid structures, leading to a decrease in identification accuracy. However, existing technologies have limited identification capabilities under dynamic changes and abnormal conditions, making it difficult to meet the real-time monitoring needs of power grids. Summary of the Invention

[0004] To address the aforementioned technical problems, this application provides a power grid topology identification method, system, terminal, and storage medium, which effectively improves the accuracy and comprehensiveness of power grid topology identification.

[0005] In a first aspect, embodiments of this application provide a power grid topology identification method, including:

[0006] Acquire real-time power grid data, which includes real-time operation data and real-time topology data of the power grid. The real-time operation data includes real-time operation information of each node and line fault information, and the real-time topology data includes the connection relationship of each node.

[0007] The real-time power grid data is input into a preset topology identification model, so that the topology identification model can perform topology identification on the power grid based on the real-time power grid data and output the topology identification result, which includes a topology diagram, node status information, fault location information and future topology status information.

[0008] The topology identification model is constructed by combining a graph convolutional network model, a time series model, and an abnormal state identification model. The graph convolutional network model and the time series model are trained based on historical power grid data.

[0009] This application provides a power grid topology identification method that uses a pre-defined topology identification model to identify the current topology of the power grid and predict its future topology. In this application, the graph convolutional network model, representing a neural network structure specifically designed to process graphical data, is particularly suitable for processing topology data because it can capture complex relationships between nodes and the structural features of the graph by performing convolution operations on the graph. Furthermore, considering that the topology of the power grid changes dynamically over time and is accompanied by abnormal states, this application further combines a time series model and an abnormal state identification model to construct the topology identification model. This enables the topology identification model to simultaneously possess time series feature recognition and abnormal state recognition capabilities, effectively improving the accuracy and comprehensiveness of power grid topology identification.

[0010] Furthermore, the acquisition of real-time power grid data includes:

[0011] The real-time operating data is acquired through a set of preset sensors;

[0012] The real-time topology data is obtained through a pre-set geographic information system.

[0013] In one possible implementation, the topology identification model performs topology identification on the power grid based on real-time power grid data and outputs a topology identification result, including:

[0014] The real-time topology data from the real-time power grid data is input into the time series model, so that the time series model can make real-time inferences based on the real-time topology data, thereby predicting the future state of the power grid and generating the current time series characteristics of the power grid and the future topology state information.

[0015] The real-time operation data in the real-time power grid data is input into the abnormal state identification model so that the abnormal state identification model can analyze the sudden abnormalities of faults or overloads in the real-time operation data, identify and locate the problem area, and then generate the fault location information.

[0016] The real-time power grid data is format-converted to obtain a corresponding graph dataset. The graph dataset includes an adjacency matrix and a feature matrix. The adjacency matrix is ​​used to store the connection relationships of each node in the power grid, and the feature matrix is ​​used to store the features of each node.

[0017] Embedding processing is performed on the graph dataset to encode the features of each node in the graph dataset and generate corresponding embedding vectors;

[0018] The embedding vector and the time series features are input into the graph convolutional network model, so that the graph convolutional network model performs feature extraction and aggregation on the embedding vector and the time series features through a multi-layer convolutional processing architecture and an adaptive feature weight optimization mechanism, thereby generating the topology graph and the node state information.

[0019] The topology identification result is output by combining the topology diagram, node status information, fault location information, and future topology status information.

[0020] This application provides a method for topology identification of a power grid using a topology identification model. First, a time series model is used to infer the future state of the power grid based on real-time topology data, generating corresponding time series features and future topology state information. Then, an anomaly identification model is used to analyze the real-time operating data of the power grid, generating fault location information. Finally, the real-time power grid data is converted into embedding vectors, and the embedding vectors and time series features are extracted and aggregated using the multi-layer convolution processing and adaptive feature weight optimization mechanism of a graph convolutional network model to generate a topology diagram and node state information. This application, by combining time series features and real-time power grid data, effectively integrates various data sources, enabling more accurate identification of the power grid topology, providing a reliable basis for power grid operation and management, reducing potential risks caused by topology identification errors, and improving the accuracy and comprehensiveness of power grid topology identification.

[0021] In one possible implementation, training the graph convolutional network model based on historical power grid data includes:

[0022] Acquire historical power grid data, which includes historical power grid operation data and historical topology data;

[0023] An initial graph convolutional network model is constructed based on a multi-layer graph convolutional network architecture and an adaptive feature weight optimization mechanism.

[0024] The graph convolutional network model is obtained by training the initial graph convolutional network model using the historical data of the power grid through a supervised learning method.

[0025] This application provides a method for training a graph convolutional network model. First, an initial graph convolutional network model is constructed through a multi-layer graph convolutional network architecture and an adaptive feature weight optimization mechanism. The multi-layer graph convolutional network (GCN) architecture can process different levels of power grid topology information layer by layer. Each layer of GCN processes the topology data of the corresponding level and transmits information between layers. The bottom layer captures the connection relationship of local nodes, and the top layer analyzes the global power grid connectivity and energy flow. The adaptive feature weight optimization mechanism can calculate the importance score of each feature and dynamically adjust the weight of the corresponding feature to improve the accuracy of power grid topology identification.

