Commutation switch control method and device based on graph neural network, equipment and medium

Abstract

The embodiment of the application discloses a commutating switch control method and device based on a graph neural network, equipment and a medium. The method comprises: acquiring state data and device connection relationship of each power device in a power system; determining graph structure data according to the state data and the device connection relationship; determining three-phase imbalance working condition information and device detection report based on the graph structure data and a graph neural network model; determining a commutating switch control strategy according to the device detection report, and controlling the commutating switch based on the commutating switch control strategy to adjust three-phase load balance. The technical scheme accurately identifies three-phase imbalance problems of each node in the power grid by using the graph structure data processing capability of the graph neural network, effectively reduces the influence of imbalance on the power grid through the rapid action of the commutating switch, and significantly improves the fault handling capability and operation stability of the power grid.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power control management, and particularly relates to a commutating switch control method and device based on a graph neural network, equipment and a medium. BACKGROUND

[0002] In the power system, three-phase imbalance is a common problem. Due to the power consumption characteristics of users and the large load variation, the distribution transformer often appears three-phase load imbalance in operation. This imbalance not only affects the service life of the transformer and increases the loss of the transformer, but also causes a series of problems such as voltage drop, flicker, low power factor, and increased line loss. Therefore, solving the three-phase imbalance problem is of great significance to ensure the safe and stable operation of the power system. The graph neural network is a neural network specially designed to process graph structure data, which can capture complex relationships and information in the graph structure. In the process of commutating switch control management under three-phase imbalance conditions, the graph neural network can utilize its powerful data processing capability and real-time response capability to ensure that the system can respond to the changes in three-phase load in real time.

[0003] In the prior art, under three-phase imbalance conditions, it is necessary to quickly respond and adjust the state of the commutating switch to restore three-phase balance. However, due to the limitation of the response speed of the equipment, the actual response speed is difficult to meet the needs of the rapid changes of the power grid, and it is difficult to quickly locate the abnormal behavior under three-phase imbalance conditions, thereby reducing the response efficiency of the system. SUMMARY

[0004] The present application provides a commutating switch control method and device based on a graph neural network, which accurately identifies three-phase imbalance problems of each node in the power grid by utilizing the graph structure data processing capability of the graph neural network, effectively reduces the impact of imbalance on the power grid through the rapid action of the commutating switch, and significantly improves the fault handling capability and operation stability of the power grid.

[0005] According to an aspect of the present application, a commutating switch control method based on a graph neural network is provided, comprising:

[0006] Obtaining state data and device connection relationship of each power device in the power system;

[0007] Determining graph structure data according to the state data and the device connection relationship;

[0008] Determining three-phase imbalance condition information and device detection report based on the graph structure data and the graph neural network model;

[0009] Determining a commutating switch control strategy according to the device detection report, and controlling the commutating switch based on the commutating switch control strategy to adjust the three-phase load balance.

[0010] According to another aspect of the present application, a phase changer control device based on a graph neural network is provided, comprising:

[0011] a data acquisition module configured to acquire state data of each power device in a power system and a device connection relationship;

[0012] a graph structure data determination module configured to determine graph structure data according to the state data and the device connection relationship;

[0013] a working condition information determination module configured to determine three-phase imbalance working condition information and a device detection report based on the graph structure data and a graph neural network model;

[0014] a phase changer control module configured to determine a phase changer control strategy according to the device detection report, and control a phase changer based on the phase changer control strategy to adjust three-phase load balance.

[0015] According to another aspect of the present application, an electronic device is provided, comprising:

[0016] at least one processor; and

[0017] a memory in communication connection with the at least one processor; wherein,

[0018] the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the graph neural network based phase changer control method according to any one of the embodiments of the present application.

[0019] According to another aspect of the present application, a computer readable storage medium is provided, which stores computer instructions for enabling a processor to execute the graph neural network based phase changer control method according to any one of the embodiments of the present application.

[0020] The technical solution of the embodiments of the present application acquires state data of each power device in a power system and a device connection relationship, determines graph structure data according to the state data and the device connection relationship, determines three-phase imbalance working condition information and a device detection report based on the graph structure data and a graph neural network model, determines a phase changer control strategy according to the device detection report, and controls a phase changer based on the phase changer control strategy to adjust three-phase load balance. The technical solution utilizes the graph structure data processing capability of the graph neural network to accurately identify three-phase imbalance problems of each node in the power grid, effectively reduces the influence of imbalance on the power grid through the rapid action of the phase changer, and significantly improves the fault handling capability and operation stability of the power grid.

[0021] It should be understood that the matters described in this detailed description are intended to be illustrative and are not intended to limit or restrict the scope of the embodiments of the present application. Other features of the present application will become apparent to those skilled in the art upon a reading of the following specification. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort based on these drawings.

[0023] Figure 1 is a flow chart of a commutator switch control method based on a graph neural network according to an embodiment of the present application;

[0024] Figure 2 is a flow chart of a commutator switch control method based on a graph neural network according to an embodiment of the present application;

[0025] Figure 3 is a structural schematic diagram of a commutator switch control device based on a graph neural network according to an embodiment of the present application;

[0026] Figure 4 is a structural schematic diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0027] In order to make the technical personnel in the art better understand the present application scheme, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without any creative effort should belong to the scope of protection of the present application.

