Generation method, device, equipment and product of voltage sag propagation characteristics of power distribution network

Abstract

The embodiment of the invention provides a method, a device, equipment and a product for generating voltage sag propagation characteristics of a power distribution network, and is applied to the technical field of power system automation. In response to a voltage sag event, acquiring an electrical parameter value of each current electrical connection point of the power distribution network and topological structure data of the power distribution network; wherein the topological structure data is used for indicating the connection relationship between the electrical connection points; generating a data characteristic matrix corresponding to the voltage sag event based on electrical parameter values of a plurality of electrical connection points and the topological structure data; inputting the data feature matrix into a pre-trained voltage sag positioning prediction model to obtain predicted fault positioning corresponding to the data feature matrix; and inputting the predicted fault location into a digital twinborn model corresponding to the power distribution network to obtain a propagation characteristic corresponding to the voltage sag event. The method achieves the technical effect of high voltage sag propagation characteristic determination efficiency.

Description

Technical Field

[0001] The present application relates to the field of power system automation technology, and in particular to a method, device, equipment and product for generating voltage sag propagation characteristics of a distribution network. Background Art

[0002] Distribution networks are shifting towards a digital structure with a high proportion of distributed renewable energy. In this process, voltage sags in distribution networks have become a major issue affecting power quality. Analyzing the propagation characteristics of voltage sags can provide a scientific basis for power quality management and fault recovery.

[0003] In the prior art, analysis of voltage sags mainly focuses on identifying the types of voltage sags and analyzing the propagation characteristics of voltage sags. The analysis of the propagation characteristics of voltage sags mainly relies on manual identification of the propagation characteristics of voltage sags.

[0004] Since the determination of voltage sag propagation characteristics in the prior art relies on manual identification, there is a technical problem in that the efficiency of determining voltage sag propagation characteristics is low. Summary of the Invention

[0005] The embodiments of the present application provide a method, apparatus, device and product for generating a distribution network voltage sag propagation feature, so as to achieve the technical effect of improving the accuracy of generating a private distribution network voltage sag propagation feature.

[0006] In a first aspect, an embodiment of the present application provides a method for generating a distribution network voltage sag propagation feature, comprising:

[0007] In response to a voltage sag event, collecting electrical parameter values of each electrical connection point of the distribution network and topological data of the distribution network; wherein the topological data is used to indicate the connection relationship between each electrical connection point;

[0008] Generate a data feature matrix corresponding to a voltage sag event based on electrical parameter values of multiple electrical connection points and topological structure data;

[0009] Input the data feature matrix into the pre-trained voltage sag location prediction model to obtain the predicted fault location corresponding to the data feature matrix;

[0010] The predicted fault location is input into the digital twin model of the distribution network to obtain the propagation characteristics corresponding to the voltage sag event;

[0011] Among them, the digital twin model is a virtual mapping of the distribution network, which is used to reflect the operating status of the distribution network; the voltage sag positioning prediction model refers to a graph neural network model obtained by model training based on the historical voltage sag data of the distribution network.

[0012] In one possible implementation, the predicted fault location is input into the digital twin model corresponding to the distribution network to obtain the propagation characteristics corresponding to the voltage sag event, including:

[0013] Input the preset fault type, preset fault impedance, preset fault duration, and predicted fault location corresponding to the voltage sag event into the digital twin model, and calculate the voltage sag ratio of each electrical connection point in the distribution network in the digital twin model;

[0014] Based on the voltage sag ratio of each electrical connection point, a feature vector matrix is constructed to obtain the propagation characteristics corresponding to the voltage sag event;

[0015] The preset fault type refers to any one of a single-phase ground fault, a phase-to-phase fault, and a three-phase ground fault; the preset fault impedance is used to indicate the severity of the voltage sag event; and the preset fault duration refers to the duration of the voltage sag event.

[0016] In one possible implementation, before collecting the current electrical parameter values of each electrical connection point of the distribution network and the topology data of the distribution network in response to a voltage sag event, the method further includes:

[0017] Obtain the electromagnetic transient model corresponding to each component in the distribution network, as well as the topological structure data of the distribution network;

[0018] Integrate topology data and electromagnetic transient models to obtain a digital twin model corresponding to the distribution network;

[0019] Among them, the components of the distribution network include: distributed power sources, energy storage systems, transmission lines, transformers, and loads.

[0020] In one possible implementation, obtaining the electromagnetic transient model corresponding to each component in the distribution network and topological structure data of the distribution network includes:

[0021] Based on a preset model category of each component in the distribution network, an electromagnetic transient model corresponding to each component is constructed; wherein the preset model category is an equivalent circuit model or a dynamic model;

[0022] Constructing topological structure data of the distribution network based on the connection relationship between multiple electrical connection points in the distribution network;

[0023] The electrical connection point refers to the connection location between the components.

[0024] In one possible implementation, based on a preset model category of each component in the distribution network, an electromagnetic transient model corresponding to each component is constructed, including:

[0025] Collecting device information corresponding to each component in the distribution network; wherein the device information includes: the electrical parameter type of the component and the rated parameter value corresponding to each electrical parameter type;

[0026] Determine the model parameters corresponding to each component according to the device information of each component;

[0027] According to the preset model category and model parameters corresponding to each component, an electromagnetic transient model corresponding to each component is constructed.

[0028] In one possible implementation, before collecting the current electrical parameter values of each electrical connection point of the distribution network and the topology data of the distribution network in response to a voltage sag event, the method further includes:

[0029] Generate a preset amount of voltage sag fault data based on the digital twin model corresponding to the distribution network;

[0030] The initial graph neural network model is trained based on the voltage sag data to obtain a model to be determined;

[0031] The loss value and accuracy of the model to be determined are calculated. When the loss value of the model to be determined is lower than the preset loss value and the accuracy is higher than the preset accuracy, the model to be determined is determined as the voltage sag positioning prediction model.

