A fault prediction method, device, equipment, computer-readable storage medium and computer program product

Abstract

The present application provides a fault prediction method, device, equipment, computer-readable storage medium and computer program product; the method includes: obtaining real-time line data of each first transmission line in the overhead line system; constructing a volt-ampere characteristic curve diagram for each first transmission line; using the fault prediction sub-model corresponding to the first transmission line, predicting based on the volt-ampere characteristic curve diagram, and obtaining a first fault prediction result of the first transmission line; obtaining a first adjacency matrix corresponding to the overhead line system, using the fault prediction total model, predicting based on the first adjacency matrix, the first fault prediction result corresponding to the first transmission line and the real-time line data, and obtaining a second fault prediction result of the overhead line system. Through the present application, it is possible to train fault prediction sub-models suitable for different line lengths, and to perform fault prediction and location in conjunction with the fault prediction total model, thereby improving the accuracy and pertinence of fault prediction and location.

Description

Technical Field

[0001] This application relates to fault prediction technology, and in particular to a fault prediction method, device, equipment, computer-readable storage medium, and computer program product. Background Art

[0002] In related technologies, the fault prediction of transmission lines relies on pre-set fault characteristics. Generally, various types of fault characteristics that may occur in transmission lines are input into an expert system in advance, and then a threshold value is set for the transmission lines. Once the characteristic value of the transmission line reaches the pre-set threshold value and a certain logical relationship holds, it is determined that the transmission line has a fault, and a fault diagnosis result is given according to the fault code pre-set by the system. However, due to the large number of overhead lines with different lengths, and different fault types and causes on each line, the pre-set fault characteristics cannot handle all fault situations, so fault prediction cannot be widely carried out, and the accuracy of fault prediction is relatively low. Summary of the Invention

[0003] Embodiments of this application provide a fault prediction method, device, computer-readable storage medium, and computer program product, which can train fault prediction sub-models applicable to different line lengths, perform fault prediction and positioning with a fault prediction total model, and improve the accuracy and pertinence of fault prediction and positioning.

[0004] The technical solution of the embodiments of this application is implemented as follows:

[0005] Embodiments of this application provide a fault prediction method, the method includes:

[0006] Obtain real-time line data of each first transmission line in the overhead line system;

[0007] For each first transmission line, based on the real-time line data of the first transmission line, construct a volt-ampere characteristic curve graph;

[0008] Use the fault prediction sub-model corresponding to the first transmission line to perform prediction based on the volt-ampere characteristic curve graph, and obtain a first fault prediction result of the first transmission line;

[0009] Obtain a first adjacency matrix corresponding to the overhead line system, and use the fault prediction total model to perform prediction based on the first adjacency matrix, the first fault prediction result corresponding to the first transmission line, and the real-time line data, and obtain a second fault prediction result of the overhead line system.

[0010] Embodiments of this application provide a fault prediction device, including:

[0011] A data acquisition module, configured to obtain real-time line data of each first transmission line in the overhead line system;

[0012] A building block for constructing a volt-ampere characteristic curve graph for each first power transmission line based on the real-time line data of the first power transmission line;

[0013] A first fault prediction module for predicting based on the volt-ampere characteristic curve graph by using the fault prediction sub-model corresponding to the first power transmission line to obtain a first fault prediction result of the first power transmission line;

[0014] A second fault prediction module for obtaining a first adjacency matrix corresponding to the overhead line system and predicting based on the first adjacency matrix, the first fault prediction result corresponding to the first power transmission line and the real-time line data by using the total fault prediction model to obtain a second fault prediction result of the overhead line system.

[0015] An embodiment of the present application provides an electronic device, including:

[0016] A memory for storing computer-executable instructions or computer programs;

[0017] A processor for implementing the fault prediction method provided by the embodiment of the present application when executing the computer-executable instructions or computer programs stored in the memory.

[0018] An embodiment of the present application provides a computer-readable storage medium storing a computer program or computer-executable instructions for implementing the fault prediction method provided by the embodiment of the present application when being executed by a processor.

[0019] An embodiment of the present application provides a computer program product including a computer program or computer-executable instructions, where when the computer program or computer-executable instructions are executed by a processor, the fault prediction method provided by the embodiment of the present application is implemented.

[0020] The embodiment of the present application has the following beneficial effects:

[0021] An embodiment of the present application provides a fault prediction method, which obtains real-time line data of each first transmission line in an overhead line system; for each first transmission line, based on the real-time line data of the first transmission line, a volt-ampere characteristic curve graph is constructed; using the fault prediction sub-model corresponding to the first transmission line, based on the volt-ampere characteristic curve graph for prediction, a first fault prediction result of the first transmission line is obtained; wherein, the fault prediction sub-model corresponding to the first transmission line is determined based on the length of the first transmission line, so as to improve the pertinence of fault prediction for transmission lines of different lengths and the accuracy of the first fault prediction result. Then, a first adjacency matrix corresponding to the overhead line system is obtained, and using the total fault prediction model, based on the first adjacency matrix, the first fault prediction result corresponding to the first transmission line and the real-time line data for prediction, a second fault prediction result of the overhead line system is obtained. In this way, after obtaining the first prediction result of each first transmission line in the overhead line system, the total fault prediction model is used to perform fault prediction on the entire overhead line system. By jointly using the fault prediction sub-model and the total fault prediction model for fault prediction and positioning, the accuracy of fault prediction is improved. Description of the Drawings

[0022] Figure 1 is a schematic structural diagram of a fault prediction system architecture provided by an embodiment of the present application;

[0023] Figure 2 is a schematic structural diagram of a fault prediction device provided by an embodiment of the present application;

[0024] Figure 3 is a first process schematic diagram of a fault prediction method provided by an embodiment of the present application;

[0025] Figure 4A is a second process schematic diagram of a first fault prediction sub-model training method provided by an embodiment of the present application;

[0026] Figure 4B is a third process schematic diagram of a second fault prediction sub-model training method provided by an embodiment of the present application;

[0027] Figure 5 is a fourth process schematic diagram of a method for determining a fault prediction sub-model corresponding to a first transmission line provided by an embodiment of the present application;

[0028] Figure 6A is a fifth process schematic diagram of a first adjacency matrix construction method provided by an embodiment of the present application;

[0029] Figure 6B is a sixth process schematic diagram of a second adjacency matrix construction method provided by an embodiment of the present application;

[0030] Figure 6C It is the seventh process schematic diagram of the third adjacent matrix construction method provided by an embodiment of the present application;

[0031] Figure 7 It is the eighth process schematic diagram of the second fault prediction result obtaining method provided by an embodiment of the present application;

[0032] Figure 8 It is the ninth process schematic diagram of the fault prediction total model training method provided by an embodiment of the present application;

[0033] Figure 9 It is the deployment diagram of the power grid line system provided by an embodiment of the present application;

[0034] Figure 10 It is the deployment diagram of the simplified power grid line system provided by an embodiment of the present application.

[0035] It should be noted that the above-mentioned "first" and "second" are only used to distinguish different solutions, and do not represent the distinction of the advantages or disadvantages of the solutions or the priority in the implementation process. Detailed implementation manners

[0036] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations on the present application. All other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present application.