[0028] It should be noted that the terms "first", "second", "target" and "original" and the like in the description, claims, and drawings of the present application are intended to distinguish between similar objects and not necessarily describe a particular chronological or sequential order. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments of the application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but can include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0029] Embodiment one

[0030] Figure 1 A flowchart of a commutation switch control method based on a graph neural network according to an embodiment one of the present application is provided. The embodiment can be applied to control commutation in a three-phase unbalanced working condition of a power system. The method can be executed by a commutation switch control device based on a graph neural network. The device can be realized in the form of hardware and / or software and can be configured in an electronic device with data processing capability. As shown in Figure 1 The method comprises:

[0031] The technical solution of the embodiment can be executed by a commutation switch control management system based on a graph neural network. The system comprises a control management platform, which is communicatively connected with a data acquisition and processing module, a graph structure construction module, a graph neural network model module, an abnormality detection and positioning module, a control strategy generation module, and an execution module. The modules are electrically connected.

[0032] S110, acquiring state data and device connection relationship of each power device in the power system.

[0033] The power devices can include sensors and power devices included in the power system. The sensors can include current sensors, voltage sensors, power sensors, and the like. The state data can be understood as state data obtained by preprocessing raw data of the power devices. The state data can include current, voltage, power, and commutation switch state information. The device connection relationship can be understood as the positions of the power devices in the power system and the connection relationship therebetween. In this embodiment, the topology information of the power system can be collected, including the positions of the devices and the connection relationship therebetween, and the rated capacity, rated voltage, and rated power parameters of the devices are obtained, wherein the devices can include busbars, generators, transformers, switch devices, and load points. The positions of the devices and the connection relationship therebetween are obtained through system design documents, field measurement, or automatic topology identification technology.

[0034] In this embodiment, the data acquisition and processing module collects raw state data from the sensors and devices of the power system, and preprocesses the collected raw data to obtain state data of the power devices in the power system.

[0035] In this embodiment, the state data of the power devices in the power system is obtained, including: obtaining raw state data of each power device in the power system respectively; preprocessing the raw state data to obtain preprocessed state data; and integrating the preprocessed state data of each power device to obtain the state data of the power device.

[0036] The raw state data can be understood as raw state information of current, voltage, power, and commutation switches collected from the sensors and state monitoring devices through a preset sampling frequency. The preprocessing operation can include data cleaning, data filtering, and data standardization operations. The preprocessed state data can be understood as state data obtained after data cleaning, data filtering, and data standardization processing. In this embodiment, the preprocessed state data of each power device is integrated to form complete state data.

[0037] In this embodiment, the data acquisition and processing module is communicatively connected with various sensors and state monitoring devices of commutation switches in the power system, and then raw state data is collected from the sensors and state monitoring devices through a preset sampling frequency and preprocessed and integrated to obtain the state data of the power devices. The raw data exists in the form of time series, recording the running state of the power grid in a certain time period.

[0038] Specifically, the collected raw data is preprocessed, and the preprocessing includes data cleaning, data filtering and data standardization. The data cleaning is used to check whether there is missing value in the data. The missing value can be processed by filling (such as mean filling, median filling, interpolation method, etc.) or deleting (if the missing proportion is too high). The abnormal value in the data is detected by using a statistical method (such as Z-score, IQR, etc.) or a model-based method (such as isolation forest). The abnormal value is corrected or deleted. The high-frequency noise is removed by data filtering, and the low-frequency signal is retained, so that the data is smoother, the data fluctuation caused by sensor noise or power grid fluctuation is reduced, the data is scaled by data standardization, and the data falls into a small specific interval (such as between 0 and 1), so that the convergence speed of the model is accelerated, and the training efficiency of the model is improved. The data is converted into a distribution with a mean of 0 and a standard deviation of 1 by calculating the mean and standard deviation of the data, so as to eliminate the influence of different dimensions on model training.

[0039] In addition, in the embodiment, data from different sensors and devices can be integrated to form a complete data set, abnormal data in the data set is labeled and marked, and the processed data is stored in a data warehouse for subsequent access and analysis.

[0040] In the embodiment, the collected raw data is preprocessed, and after the preprocessing step, the noise and abnormal values in the raw data are eliminated, and the data quality is further improved. At the same time, the standardization of the data can make the data of different dimensions on the same scale for comparison and analysis, and provide clean and effective input data for the subsequent graph neural network model. In the embodiment, the preprocessing includes cleaning, filtering and standardization operations to eliminate noise and improve data quality, and provide necessary input data for the system, ensure that the control decision is based on the latest power grid state, and ensure that the data input to the graph neural network is clean and effective, thereby improving the accuracy and stability of the model.