[0032] In a second aspect, an embodiment of the present application provides a device for generating a distribution network voltage sag propagation feature, comprising:

[0033] an acquisition module, configured to collect, in response to a voltage sag event, electrical parameter values of each electrical connection point of the distribution network and topological data of the distribution network; wherein the topological data is used to indicate a connection relationship between each electrical connection point;

[0034] A first processing module is configured to generate a data feature matrix corresponding to a voltage sag event based on electrical parameter values of a plurality of electrical connection points and topological structure data;

[0035] The second processing module is used to input the data feature matrix into a pre-trained voltage sag location prediction model to obtain the predicted fault location corresponding to the node feature matrix;

[0036] The third processing module is used to input the predicted fault location into the digital twin model corresponding to the distribution network to obtain the propagation characteristics corresponding to the voltage sag event;

[0037] Among them, the digital twin model is a virtual mapping of the distribution network, which is used to reflect the operating status of the distribution network; the voltage sag positioning prediction model refers to a graph neural network model obtained by model training based on the historical voltage sag data of the distribution network.

[0038] In a possible implementation, the third processing module is further configured to:

[0039] Input the preset fault type, preset fault impedance, preset fault duration, and predicted fault location corresponding to the voltage sag event into the digital twin model, and calculate the voltage sag ratio of each electrical connection point in the distribution network in the digital twin model;

[0040] Based on the voltage sag ratio of each electrical connection point, a feature vector matrix is constructed to obtain the propagation characteristics corresponding to the voltage sag event;

[0041] The preset fault type refers to any one of a single-phase ground fault, a phase-to-phase fault, and a three-phase ground fault; the preset fault impedance is used to indicate the severity of the voltage sag event; and the preset fault duration refers to the duration of the voltage sag event.

[0042] In a possible implementation, the acquisition module is further configured to:

[0043] Obtain the electromagnetic transient model corresponding to each component in the distribution network, as well as the topological structure data of the distribution network;

[0044] Integrate topology data and electromagnetic transient models to obtain a digital twin model corresponding to the distribution network;

[0045] Among them, the components of the distribution network include: distributed power sources, energy storage systems, transmission lines, transformers, and loads.

[0046] In a possible implementation, the acquisition module is further configured to:

[0047] Based on a preset model category of each component in the distribution network, an electromagnetic transient model corresponding to each component is constructed; wherein the preset model category is an equivalent circuit model or a dynamic model;

[0048] Constructing topological structure data of the distribution network based on the connection relationship between multiple electrical connection points in the distribution network;

[0049] The electrical connection point refers to the connection location between the components.

[0050] In a possible implementation, the acquisition module is further configured to:

[0051] Collecting device information corresponding to each component in the distribution network; wherein the device information includes: the electrical parameter type of the component and the rated parameter value corresponding to each electrical parameter type;

[0052] Determine the model parameters corresponding to each component according to the device information of each component;

[0053] According to the preset model category and model parameters corresponding to each component, an electromagnetic transient model corresponding to each component is constructed.

[0054] In a possible implementation, the acquisition module is further configured to:

[0055] Generate a preset amount of voltage sag fault data based on the digital twin model corresponding to the distribution network;

[0056] The initial graph neural network model is trained based on the voltage sag data to obtain a model to be determined;

[0057] The loss value and accuracy of the model to be determined are calculated. When the loss value of the model to be determined is lower than the preset loss value and the accuracy is higher than the preset accuracy, the model to be determined is determined as the voltage sag positioning prediction model.

[0058] In a third aspect, an embodiment of the present application provides an electronic device, comprising: a memory, a processor;

[0059] Memory stores computer-executable instructions;

[0060] The processor executes the computer-executable instructions stored in the memory, so that the processor executes the above first aspect and various possible implementations of the first aspect.

[0061] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the above-mentioned first aspect and various possible implementation methods of the first aspect.

[0062] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the above first aspect and various possible implementation methods of the first aspect.

[0063] The embodiments of the present application provide a method, device, equipment and product for generating the propagation characteristics of voltage sag in a distribution network. The method collects the electrical parameter values of each electrical connection point in the distribution network and the topological data structure corresponding to the distribution network when a voltage sag event occurs; generates a data feature matrix based on the electrical parameter values and the topological data structure; inputs the data feature matrix into a pre-trained voltage sag location prediction model to obtain the predicted fault location corresponding to the voltage sag event; and feeds the predicted fault location back to the digital twin model for inference of the propagation characteristics to obtain the propagation characteristics corresponding to the voltage sag event. Compared with the prior art, the present application uses a combination of deep learning models and digital twin models to determine the positioning nodes and propagation characteristics corresponding to the voltage sag event, thereby improving the technical effect of the efficiency of determining the voltage sag propagation characteristics. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0065] Figure 1 Schematic diagram of the process of generating the distribution network voltage sag propagation characteristics provided by this application Figure 1 ;

[0066] Figure 2 A flowchart of a method for training a voltage sag location prediction model provided in this application;

[0067] Figure 3 A structural diagram of a graph neural network layer is provided for an embodiment of the present application;

[0068] Figure 4 A flow chart of the method for generating a digital twin model of a distribution network provided in this application;

[0069] Figure 5 Schematic diagram of the structure of the electromagnetic transient model of the distributed power supply provided in the embodiment of the present application Figure 1 ;

[0070] Figure 6 Schematic diagram of the structure of the electromagnetic transient model of the distributed power supply provided in the embodiment of the present application Figure 2 ;

[0071] Figure 7 Schematic diagram of the structure of the electromagnetic transient model of the distributed power supply provided in the embodiment of the present application Figure 3 ;

[0072] Figure 8 A schematic structural diagram of an electromagnetic transient model of an energy storage system provided in an embodiment of the present application;

[0073] Figure 9 A schematic structural diagram of an equivalent circuit model of a digital twin model provided in an embodiment of the present application;

[0074] Figure 10 A schematic diagram of the structure of the load electromagnetic transient model provided in an embodiment of the present application;

[0075] Figure 11 A schematic diagram of the structure of the transformer electromagnetic transient model provided in an embodiment of the present application;

[0076] Figure 12 Schematic diagram of the process of generating the distribution network voltage sag propagation characteristics provided by this application Figure 2 ;

[0077] Figure 13A schematic diagram of the structure of a digital twin model provided in an embodiment of the present application;

[0078] Figure 14 A schematic diagram of the iterative process of an initial graph neural network model provided in an embodiment of the present application;

[0079] Figure 15 A schematic diagram of the distribution of voltage sag propagation characteristics provided in an embodiment of the present application;

[0080] Figure 16 This is a schematic diagram of the structure of the device for generating the distribution network voltage sag propagation characteristics provided by this application;

[0081] Figure 17 This is a schematic diagram of the structure of the electronic device provided in this application.