[0037] In the following description, "some embodiments" are involved, which describe a subset of all possible embodiments. However, it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict.

[0038] In the following description, the terms "first / second / third" only distinguish similar objects and do not represent a specific order for the objects. It can be understood that "first / second / third" can be interchanged with a specific order or sequence when allowed, so that the embodiments of the present application described here can be implemented in an order other than that illustrated or described here.

[0039] In the embodiments of the present application, the term "module" or "unit" refers to a computer program with a predetermined function or a part of a computer program, which works together with other related parts to achieve a predetermined goal, and can be fully or partially implemented by using software, hardware (such as a processing circuit or a memory), or a combination thereof. Similarly, one processor (or multiple processors or memories) can be used to implement one or more modules or units. In addition, each module or unit can be a part of the overall module or unit that includes the function of the module or unit.

[0040] Unless otherwise defined, all technical and scientific terms used in the embodiments of this application have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used in the embodiments of this application are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0041] In the practical application of data collection and processing in the embodiments of this application, it should strictly comply with the requirements of relevant laws and regulations, obtain the informed consent or separate consent of the personal information subject, and carry out subsequent data use and processing within the scope authorized by laws and regulations and the personal information subject.

[0042] Before further elaborating on the embodiments of this application, the nouns and terms involved in the embodiments of this application are explained, and the nouns and terms involved in the embodiments of this application are applicable to the following explanations.

[0043] 1) Overhead line: mainly refers to overhead open wire, which is a bare conductor telecommunication line supported by electric poles and erected above the ground.

[0044] 2) Insulator: a device installed between conductors at different potentials or between a conductor and a grounding member, which can withstand voltage and mechanical stress. Insulators are installed on the poles of transmission lines to support and fix conductors and provide electrical insulation at the same time.

[0045] 3) Transmission corridor: a strip-shaped area below the line that extends a specified width on both sides along the outer conductor of the overhead line.

[0046] 4) Secondary side three-phase voltage: the three-phase voltage generated on the secondary side, i.e., the output side, of electrical equipment such as power transformers or generators. Three-phase voltage refers to three AC voltages with a phase difference of 120° from each other, which act together in a three-phase system to provide electrical energy for various three-phase loads.

[0047] 5) Oscillogram data: the waveform data of electrical quantities such as voltage and current of a power system or electrical equipment recorded by an oscillograph device within a specific time. It usually includes the time-domain waveform data of voltage and current, as well as possible spectrum analysis data, harmonic analysis data, etc. These data can reflect the changes in electrical quantities of the power system within a specific time, including characteristics such as the amplitude, phase, and frequency of the waveform.

[0048] 6) Zero-sequence voltage: in a three-phase three-wire circuit, when the three-phase voltages are unbalanced, the vector sum of the three phase voltages is not zero, and this non-zero vector sum is the zero-sequence voltage. It is an abnormal voltage state in the power system and is usually related to conditions such as three-phase imbalance and grounding faults.

[0049] 7) Three-phase three-wire system: A circuit system that uses three AC power sources with a phase difference of 120° from each other and conducts power transmission and distribution through three wires, namely phase wires. In a three-phase three-wire system, the ends of the three power sources, or the so-called neutral points, are usually connected together, but no neutral wire or zero wire is drawn out. Therefore, only three phase wires are used to transmit electrical energy.

[0050] 8) Zero-sequence current: In a three-phase AC power system, when the three-phase currents are unbalanced due to certain reasons such as system imbalance, grounding fault, or equipment failure, the common current generated through all wires or equipment relative to the ground.

[0051] 9) Volt-ampere characteristic curve graph: The I-U graph or IV curve, with the vertical axis representing the current I and the horizontal axis representing the voltage U. By drawing the I-U graph, it shows the current response of electrical components at different voltages.

[0052] 10) Fully connected layer: A basic layer in a neural network. In this layer, each neuron in the network is connected to all neurons in the previous layer. This connection method enables this layer to capture global features or patterns from the input data and pass these features to subsequent layers for further processing.

[0053] 11) Activation function: A key component in a neural network that determines whether a neuron in the network should be activated, including linear and non-linear activation functions. The linear activation function is the simplest activation function, and its output is a linear transformation of the input. For example, a common linear activation function is . The non-linear activation function can introduce non-linearity, enabling the neural network to learn and represent complex functions and decision boundaries. Common non-linear activation functions include Sigmoid, ReLU, Tanh, etc.

[0054] 12) Degree matrix: A diagonal matrix, where the elements on the diagonal represent the degrees of the vertices in the graph. In an undirected graph, the degree of a vertex is the number of edges connected to that vertex; in a directed graph, the degree of a vertex is divided into out-degree and in-degree. The out-degree is the number of directed edges starting from that vertex, and the in-degree is the number of directed edges entering that vertex.

[0055] 13) Graph Convolutional Network (GCN): A deep learning model specifically designed to process graph-structured data, which extracts useful feature information from graph data and can then be used for tasks such as node classification, graph classification, edge prediction, and obtaining the embedding representation of the graph.

[0056] 14) Classifier: A mathematical model designed for specific problems. Through some mathematical calculations, it can classify the features of target samples to be identified and classified and assign corresponding labels. In data mining, classifiers are the general term for sample classification methods, which include various algorithms such as decision trees, logistic regression, naive Bayes, neural networks, etc.

[0057] 15) Disconnection node: Refers to a specific point pre-marked in sample data, which represents the possible position of disconnection in the circuit. Disconnection is a state where a part of the circuit is disconnected, resulting in the interruption of the circuit and the inability of current to flow.

[0058] 16) Transfer learning: An important method in machine learning that allows a model to transfer the knowledge learned in one task or domain to another related task or domain to accelerate and improve the ability to learn and solve problems in the new domain.

[0059] 17) Current trend: The variation law of current in a circuit with time or other variables. This trend can be affected by various factors, such as the properties of circuit elements such as power supply voltage, resistance, capacitance, and inductance, as well as changes in external conditions such as temperature and pressure. The positive and negative of the current trend are usually used in electricity to represent the relationship between the actual direction of the current and the reference direction. When the current trend is positive, it means that the actual direction of the current is the same as the reference direction, that is, flowing from the positive pole to the negative pole. When the current trend is negative, it means that the actual direction of the current is opposite to the reference direction, that is, the current actually flows from the negative pole to the positive pole.

[0060] 18) Voltage trend: The direction and rate of change of voltage in a circuit with time or spatial position. When the voltage trend is positive, it means that the electric potential at this point or this section of the circuit is higher than the reference point. When the voltage trend is negative, it means that the electric potential at this point or this section of the circuit is lower than the reference point.

[0061] In related technologies, the fault prediction of transmission lines relies on pre-set fault features. However, due to the large number of overhead lines with different lengths and different fault types and causes on each line, the pre-set fault features cannot handle all fault situations. Therefore, fault prediction cannot be widely carried out, and the accuracy of fault prediction is relatively low.

[0062] Embodiments of the present application provide a fault prediction method, apparatus, device, computer-readable storage medium, and computer program product, which can train fault prediction sub-models applicable to different line lengths, and jointly perform fault prediction and location with a fault prediction total model, improving the accuracy and pertinence of fault prediction and location. The following describes an exemplary application of the fault prediction device provided by the embodiments of the present application. The device provided by the embodiments of the present application can be implemented as a terminal such as a desktop computer or a laptop computer, or can also be implemented as a server. Next, the exemplary application when the device is implemented as a server will be described.