[0041] In the embodiment, the collected raw data is preprocessed, and after the preprocessing step, the noise and abnormal values in the raw data are eliminated, and the data quality is further improved. At the same time, the standardization of the data can make the data of different dimensions on the same scale for comparison and analysis, and provide clean and effective input data for the subsequent graph neural network model. In the embodiment, the preprocessing includes cleaning, filtering and standardization operations to eliminate noise and improve data quality, and provide necessary input data for the system, ensure that the control decision is based on the latest power grid state, and ensure that the data input to the graph neural network is clean and effective, thereby improving the accuracy and stability of the model.

[0041] S120, determining graph structure data according to the state data and the device connection relationship.

[0042] The graph structure data can be understood as graph data constructed according to the connection relationship between the power devices and the corresponding state data. In the embodiment, the graph structure construction module can convert the physical structure and device relationship of the power grid into a graph structure according to the topology of the power system and the connection relationship between the devices, and construct corresponding graph structure data, wherein the node represents the device, the edge represents the connection relationship between the devices, and the state data can be the attribute information of the node.

[0043] In this embodiment, optionally, the graph structure data is determined according to the state data and the device connection relationship, including: determining each power device as a node, and defining the connection relationship between the devices as an edge; assigning corresponding attribute data to the nodes and edges based on the state data; and constructing the graph structure data based on the nodes, edges and attribute information.

[0044] The device connection relationship includes the positions of the power devices and the connection relationship between the devices, and the graph structure data includes nodes and edges. Specifically, the nodes can represent the devices in the power system, and the edges represent the physical connections between the devices. The attribute data corresponding to the nodes can include the physical positions and operating states of the devices, and the attribute data corresponding to the edges can include the connection strength and transmission capacity.

[0045] Specifically, in this embodiment, the graph structure elements are defined, the nodes and edges in the graph are determined, a unique identifier is assigned to each device as a node according to the device list of the power system, and an edge is defined for each pair of connected devices according to the connection relationship of the devices. The node includes a unique identifier, a type and parameter information of the device, each edge can include a connected node ID, a connection type and connection parameter information, and each node and edge is assigned corresponding attribute information according to the state data. Then, the graph is initialized using the graph database, an empty graph structure is created, all devices and connections are traversed, and the nodes and edges are added to the graph, and the graph structure data is constructed in combination with the attribute information of the nodes and edges.

[0046] In this embodiment, the graph structure data can also be compared with the original topology information to check whether each node and edge in the graph is accurate, the graph structure data is preprocessed, and the preprocessed graph structure data is stored in the graph database to facilitate access and processing of the graph neural network.

[0047] In this embodiment, by such a setting, the graph structure data can be constructed from the state data of the devices and the connection relationship between the devices, the complex relationship of the power system can be represented in the form of a graph, the topology information of the power grid can be provided for the graph neural network, and the graph neural network can be efficiently processed.

[0048] S130, determining three-phase unbalanced working condition information and a device detection report based on the graph structure data and the graph neural network model.

[0049] The graph neural network model can be a trained neural network model and can be used to process the graph structure data to output corresponding three-phase unbalanced working condition information and corresponding device detection reports. The three-phase unbalanced working condition information can be obtained by identifying the preprocessed graph structure data using the graph neural network model. The device detection report can be generated by using the graph neural network to quickly detect and identify the cause of the three-phase unbalance based on the three-phase unbalanced working condition information.

[0050] Specifically, in the embodiment, the preprocessed graph structure data can be learned and inferred by the graph neural network model module using the trained graph neural network model to identify the three-phase imbalance working condition, and the optimal phase-change switch control strategy can also be predicted. Through the abnormality detection positioning module, the graph neural network is used to quickly detect and locate the abnormal behavior in the power system, identify the cause of the three-phase imbalance, and generate a corresponding device detection report.

[0051] S140, determining a phase-change switch control strategy according to the device detection report, and controlling the phase-change switch based on the phase-change switch control strategy to adjust the three-phase load balance.

[0052] The phase-change switch control strategy can be a strategy for generating a phase-change switch operation according to the device detection report. Specifically, in the embodiment, the three-phase imbalance working condition information output by the graph neural network model is used by the control strategy generation module to identify abnormal behavior and locate the device detection report based on the three-phase imbalance working condition information. The specific phase-change switch control strategy is generated according to the device detection report, and then the control strategy generation module in the execution module is used to convert the phase-change switch control strategy into a specific operation command to control the actual action of the phase-change switch to adjust the three-phase load balance.

[0053] The technical scheme of the embodiment of the application comprises the following steps: obtaining state data and device connection relationships of each power device in a power system; determining graph structure data according to the state data and the device connection relationships; determining three-phase imbalance working condition information and a device detection report based on the graph structure data and a graph neural network model; determining a phase-change switch control strategy according to the device detection report, and controlling the phase-change switch based on the phase-change switch control strategy to adjust the three-phase load balance. The technical scheme uses the graph structure data processing capability of the graph neural network to accurately identify the three-phase imbalance problem of each node in the power grid, effectively reduces the impact of imbalance on the power grid through the rapid action of the phase-change switch, and significantly improves the fault handling capability and operation stability of the power grid.