[0082] The above drawings illustrate specific embodiments of the present application, which will be described in more detail below. These drawings and the textual description are not intended to limit the scope of the present application in any way, but rather to illustrate the concepts of the present application to those skilled in the art by reference to specific embodiments. DETAILED DESCRIPTION

[0083] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0084] In existing technologies, propagation characteristics analysis of voltage sags is necessary to fully understand the propagation patterns of voltage sags in distribution networks and provide a scientific basis for subsequent power quality management and fault recovery. This propagation characteristic analysis primarily locates the fault causing the current voltage sag. Based on the fault location and the topology of the distribution network, operations and maintenance personnel search for monitoring data from various components and electrical connection points in the distribution network. Based on this data, they derive propagation characteristics to determine the propagation characteristics of the voltage sag.

[0085] However, the propagation characteristic analysis method in the prior art relies on manual search. If the topology of the distribution network is complex and the coverage area is large, the number of nodes for locating the corresponding voltage sag fault will increase. The use of manual search to analyze the propagation characteristics of the voltage sag requires a lot of manpower and time costs, which leads to the technical problem of low efficiency in determining the voltage sag propagation characteristics in the prior art.

[0086] In response to the above technical problems, the present application proposes the following technical concept: using a combination of deep learning models and digital twin models to determine the fault location and propagation characteristics of voltage sag events. Specifically: collecting the electrical parameter values of each electrical connection point in the distribution network and the topological data structure corresponding to the distribution network when the voltage sag event occurs; generating a data feature matrix based on the electrical parameter values and the topological data structure; inputting the data feature matrix into a pre-trained voltage sag location prediction model to obtain the predicted fault location corresponding to the voltage sag event; feeding the predicted fault location back to the digital twin model for propagation feature reasoning to obtain the propagation features corresponding to the voltage sag event. Compared with the existing technology, the present application uses a deep learning model to achieve accurate, fast and efficient guzheng positioning, and uses a digital twin model to perform reasoning and analysis of propagation features, thereby obtaining the positioning nodes and propagation features corresponding to the voltage sag event, achieving the technical effect of improving the efficiency of determining the voltage sag propagation features.

[0087] The following specific embodiments describe in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0088] Figure 1 Schematic diagram of the process of generating the distribution network voltage sag propagation characteristics provided by this application Figure 1 ,like Figure 1 As shown, the method includes:

[0089] S101 . In response to a voltage sag event, collect electrical parameter values of each electrical connection point of a distribution network and topological structure data of the distribution network.

[0090] In this step, the topology data is used to indicate the connection relationship between each electrical connection point. The electrical parameters include: voltage-related parameters, current-related parameters, power-related parameters, frequency-related parameters, and harmonic impedance-related parameters.

[0091] Optionally, electrical parameter values for each electrical connection point in the distribution network may be collected using a smart meter, a synchronized phasor measurement unit, a current transformer, a voltage transformer, a power supply monitoring device, or a fault recorder. The tool used to collect the parameters for each electrical connection point is determined by the location and type of the electrical connection point. Therefore, it is necessary to select an appropriate tool for collecting electrical parameter values based on the characteristics of each electrical connection point.

[0092] Exemplarily, there is an electrical connection point, which is a busbar of a distribution line, connecting a distributed power source, a load, and a transmission line.

[0093] The electrical parameters of the electrical connection point include: voltage amplitude, harmonic voltage distortion rate, current amplitude, current phase angle, harmonic current content, active power, reactive power, power factor, system frequency, and load-side harmonic impedance.

[0094] S102: Generate a data feature matrix corresponding to a voltage sag event based on electrical parameter values of a plurality of electrical connection points and topology data.

[0095] In this step, the data feature matrix is a two-dimensional matrix used to describe the electrical parameter values and topological relationships of each electrical connection point in the distribution network.

[0096] Optionally, a possible implementation of data feature matrix generation is:

[0097] S1021. Determine the connection relationship between the electrical connection points based on the topological data structure, and construct an adjacency matrix according to the connection relationship.

[0098] S1022. Arrange the electrical parameter values of each electrical connection point according to the electrical connection points in the adjacency matrix to obtain a node feature matrix.

[0099] S1023. Determine a combination of the adjacency matrix and the node feature matrix as a data feature matrix.

[0100] For example, a distribution network has four electrical connection points, and the corresponding topological structure data is: electrical connection point 1 is connected to electrical connection point 2 and electrical connection point 3, and electrical connection point 2 is connected to electrical connection point 4. The electrical parameter value of each electrical connection point is:

[0101] Electrical connection point 1: voltage 10.0 kV, current 200 A, active power 1.0 MW, reactive power 0.5 MVar, frequency 50.0 Hz, harmonic distortion rate 2%.

[0102] Electrical connection point 2: voltage 9.5 kV, current 250 A, active power 0.8 MW, reactive power 0.6 MVar, frequency 49.9 Hz, harmonic distortion rate 5%.

[0103] Electrical connection point 3: voltage 8.5 kV, current 300 A, active power 0.5 MW, reactive power 1.2 MVar, frequency 49.8 Hz, harmonic distortion rate 8%.

[0104] Electrical connection point 4: voltage 9.0 kV, current 220 A, active power 0.7 MW, reactive power 0.8 MVar, frequency 49.9 Hz, harmonic distortion rate 4%.