[0063] See Figure 1 , Figure 1 is a schematic diagram of the architecture of the fault prediction system 100 provided by the embodiments of the present application. The fault prediction system 100 includes a plurality of sensors (the first sensor 400-1 and the second sensor 400-2 are exemplarily shown in Figure 1 ), and a server 200. To support a fault prediction application, the first sensor 400-1 and the second sensor 400-2 are connected to the server 200 through a network 300. The network 300 can be a wide area network, a local area network, or a combination of the two.

[0064] The first sensor 400-1 and the second sensor 400-2 can be arranged on the transmission line in the overhead line system to collect real-time line data of the transmission line and send the collected real-time line data to the server 200. A plurality of fault prediction sub-models can be deployed in the server 200. A fault prediction sub-model corresponds to one sensor, which can be understood as a fault prediction sub-model corresponding to one transmission line. After the server 200 obtains the real-time line data collected by the sensor, it constructs a volt-ampere characteristic curve graph according to the real-time line data, and uses the fault prediction sub-model corresponding to the transmission line where the sensor is located to perform fault prediction based on the volt-ampere characteristic curve graph to obtain the fault prediction result corresponding to the transmission line. A fault prediction total model is also deployed in the server 200. The server 200 uses the fault prediction total model to perform fault prediction based on the pre-constructed adjacency matrix, the fault prediction results corresponding to each transmission line, and the real-time line data to obtain the fault prediction result corresponding to the overhead line system. Among them, the fault prediction result can include the fault type and the fault location. The fault type indicates whether a fault occurs, and after a fault occurs, the specific type of the fault, such as phase-to-phase short circuit, circuit open circuit, line grounding fault, etc.; the fault location refers to the location where the fault occurs, such as a certain tower, or a line segment between any two towers.

[0065] In some embodiments, in addition to multiple sensors and a server, the fault prediction system may further include line prediction terminals. The line prediction terminals are arranged on the power transmission lines in the overhead line system. The line prediction terminals correspond to the sensors one by one, that is, a line prediction terminal configured with a fault prediction sub-model adapted to the line length will be correspondingly configured at the position where the sensor is located. It should be noted that the corresponding sensors and line prediction terminals are deployed on the same tower pole. And a fault prediction sub-model is deployed in the line prediction terminal. After the sensor collects the real-time line data of the power transmission line, it sends the real-time line data to the line prediction terminal. The line prediction terminal constructs a volt-ampere characteristic curve based on the real-time line data, and thus performs fault prediction based on the fault prediction sub-model to obtain the fault prediction result of the power transmission line. The line prediction terminal sends the fault prediction result of the power transmission line to the server. The server uses the total fault prediction model to perform fault prediction based on the pre-constructed adjacency matrix, the fault prediction results of each power transmission line, and the real-time line data to obtain the final fault type and fault location.

[0066] Here, the pre-constructed adjacency matrix is the sum of the adjacency matrices corresponding to each line segment. The specific construction method of the adjacency matrix is described in the following steps 401 to 403, steps 4021 to 4023, and steps 4031 to 4032, and will not be elaborated here.

[0067] In some embodiments, the server 200 may be an independent physical server, or a server cluster or distributed system composed of multiple physical servers. It may also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery network (Content Delivery Network, CDN), and big data and artificial intelligence platforms. The terminal and the server can be connected by wired or wireless communication methods, which are not limited in the embodiments of the present application.

[0068] See Figure 2 , Figure 2 is a schematic structural diagram of the server 200 provided by the embodiments of the present application. Figure 2 The server 200 shown includes: at least one processor 210, a memory 250, at least one network interface 220, and a user interface 230. Each component in the server 200 is coupled together through a bus system 240. It can be understood that the bus system 240 is used to realize the connection and communication between these components. In addition to the data bus, the bus system 240 also includes a power bus, a control bus, and a status signal bus. However, for the sake of clear illustration, in Figure 2 all kinds of buses are labeled as the bus system 240.

[0069] The processor 210 may be an integrated circuit chip with signal processing capabilities, such as a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or any conventional processor, etc.

[0070] The user interface 230 includes one or more output devices 231 that enable the presentation of media content, including one or more speakers and / or one or more visual display screens. The user interface 230 also includes one or more input devices 232, including user interface components that facilitate user input, such as a keyboard, a mouse, a microphone, a touch screen display, a camera, other input buttons, and controls.

[0071] The memory 250 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid-state memory, hard disk drives, optical disc drives, etc. The memory 250 optionally includes one or more storage devices that are physically remote from the processor 210.

[0072] The memory 250 includes volatile memory or non-volatile memory, and may also include both volatile and non-volatile memory. The non-volatile memory may be a read-only memory (ROM), and the volatile memory may be a random access memory (RAM). The memory 250 described in the embodiments of the present application is intended to include any suitable type of memory.

[0073] In some embodiments, the memory 250 is capable of storing data to support various operations. Examples of such data include programs, modules, and data structures, or subsets or supersets thereof, which are illustrated below.

[0074] The operating system 251 includes system programs for processing various basic system services and performing hardware-related tasks, such as a framework layer, a core library layer, a driver layer, etc., for implementing various basic services and processing hardware-based tasks;

[0075] The network communication module 252 is used to reach other electronic devices via one or more (wired or wireless) network interfaces 220. Exemplary network interfaces 220 include: Bluetooth, wireless compatibility certification (WiFi), and universal serial bus (USB), etc.;

[0076] A presentation module 253 for enabling the presentation of information (e.g., a user interface for operating peripheral devices and displaying content and information) via one or more output devices 231 associated with the user interface 230 (e.g., a display screen, a speaker, etc.);

[0077] An input processing module 254 for detecting and translating one or more user inputs or interactions from one of one or more input devices 232.

[0078] In some embodiments, the device provided by the embodiments of the present application may be implemented in software. Figure 2 A fault prediction device 255 stored in the memory 250 is shown, which may be software in the form of a program and a plug-in, etc., including the following software modules: a data acquisition module 2551, a construction module 2552, a first fault prediction module 2553, and a second fault prediction module 2554. These modules are logical, and thus can be arbitrarily combined or further split according to the implemented functions. The functions of each module will be described below.

[0079] In other embodiments, the device provided by the embodiments of the present application may be implemented in hardware. As an example, the device provided by the embodiments of the present application may be a processor in the form of a hardware decoding processor, which is programmed to execute the fault prediction method provided by the embodiments of the present application. For example, a processor in the form of a hardware decoding processor may employ one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), or other electronic components.

[0080] In some embodiments, a terminal or a server may implement the fault prediction method provided in the embodiments of the present application by running various computer-executable instructions or computer programs. For example, the computer-executable instructions may be microprogram-level commands, machine instructions, or software instructions. The computer program may be a native program or a software module in an operating system; it may be a local application (APPlication, APP), that is, a program that needs to be installed in the operating system to run, such as a fault prediction APP. In short, the above computer-executable instructions may be instructions in any form, and the above computer programs may be application programs, modules, or plug-ins in any form.