[0026] Furthermore, the step of employing a supervised learning method to train the initial graph convolutional network model based on the historical power grid data to obtain the graph convolutional network model includes:

[0027] The historical power grid data is labeled to obtain several corresponding labeled training data.

[0028] Each training data is input into the initial graph convolutional network model so that the initial graph convolutional network model outputs corresponding prediction results. The difference between each prediction result and the corresponding label is calculated according to the preset cross-entropy loss function and mean squared error loss function. Then, the model parameters are updated according to the difference between each prediction result and the corresponding label to obtain the first graph convolutional network model.

[0029] If the performance of the first graph convolutional network model meets the preset conditions, then the first graph convolutional network model is used as the graph convolutional network model; otherwise, the first graph convolutional network model continues to be trained until the trained model meets the preset conditions.

[0030] In this embodiment, an initial graph convolutional network (GCNN) model is trained using a supervised learning method to obtain the GCNN model. Specifically, an iterative method for updating model parameters is employed during training. In each iteration, the model makes predictions on the training data, and then the differences between each prediction result and its corresponding label are calculated based on the cross-entropy loss function and the mean squared error loss function. The model parameters are then updated based on these differences until the model's performance meets preset conditions, thus obtaining the GCNN model. The iterative training method proposed in this embodiment fully utilizes the training data, avoids data waste, and ensures that the final GCNN model accurately reflects the characteristics of the power grid, thereby improving the accuracy and efficiency of the entire power grid topology identification process.

[0031] In one possible implementation, training the time series model based on historical power grid data includes:

[0032] Obtain historical topology data from the power grid historical data;

[0033] An initial time series model was constructed based on a long short-term memory network and gated recursive units;

[0034] Based on the state information of each node and edge in the historical topology data, a time series input dataset is constructed;

[0035] The time series input dataset is input into the initial time series model so that the initial time series model can capture the dependence of the power grid topology on the changes over time based on the time series input dataset, thereby obtaining the time series model.

[0036] This application provides a training method for a time series model. First, a Long Short-Term Memory (LSTM) network and gated recursive units (GRUs) are selected to construct an initial time series model. This is because LSMs and GRUs are better suited for handling time series data and can effectively solve the gradient vanishing or exploding problems during training. They can also learn long-term dependencies, improving the efficiency of model training. During training, the model learns patterns from historical data to identify the dynamic evolution trend of the power grid structure. A fully trained model can perform real-time inference on newly input power grid data, thereby predicting the future state of the power grid topology and improving the comprehensiveness of power grid topology identification.

[0037] Furthermore, when the real-time power grid data is missing, incomplete, or contains noise, a preset noise reduction algorithm and uncertainty processing technology are used to complete the real-time power grid data.

[0038] Secondly, correspondingly, embodiments of this application provide a power grid topology identification system, including an acquisition module and an identification module;

[0039] The acquisition module is used to acquire real-time power grid data, which includes real-time operation data and real-time topology data of the power grid. The real-time operation data includes real-time operation information of each node and line fault information, and the real-time topology data includes the connection relationship of each node.

[0040] The identification module is used to input the real-time data of the power grid into a preset topology identification model, so that the topology identification model can perform topology identification on the power grid based on the real-time data of the power grid and output the topology identification result. The topology identification result includes a topology diagram, node status information, fault location information and future topology status information.

[0041] The topology identification model is constructed by combining a graph convolutional network model, a time series model, and an abnormal state identification model. The graph convolutional network model and the time series model are trained based on historical power grid data.

[0042] Thirdly, embodiments of this application provide a terminal, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements any of the power grid topology identification methods described in embodiments of this application.

[0043] Fourthly, embodiments of this application provide a computer-readable storage medium, the computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform any of the power grid topology identification methods described in embodiments of this application. Attached Figure Description

[0044] Figure 1 This is a flowchart illustrating a power grid topology identification method provided in an embodiment of this application.

[0045] Figure 2 This is a schematic diagram illustrating the process of using a topology identification model to perform topology identification in a power grid topology identification method provided in an embodiment of this application.

[0046] Figure 3 This is a schematic diagram of the structure of a graph convolution topology identification system provided in an embodiment of this application.

[0047] Figure 4 This is a schematic diagram of a power grid topology identification system provided in an embodiment of this application. Detailed Implementation

[0048] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0049] It should be noted that the step numbers in this document are only for the convenience of explaining the specific embodiments and are not intended to limit the order in which the steps are performed. In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include one or more of that feature.

[0050] Example 1:

[0051] like Figure 1 As shown, Embodiment 1 provides a power grid topology identification method, including steps S1 and S2:

[0052] Step S1: Obtain real-time power grid data, which includes real-time operation data and real-time topology data of the power grid. The real-time operation data includes real-time operation information of each node and line fault information, and the real-time topology data includes the connection relationship of each node.

[0053] Step S2: Input the real-time power grid data into a preset topology identification model so that the topology identification model can perform topology identification on the power grid based on the real-time power grid data and output the topology identification result. The topology identification result includes a topology diagram, node status information, fault location information, and future topology status information.