[0105] The adjacency matrix determined based on the topological structure data is: A=[(0 1 1 0),(1 0 0 1),(1 0 0 0),(01 0 0)]; the adjacency matrix is a 4×4 two-dimensional matrix.

[0106] The node characteristic matrix determined based on the adjacency matrix and the electrical parameter values of each electrical connection point is: X=[(10.0 9.5 8.5 9.0),(200 250 300 220),(1.0 0.8 0.5 0.7),(0.5 0.6 1.2 0.8),(50.0 49.9 49.8 49.9),(2 5 8 4)]; the node characteristic matrix is a 4×6 two-dimensional matrix.

[0107] S103 , inputting the data feature matrix into a pre-trained voltage sag location prediction model to obtain a predicted fault location corresponding to the data feature matrix.

[0108] In this step, the voltage sag location prediction model refers to a graph neural network model obtained by model training based on historical voltage sag data of the distribution network.

[0109] Predictive fault location refers to the electrical connection point that caused the voltage sag event. Optionally, predictive fault location includes not only the location of the electrical connection point but also the fault type corresponding to the voltage sag event. The fault type can be any of single-phase ground fault, phase-to-phase fault, and three-phase ground fault.

[0110] It should be noted that the training method of the voltage sag location prediction model in this step is as follows Figure 2 Further explanation is given in the embodiment shown and no redundant description is given here.

[0111] S104: Input the predicted fault location into the digital twin model corresponding to the distribution network to obtain the propagation characteristics corresponding to the voltage sag event.

[0112] In this step, the digital twin model is a virtual mapping of the distribution network, which is used to reflect the operating status of the distribution network.

[0113] Optionally, a possible implementation method for determining the propagation characteristics is:

[0114] S1041. Input the preset fault type, preset fault impedance, preset fault duration, and predicted fault location corresponding to the voltage sag event into the digital twin model, and calculate the voltage sag ratio of each electrical connection point of the distribution network in the digital twin model.

[0115] In this step, the preset fault type refers to any one of a single-phase ground fault, a phase-to-phase fault, and a three-phase ground fault; the preset fault impedance is used to indicate the severity of the voltage sag event; and the preset fault duration refers to the duration of the voltage sag event.

[0116] Optionally, the voltage sag ratio is calculated as follows:

[0117] a1. Input the preset fault type, preset fault impedance, preset fault duration, and predicted fault location into the digital twin model.

[0118] a2. Applying the preset fault type, preset fault impedance, and preset fault duration to the determined predicted fault location.

[0119] For example, there are 20 electrical connection points in the distribution network, the predicted fault location is electrical connection point 15, the preset fault type is a single-phase grounding fault, and the preset fault impedance is 10Ω. Then, in the digital twin model, electrical connection point 15 is set as a single-phase grounding fault, and the fault impedance is set to 10Ω.

[0120] a3. Run dynamic simulation based on the digital twin model to simulate the grid behavior after a voltage sag fault occurs, collect the electrical parameter values of each electrical connection point in real time, and obtain simulation results.

[0121] a4. Extract the voltage amplitude change data of each electrical connection point from the simulation results.

[0122] a5. Based on the voltage amplitude change data of each electrical connection point, calculate the voltage amplitude drop ratio of each electrical connection point to obtain the voltage sag ratio of each electrical connection point.

[0123] Exemplarily, the voltage amplitude drop ratio is calculated by calculating the difference between the rated voltage value and the fault voltage value of the electrical connection point, and calculating the ratio between the difference and the rated voltage value to obtain the voltage amplitude drop ratio.

[0124] S1042: Construct a characteristic vector matrix based on the voltage sag ratio of each electrical connection point to obtain propagation characteristics corresponding to the voltage sag event.

[0125] In this step, the characteristic vector matrix based on the voltage sag ratio can be constructed by arranging the voltage sag ratios of each electrical connection point according to the node arrangement order in the digital twin model to obtain a propagation feature. The propagation feature is a vector that identifies the voltage sag ratio of each electrical connection point.

[0126] The embodiment of the present application provides a method for generating propagation characteristics of voltage sags in a distribution network. The method collects the electrical parameter values of each electrical connection point in the distribution network and the corresponding topological data structure of the distribution network when a voltage sag event occurs; generates a data feature matrix based on the electrical parameter values and the topological data structure; inputs the data feature matrix into a pre-trained voltage sag location prediction model to obtain a predicted fault location corresponding to the voltage sag event; and feeds the predicted fault location back to a digital twin model for propagation feature inference to obtain the propagation characteristics corresponding to the voltage sag event. Compared with the prior art, the present application utilizes a combination of deep learning models and digital twin models to determine the location nodes and propagation characteristics corresponding to the voltage sag event, thereby improving the technical effect of the efficiency of determining voltage sag propagation characteristics.

[0127] Figure 2 This is a flow chart of the training method of the voltage sag location prediction model provided in this application. Figure 1 Based on the embodiment shown, the training of the voltage sag location prediction model in step S103 of this embodiment is further explained, as shown in FIG. Figure 2 As shown, the method includes:

[0128] S201. Generate a preset amount of voltage sag fault data based on the digital twin model corresponding to the distribution network.

[0129] In this step, a possible implementation method for generating a preset amount of voltage sag fault data based on the digital twin model is:

[0130] S2011. Inject the fault type and fault location into the digital twin model, run the digital twin model simulation, and obtain the electrical parameter values of each electrical connection point.

[0131] S2012. Perform data cleaning and normalization processing on the acquired electrical parameter values of the multiple electrical connection points.

[0132] S2013. Extract the connection relationship between each electrical connection point from the digital twin model and construct an adjacency matrix.

[0133] S2014. Construct a node feature matrix corresponding to each voltage sag fault data based on the adjacency matrix and the electrical parameter values of each electrical connection point.

[0134] S2015 : Generate a label corresponding to each piece of voltage sag fault data based on the fault type and fault location corresponding to the piece of voltage sag fault data.