[0081] The exemplary applications and implementations on the server side provided in the embodiments of the present application will be combined to illustrate the fault prediction method provided in the embodiments of the present application.

[0082] Next, the fault prediction method provided in the embodiments of the present application will be described. It can be applied to scenarios such as transmission line fault prediction. The fault prediction method provided in the embodiments of the present application can accurately predict faults on transmission lines of different lengths. As described above, the electronic device implementing the fault prediction method in the embodiments of the present application may be a terminal, a server, or a combination of both. Therefore, the execution entity of each step will not be repeated hereinafter.

[0083] See Figure 3 , Figure 3 is a schematic flowchart of the image processing method provided in the embodiments of the present application, and will be described in combination with Figure 3 the steps shown. Figure 3 The entity of the step is the server.

[0084] In step 101, real-time line data of each first transmission line in the overhead line system is acquired.

[0085] In some embodiments, the real-time data on each first transmission line is collected by sensors, and the collected data includes the recorded data of the secondary side three-phase voltage and current signals at both ends of each first transmission line, and the recorded data of the zero-sequence voltage and zero-sequence current signals.

[0086] In step 102, for each first transmission line, a volt-ampere characteristic curve graph is constructed based on the real-time line data of the first transmission line.

[0087] It should be noted that each volt-ampere characteristic curve graph uniquely corresponds to a first transmission line.

[0088] In some embodiments, for each first power transmission line, a volt-ampere characteristic curve graph needs to be constructed. In the volt-ampere characteristic curve graph, the abscissa represents voltage U with the unit of volt (V), and the ordinate represents current I with the unit of ampere (A). To construct the volt-ampere characteristic curve graph, the voltage or current in the first power transmission line needs to be changed, and the corresponding voltage and current values are recorded. Usually, multiple measurements are required to obtain sufficient data points to plot the volt-ampere characteristic curve graph.

[0089] In step 103, using the fault prediction sub-model corresponding to the first power transmission line, prediction is performed based on the volt-ampere characteristic curve graph to obtain the first fault prediction result of the first power transmission line.

[0090] In some embodiments, each volt-ampere characteristic curve graph is input into the fault prediction sub-model closest to the distance sensor in the corresponding first power transmission line. Fault prediction is performed based on the fault prediction sub-model to obtain the fault prediction result, which includes the fault type and the fault location.

[0091] Here, the fault prediction sub-model corresponds one-to-one with the sensor. It can be understood that a fault prediction sub-model adapted to the length of the power transmission line is configured at the location where the sensor is located. Therefore, there is no situation where there are multiple fault prediction sub-models closest to the distance sensor. The fault type includes whether a fault occurs, and the specific type of the fault after it occurs, such as phase-to-phase short circuit, circuit open circuit, line grounding fault, etc. The fault location refers to the location where the fault occurs, such as a certain tower pole, or the power transmission line between any two tower poles.

[0092] As an example of step 103, the volt-ampere characteristic curve graph corresponding to the first power transmission line is input into the fault prediction sub-model corresponding to the first power transmission line for fault prediction. The first fault prediction result of the first power transmission line obtained is: a fault occurs, specifically an open circuit fault, and the location where the fault occurs is tower pole A.

[0093] Through step 103, for first power transmission lines of different lengths, the fault prediction sub-model suitable for first power transmission lines of different lengths can be used to predict faults, improving the pertinence of the fault prediction by the fault prediction sub-model. The first fault prediction result obtained by the fault prediction sub-model can be used as the basis for the fault prediction by the total fault prediction model, further improving the accuracy of the fault prediction.

[0094] In some embodiments, refer to Figure 4A , before Figure 3 the steps 103 shown, steps 201 to 205 can also be executed to train the fault prediction sub-model, which is specifically described below.

[0095] In step 201, obtain the fault prediction sub-model to be trained and the fault waveform data corresponding to the second transmission lines with multiple different candidate lengths.

[0096] As an example of step 201, the fault waveform data corresponding to the second transmission lines with multiple different candidate lengths can be the fault waveform data of a 500-km overhead line, the fault waveform data of a 200-km overhead line, the fault waveform data of a 100-km overhead line, the fault waveform data of a 50-km overhead line, and the fault waveform data of a 12-km overhead line.

[0097] In step 202, for the fault waveform data corresponding to each of the second transmission lines with the candidate length, preprocess the fault waveform data to obtain the training data corresponding to the second transmission lines with the candidate length.

[0098] In some embodiments, before training the fault prediction sub-models corresponding to the second transmission lines with multiple different candidate lengths, it is necessary to preprocess the obtained fault waveform data corresponding to the second transmission lines with multiple different candidate lengths, including denoising, normalization, etc., to obtain the training data corresponding to each of the second transmission lines with the candidate length.

[0099] In step 203, based on the average length among the multiple different candidate lengths, determine the first length.

[0100] In some embodiments, in order to train the fault prediction sub-models corresponding to different candidate lengths, it is necessary to pre-train the fault prediction sub-model corresponding to the second transmission line with the first length. In order to make the adaptation gap between the pre-trained fault prediction sub-model and the fault prediction sub-models corresponding to other different candidate lengths smaller, it is necessary to select the candidate length close to the average value among the multiple different candidate lengths as the first length. The pre-trained fault prediction sub-model corresponds to the first length. In practical applications, the first length can be the candidate length closest to the average value among the multiple different candidate lengths, or any one of the multiple different candidate lengths, which is not limited here.

[0101] As an embodiment of step 203, referring to the example of step 201, the lengths of the overhead lines are 500 km, 200 km, 100 km, 50 km, and 12 km, and the average length of the overhead lines is 172.4 km. Therefore, the first length is determined to be 200 km.

[0102] In step 204, use the training data corresponding to the second transmission line with the first length to train the fault prediction sub-model to be trained, and obtain the trained fault prediction sub-model corresponding to the second transmission line with the first length.

[0103] In some embodiments, the fault prediction sub-model is trained with the training data corresponding to the fault waveform data of the second power transmission line based on the first length, and the trained fault prediction sub-model corresponding to the second power transmission line with the first length is obtained, where the fault prediction sub-model includes a plurality of convolutional layers and fully connected layers.

[0104] In step 205, the parameters of the trained fault prediction sub-model corresponding to the second power transmission line with the first length are adjusted using the training data corresponding to the second power transmission line with other lengths, and the trained fault prediction sub-model corresponding to the second power transmission line with the other lengths is obtained.

[0105] Wherein, the other length is any candidate length among the plurality of different candidate lengths except the first length.

[0106] In some embodiments, referring to Figure 4B , Figure 4A The steps shown in step 205 can be implemented by the following steps 2051 to 2053, which will be described below in conjunction with Figure 4B for illustration.

[0107] In step 2051, the parameters of the convolutional layers in the trained fault prediction sub-model corresponding to the second power transmission line with the first length are kept unchanged.

[0108] In step 2052, the parameters of the fully connected layers in the trained fault prediction sub-model corresponding to the second power transmission line with the first length are set to a fourth value to obtain an intermediate fault prediction sub-model.

[0109] In some embodiments, the parameters of the convolutional layers in the trained fault prediction sub-model corresponding to the second power transmission line with the first length are retained, and the parameters of the fully connected layers are cleared, that is, the fourth value is 0, to obtain an intermediate fault prediction sub-model.