[0054] The topology identification model is constructed by combining a graph convolutional network model, a time series model, and an abnormal state identification model. The graph convolutional network model and the time series model are trained based on historical power grid data.

[0055] In a preferred embodiment, the topology diagram of the power grid is used to display an overall diagram of the power grid nodes and their connections; the node status information is used to reflect the real-time operating status of each node, including parameters such as voltage and current; and the fault location information is used to identify the location of a faulty node or line in the power grid.

[0056] This application provides a power grid topology identification method that uses a pre-defined topology identification model to identify the current topology of the power grid and predict its future topology. In this application, the graph convolutional network model, representing a neural network structure specifically designed to process graphical data, is particularly suitable for processing topology data because it can capture complex relationships between nodes and the structural features of the graph by performing convolution operations on the graph. Furthermore, considering that the topology of the power grid changes dynamically over time and is accompanied by abnormal states, this application further combines a time series model and an abnormal state identification model to construct the topology identification model. This enables the topology identification model to simultaneously possess time series feature recognition and abnormal state recognition capabilities, effectively improving the accuracy and comprehensiveness of power grid topology identification.

[0057] Furthermore, in step S1, acquiring real-time power grid data includes:

[0058] The real-time operating data is acquired through a set of preset sensors;

[0059] The real-time topology data is obtained through a pre-set geographic information system.

[0060] In one possible implementation, in step S2, the topology identification model performs topology identification on the power grid based on the real-time power grid data and outputs the topology identification result, such as... Figure 2 As shown, steps S201-S206 are included:

[0061] Step S201: Input the real-time topology data from the real-time power grid data into the time series model, so that the time series model can make real-time inferences based on the real-time topology data, thereby predicting the future state of the power grid and generating the current time series characteristics of the power grid and the future topology state information.

[0062] Step S202: Input the real-time operation data in the real-time power grid data into the abnormal state identification model, so that the abnormal state identification model can analyze the sudden abnormalities of faults or overloads in the real-time operation data, identify and locate the problem area, and then generate the fault location information.

[0063] Step S203: Convert the format of the real-time power grid data to obtain the corresponding graph dataset. The graph dataset includes an adjacency matrix and a feature matrix. The adjacency matrix is ​​used to store the connection relationship of each node in the power grid, and the feature matrix is ​​used to store the features of each node.

[0064] Step S204: Perform embedding processing on the graph dataset, encode the features of each node in the graph dataset, and generate corresponding embedding vectors;

[0065] Step S205: Input the embedding vector and the time series features into the graph convolutional network model, so that the graph convolutional network model can extract and aggregate features from the embedding vector and the time series features through a multi-layer convolutional processing architecture and an adaptive feature weight optimization mechanism, thereby generating the topology graph and the node state information;

[0066] Step S206: Combining the topology diagram, node status information, fault location information, and future topology status information, output the topology identification result.

[0067] This application provides a method for topology identification of a power grid using a topology identification model. First, a time series model is used to infer the future state of the power grid based on real-time topology data, generating corresponding time series features and future topology state information. Then, an anomaly identification model is used to analyze the real-time operating data of the power grid, generating fault location information. Finally, the real-time power grid data is converted into embedding vectors, and the embedding vectors and time series features are extracted and aggregated using the multi-layer convolution processing and adaptive feature weight optimization mechanism of a graph convolutional network model to generate a topology diagram and node state information. This application, by combining time series features and real-time power grid data, effectively integrates various data sources, enabling more accurate identification of the power grid topology, providing a reliable basis for power grid operation and management, reducing potential risks caused by topology identification errors, and improving the accuracy and comprehensiveness of power grid topology identification.

[0068] In a preferred embodiment, in step S203, constructing the adjacency matrix and feature matrix of the power grid based on real-time power grid data first requires determining the nodes and edges of the power grid. Nodes represent various devices or stations in the power grid, while edges represent the connection relationships between nodes. The adjacency matrix is ​​a square matrix, where each element indicates whether there is a direct connection between two nodes, typically represented by 0 and 1. The feature matrix stores the characteristics of each node, such as load, generation capacity, and voltage information, with each row corresponding to a node and each column corresponding to a feature. By combining these two matrices, the topology of the power grid and its related characteristics can be effectively represented, providing basic data for subsequent graph convolutional network analysis. Furthermore, while acquiring data, the real-time status of the power grid is periodically monitored, the current status is analyzed in real time, and the node and edge features are adjusted according to the monitoring results to achieve adaptive adjustment.

[0069] In a preferred embodiment, in step S204, the embedding vector of the node is specifically represented as a vector that encodes the features of each node through a graph convolutional network, compressing the high-dimensional node features into a low-dimensional vector representation. These vectors not only retain the node's own information but also contain the structural information and feature relationships of other nodes connected to it. The final embedding form is usually a low-dimensional dense vector, where each vector represents the node's position in the graph network and its feature relevance, thus facilitating subsequent model training and inference.

[0070] In a preferred embodiment, in step S205, the time series features output by the time series model are used as input to the graph convolutional network model, combined with the embedding vector, to dynamically adjust the identification process of the power grid topology, thereby generating the topology graph and the node state information.