[0054] Embodiment two

[0055] Figure 2 is a flowchart of a phase-change switch control method based on a graph neural network according to the second embodiment of the application. The embodiment is optimized based on the above-mentioned embodiment. The specific optimization is: determining three-phase imbalance working condition information and a device detection report based on graph structure data and a graph neural network model, comprising: inputting the graph structure data into the graph neural network model to obtain the three-phase imbalance working condition information; and generating the device detection report based on the three-phase imbalance working condition information and the graph neural network model. As shown in Figure 2 the method comprises:

[0056] S210, acquire state data and device connection relationship of each power device in the power system.

[0057] S220, determine graph structure data according to the state data and the device connection relationship.

[0058] S230, input the graph structure data into a graph neural network model to obtain three-phase unbalanced working condition information.

[0059] In the embodiment, the graph structure data can be input into the graph neural network model, node features and edge features are extracted through the graph neural network model, then node vectors and edge vectors are determined according to the node features and the edge features, unbalanced index data is determined based on the node vectors and the edge vectors, and three-phase unbalanced information of the node is identified according to the unbalanced index data.

[0060] In the embodiment, the graph structure data is input into the graph neural network model to obtain the three-phase unbalanced working condition information, including: inputting the graph structure data into the graph neural network model to extract node features and edge features; determining a feature vector of each node and a feature vector of each edge based on the node features and the edge features; performing aggregation processing based on neighbor node feature vectors of each node to obtain a target node feature vector; performing aggregation processing based on neighbor edge feature vectors of each edge to obtain a target edge feature vector; determining an unbalanced index according to the target node feature vector and the target edge feature vector, and identifying three-phase unbalanced working condition information of the node based on the unbalanced index.

[0061] The node features can include voltage, current, power factor and other features of the node. The edge features can include resistance, reactance and transmission capacity of the edge. The feature vector of each node can be the node features extracted and input into the graph neural network model to determine the corresponding node feature vector. The feature vector of the edge can be the edge features extracted and input into the graph neural network model for processing to determine the corresponding edge feature vector. The neighbor node feature vector can be understood as the feature vector of the neighbor node of the node. The neighbor edge feature vector can be the feature vector of the neighbor edge of the edge. The aggregation processing can be aggregation processing of the neighbor edge features of each edge by the graph neural network. The target node feature vector can be a new node feature vector obtained by the graph neural network performing aggregation processing on the neighbor node features of each node. The target edge feature vector can be a new edge feature vector obtained by performing aggregation processing on the neighbor edge features of each edge. The unbalanced index can be an unbalanced index of the node, which is used to indicate whether the node has an unbalanced problem.

[0062] Specifically, in this embodiment, preprocessed graph structure data is extracted from a graph database, and features are extracted for each node and edge. During the model training phase, this embodiment can label the nodes and edges in the graph structure to indicate whether a three-phase imbalance problem exists. The layer structure of the graph neural network model is designed, including an input layer, hidden layers, and an output layer. The hidden layers are used to extract high-level features of the graph, and the output layer is used to generate prediction results. The preprocessed graph structure data and the extracted node and edge features are input into the graph convolutional network model. The ReLU activation function and the Adam optimizer are selected, and the labeled graph structure data is used as the training set for model training to construct the graph neural network model.

[0063] In this embodiment, the graph structure data to be processed can be input into a trained graph neural network model, and the graph structure data can be traversed. For each node i, its feature vector is determined as follows. including voltage Current Power factor For each edge Its eigenvectors are determined as follows Including resistors Reactance Transmission capability By using a graph neural network model to aggregate the features of each node i's neighboring nodes, a new node feature vector is calculated. That is, the feature vector of the target node And aggregate each edge through a graph neural network model. Calculate the new edge feature vector from the neighbor edge features. That is, the target edge feature vector .

[0064] In this embodiment, the imbalance index can be determined based on the target node feature vector and the target edge feature vector. The specific method for identifying the three-phase imbalance condition information of the node based on the imbalance index can be to calculate the node's three-phase imbalance condition information using the target node feature vector and the target edge feature vector. Imbalance indicators Analyze the characteristic differences and average transmission capacity between nodes and their neighbors, and set a threshold. ,like Greater than If a three-phase imbalance problem is identified at node i, then the three-phase imbalance control information can be determined. In this embodiment, the imbalance index z for each node i can also be... i Calculate the optimal commutator control strategy.

[0065] Furthermore, the target node feature vector The calculation formula is:

[0066] ;

[0067] in, For the new node feature vector, To modify the activation function of the linear unit, For nodes The set of neighboring nodes, The weight matrix is ​​a learnable matrix. For nodes eigenvectors, For nodes eigenvectors, For the edge eigenvectors;

[0068] Target edge feature vector The calculation formula is:

[0069] ;

[0070] in, For the new edge feature vector, For nodes Feature vectors and nodes Concatenation and merging of eigenvectors The weight matrix used for edge feature updates. For the edge eigenvectors;

[0071] Imbalance Indicators The calculation formula is:

[0072] ;

[0073] in, For nodes The imbalance index is used to assess whether a node has a three-phase imbalance problem. For nodes New node feature vectors For nodes New node feature vectors For the edge New edge feature vectors In the graph structure, this represents the number of nodes directly connected to the target node i. For the edge New edge feature vector, when When it is larger, it represents a node. There is a three-phase imbalance problem.