[0110] In step 2053, the parameters of the fully connected layers in the intermediate fault prediction sub-model are adjusted using the training data corresponding to the second power transmission line with the other lengths, and the trained fault prediction sub-model corresponding to the second power transmission line with the other lengths is obtained.

[0111] Here, adjusting the parameters of the fully connected layers in the intermediate fault prediction sub-model using the training data corresponding to the second power transmission line with other lengths means retraining the fully connected layers in the intermediate fault prediction sub-model with the training data corresponding to the second power transmission line with other lengths to obtain the corresponding multiple fault prediction sub-models.

[0112] In the above steps 2051 to 2053, while keeping the parameters of the convolutional layer in the trained fault prediction sub-model unchanged, using the training data corresponding to the second transmission lines of other lengths to adjust the parameters of the fully connected layer can quickly train various types of models with limited data, which can not only improve the training speed but also facilitate the subsequent targeted layout of different fault prediction sub-models suitable for different transmission line lengths for different lengths of transmission lines, so as to improve the accuracy of fault detection.

[0113] Through steps 201 to 205, it is possible to train fault prediction sub-models suitable for different lengths of transmission lines for different lengths of transmission lines, improving the pertinence and accuracy of the fault prediction sub-models in predicting faults.

[0114] In some embodiments, after obtaining the fault prediction sub-models corresponding to different candidate lengths of transmission lines through the above steps 201 to 205, in the actual application process, the length of the transmission line in the overhead line system may be different from the candidate length. At this time, it is necessary to determine the fault prediction sub-model that matches the actual transmission line from multiple fault prediction sub-models. Refer to Figure 5 In Figure 4A After step 205 shown, steps 301 to 303 can also be executed to determine the fault prediction sub-model corresponding to the first transmission line, which will be specifically described below.

[0115] In step 301, obtain the second length of each of the first transmission lines.

[0116] Here, the second length of the first transmission line is the actual length of the first transmission line measured during actual application. Exemplarily, the second length can be 150 km.

[0117] In step 302, for each of the first transmission lines, determine the first length with the smallest difference from the second length among the multiple candidate lengths.

[0118] In some embodiments, when determining the fault prediction sub-model corresponding to the second length based on the trained fault prediction sub-models corresponding to different candidate lengths, it is necessary to determine the candidate difference length closest to the second length among the multiple candidate lengths, that is, the difference between this candidate length and the second length is the smallest.

[0119] In step 303, determine the trained fault prediction sub-model corresponding to the first length as the fault prediction sub-model corresponding to the first transmission line.

[0120] As an example of steps 301 to 303, for instance, multiple candidate lengths are 200 km, 180 km, 50 km, and 100 km respectively; the second length of a certain first transmission line is 185 km, and the first length with the smallest difference from the second length among the multiple candidate lengths is 180 km. Therefore, the fault prediction sub-model corresponding to the second length of 185 km is the fault prediction sub-model corresponding to the first length of 180 km.

[0121] Through steps 301 to 303, it is possible to determine the fault prediction sub-model that best matches the length of the first transmission line from multiple trained fault prediction sub-models, thereby ensuring the accuracy of the first fault prediction result.

[0122] Continue to refer to Figure 3 , in step 104, obtain the first adjacency matrix corresponding to the overhead line system. Here, the element values in the first adjacency matrix represent the degree of influence of the line segments between different sensors by other line segments. The larger the element value in the first adjacency matrix, the greater the degree of influence of the line segment corresponding to the element between different sensors by other line segments.

[0123] In some embodiments, refer to Figure 6A , before Figure 3 shown in step 104, steps 401 to 403 may also be executed to construct the first adjacency matrix, which is specifically described below.

[0124] In step 401, obtain the electrical diagram of the overhead line system.

[0125] Among them, the electrical diagram includes: at least one line segment included in each of the first transmission lines, the connection relationships between the line segments, and the deployment positions of the sensors, where the sensors are used to collect real-time line data of the first transmission line.

[0126] In step 402, for each line segment, based on the connection relationships between the line segments and the deployment positions of the sensors, construct the second adjacency matrix corresponding to the line segment.

[0127] In some embodiments, it is necessary to abstract the electrical diagram to obtain an abstract electrical diagram. The generation process of the abstract electrical diagram is as follows: regard the intersection of transmission lines as a node, and divide the transmission lines into multiple segments with the position of each node as the boundary. Since the sensors and the fault prediction sub-models are deployed at the same position in the electrical diagram, in the abstract electrical diagram, the sensors and the fault prediction sub-models are abstracted into one entity, which can be regarded as a sensor.

[0128] In some embodiments, after the fault prediction sub-model predicts the faults of a certain transmission line section, it is necessary to perform another prediction in the overall fault prediction model. Therefore, it is necessary to construct the state information of the sensors on all transmission line sections in the abstract electrical diagram, that is, to construct the second adjacency matrix corresponding to the line sections. The elements in the matrix are used to indicate whether there is an available line section between two sensors when a fault occurs in the line section.

[0129] In some embodiments, referring to Figure 6B , Figure 6A The step 402 shown can be implemented by the following steps 4021 to 4023. The following is described in conjunction with Figure 6B for illustration.

[0130] In step 4021, for each of the line sections, it is determined whether there is an available line section between the i-th sensor and the j-th sensor in the case where a fault occurs in the line section.

[0131] In some embodiments, it is necessary to determine the situation of other line sections in the case where a certain line section has a fault according to the current direction. For example, the current passes through line sections A, B, C, and D in sequence. At this time, it is assumed that line section C has a fault. Since the current is from line section C to line section D, line section D also has a fault. At this time, it is necessary to determine whether there is an available line section between every two sensors, that is, the line sections that have not failed.

[0132] Among them, if there is an available line section between the i-th sensor and the j-th sensor, step 4022 is entered; if there is no available line section between the i-th sensor and the j-th sensor, step 4023 is entered.

[0133] In step 4022, the element in the i-th row and j-th column of the second adjacency matrix is determined as the first value.

[0134] Wherein, i = 1, 2,..., N, j = 1, 2,..., N, and N is the total number of sensors in the overhead line system.

[0135] In some embodiments, as long as there is an available line section, that is, a line section that has not failed, between sensor i and sensor j, it is considered that sensor i has no influence on sensor j, and the value corresponding to the i-th row and j-th column of sensor i and sensor j in the second adjacency matrix is set as the first value, and the first value is 0.

[0136] As an example of step 4022, the line segments between the first sensor and the second sensor are A, B, C, and D. At this time, it is assumed that the line segment C fails. Since the current flows from the line segment C to the line segment D, the line segment D also fails. At this time, there is an available line segment A between the first sensor and the second sensor. Therefore, the first sensor has no influence on the second sensor, and the value in the first row and the second column of the second adjacency matrix is 0.

[0137] In step 4023, the element in the i-th row and the j-th column of the second adjacency matrix is determined as the second value.

[0138] Where i = 1, 2,..., N, j = 1, 2,..., N, and N is the total number of sensors in the overhead line system.

[0139] In some embodiments, if there is no available line segment in the line segments between sensor i and sensor j, that is, all line segments fail, it is considered that sensor i has an influence on sensor j, and the value corresponding to the i-th row and the j-th column of sensor i and sensor j in the second adjacency matrix is set as the second value, and the second value is 1.