[0071] The topology identification model provided in this application integrates time series models, abnormal state identification models, and graph convolutional network models. It can support real-time monitoring and management, and output the current state information of the power grid, including the identification results of the topology diagram, node state information, and fault location information, so as to help operation and maintenance personnel quickly understand the current state of the power grid and take corresponding measures. At the same time, the topology identification model can also perform real-time inference on the input power grid data, thereby predicting the future state of the power grid topology, including potential topology changes, node state trends, and possible fault locations, and output these prediction results together with the current state information.

[0072] In a preferred embodiment, the topology identification model can be further combined with spatiotemporal feature modeling to perform convolution processing on the spatiotemporal data of the power grid in order to capture the characteristics of the power grid changing with time and space; specifically, time series data is combined with spatial data, and spatiotemporal features are extracted through convolution operations, thereby more accurately identifying dynamic changes.

[0073] In one possible implementation, training the graph convolutional network model based on historical power grid data includes:

[0074] Acquire historical power grid data, which includes historical power grid operation data and historical topology data;

[0075] An initial graph convolutional network model is constructed based on a multi-layer graph convolutional network architecture and an adaptive feature weight optimization mechanism.

[0076] The graph convolutional network model is obtained by training the initial graph convolutional network model using the historical data of the power grid through a supervised learning method.

[0077] This application provides a method for training a graph convolutional network model. First, an initial graph convolutional network model is constructed through a multi-layer graph convolutional network architecture and an adaptive feature weight optimization mechanism. The multi-layer graph convolutional network (GCN) architecture can process different levels of power grid topology information layer by layer. Each layer of GCN processes the topology data of the corresponding level and transmits information between layers. The bottom layer captures the connection relationship of local nodes, and the top layer analyzes the global power grid connectivity and energy flow. The adaptive feature weight optimization mechanism can calculate the importance score of each feature and dynamically adjust the weight of the corresponding feature to improve the accuracy of power grid topology identification.

[0078] In a preferred embodiment, further, the step of employing a supervised learning method to train the initial graph convolutional network model based on the historical power grid data to obtain the graph convolutional network model includes:

[0079] The historical power grid data is labeled to obtain several corresponding labeled training data.

[0080] Each training data is input into the initial graph convolutional network model so that the initial graph convolutional network model outputs corresponding prediction results. The difference between each prediction result and the corresponding label is calculated according to the preset cross-entropy loss function and mean squared error loss function. Then, the model parameters are updated according to the difference between each prediction result and the corresponding label to obtain the first graph convolutional network model.

[0081] If the performance of the first graph convolutional network model meets the preset conditions, then the first graph convolutional network model is used as the graph convolutional network model; otherwise, the first graph convolutional network model continues to be trained until the trained model meets the preset conditions.

[0082] In this embodiment, an initial graph convolutional network (GCNN) model is trained using a supervised learning method to obtain the GCNN model. Specifically, an iterative method for updating model parameters is employed during training. In each iteration, the model makes predictions on the training data, and then the differences between each prediction result and its corresponding label are calculated based on the cross-entropy loss function and the mean squared error loss function. The model parameters are then updated based on these differences until the model's performance meets preset conditions, thus obtaining the GCNN model. The iterative training method proposed in this embodiment fully utilizes the training data, avoids data waste, and ensures that the final GCNN model accurately reflects the characteristics of the power grid, thereby improving the accuracy and efficiency of the entire power grid topology identification process.

[0083] In one possible implementation, training the time series model based on historical power grid data includes:

[0084] Obtain historical topology data from the power grid historical data;

[0085] An initial time series model was constructed based on a long short-term memory network and gated recursive units;

[0086] Based on the state information of each node and edge in the historical topology data, a time series input dataset is constructed;

[0087] The time series input dataset is input into the initial time series model so that the initial time series model can capture the dependence of the power grid topology on the changes over time based on the time series input dataset, thereby obtaining the time series model.

[0088] This application provides a training method for a time series model. First, a Long Short-Term Memory (LSTM) network and gated recursive units (GRUs) are selected to construct an initial time series model. This is because LSMs and GRUs are better suited for handling time series data and can effectively solve the gradient vanishing or exploding problems during training. They can also learn long-term dependencies, improving the efficiency of model training. During training, the model learns patterns from historical data to identify the dynamic evolution trend of the power grid structure. A fully trained model can perform real-time inference on newly input power grid data, thereby predicting the future state of the power grid topology and improving the comprehensiveness of power grid topology identification.

[0089] Furthermore, when the real-time power grid data is missing, incomplete, or contains noise, a preset noise reduction algorithm and uncertainty processing technology are used to complete the real-time power grid data.

[0090] In summary, the embodiments of this application have the following beneficial effects:

[0091] 1. The embodiments of this application use an adaptive feature weight optimization mechanism to calculate feature importance scores and adjust weights; at the same time, the multi-level topology structure processes information at different levels, analyzes from local to global perspectives, effectively integrates various data sources, can more accurately identify the power grid topology structure, provide a reliable basis for power grid operation and management, and reduce potential risks caused by topology identification errors.