[0074] The calculation formula for the optimal commutator control strategy in this embodiment can be:

[0075] ;

[0076] wherein, is a commutation switch control strategy of the node , and is an adjustable parameter in the control strategy, is a function for ensuring the output probability distribution, is an imbalance index of the node , the value range of the imbalance index is between 0 and 1, and when the imbalance index is close to 1, it indicates that the possibility of taking a certain control strategy is high. In this embodiment, by such a setting, three-phase imbalance industrial control information can be recognized through a graph neural network, intelligent decision-making is realized, three-phase imbalance problems can be quickly responded to, and the automation and intelligent level of the system is improved.

[0077] S240, generating a device detection report based on the three-phase imbalance working condition information and the graph neural network model.

[0078] In this embodiment, according to the three-phase imbalance industrial control information, abnormal detection features in the nodes can be extracted, the abnormal detection features are input into the graph neural network model to recognize abnormal scores corresponding to the nodes and edges, an abnormal detection coefficient is determined according to the abnormal scores, and the abnormal nodes and three-phase imbalance cause analysis are determined, and a corresponding device detection report is generated according to the abnormal nodes and three-phase imbalance cause analysis.

[0079] In this embodiment, optionally, generating a device detection report based on the three-phase imbalance working condition information and the graph neural network model includes: extracting abnormal detection features in the three-phase imbalance working condition information; inputting the abnormal detection features into the graph neural network model to obtain abnormal scores of each node; determining an abnormal detection coefficient based on the abnormal scores of each node; and forming a device detection report according to the abnormal detection coefficient and the three-phase imbalance working condition information.

[0080] wherein, the abnormal detection features can be features obtained by performing abnormal detection on the features of the nodes and edges. The abnormal detection features of this embodiment can include voltage deviation, current fluctuation, and power abnormality. The abnormal score can be an abnormal score of the nodes and edges obtained by inputting the abnormal detection features into the graph neural network model. The abnormal detection coefficient can be an abnormal detection system determined by comprehensively considering the abnormal scores of the nodes and edges.

[0081]

[0082] ​In this embodiment, graph structure data of the power system can be extracted from a graph database. This graph structure data may include the features of nodes and edges. The features of nodes and edges are processed to extract anomaly detection features, which are then input into a graph neural network model to identify normal and abnormal power system behavior patterns. Furthermore, in this embodiment, historical data of the anomaly detection features can also be input into the graph neural network model to train it, enabling it to recognize both normal and abnormal power system behavior patterns.

[0083] In this embodiment, real-time anomaly detection feature data can be input into a trained graph neural network model. The graph neural network model outputs anomaly scores for each node and edge, and the anomaly detection coefficients are obtained by combining the output results of the anomaly scores to analyze the correlation between abnormal behavior and three-phase imbalance. The fundamental embodiment can generate an equipment inspection report based on the results of anomaly detection and localization. The equipment inspection report includes information about the abnormal nodes (such as location, type, and anomaly score), an analysis of the causes of the three-phase imbalance, and suggested remedial measures.

[0084] Furthermore, the formula for calculating the anomaly score in this embodiment can be:

[0085] ;

[0086] in, For graph neural network models, nodes Output of abnormal scores, For nodes eigenvectors, For nodes The set of neighboring nodes, The weight matrix represents the feature changes. Here is the weight matrix of the graph convolutional layer. The bias vector of the graph convolutional layer. For the activation function, use Modify the linear unit activation function. This is an aggregate function that represents the summation operation. The higher the value, the greater the difference between the node's behavior and the normal pattern, and thus the more likely it is to be in an abnormal state.

[0087] In this embodiment, the formula for calculating the anomaly detection coefficient can be:

[0088] ;

[0089] Where Y is the anomaly detection coefficient, used to comprehensively consider the anomaly scores of all nodes in the entire power system; N is the total number of nodes in the graph. The weight of point i reflects the importance of that node in the power grid; is an abnormal score output by the graph neural network model for node i, is an abnormal score output by the graph neural network model for node j, is a set of neighbor nodes of point i, k is the number of neighbor nodes, used to normalize the abnormal score difference of neighbor nodes; is a regulation term, used to balance the influence of voltage and power factor abnormal scores on the final coefficient; is a regulation factor related to node i, determined based on historical abnormal records; is the power factor of node i, is a regulation index, used to control the degree of contribution of the power factor abnormality to the abnormality detection coefficient.

[0090] In this embodiment, through such a setting, the graph neural network can be used to quickly detect and locate abnormal behavior in the power system, identify the cause of three-phase imbalance, and generate a detection report, thereby improving the troubleshooting capability of the system, reducing troubleshooting time, and providing strong support for quickly restoring three-phase balance.

[0091] S250, determining a commutation switch control strategy according to the device detection report, and controlling the commutation switch based on the commutation switch control strategy to adjust the three-phase load balance.

[0092] In this embodiment, optionally, determining a commutation switch control strategy according to the device detection report includes: determining a to-be-adjusted node, an unbalance degree evaluation factor, and an influence factor based on the device detection report; sorting the to-be-adjusted node according to the unbalance degree evaluation factor and the influence factor to obtain a sorting result, and setting a three-phase balance target for the to-be-adjusted node; and determining a corresponding commutation switch control strategy according to the sorting result and the three-phase balance target.