[0140] As an example of step 4023, the line segments between the first sensor and the second sensor are A and B. At this time, it is assumed that the line segment A fails. Since the current flows from the line segment A to the line segment B, the line segment B also fails. At this time, there is no available line segment between the first sensor and the second sensor. Therefore, the first sensor has an influence on the second sensor, and the value in the first row and the second column of the second adjacency matrix is 1.

[0141] Continue to refer to Figure 6A In step 403, based on the second adjacency matrix corresponding to each line segment, the first adjacency matrix corresponding to the overhead line system is determined.

[0142] In some embodiments, refer to Figure 6C , Figure 6A The step 403 shown can be implemented by the following steps 4031 to 4032, which will be described below in conjunction with Figure 6C for illustration.

[0143] In step 4031, the second adjacency matrices corresponding to each line segment are summed bit by bit to obtain a third adjacency matrix.

[0144] It should be noted that during the training process of the overall fault prediction model, only one adjacency matrix is required as the convolutional kernel for training. Therefore, it is necessary to sum the second adjacency matrices corresponding to each line segment bit by bit, and the obtained third adjacency matrix is used as the convolutional kernel during the training process of the overall fault prediction model. Summing bit by bit here can unify the adjacency matrix. In the obtained third adjacency matrix, the larger the value of an element, the more it indicates that the line segment between two sensors is susceptible to the influence of other line segments.

[0145] As an example of step 4031, for instance, the line segments are A, B, and C. The second adjacency matrix corresponding to line segment A is , the second adjacency matrix corresponding to line segment B is , and the second adjacency matrix corresponding to line segment C is , then the third adjacency matrix is .

[0146] In step 4032, set the elements in the third adjacency matrix that are less than the preset threshold to a third value to obtain the first adjacency matrix.

[0147] Here, the third value can be 0. Appropriately clearing the values in the third adjacency matrix corresponding to the line segments that are not easily affected by other line segments between two sensors can reduce the computational amount. The value of the preset threshold is related to the actual requirements and is not limited here.

[0148] Through steps 4021 to 4023, the constructed second adjacency matrix reflects whether there is an influence between sensors when each line fails. The third adjacency matrix constructed through steps 4031 to 4032 can reflect the influence state between sensors in the entire system and can clarify whether a certain line segment is more easily affected by other line segments. Setting the elements in the third adjacency matrix that are less than the preset threshold to 0 to obtain the first adjacency matrix, and using the status information of each node in the first adjacency matrix as the status matrix and taking the status matrix as the input of the overall fault prediction model can improve the accuracy of fault prediction.

[0149] Continue to refer to Figure 3 , in step 105, use the overall fault prediction model to perform prediction based on the first adjacency matrix, the first fault prediction result corresponding to the first transmission line, and real-time line data to obtain the second fault prediction result of the overhead line system.

[0150] In some embodiments, refer to Figure 7 , Figure 3 The shown step 105 can be implemented through the following steps 1051 to 1054. The following is described in conjunction with Figure 7 .

[0151] In step 1051, a state matrix is constructed based on the first fault prediction result corresponding to the first transmission line and the real-time line data.

[0152] Here, the information in the state matrix includes the sensor / corresponding fault prediction sub-model, current setting value, real-time current value, current trend, voltage setting value, real-time voltage value, voltage trend, zero-sequence voltage, zero-sequence current, whether there is a fault, and the fault type, etc.

[0153] It should be noted that in whether there is a fault, 1 indicates a fault and 0 indicates no fault; for the fault type, 1 corresponds to a short circuit, 2 corresponds to an open circuit, and 3 corresponds to a ground fault. In the current trend and voltage trend, 1 indicates a positive current or positive voltage, and 2 indicates a reverse current or reverse voltage. Normally, the current trend and voltage trend are in the same direction. The value range of the real-time voltage is -380 to +380V. The real-time current is determined according to the voltage and power factor of the three-phase electricity. Assuming U = 380V, = 0.9, and the power P is approximately between 950W and 1300W, then the current I can be obtained to be approximately between 1.5 - 2A according to the following formula (1). Taking 2A as an example, in the zero-sequence voltage, 1 indicates the existence of zero-sequence voltage and 0 indicates the non-existence. In the zero-sequence current, 1 indicates the existence of zero-sequence current and 0 indicates the non-existence.

[0154] (1)

[0155] Where I represents the current, P represents the power, and U represents the voltage, is the power factor, which is used to convert the line voltage to the phase voltage.

[0156] In some embodiments, the state matrix takes the state information of each node in the first adjacency matrix as the state matrix, and the state matrix is used as the input of the overall fault prediction model and input into the overall fault prediction model to predict the fault.

[0157] In step 1052, the first adjacency matrix is determined as the convolution kernel of the overall fault prediction model.

[0158] In step 1053, the overall fault prediction model is used to extract features from the state matrix to obtain a feature matrix.

[0159] In some embodiments, the feature extraction is implemented using the graph convolutional neural network GCN. GCN has a total of N layers, where N is greater than or equal to 2. The feature extraction of the i-th layer is performed according to the following formula (2):

[0160] (2)

[0161] Where denotes a non - linear activation function, D denotes the degree matrix of A, and A denotes an undirected adjacency matrix, i.e., the first - order adjacency matrix. denotes the input of the feature matrix of the i.e., the state matrix. denotes the parameter matrix of the i - th layer. The parameter matrix W needs to be determined through learning during model training. denotes the feature matrix output by the

[0162] i - th layer, and at the same time, it also serves as the feature matrix input to the (i + 1)-th layer. After the feature extraction of the N - th layer, the feature matrix extracted by the GCN is obtained.

[0163] Through steps 1051 to 1054, the fault prediction total model can further perform fault prediction on the basis of the fault prediction of the fault prediction sub - models, improving the accuracy of fault prediction.

[0164] In some embodiments, referring to Figure 8 , before Figure 3 step 105 shown, steps 501 to 505 can also be executed to train the fault prediction total model, which is specifically described below.

[0165] In step 501, obtain the fault prediction total model to be trained, the historical fault prediction results and historical line data corresponding to each of the first transmission lines, and the actual fault lines.

[0166] It should be noted that after the fault prediction sub - models are trained, they can be deployed first without deploying the fault prediction total model. After collecting enough data, the fault prediction total model can be trained. The historical state data refers to the state data generated by the fault prediction sub - models in the transmission lines at previous times.

[0167] In some embodiments, the feature matrix extracted by the GCN needs to be input into a classifier. The classifier includes a fully - connected layer and an activation function such as softmax, and the probability of whether each node has a fault is obtained. The transmission lines with fault probabilities higher than the threshold are regarded as the lines that may have faults, i.e., the actual fault lines.

[0168] It should be noted that the GCN training steps are the same as those in step 1053 and will not be elaborated here.

[0169] In step 502, based on the historical fault prediction results and the historical line data, construct a historical state matrix.

[0170] In some embodiments, when training the overall fault prediction model, the historical state information of each node in the first adjacency matrix needs to be used as the historical state matrix. The nodes in the first adjacency matrix may include sensors, and the historical state information of the sensors includes the historical prediction results and historical line data of their corresponding sub-fault prediction models. The requirements for the historical state matrix are the same as those for the state matrix constructed in step 1051, which will not be elaborated here.