[0092] 2. The embodiments of this application combine spatiotemporal feature modeling, time series modeling, and abnormal state identification capabilities; they can accurately capture the dynamic changes of the power grid state over time and space, and effectively handle abnormal situations such as data loss and noise using relevant technologies to maintain efficient topology identification, thereby ensuring the effectiveness and stability of real-time monitoring and management of the power grid.

[0093] 3. The embodiments of this application construct adjacency matrices and feature matrices by acquiring data through multiple channels, comprehensively covering power grid-related information; the collaboration of various modules enables these data to be fully utilized at different processing stages, such as accurately reflecting power grid characteristics when training models, avoiding data waste, and thus improving the accuracy and efficiency of the entire power grid topology identification process.

[0094] Example 2:

[0095] like Figure 3 As shown in the figure, Embodiment 2 provides a graph convolutional topology identification system that integrates the power grid topology identification method described in Embodiment 1. It includes a data collection module, a construction and training module, and a topology identification module. By combining graph convolutional networks, adaptive feature weight optimization mechanisms, multi-level topology processing, and time series modeling, it can accurately identify and predict the topology of the power grid and its dynamic evolution.

[0096] The graph convolutional topology identification system described in Example 2 acquires data through a data collection module. This module deploys a large number of sensors and combines historical operational data with a Geographic Information System (GIS) to comprehensively collect node and edge features in the power grid. Nodes typically represent power plants, substations, and load centers in the power grid, while edges represent the connection relationships between these nodes. To characterize the topology of the power grid, the data collection module constructs two core data structures: an adjacency matrix and a feature matrix. The adjacency matrix is ​​a square matrix used to describe the connection relationships between power grid nodes, where 1 indicates that there is a connection between nodes, and 0 indicates that there is no connection. The feature matrix is ​​used to store the feature information of each node, such as voltage, current, and power. At the same time, the data collection module cleans and standardizes the collected data, removing outliers and noise to ensure the accuracy and consistency of the data.

[0097] After data collection, the data is input into the construction and training module. The core of this module is a multi-level graph convolutional network architecture. First, the adjacency matrix and feature matrix are input into the graph convolutional network for feature extraction. The graph convolutional network consists of multiple convolutional layers. Each layer captures local information of nodes and edges through convolution operations and passes global information layer by layer. Activation functions are introduced into the convolutional layers to enhance the nonlinear expressive power of the model. During the convolution process, an adaptive feature weight optimization mechanism dynamically adjusts the feature weights of each node and edge to improve the model's ability to identify the power grid topology. The working principle of this optimization mechanism is to calculate the importance score of each feature through an attention mechanism and dynamically adjust the weights based on these scores, so that the model can flexibly adjust its focus under different power grid conditions.

[0098] Then, the supervised learning model training phase begins. The construction and training module uses the cross-entropy loss function or the mean squared error loss function to train the model. By comparing the difference between the predicted output and the actual topology label, the model parameters are continuously adjusted. The training process is an iterative process. In each iteration, the model learns more features of the power grid structure and updates the parameters in real time until the model error converges to an acceptable range. Through this training, the model can learn how to infer the topology of the power grid from the input data and identify its dynamic evolution.

[0099] After training, the graph convolutional topology identification system integrates the model into the topology recognition module. This module combines time series modeling and anomaly identification capabilities to monitor and analyze the power grid topology in real time. It can not only output the current power grid topology but also capture the dynamic changes of the power grid over time. By analyzing historical topology data, it can predict the future state of the power grid. Especially in the event of sudden situations or anomalies, this module can quickly identify problems and output topology diagrams, node status information, and fault location information. This helps maintenance personnel quickly understand the current state of the power grid and take corresponding measures to prevent potential power grid failures.

[0100] During the identification process of the topology identification module, various data from the power grid need to be acquired in real time, including operating parameters such as node voltage, current, power and load information, edge connection status, load status and topology change data, as well as environmental data such as temperature, weather impact and line fault information. These data are updated in real time and transmitted to the identification module through sensor networks and monitoring systems.

[0101] During the topology recognition module's recognition process, the power grid data obtained from the data collection module is fed into a pre-trained graph convolutional network model. First, the model undergoes format conversion, transforming the real-time power grid data into a graph data structure, including an adjacency matrix and a feature matrix. The graph structure is then updated based on changes in nodes and edges. The model performs embedding processing on this input data, converting the state features of nodes and edges into representation vectors in a high-dimensional space. Subsequently, through multi-layer convolution processing and an adaptive feature weight optimization mechanism, the model extracts and aggregates features from the input graph data, outputting the recognition results, including the current topology and node states.

[0102] When combining graph convolutional network models with time series modeling and anomaly identification capabilities, LSTM and GRU time series models are first used to model historical data and capture the time dependence of power grid operation status. Then, the output time series features are used as input to the graph convolutional network and combined with real-time power grid data to dynamically adjust the identification process of power grid topology. The anomaly identification part analyzes sudden anomalies of faults or overloads in the input data to identify and locate problem areas, thereby generating a comprehensive identification result including topology diagram, node status information and fault location information.

[0103] Example 3:

[0104] The method used in Example 3 is similar to that in Example 2, except that the graph convolution topology identification system in Example 3 is particularly suitable for processing the multi-level features of regional power grids, urban power grids and local distribution networks.