[0093] The to-be-adjusted node can be understood as a node that needs to be adjusted. The unbalance degree evaluation factor can be an evaluation factor determined based on the unbalance condition of the node. The influence factor can be understood as a factor determined based on the emergency degree influence of the node. The sorting result can be a result of priority sorting of the node according to the unbalance degree factor and the influence factor. The three-phase balance target can be a target for adjusting the node to achieve three-phase balance. For example, the three-phase balance target can include balance standards of current, voltage, or power factor.

[0094] In this embodiment, the abnormal node identification, imbalance degree evaluation and influence range analysis of each node or edge can be obtained from the graph neural network model. The abnormal node identification indicates which nodes (such as loads or generators) have three-phase imbalance problems, the imbalance degree evaluation indicates the quantitative indicators of three-phase imbalance of each abnormal node, and the influence range analysis indicates the areas or node sets that may be affected by the imbalance problem. Then, according to the output content based on the graph neural network model, the nodes to be adjusted are prioritized according to the imbalance degree and emergency level, and specific three-phase balance targets are set for each node to be adjusted. Then, based on the prioritization results and the three-phase balance targets, the control strategy of the phase-changing switch is preliminarily planned, and the switch control instructions are generated and sent to the execution module for execution.

[0095] Further, in this embodiment, the process of controlling the actual action of the phase-changing switch in the execution module includes: receiving control instructions from the control strategy generation module, the control instructions including the opening and closing state, operation sequence and timing of the phase-changing switch; the execution module converts the received control instructions into switch operation instructions recognizable by the phase-changing switch actuator, specifies the phase-changing switch to be opened or closed, and the operation sequence and timing; the instructions are transmitted to the phase-changing switch actuator of the phase-changing switch through the communication network, the execution module continuously listens to the feedback information from the phase-changing switch actuator, confirms whether the instructions have been correctly executed, and the feedback information includes whether the operation is successfully executed, execution time, and key information of execution result; the execution module compares and analyzes the received feedback information with the expected effect, evaluates the adjustment effect of three-phase load balance; if the comparison and analysis finds that there is a deviation between the actual execution result and the expected effect, the execution module feeds back the deviation information of the comparison and analysis to the control strategy generation module, and the control strategy generation module adjusts the control strategy according to the feedback information to optimize the subsequent control operation.

[0096] In this embodiment, through such a setting, through the control strategy generation module, specific phase-changing switch control strategies can be generated according to the output results of the graph neural network model to adjust the three-phase load balance, and the abstract decision results are converted into executable control instructions to guide the actual operation of the phase-changing switch; through the execution module, the control instructions generated by the control strategy generation module are converted into specific operation commands to control the actual action of the phase-changing switch, which are sent to the phase-changing switch actuator in real time, and the feedback information after the execution of the phase-changing switch is received, compared and analyzed with the expected effect, and the control strategy is adjusted as needed to realize closed-loop control, continuously optimize the control strategy, and improve the stability and reliability of the system.

[0097] The three-phase imbalance condition commutating switch control management system based on a graph neural network in this embodiment can accurately identify the three-phase imbalance problems of each node in the power grid by using the graph structure data processing capability of the graph neural network, in-depth analyze the connection relationship and electrical parameters between nodes, quickly locate the imbalance source, and evaluate the imbalance degree. Once the imbalance problem is detected, control instructions can be immediately generated, and the influence of imbalance on the power grid can be effectively reduced through the rapid action of the commutating switch, thereby significantly improving the fault handling capability and operation stability of the power grid. By using the graph neural network, the three-phase imbalance problem can be quickly identified and responded to. The graph neural network model analyzes the data collected from the power system sensors in real time to identify abnormal nodes and imbalance areas. Compared with traditional monitoring and control systems, the dependence on manual intervention is reduced, the automation level is improved, and the system can take measures before the imbalance problem worsens, thereby reducing equipment damage and power quality degradation caused by imbalance.

[0098] The technical scheme of the embodiment of the present application acquires the state data and device connection relationship of each power device in the power system, determines the graph structure data according to the state data and device connection relationship, inputs the graph structure data into the graph neural network model to obtain three-phase imbalance condition information, generates a device detection report based on the three-phase imbalance condition information and the graph neural network model, determines a commutating switch control strategy according to the device detection report, and controls the commutating switch based on the commutating switch control strategy to adjust the three-phase load balance. The technical scheme uses the graph structure data processing capability of the graph neural network to accurately identify the three-phase imbalance problems of each node in the power grid, effectively reduces the influence of imbalance on the power grid through the rapid action of the commutating switch, and significantly improves the fault handling capability and operation stability of the power grid.

[0099] Embodiment three

[0100] Figure 3 is a structural schematic diagram of a commutating switch control device based on a graph neural network according to Embodiment three of the present application. As shown in Figure 3 , the device includes:

[0101] The data acquisition module 310 is configured to acquire the state data and device connection relationship of each power device in the power system.