[0135] For example, the preset number is 3000, and the number of electrical connection points in the distribution network corresponding to the digital twin model is 10. The fault types include single-phase ground fault, phase-to-phase fault, and three-phase short circuit. At least 50 faults are simulated at each electrical connection point. The fault impedances are low impedance, medium impedance, and high impedance. 500 faults are simulated at each electrical connection point to obtain 500 data points. Each fault type covers all nodes, resulting in 3 times 500 data points, or 1500 data points. Three impedances are set for each fault scenario, resulting in 3 times 1500 data points, and the final voltage sag fault data is 4500. 3000 data points are randomly selected from the 4500 data points as the voltage sag data for the final model training. Each voltage sag data point has a corresponding label, and the label of each voltage sag data point includes at least one of the following: fault location, fault type, and fault impedance.

[0136] S202: Perform model training on the initial graph neural network model based on the voltage sag data to obtain a model to be determined.

[0137] In this step, the initial graph neural network model used mainly includes a graph neural network layer and a convolutional layer, where the graph neural network layer consists of multiple nodes and edges.

[0138] For example, Figure 3 A structural diagram of a graph neural network layer is provided for the embodiment of the present application, such as Figure 3 As shown, the graph neural network layer includes multiple nodes and multiple edges. The representation of each node is updated by the features of its adjacent nodes. For a graph, the operation of each layer of GCN can be expressed by Formula 1:

[0139]

[0140] Among them, H (l) is the node feature matrix of the lth layer, H (0) is the input feature matrix. is the normalized adjacency matrix, W (l) is the weight matrix of the lth layer, σ is the activation function, H (l+1) It is the node feature after convolution operation.

[0141] in, It is obtained by adding self-loops, which represents the adjacency relationship between nodes. Refer to Formula 2:

[0142]

[0143] Where A is the original adjacency matrix, I is the identity matrix, and D is the degree matrix. The calculation formula of the degree matrix is shown in Formula 3:

[0144]

[0145] Among them, D ii is the diagonal element in the degree matrix, indicating the degree of node i, A ij Indicates whether there is an edge between node i and node j.

[0146] S203 , calculating the loss value and accuracy of the model to be determined, and when the loss value of the model to be determined is lower than a preset loss value and the accuracy is higher than the preset accuracy, determining the model to be determined as a voltage sag location prediction model.

[0147] In this step, the loss value calculation of the voltage sag location prediction model can use various types of loss functions. The accuracy refers to the consistency between the fault location predicted by the calculation model and the actual label of each data.

[0148] For example, the cross entropy loss function is used to calculate the loss value, as shown in Formula 4:

[0149]

[0150] Among them, N is the number of samples, C is the number of categories, and y i,c is the true label of sample i belonging to category c, 1 means it belongs to category c, and 0 means it does not belong to category c; The probability of category c predicted by the model.

[0151] Formula 5 can be used to calculate the accuracy:

[0152]

[0153] In order to prevent overfitting during model training, a regularization formula is needed to simplify the model structure so that the model is more inclined to learn the overall laws of the data rather than local noise. The regularization formula is shown in Formula 6:

[0154]

[0155] Among them, λ is the regularization coefficient, W is the parameter of the model, and L_reg is the regularization processing result.

[0156] In this embodiment, the digital twin model is used to generate model training data, which does not require high-frequency collection of distribution network data. At the same time, a large amount of training data can be generated quickly and efficiently, reducing the overfitting phenomenon in the model training process. At the same time, the classification of the generated data in multiple data categories remains consistent, which can effectively improve the accuracy of the voltage sag positioning prediction model.

[0157] Figure 4This is a flow chart of the method for generating a digital twin model of a distribution network provided in this application. Based on the above embodiment, the construction of the digital twin model in step S104 of this embodiment is further explained in detail. Figure 4 As shown, the method includes:

[0158] S401. Obtain an electromagnetic transient model corresponding to each component in the distribution network, as well as topological structure data of the distribution network.

[0159] In this step, the components of the distribution network include: distributed power sources, energy storage systems, transmission lines, transformers, and loads.

[0160] Optionally, a possible implementation method for obtaining the electromagnetic transient model and topological data structure is:

[0161] S4011. Based on the preset model category of each component in the distribution network, construct an electromagnetic transient model corresponding to each component.

[0162] In this step, the preset model type is an equivalent circuit model or a dynamic model.

[0163] Alternatively, a possible implementation method of constructing the electromagnetic transient model corresponding to each component is:

[0164] b1. Collect the device information corresponding to each component in the distribution network.

[0165] In this step, the device information includes: the electrical parameter types of the component components, and the rated parameter values corresponding to each electrical parameter type.

[0166] For example, when the component is a digital display circuit, the device information includes: resistance, inductance, capacitance, and their corresponding rated parameter values.

[0167] b2. Determine the model parameters corresponding to each component based on the device information of each component.

[0168] For example, when the component is a transformer, the model parameters determined based on the device information may be leakage impedance, resistance component, and reactance component. The model parameters are used to describe the electrical characteristics of the component.

[0169] b3. Construct an electromagnetic transient model corresponding to each component based on the preset model category and model parameters corresponding to each component.

[0170] In this step, the preset model type may be an equivalent circuit model or a dynamic model.

[0171] Exemplarily, the preset model category corresponding to the transmission line is an equivalent circuit model, and the model parameters are inductance and capacitance; the preset model category corresponding to the transformer is an equivalent circuit model, and the model parameters are leakage impedance, resistance component, and reactance component.