[0105] The main difference lies in the fact that this application's embodiments employ a multi-level graph convolutional network model to process the multi-level features of regional power grids, urban power grids, and local distribution networks. In specific operation, an independent graph convolutional network layer is designed for each level to process power grid information at different levels: for the global power grid layer, the model focuses on capturing the connectivity of the power grid over a large area; for the urban power grid layer, the model pays more attention to the connection relationships of medium-sized nodes and the steady-state properties of the local network; and for the local distribution network layer, the model refines the detailed connections and load characteristics between local nodes. Furthermore, the model dynamically adjusts the feature weights of nodes and edges in each level through an adaptive feature weight optimization mechanism to adapt to the needs of different levels of the power grid, thereby improving the accuracy of identifying multi-level features.

[0106] The graph convolutional topology identification system provided in this application embodiment uses a data collection module deployed at different levels of sensors to comprehensively monitor the power grid and collect topology data of regional power grids, urban power grids, and local distribution networks. The data collection module adopts a periodic monitoring method to ensure that the power grid status at each level can be acquired in real time. By combining sensor data, historical operation records, and geographic information system (GIS) data, the adjacency matrix and feature matrix of the power grid can be effectively constructed. Similar to the previous embodiment, the adjacency matrix is ​​used to describe the connection relationship of nodes in the power grid, while the feature matrix is ​​used to store the voltage, power, and load attributes of each node. During the data collection process, data cleaning and standardization are automatically performed to ensure that the power grid data at each level has good quality.

[0107] The graph convolutional topology identification system then inputs the collected data into the construction and training module, where a multi-level GCN model is used to process the topology information of the power grid. First, the adjacency matrix and feature matrix are input into the bottom-level GCN network, which is responsible for capturing the node connections of the local distribution network. As the convolutional layers are passed layer by layer, higher-level power grid topology information, such as the connectivity of urban and regional power grids, is gradually processed. The characteristic of the multi-level GCN model is that it can simultaneously capture the fine-grained connections between local nodes and the global features of the entire power grid. Each layer of the GCN network focuses on power grid data at a specific level and integrates global and local information through inter-layer information transfer. In this way, it can maintain efficient identification capabilities when processing complex multi-level topologies.

[0108] The adaptive feature weight optimization mechanism in this module further improves the model's recognition accuracy. By calculating the feature importance scores of nodes and edges at each level, the weights of the corresponding features can be adjusted according to the scores, thereby flexibly adjusting the model's focus at different power grid levels. In this way, when dealing with regional power grids, more attention can be paid to the global connectivity of the power grid, while when dealing with local distribution networks, more attention can be paid to the specific connection relationships between nodes.

[0109] Supervised learning was employed during training, using labeled power grid topology data to train the model. The model parameters were adjusted using cross-entropy loss or mean squared error loss functions to gradually improve its predictive ability. After each iteration, the model updated its parameters based on the difference between the predicted and actual results. Ultimately, the model was able to accurately identify the topology of multi-level power grids.

[0110] After training, the topology recognition module is responsible for real-time identification and management of the power grid's topology. This module combines time series modeling to capture dynamic changes in the power grid by analyzing topology data that changes over time. By modeling historical data, it can predict future power grid topology evolution. In addition, the module also has the ability to identify abnormal states, enabling it to react quickly when a sudden fault occurs in the power grid and output detailed topology diagrams, node status information, and fault location information. This information helps power grid operation and maintenance personnel to promptly identify problems, perform power grid maintenance and management, and ensure the safe operation of the power grid.

[0111] Example 4:

[0112] like Figure 4 As shown, Embodiment 4 provides a power grid topology identification system, including an acquisition module 10 and an identification module 20;

[0113] The acquisition module 10 is used to acquire real-time power grid data, which includes real-time operation data and real-time topology data of the power grid. The real-time operation data includes real-time operation information of each node and line fault information, and the real-time topology data includes the connection relationship of each node.

[0114] The identification module 20 is used to input the real-time data of the power grid into a preset topology identification model, so that the topology identification model can perform topology identification on the power grid based on the real-time data of the power grid and output the topology identification result. The topology identification result includes a topology diagram, node status information, fault location information and future topology status information.

[0115] The topology identification model is constructed by combining a graph convolutional network model, a time series model, and an abnormal state identification model. The graph convolutional network model and the time series model are trained based on historical power grid data.

[0116] Furthermore, the acquisition module 10 acquires real-time power grid data, including:

[0117] The real-time operating data is acquired through a set of preset sensors;

[0118] The real-time topology data is obtained through a pre-set geographic information system.

[0119] In one possible implementation, the identification module 20 performs topology identification on the power grid based on the real-time power grid data and outputs the topology identification result, including:

[0120] The real-time topology data from the real-time power grid data is input into the time series model, so that the time series model can make real-time inferences based on the real-time topology data, thereby predicting the future state of the power grid and generating the current time series characteristics of the power grid and the future topology state information.

[0121] The real-time operation data in the real-time power grid data is input into the abnormal state identification model so that the abnormal state identification model can analyze the sudden abnormalities of faults or overloads in the real-time operation data, identify and locate the problem area, and then generate the fault location information.