[0171] In step 503, using the to-be-trained overall fault prediction model, based on the historical state matrix, a third prediction result is obtained through prediction.

[0172] In step 504, based on the third prediction result and the actual fault line, a loss value is determined.

[0173] When training the GCN and the classifier, the tripped nodes in the sample data are pre-marked. According to the tripping probability of each node output by the classifier and the actual tripped nodes, the loss value is calculated according to the loss function. The present application does not impose any restrictions on the setting of the loss function, and the loss function will be selected according to the actual application needs.

[0174] In step 505, using the loss value, backpropagation training is performed on the to-be-trained overall fault prediction model to obtain the trained overall fault prediction model.

[0175] It should be noted that when performing backpropagation training on the trained overall fault prediction model using the loss value, the GCN and the classifier need to be updated according to the loss value.

[0176] Through steps 501 to 505, the training of the overall fault prediction model is achieved. By updating the GCN and the classifier, the training effect of the overall fault prediction model can be further improved, and the accuracy of fault prediction can be enhanced.

[0177] Next, an exemplary application of the embodiments of the present application in an actual application scenario will be described.

[0178] Overhead lines mainly refer to overhead open wires, which are erected above the ground. The transmission wires are fixed on the poles standing upright on the ground by insulators and are used to transmit electric energy. This kind of erection and maintenance is relatively convenient and the cost is relatively low. However, it is easily affected by meteorology and the environment such as strong winds, lightning strikes, pollution, ice and snow, etc., which may cause faults. At the same time, the entire transmission corridor occupies a large amount of land area and is prone to causing electromagnetic interference to the surrounding environment.

[0179] When a fault occurs in the overhead line, it is necessary to intervene in a timely manner and make timely, rapid and accurate handling of the fault to enhance the power supply reliability and improve the satisfaction of users.

[0180] In order to improve the accuracy of overhead line fault prediction, this application deploys sensors and fault prediction sub-models at multiple key nodes of the power grid line, and uses the fault prediction sub-models matching the power grid line to perform regional fault prediction, which can effectively improve the fault prediction accuracy and prediction efficiency. At the same time, using the total fault prediction model, the total fault prediction is carried out based on the deployed multiple sensors and fault prediction sub-models to further improve the prediction accuracy.

[0181] The entire fault prediction system includes multiple sensors, fault prediction sub-models, and a total fault prediction model. Refer to Figure 9 , the deployment of the fault prediction system in the power grid line system is as follows:

[0182] The main line 600 with a total length of 500 km includes four branch lines, namely the first branch line 601, the second branch line 602, the third branch line 603, and the fourth branch line 604. The length of the first branch line 601 is 100 km, the length of the second branch line 602 is 200 km, the length of the third branch line 603 is 50 km, and the length of the fourth branch line 604 is 12 km. There are poles 605 erected at regular intervals on each line.

[0183] Due to the different distances between the location where the fault occurs and the location of the sensor, the collected fault waveform data is also different. If a single fault prediction model is uniformly used for fault prediction, the accuracy of the fault prediction is relatively poor. Therefore, in this application, in order to improve the accuracy of fault prediction, fault prediction sub-models suitable for different lengths are deployed for lines of different lengths.

[0184] Refer to Figure 9 , in the first branch line 601, a fault prediction sub-model 606 suitable for 100-km fault prediction is deployed; in the second branch line 602, a fault prediction sub-model 607 suitable for 200-km fault prediction is deployed; in the third branch line 603, a fault prediction sub-model 608 suitable for 50-km fault prediction is deployed; in the fourth branch line 604, a fault prediction sub-model 609 suitable for 12-km fault prediction is deployed. At the same time, a total fault prediction model is deployed in the monitoring center 610, and the monitoring center 610 is communicatively connected to each fault prediction sub-model respectively. Sensors are also deployed at the poles where each fault prediction sub-model is located to obtain line information in real time.

[0185] Refer to Figure 9 , Figure 9In the text, ABC is three-phase alternating current. A three-phase alternating current power system consists of three AC circuits with the same frequency, equal potential amplitudes, and a phase difference of 120 degrees. In the industrial three-phase four-wire power supply line, the line voltage, that is, the voltage between the live wires, is 380V, and the phase voltage, that is, the voltage between the live wire and the neutral wire, is 220V. Here, the three phases can all be understood as live wires.

[0186] In the model inference stage, first, for each line, each sensor obtains real-time line data. The real-time line data includes the recorded wave data of the secondary side three-phase voltage and current signals at both ends of the transmission line, and the recorded wave data of the zero-sequence voltage and zero-sequence current signals.

[0187] Secondly, a volt-ampere characteristic curve graph is constructed based on the line data, and the volt-ampere characteristic curve graph is input into the fault prediction sub-model closest to the sensor. Fault prediction is carried out based on the fault prediction sub-model to obtain the fault prediction result. The fault prediction result includes the fault type and the distance of the fault location. Here, a volt-ampere characteristic curve graph is constructed respectively for the line data collected by each sensor. The fault prediction sub-model corresponds to the sensor, and a fault prediction sub-model adapted to the line length will be configured corresponding to the location where the sensor is located, so there is no problem of multiple closest sensors. The fault type includes whether a fault occurs, and after a fault occurs, the specific type of the fault, such as phase-to-phase short circuit, circuit break, line grounding fault, etc. The fault location refers to the location where the fault occurs, such as a certain tower, or the line between any two towers, etc.

[0188] Finally, based on the fault prediction total model, the pre-constructed adjacency matrix, the results of each fault prediction sub-model, and the real-time line data, the monitoring center conducts fault prediction to obtain the final fault type and fault location. The pre-constructed adjacency matrix is the final adjacency matrix mentioned below.

[0189] The monitoring center obtains the electrical diagram of the overhead line. The electrical diagram includes the connection relationship of the transmission line and the deployment location of the fault prediction sub-model. An adjacency matrix is constructed based on the transmission line connection relationship, the sensor, and the fault prediction sub-model.

[0190] A state matrix is constructed based on the state information of each node in the adjacency matrix. The nodes in the adjacency matrix can include sensors. The state information of the sensor includes the prediction result of its corresponding fault prediction sub-model and the real-time line data. The state matrix is used as the input of the fault prediction total model, and the adjacency matrix is used as the convolution kernel of the graph neural network in the fault prediction total model. The fault prediction total model is used for fault prediction to obtain the final fault prediction result. The final fault prediction result indicates whether a fault occurs and the fault location when a fault occurs.

[0191] For the specific state information, please refer to Table 1.

[0192] Table 1, Status Information Table

[0193]

[0194] Whether there is a fault in Table 1, 1 indicates a fault, 0 indicates no fault; fault type 1 corresponds to a short circuit, 2 corresponds to an open circuit, and 3 corresponds to a ground fault. In the current trend and voltage trend, 1 indicates a positive current or positive voltage, 2 indicates a reverse current or reverse voltage. Under normal circumstances, the current trend and voltage trend are in the same direction. The value range of the real-time voltage is -380V to +380V. The real-time current is determined according to the voltage and power factor of the three-phase electricity. Assuming U = 380V, = 0.9, and the power P is approximately between 950W and 1300W, then the current I can be obtained to be approximately between 1.5 - 2A according to the following formula (3). Taking 2A as an example, in the zero-sequence voltage, 1 indicates the existence of zero-sequence voltage, 0 indicates the non-existence. In the zero-sequence current, 1 indicates the existence of zero-sequence current, 0 indicates the non-existence.