[0102] The graph structure data determination module 320 is configured to determine the graph structure data according to the state data and device connection relationship.

[0103] The condition information determination module 330 is configured to determine the three-phase imbalance condition information and the device detection report based on the graph structure data and the graph neural network model.

[0104] The phase change switch control module 340 is configured to determine a phase change switch control strategy according to the device detection report, and control the phase change switch based on the phase change switch control strategy to adjust three-phase load balance.

[0105] Optionally, the data acquisition module 310 is specifically configured to acquire original state data of each power device in the power system respectively, perform a preprocessing operation on the original state data to obtain preprocessed state data, and integrate the preprocessed state data of each power device to obtain the state data of the power device.

[0106] Optionally, the device connection relationship includes positions of the power devices and connection relationships between the devices, and the graph structure data includes nodes and edges.

[0107] The graph structure data determination module 320 is specifically configured to determine each power device as a node, define the connection relationship between the devices as an edge, assign corresponding attribute data to the nodes and the edges based on the state data, and construct the graph structure data based on the nodes, the edges and the attribute information.

[0108] Optionally, the working condition information determination module 330 includes:

[0109] The three-phase imbalance working condition information obtaining unit is configured to input the graph structure data into a graph neural network model to obtain the three-phase imbalance working condition information.

[0110] The device detection report generation unit is configured to generate a device detection report based on the three-phase imbalance working condition information and the graph neural network model.

[0111] Optionally, the three-phase imbalance working condition information obtaining unit is specifically configured to input the graph structure data into the graph neural network model to extract node features and edge features, determine a feature vector of each node and a feature vector of each edge based on the node features and the edge features, perform aggregation processing on neighbor node feature vectors of each node to obtain a target node feature vector, perform aggregation processing on neighbor edge feature vectors of each edge to obtain a target edge feature vector, determine an imbalance index according to the target node feature vector and the target edge feature vector, and identify the three-phase imbalance working condition information of the node based on the imbalance index.

[0112] Optionally, the device detection report generation unit is specifically configured to extract an abnormality detection feature in the three-phase imbalance working condition information, input the abnormality detection feature into the graph neural network model to obtain an abnormality score of each node, determine an abnormality detection coefficient based on the abnormality score of each node, and form the device detection report according to the abnormality detection coefficient and the three-phase imbalance working condition information.

[0113] Optionally, the commutation switch control module 340 is specifically configured to determine the to-be-adjusted node, the unbalance degree evaluation factor and the influence factor based on the device detection report; sort the to-be-adjusted node according to the to-be-adjusted node, the unbalance degree evaluation factor and the influence factor to obtain a sorting result, and set a three-phase balance target for the to-be-adjusted node; and determine a corresponding commutation switch control strategy according to the sorting result and the three-phase balance target.

[0114] The commutation switch control device based on the graph neural network provided in the embodiment of the application can execute the commutation switch control method based on the graph neural network provided in any embodiment of the application, and has the corresponding function modules and beneficial effects of the execution method.

[0115] Embodiment four

[0116] Figure 4 is a structural schematic diagram of an electronic device according to Embodiment Four of the present application. The electronic device 10 is intended to represent various forms of digital computers, such as laptops, desktops, tablets, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular telephones, smartphones, wearable devices (such as headsets, glasses, watches, etc.), and other similar computing devices. The components shown here, their connections, and their functions, are meant to be examples only, and are not intended to limit the implementations of the present application as described and / or claimed in this document.

[0117] As shown in Figure 4 The electronic device 10 includes at least one processor 11, and a memory, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., which is communicatively connected to the at least one processor 11, wherein the memory stores a computer program that can be executed by the at least one processor. The processor 11 can execute various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0118] A plurality of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network, such as the Internet, and / or various telecommunication networks.

[0119] The processor 11 can be various general and / or special purpose processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 performs various methods and processes described above, such as the phase-change-over switch control method based on a graph neural network.

[0120] In some embodiments, the phase-change-over switch control method based on a graph neural network can be implemented as a computer program tangibly embodied in a computer readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded onto the RAM 13 and executed by the processor 11, one or more steps of the phase-change-over switch control method based on a graph neural network described above can be performed. Alternatively, in other embodiments, the processor 11 can be configured to perform the phase-change-over switch control method based on a graph neural network by any other appropriate means, such as by means of firmware.

[0121] Various implementations of the systems and techniques described above can be realized in digital electronic circuitry, integrated circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a system on a chip (SOC), a programmable logic device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0122] Computer programs for implementing the methods of the present application can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer program, when executed, enables the functions / acts specified in the flowcharts and / or block diagrams to be implemented. The computer program can be executed entirely on a machine, partially on a machine, partially on a machine as a standalone software package and partially on a remote machine or entirely on a remote machine or server.

[0123] In the context of the present application, a computer-readable storage medium can be a tangible medium that can contain or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. A computer-readable storage medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium can be a machine-readable signal medium. More specific examples of a machine-readable storage medium will include one or more lines of a program of instructions in a transitory signal, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0124] To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

[0125] The systems and techniques described herein can be implemented in a computing system that includes a back end component, e.g., as a data server, or that includes a middleware component, e.g., an application server, or that includes a front end component, e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described herein, or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication, e.g., a communication network. Examples of communication networks include a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0126] The computing system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. A server can be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system, to solve the defects of large management difficulty and weak business scalability in traditional physical host and VPS service.