[0122] The real-time power grid data is format-converted to obtain a corresponding graph dataset. The graph dataset includes an adjacency matrix and a feature matrix. The adjacency matrix is ​​used to store the connection relationships of each node in the power grid, and the feature matrix is ​​used to store the features of each node.

[0123] Embedding processing is performed on the graph dataset to encode the features of each node in the graph dataset and generate corresponding embedding vectors;

[0124] The embedding vector and the time series features are input into the graph convolutional network model, so that the graph convolutional network model performs feature extraction and aggregation on the embedding vector and the time series features through a multi-layer convolutional processing architecture and an adaptive feature weight optimization mechanism, thereby generating the topology graph and the node state information.

[0125] The topology identification result is output by combining the topology diagram, node status information, fault location information, and future topology status information.

[0126] In one possible implementation, training the graph convolutional network model based on historical power grid data includes:

[0127] Acquire historical power grid data, which includes historical power grid operation data and historical topology data;

[0128] An initial graph convolutional network model is constructed based on a multi-layer graph convolutional network architecture and an adaptive feature weight optimization mechanism.

[0129] The graph convolutional network model is obtained by training the initial graph convolutional network model using the historical data of the power grid through a supervised learning method.

[0130] Furthermore, the step of employing a supervised learning method to train the initial graph convolutional network model based on the historical power grid data to obtain the graph convolutional network model includes:

[0131] The historical power grid data is labeled to obtain several corresponding labeled training data.

[0132] Each training data is input into the initial graph convolutional network model so that the initial graph convolutional network model outputs corresponding prediction results. The difference between each prediction result and the corresponding label is calculated according to the preset cross-entropy loss function and mean squared error loss function. Then, the model parameters are updated according to the difference between each prediction result and the corresponding label to obtain the first graph convolutional network model.

[0133] If the performance of the first graph convolutional network model meets the preset conditions, then the first graph convolutional network model is used as the graph convolutional network model; otherwise, the first graph convolutional network model continues to be trained until the trained model meets the preset conditions.

[0134] In one possible implementation, training the time series model based on historical power grid data includes:

[0135] Obtain historical topology data from the power grid historical data;

[0136] An initial time series model was constructed based on a long short-term memory network and gated recursive units;

[0137] Based on the state information of each node and edge in the historical topology data, a time series input dataset is constructed;

[0138] The time series input dataset is input into the initial time series model so that the initial time series model can capture the dependence of the power grid topology on the changes over time based on the time series input dataset, thereby obtaining the time series model.

[0139] In one possible implementation, the power grid topology identification system further includes a data completion module, which is used to complete the real-time power grid data using a preset noise reduction algorithm and uncertainty processing technology when the real-time power grid data has missing, incomplete, or noisy data.

[0140] This application provides a power grid topology identification system that identifies the current topology of the power grid and predicts its future topology through a preset topology identification model. In this application, the graph convolutional network model represents a neural network structure specifically designed to process graphical data. It can capture complex relationships between nodes and the structural features of the graph by performing convolution operations on the graph, making it particularly suitable for processing topology data. Furthermore, considering that the topology of the power grid changes dynamically over time and is accompanied by abnormal states, this application further combines a time series model and an abnormal state identification model to construct the topology identification model. This enables the topology identification model to simultaneously possess time series feature recognition and abnormal state recognition capabilities, effectively improving the accuracy and comprehensiveness of power grid topology identification.

[0141] For a more detailed explanation of the working principle and procedures of this embodiment, please refer to the relevant description in Embodiment 1.

[0142] Example 5:

[0143] Embodiment 5 provides a terminal including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the power grid topology identification method described in this application.

[0144] The processor can 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. The general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal, connecting various parts of the terminal via various interfaces and lines.

[0145] The memory can be used to store the computer program. The processor implements various functions of the terminal by running or executing the computer program stored in the memory and calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0146] Example 6:

[0147] Embodiment Six provides a computer-readable storage medium, the computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute the power grid topology identification method described in this application.

[0148] If the integrated modules of the power grid topology identification method 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 above embodiments of the present invention 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, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0149] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above descriptions are merely specific embodiments of this application and are not intended to limit the scope of protection of this application. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application for those skilled in the art.

Claims

1. A method for identifying power grid topology, characterized in that, include: Acquire real-time power grid data, which includes real-time operation data and real-time topology data of the power grid. The real-time operation data includes real-time operation information of each node and line fault information, and the real-time topology data includes the connection relationship of each node. The real-time power grid data is input into a preset topology identification model, so that the topology identification model can perform topology identification on the power grid based on the real-time power grid data and output the topology identification result, which includes a topology diagram, node status information, fault location information and future topology status information. The topology identification model performs topology identification on the power grid based on real-time power grid data and outputs topology identification results, including: inputting real-time topology data from the real-time power grid data into a time series model, so that the time series model can make real-time inferences based on the real-time topology data, thereby predicting the future state of the power grid and generating the current time series characteristics of the power grid and the future topology state information; inputting real-time operating data from the real-time power grid data into an anomaly state identification model, so that the anomaly state identification model can analyze sudden anomalies of faults or overloads in the real-time operating data, identify and locate problem areas, thereby generating the fault location information; and converting the format of the real-time power grid data to obtain a corresponding graph dataset. The graph dataset includes an adjacency matrix and a feature matrix. The adjacency matrix stores the connection relationships between nodes in the power grid, and the feature matrix stores the features of each node. Embedding processing is performed on the graph dataset to encode the features of each node and generate corresponding embedding vectors. The embedding vectors and the time-series features are input into a graph convolutional network model. The graph convolutional network model uses a multi-layer convolutional processing architecture and an adaptive feature weight optimization mechanism to extract and aggregate features from the embedding vectors and the time-series features, thereby generating the topology graph and the node state information. Combining the topology graph, node state information, fault location information, and future topology state information, the topology identification result is output. The topology identification model is constructed by combining a graph convolutional network model, a time series model, and an abnormal state identification model. The graph convolutional network model and the time series model are trained based on historical power grid data.