[0172] For example, Figures 5 to 7 Schematic diagram of the structure of the electromagnetic transient model of three distributed power sources provided in this application. Figure 5 Schematic diagram of the structure of the electromagnetic transient model of the distributed power supply provided in the embodiment of the present application Figure 1 ,like Figure 5 As shown, an embodiment of the present application provides an electromagnetic transient model of a direct-drive wind turbine, which includes: a permanent magnet synchronous motor, a pitch angle control model, a machine-side converter, a DC circuit, a grid-side converter, a filter circuit, a machine-side converter control model, and a grid-side converter control model. Figure 6 Schematic diagram of the structure of the electromagnetic transient model of the distributed power supply provided in the embodiment of the present application Figure 2 ,like Figure 6 As shown, an embodiment of the present application provides an electromagnetic transient model of a doubly fed wind turbine, which includes: a pitch angle control model, an asynchronous motor, a machine-side converter, a machine-side converter control model, a DC circuit, a grid-side converter, a grid-side converter control model, and a filter circuit. Figure 7 Schematic diagram of the structure of the electromagnetic transient model of the distributed power supply provided in the embodiment of the present application Figure 3 ,like Figure 7 As shown, an embodiment of the present application provides an electromagnetic transient model of a photovoltaic power generation system, which includes: a photovoltaic cell, a maximum power point tracking control, a boost / buck converter, a boost / buck converter control, a converter, a converter control model, and a filter circuit.

[0173] For example, Figure 8 A schematic diagram of the structure of an electromagnetic transient model of an energy storage system provided in an embodiment of the present application is shown in FIG. Figure 8 As shown, the model includes: a battery, a converter, a converter control model, and a filter circuit.

[0174] S4012. Construct topological structure data of the distribution network based on the connection relationship between multiple electrical connection points in the distribution network.

[0175] In this step, the electrical connection point refers to the connection position between the component elements.

[0176] S402: Integrate the topology data and the electromagnetic transient model to obtain a digital twin model corresponding to the distribution network.

[0177] Optionally, in the process of building a digital twin model of the distribution network, it is necessary to accurately simulate and analyze the harmonic problems in the distribution network. Therefore, a harmonic impedance model can be constructed and integrated into the digital twin model to describe the dynamic behavior of nonlinear compliance.

[0178] For example, Figure 9 A schematic diagram of the structure of an equivalent circuit model of a digital twin model provided in an embodiment of the present application is shown in FIG. Figure 9 As shown, the model includes the distribution network side and the load side; I s , Z s , I l , Z l and U PCC They represent the short-circuit current on the distribution network side, the distribution network input impedance, the short-circuit current on the load side, the load input impedance and the voltage at the point of common coupling (PCC). The load harmonic impedance is calculated based on the voltage and current measurements at the PCC point.

[0179] The relationship at the PCC point is shown in Formula 7:

[0180]

[0181] Among them, U PCC Refers to the voltage at PCC point, I PCC Refers to the current at the PCC point, Z l Refers to the load input impedance, I l Refers to the short-circuit current on the load side.

[0182] Then the relationship of the Kth sampling point is shown in Formula 8:

[0183]

[0184] Among them, Z l (k) refers to the grid-side equivalent impedance of the Kth sampling point, U l (k) refers to the grid side voltage at the Kth sampling point, U PCC (k) refers to the grid side voltage at the Kth sampling point, I PCC (k) refers to the PCC point current at the Kth sampling point.

[0185] Since the impedance of the power grid does not change much in a short time, an optimization objective function is established as shown in Formula 9:

[0186]

[0187] Where J refers to the optimization objective function, represents the grid-side equivalent impedance of the Kth sampling point; m represents the total number of sampling points, Ul (k) refers to the grid side voltage at the Kth sampling point, U PCC (k) refers to the grid side voltage at the Kth sampling point, I PCC (k) refers to the PCC current of the Kth sampling point, U l (i) refers to the grid side voltage at the i-th sampling point, U PCC (i) refers to the grid side voltage at the i-th sampling point, I PCC (i) refers to the PCC point current at the i-th sampling point; the estimated value of the grid side impedance is optimized by optimizing the objective function to keep it consistent between different sampling points.

[0188] The equivalent voltage U on the load side is calculated using the optimization objective function l , the load input impedance Z is calculated based on formula 8 l .

[0189] After calculating the load harmonic impedance, it is necessary to construct the electromagnetic transient model corresponding to the load and the electromagnetic transient model of the transformer.

[0190] For example, Figure 10 This is a schematic diagram of the structure of the load electromagnetic transient model provided in the embodiment of the present application. Figure 11 This is a schematic diagram of the structure of the transformer electromagnetic transient model provided in the embodiment of the present application. Figure 10 As shown, according to the actual load impedance characteristics, the load can be equivalent to the inductive impedance jX s Series resistance R f Then with the inductive impedance jX p In parallel, the power consumption is P+jQ. Figure 11 As shown, the transformer can be equivalent to a resistor R p Parallel inductive impedance jhX T Then with the resistor R s Series connection.

[0191] In combination with the above embodiments, the present application also provides a method for generating voltage sag propagation characteristics based on a digital twin model. Figure 12 Schematic diagram of the process of generating the distribution network voltage sag propagation characteristics provided by this application Figure 2 ,like Figure 12 As shown, the method includes

[0192] A1. Estimate the harmonic impedance of the unknown load.

[0193] A2. Build a digital twin model with the same external impedance characteristics as the world's distribution network.

[0194] A3. Build a graph convolutional neural network model.

[0195] A4. Generate training samples, verification samples, and test samples based on the digital twin model.

[0196] A5. Train, verify, and test the graph convolutional neural network model based on training samples, verification samples, and test samples to obtain a voltage sag location prediction model.

[0197] A6. Based on the voltage sag location prediction model, the real-time collected distribution network electrical parameters are analyzed to obtain the predicted fault location.

[0198] A7. Input the predicted fault location into the digital twin model to obtain the propagation characteristic vector of the voltage sag.

[0199] Based on the above embodiments, an embodiment of the present application provides a digital twin model of a distribution network with 10 machines and 39 nodes. Figure 13 A schematic diagram of the structure of a digital twin model provided in an embodiment of the present application is shown in FIG. Figure 13 As shown in the figure, 39 nodes represent 39 busbars and electrical connection points of the distribution network, and 10 machines refer to 10 synchronous generators. 10 of the 39 nodes are traversed, and 500 samples are generated for each node based on the impedance size and fault type. The final number of batch simulation samples is 5000, and a 10-classification problem is constructed based on the positioning of the 10 nodes.