[0127] It should be understood that the various forms of flow shown above can be re-ordered, added to, or deleted from without departing from the scope of the present disclosure. For example, the steps recited in the present disclosure can be executed in parallel, executed in series, or executed in different orders, as long as the desired results of the technical solutions of the present disclosure can be achieved, and the present disclosure is not limited herein.

[0128] The above detailed description does not constitute a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent replacements, and improvements made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A graph neural network-based control method for a phase-change switch, characterized in that, The method comprises: obtaining state data and device connection relationship of each power device in a power system; determining graph structure data according to the state data and the device connection relationship; determining three-phase imbalance working condition information and device detection report based on the graph structure data and a graph neural network model; determining a commutation switch control strategy according to the device detection report, and controlling a commutation switch based on the commutation switch control strategy to adjust three-phase load balance; determining three-phase imbalance working condition information and device detection report based on the graph structure data and a graph neural network model, comprising: inputting the graph structure data into a graph neural network model to obtain three-phase imbalance working condition information; generating a device detection report based on the three-phase imbalance working condition information and the graph neural network model; inputting the graph structure data into a graph neural network model to obtain three-phase imbalance working condition information, comprising: inputting the graph structure data into a graph neural network model to extract node features and edge features; determining a feature vector of each node and a feature vector of each edge based on the node features and the edge features; performing aggregation processing on neighbor node feature vectors of each node to obtain a target node feature vector; performing aggregation processing on neighbor edge feature vectors of each edge to obtain a target edge feature vector; determining an imbalance index according to the target node feature vector and the target edge feature vector, and identifying three-phase imbalance working condition information of a node based on the imbalance index.

2. The method of claim 1, wherein, Obtaining state data of each power device in a power system, comprising: respectively obtaining original state data of each power device in a power system; performing preprocessing operations on the original state data to obtain preprocessed state data; integrating the preprocessed state data of each power device to obtain state data of the power device.

3. The method of claim 1, wherein, The device connection relationship includes the positions of each power device and the connection relationship between devices; the graph structure data includes nodes and edges; determining graph structure data according to the state data and the device connection relationship, comprising: determining each power device as a node, and defining the connection relationship between the devices as an edge; assigning corresponding attribute data to the nodes and edges based on the state data; constructing graph structure data based on the nodes, edges, and attribute information.

4. The method of claim 1, wherein, Generating a device detection report based on the three-phase imbalance working condition information and the graph neural network model, comprising: extracting abnormal detection features in the three-phase imbalance working condition information; inputting the abnormal detection features into the graph neural network model to obtain an abnormal score of each node; determining an abnormal detection coefficient based on the abnormal score of each node; forming a device detection report according to the abnormal detection coefficient and the three-phase imbalance working condition information.

5. The method of claim 1, wherein, Determining a commutation switch control strategy according to the device detection report, comprising: determining a to-be-adjusted node, an imbalance degree evaluation factor, and an influence factor based on the device detection report; sorting the to-be-adjusted node according to the to-be-adjusted node, the imbalance degree evaluation factor, and the influence factor to obtain a sorting result, and setting a three-phase balance target for the to-be-adjusted node; According to the sorting result and the three-phase balance target, a corresponding phase change switch control strategy is determined.

6. A commutator switch control device based on a graph neural network, characterized by, Comprise: A data acquisition module for acquiring state data and device connection relationships of each power device in a power system; A graph structure data determination module for determining graph structure data according to the state data and the device connection relationships; A working condition information determination module for determining three-phase unbalanced working condition information and a device detection report based on the graph structure data and a graph neural network model; A phase change switch control module for determining a phase change switch control strategy according to the device detection report, and controlling a phase change switch based on the phase change switch control strategy to adjust three-phase load balance; The working condition information determination module comprises: A three-phase unbalanced working condition information obtaining unit for inputting the graph structure data into a graph neural network model to obtain three-phase unbalanced working condition information; A device detection report generation unit for generating a device detection report based on the three-phase unbalanced working condition information and the graph neural network model; The three-phase unbalanced working condition information obtaining unit is specifically configured to input the graph structure data into a graph neural network model to extract node features and edge features; determine a feature vector of each node and a feature vector of each edge based on the node features and the edge features; perform aggregation processing on neighbor node feature vectors of each node to obtain a target node feature vector; perform aggregation processing on neighbor edge feature vectors of each edge to obtain a target edge feature vector; determine an unbalance index according to the target node feature vector and the target edge feature vector, and identify three-phase unbalanced working condition information of the node based on the unbalance index.

7. An electronic device, comprising: The electronic device comprises: At least one processor; and A memory connected in communication with the at least one processor; wherein The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the graph neural network-based phase change switch control method of any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer instructions for enabling the processor to execute the graph neural network-based phase change switch control method of any one of claims 1-5 when executed.

Images