2. The power grid topology identification method as described in claim 1, characterized in that, The acquisition of real-time power grid data includes: The real-time operating data is acquired through a set of preset sensors; The real-time topology data is obtained through a pre-set geographic information system.

3. The power grid topology identification method as described in claim 1, characterized in that, The process of training the graph convolutional network model based on historical power grid data includes: Acquire historical power grid data, which includes historical power grid operation data and historical topology data; An initial graph convolutional network model is constructed based on a multi-layer graph convolutional network architecture and an adaptive feature weight optimization mechanism. The graph convolutional network model is obtained by training the initial graph convolutional network model using the historical data of the power grid through a supervised learning method.

4. The power grid topology identification method as described in claim 3, characterized in that, The method employs supervised learning to train the initial graph convolutional network model based on historical power grid data to obtain the graph convolutional network model, including: The historical power grid data is labeled to obtain several corresponding labeled training data. Each training data is input into the initial graph convolutional network model so that the initial graph convolutional network model outputs corresponding prediction results. The difference between each prediction result and the corresponding label is calculated according to the preset cross-entropy loss function and mean squared error loss function. Then, the model parameters are updated according to the difference between each prediction result and the corresponding label to obtain the first graph convolutional network model. If the performance of the first graph convolutional network model meets the preset conditions, then the first graph convolutional network model is used as the graph convolutional network model; otherwise, the first graph convolutional network model continues to be trained until the trained model meets the preset conditions.

5. The power grid topology identification method as described in claim 1, characterized in that, The process of training the time series model based on historical power grid data includes: Obtain historical topology data from the power grid historical data; An initial time series model was constructed based on a long short-term memory network and gated recursive units; Based on the state information of each node and edge in the historical topology data, a time series input dataset is constructed; The time series input dataset is input into the initial time series model so that the initial time series model can capture the dependence of the power grid topology on the changes over time based on the time series input dataset, thereby obtaining the time series model.

6. The power grid topology identification method as described in claim 1, characterized in that, When the real-time power grid data is missing, incomplete, or contains noise, a preset noise reduction algorithm and uncertainty processing technology are used to complete the real-time power grid data.

7. A power grid topology identification system, characterized in that, Includes an acquisition module and a recognition module; The acquisition module is used to acquire real-time power grid data, which includes real-time operation data and real-time topology data of the power grid. The real-time operation data includes real-time operation information of each node and line fault information, and the real-time topology data includes the connection relationship of each node. The identification module is used to input the real-time data of the power grid into a preset topology identification model, so that the topology identification model can perform topology identification on the power grid based on the real-time data of the power grid and output the topology identification result. The topology identification result includes a topology diagram, node status information, fault location information and future topology status information. The topology identification model performs topology identification on the power grid based on real-time power grid data and outputs topology identification results, including: inputting real-time topology data from the real-time power grid data into a time series model, so that the time series model can make real-time inferences based on the real-time topology data, thereby predicting the future state of the power grid and generating the current time series characteristics of the power grid and the future topology state information; inputting real-time operating data from the real-time power grid data into an anomaly state identification model, so that the anomaly state identification model can analyze sudden anomalies of faults or overloads in the real-time operating data, identify and locate problem areas, thereby generating the fault location information; and converting the format of the real-time power grid data to obtain a corresponding graph dataset. The graph dataset includes an adjacency matrix and a feature matrix. The adjacency matrix stores the connection relationships between nodes in the power grid, and the feature matrix stores the features of each node. Embedding processing is performed on the graph dataset to encode the features of each node and generate corresponding embedding vectors. The embedding vectors and the time-series features are input into a graph convolutional network model. The graph convolutional network model uses a multi-layer convolutional processing architecture and an adaptive feature weight optimization mechanism to extract and aggregate features from the embedding vectors and the time-series features, thereby generating the topology graph and the node state information. Combining the topology graph, node state information, fault location information, and future topology state information, the topology identification result is output. The topology identification model is constructed by combining a graph convolutional network model, a time series model, and an abnormal state identification model. The graph convolutional network model and the time series model are trained based on historical power grid data.

8. A terminal, characterized in that, The device includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements a power grid topology identification method as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device containing the computer-readable storage medium to perform a power grid topology identification method as described in any one of claims 1-6.

Images