[0200] The initial graph neural network model was iteratively trained based on 5,000 simulation samples, and the changes in model loss value and accuracy during the iterative training process were recorded. Figure 14 A schematic diagram of the iterative process of an initial graph neural network model provided in an embodiment of the present application is shown as follows: Figure 14 As shown in the figure, the model was iterated 1000 times. The left vertical axis represents the loss value (Loss) of the loss function (LossFunction), and the right vertical axis represents the accuracy (Accuracy). Iteration ends when the model accuracy reaches 91% and the loss function gradually decreases as the accuracy increases.

[0201] After the initial graph neural network model training is completed and the voltage sag location prediction model is obtained, the voltage sag location prediction model is used to calculate the predicted fault location corresponding to the real-time collected distribution network electrical parameters, and the predicted fault location is fed back to the Figure 13 The voltage sag propagation characteristics are obtained from the digital twin model shown in the figure. Figure 15 A distribution diagram of voltage sag propagation characteristics provided in an embodiment of the present application is shown in FIG. Figure 15 As shown in Figure 1, by analyzing the voltage sag propagation characteristics, the voltage sag conditions of 39 buses corresponding to 39 nodes are obtained.

[0202] Figure 16This is a schematic diagram of the structure of the device for generating the distribution network voltage sag propagation characteristics provided by this application, as shown in FIG. Figure 16 As shown, the device for generating the distribution network voltage sag propagation characteristics provided by this embodiment includes:

[0203] The acquisition module 1601 is configured to collect the electrical parameter values of each electrical connection point of the distribution network and the topological structure data of the distribution network in response to a voltage sag event, wherein the topological structure data is used to indicate the connection relationship between each electrical connection point.

[0204] The first processing module 1602 is configured to generate a data feature matrix corresponding to a voltage sag event based on electrical parameter values of a plurality of electrical connection points and topology data.

[0205] The second processing module 1603 is configured to input the data feature matrix into a pre-trained voltage sag location prediction model to obtain a predicted fault location corresponding to the node feature matrix.

[0206] The third processing module 1604 is used to input the predicted fault location into the digital twin model corresponding to the distribution network to obtain the propagation characteristics corresponding to the voltage sag event.

[0207] The digital twin model is a virtual representation of the distribution network, used to reflect its operating status. The voltage sag location prediction model is a graph neural network model trained based on historical voltage sag data from the distribution network.

[0208] In a possible implementation, the third processing module 1604 is further configured to:

[0209] The preset fault type, preset fault impedance, preset fault duration, and predicted fault location corresponding to the voltage sag event are input into the digital twin model, and the voltage sag ratio of each electrical connection point in the distribution network in the digital twin model is calculated.

[0210] The characteristic vector matrix is constructed based on the voltage sag ratio of each electrical connection point to obtain the propagation characteristics corresponding to the voltage sag event.

[0211] The preset fault type is any of a single-phase ground fault, a phase-to-phase fault, and a three-phase ground fault. The preset fault impedance indicates the severity of the voltage sag event. The preset fault duration is the duration of the voltage sag event.

[0212] In a possible implementation, the obtaining module 1601 is further configured to:

[0213] Obtain the electromagnetic transient model corresponding to each component in the distribution network, as well as the topological structure data of the distribution network.

[0214] The topology data and electromagnetic transient model are integrated to obtain the digital twin model corresponding to the distribution network.

[0215] Among them, the components of the distribution network include: distributed power sources, energy storage systems, transmission lines, transformers, and loads.

[0216] In a possible implementation, the obtaining module 1601 is further configured to:

[0217] Based on the preset model category of each component in the distribution network, an electromagnetic transient model corresponding to each component is constructed, wherein the preset model category is an equivalent circuit model or a dynamic model.

[0218] Based on the connection relationship between multiple electrical connection points in the distribution network, the topological structure data of the distribution network is constructed.

[0219] The electrical connection point refers to the connection location between the components.

[0220] In a possible implementation, the obtaining module 1601 is further configured to:

[0221] Collect the device information corresponding to each component in the distribution network. The device information includes the electrical parameter type of the component and the rated parameter value corresponding to each electrical parameter type.

[0222] The model parameters corresponding to each component are determined based on the device information of each component.

[0223] According to the preset model category and model parameters corresponding to each component, an electromagnetic transient model corresponding to each component is constructed.

[0224] In a possible implementation, the obtaining module 1601 is further configured to:

[0225] Generate a preset amount of voltage sag fault data based on the digital twin model corresponding to the distribution network.

[0226] The initial graph neural network model is trained based on the voltage sag data to obtain the model to be determined.

[0227] The loss value and accuracy of the model to be determined are calculated. When the loss value of the model to be determined is lower than the preset loss value and the accuracy is higher than the preset accuracy, the model to be determined is determined as the voltage sag positioning prediction model.

[0228] The device for generating the distribution network voltage sag propagation characteristics provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effects are similar and will not be described in detail in this embodiment.

[0229] Figure 17 This is a schematic diagram of the structure of the electronic device provided in this application. Figure 17 As shown, the electronic device provided by this embodiment includes: at least one processor 1701 and a memory 1702. Optionally, the device also includes a communication component 1703. The processor 1701, the memory 1702 and the communication component 1703 are connected via a bus 1704.

[0230] In a specific implementation process, at least one processor 1701 executes computer-executable instructions stored in the memory 1702 , so that the at least one processor 1701 executes the above-mentioned method for generating the propagation characteristics of the voltage sag in the distribution network.

[0231] The specific implementation process of the processor 1701 can be found in the above method embodiment. Its implementation principle and technical effects are similar and will not be repeated here in this embodiment.

[0232] In the above embodiments, it should be understood that the processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASICs), etc. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in the present invention may be directly executed by a hardware processor or by a combination of hardware and software modules within the processor.

[0233] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage.

[0234] A bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. Buses can be categorized as address buses, data buses, and control buses. For ease of illustration, the buses in the drawings of this application are not limited to just one bus or just one type of bus.

[0235] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the above-mentioned method for generating distribution network voltage sag propagation characteristics.

[0236] The present application also provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the above-mentioned method for generating the distribution network voltage sag propagation characteristics is implemented.

[0237] The readable storage medium may be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium may be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0238] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Alternatively, the readable storage medium may be an integral part of the processor. The processor and the readable storage medium may reside in an application-specific integrated circuit (ASIC). Alternatively, the processor and the readable storage medium may reside in a device as discrete components.

[0239] The division of units is merely a logical functional division; actual implementations may employ alternative divisions, such as combining or integrating multiple units or components into another system, or omitting or disabling certain features. Furthermore, any coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between devices or units, whether electrical, mechanical, or otherwise, through some interface.

[0240] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0241] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0242] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of the present invention. The aforementioned storage medium includes various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0243] Those skilled in the art will appreciate that all or part of the steps in the above-described method embodiments can be implemented using hardware associated with program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0244] Finally, it should be noted that those skilled in the art will readily identify other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art not disclosed herein. The present invention is not limited to the precise structure described above and illustrated in the accompanying drawings, and various modifications and variations may be made without departing from the scope thereof. The scope of the present invention is limited solely by the appended claims.

Claims

1. A method for generating propagation characteristics of voltage sag in a distribution network, characterized in that: The method comprises: In response to a voltage sag event, collecting electrical parameter values of each electrical connection point of the distribution network and topological structure data of the distribution network; wherein the topological structure data is used to indicate the connection relationship between each electrical connection point; generating a data feature matrix corresponding to the voltage sag event based on electrical parameter values of a plurality of electrical connection points and the topology structure data; Inputting the data feature matrix into a pre-trained voltage sag location prediction model to obtain a predicted fault location corresponding to the data feature matrix; Inputting the predicted fault location into a digital twin model corresponding to the distribution network to obtain propagation characteristics corresponding to the voltage sag event; Among them, the digital twin model is a virtual mapping of the distribution network, which is used to reflect the operating status of the distribution network; the voltage sag positioning prediction model refers to a graph neural network model obtained by model training based on the historical voltage sag data of the distribution network.

2. The method according to claim 1, characterized in that Inputting the predicted fault location into the digital twin model corresponding to the distribution network to obtain the propagation characteristics corresponding to the voltage sag event includes: Inputting a preset fault type, a preset fault impedance, a preset fault duration, and the predicted fault location corresponding to the voltage sag event into the digital twin model, and calculating the voltage sag ratio of each electrical connection point of the distribution network in the digital twin model; Constructing a characteristic vector matrix based on the voltage sag ratio of each electrical connection point to obtain the propagation characteristics corresponding to the voltage sag event; The preset fault type refers to any one of a single-phase grounding fault, a phase-to-phase fault, and a three-phase grounding fault; the preset fault impedance is used to indicate the severity of a voltage sag event; and the preset fault duration refers to the duration of a voltage sag event.

3. The method according to claim 1, characterized in that Before collecting the current electrical parameter values of each electrical connection point of the distribution network and the topology data of the distribution network in response to the voltage sag event, the method further includes: Obtaining an electromagnetic transient model corresponding to each component in the distribution network and topological structure data of the distribution network; Integrating the topology data and the electromagnetic transient model to obtain a digital twin model corresponding to the distribution network; The components of the distribution network include: distributed power sources, energy storage systems, transmission lines, transformers, and loads.

4. The method according to claim 3, characterized in that The obtaining of the electromagnetic transient model corresponding to each component in the distribution network and the topological structure data of the distribution network includes: Based on a preset model category of each component in the distribution network, construct an electromagnetic transient model corresponding to each component; wherein the preset model category is an equivalent circuit model or a dynamic model; Constructing topological structure data of the distribution network based on the connection relationship between multiple electrical connection points in the distribution network; The electrical connection point refers to the connection position between the constituent elements.

5. The method according to claim 4, characterized in that The step of constructing an electromagnetic transient model corresponding to each component in the distribution network based on a preset model category of each component includes: Collecting device information corresponding to each component in the distribution network; wherein the device information includes: the electrical parameter type of the component and the rated parameter value corresponding to each electrical parameter type; Determine the model parameters corresponding to each component according to the device information of each component; According to the preset model category and model parameters corresponding to each component, an electromagnetic transient model corresponding to each component is constructed.

6. The method according to claim 5, characterized in that Before collecting the current electrical parameter values of each electrical connection point of the distribution network and the topology data of the distribution network in response to the voltage sag event, the method further includes: Generate a preset amount of voltage sag fault data based on the digital twin model corresponding to the distribution network; Performing model training on the initial graph neural network model based on the voltage sag data to obtain a model to be determined; The loss value and accuracy of the model to be determined are calculated, and when the loss value of the model to be determined is lower than a preset loss value and the accuracy is higher than a preset accuracy, the model to be determined is determined as a voltage sag positioning prediction model.

7. A device for generating propagation characteristics of voltage sag in a distribution network, characterized in that: include: an acquisition module, configured to collect, in response to a voltage sag event, electrical parameter values of each electrical connection point of the distribution network and topological structure data of the distribution network; wherein the topological structure data is used to indicate a connection relationship between each electrical connection point; A first processing module is configured to generate a data feature matrix corresponding to the voltage sag event based on electrical parameter values of a plurality of electrical connection points and the topology data; A second processing module is configured to input the data feature matrix into a pre-trained voltage sag location prediction model to obtain a predicted fault location corresponding to the node feature matrix; a third processing module, configured to input the predicted fault location into a digital twin model corresponding to the distribution network to obtain a propagation feature corresponding to the voltage sag event; Among them, the digital twin model is a virtual mapping of the distribution network, which is used to reflect the operating status of the distribution network; the voltage sag positioning prediction model refers to a graph neural network model obtained by model training based on the historical voltage sag data of the distribution network.

8. An electronic device, characterized in that: include: Memory, processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the processor performs the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 6 when executed by a processor.

10. A computer program product, characterized in that The invention comprises a computer program, which implements the method according to any one of claims 1 to 6 when the computer program is executed by a processor.

Images