[0195] (3)

[0196] where I represents the current, P represents the power, U represents the voltage, is the power factor, used to convert the line voltage to the phase voltage.

[0197] Convert Figure 9 Simplify. The intersection of the transmission lines is represented by a triangle. Taking the triangle as the boundary, Figure 9 The 5 transmission lines shown can be divided into 9 segments, resulting in 9 line segments (corresponding to the line segments in other embodiments). The sensors, i.e., the fault prediction sub-models, are represented by circles, and five circles A, B, C, D, and E are obtained respectively. After simplification, Figure 9 See Figure 10 .

[0198] For the first line segment 701, taking whether the subsequent line segments will be affected after a fault occurs in this line segment as the standard, an adjacency matrix between the sensors is constructed. This adjacency matrix is shown in Table 2:

[0199] Table 2, Adjacency matrix constructed under the condition that the first line segment 701 has a fault

[0200]

[0201] When a fault occurs in the first line segment 701, the subsequent second line segment 702, third line segment 703, fourth line segment 704, fifth line segment 705, sixth line segment 706, seventh line segment 707, eighth line segment 708, and ninth line segment 709 will all be affected. That is, the second line segment 702, sixth line segment 706, third line segment 703, seventh line segment 707, and fourth line segment 704 between sensors A and B are all affected. That is, the number in the B column of row A is 1. At the same time, the sixth line segment 706 where row A is located will also be affected, that is, the number in the A column of row A is also 1. The same applies to others.

[0202] For the second line segment 702, taking the fact that a fault in this line segment will affect subsequent line segments as the standard, an adjacency matrix between sensors is constructed. This adjacency matrix is shown in Table 3:

[0203] Table 3, Adjacency matrix constructed in the case of a fault in the second line segment 702

[0204]

[0205] When a fault occurs in the second line segment 702, the subsequent third line segment 703, fourth line segment 704, fifth line segment 705, seventh line segment 707, eighth line segment 708, and ninth line segment 709 will be affected. The sixth line segment 706 where sensor A is located and the first line segment 701 are not affected. Therefore, the number in the A column of row A is 0, and the line segments from A to B are the sixth line segment 706, second line segment 702, third line segment 703, and fourth line segment 704. Because there is an unaffected sixth line segment 706, the value in the B column of row A is also 0. That is, as long as there is an unaffected line segment, the corresponding value is 0. The same applies to others.

[0206] For the third line segment 703, taking the fact that a fault in this line segment will affect subsequent line segments as the standard, an adjacency matrix between sensors is constructed. This adjacency matrix is shown in Table 4:

[0207] Table 4, Adjacency matrix constructed in the case of a fault in the third line segment 703

[0208]

[0209] When a fault occurs in the third line segment 703, the subsequent fourth line segment 704 and fifth line segment 705 will be affected. The same applies to others.

[0210] For the fourth line segment 704, taking the fact that a fault in this line segment will affect subsequent line segments as the standard, an adjacency matrix between sensors is constructed. This adjacency matrix is shown in Table 5:

[0211] Adjacency matrix constructed in the case of a fault occurring in the fourth line segment 704, Table 5

[0212]

[0213] Taking the above adjacency matrix as an example, for each line segment from the first line segment 701 to the ninth line segment 709, an adjacency matrix is constructed after a fault occurs. Nine adjacency matrices corresponding to the nine line segments can be obtained. By adding the nine adjacency matrices, the final adjacency matrix is obtained. Optionally, relationships less than 3 can be deleted to obtain the final adjacency matrix. Here, addition is because only one is used as the convolutional kernel in the final adjacency matrix, and there are nine adjacency matrices here. On the one hand, it is necessary to unify them, and on the other hand, the larger the value, the more it indicates that the line segment between two sensor nodes is easily affected by other line segments. Deleting those less than 3 is to reduce the computational complexity and appropriately delete the values with less influence on the line segment without calculation.

[0214] The total fault prediction model is a pre-trained graph neural network, specifically it can be a graph convolutional neural network GCN.

[0215] In the model training stage, first is the training of the fault prediction sub-model. Fault waveform data of various transmission lines are obtained, including fault waveform data of 500km overhead lines, 200km overhead lines, 100km overhead lines, 50km overhead lines, and 12km overhead lines. The fault waveform data is the same as the data in the inference stage, and also includes the recorded wave data of three-phase voltage and current signals. The fault waveform data of various transmission lines is preprocessed, including denoising, normalization, etc., to obtain training data.

[0216] Based on the training data corresponding to the 200km overhead line fault waveform data, a convolutional neural network CNN is trained to obtain a 200km fault prediction sub-model. The convolutional neural network includes multiple convolutional layers and fully connected layers. The parameters of the convolutional layers in the 200km fault prediction sub-model are retained, and the parameters of the fully connected layers are cleared to obtain an intermediate fault prediction sub-model. This operation is to quickly train various types of models using limited data. On the one hand, it can improve the training speed, and on the other hand, it can facilitate the subsequent targeted layout of different fault models for different lines to improve the accuracy of fault detection. Selecting 200km is because an intermediate value needs to be selected so that the gap with other data is relatively moderate.

[0217] Respectively use the training data corresponding to the fault waveform data of 500 km overhead lines, the training data corresponding to the fault waveform data of 100 km overhead lines, the training data corresponding to the fault waveform data of 50 km overhead lines, and the training data corresponding to the fault waveform data of 12 km overhead lines to retrain the fully connected layer in the intermediate fault prediction sub-model to obtain multiple corresponding fault prediction sub-models.

[0218] Secondly is the training of the overall fault prediction model. Obtain the electrical diagram of the overhead line. The electrical diagram includes the connection relationship of the transmission line and the deployment location of the fault prediction sub-model. Based on the transmission line connection relationship, sensors, and the fault prediction sub-model, construct an adjacency matrix. Use the historical state information of each node in the adjacency matrix as the historical state matrix. The nodes in the adjacency matrix can include sensors. The historical state information of the sensors includes the historical prediction results and historical line data of their corresponding fault prediction sub-models. After the fault prediction sub-model is trained, it can be deployed first, and the overall fault prediction model is not deployed. Wait until enough data is collected and then train the overall fault prediction model. The historical state data refers to the state data generated by the fault prediction sub-model in the past time.

[0219] Use the graph convolutional neural network GCN for feature extraction. GCN includes N layers, where N is greater than or equal to 2. The th layer performs feature extraction according to the following formula (4).

[0220] (4)

[0221] Among them, represents the non-linear activation function, D represents the degree matrix of A, A represents the undirected adjacency matrix, represents the feature matrix input to the i-th layer, W i represents the parameter matrix of the i-th layer. The parameter matrix W needs to be determined through learning during model training. represents the feature matrix output by the i-th layer, and at the same time serves as the feature matrix input to the i + 1-th layer. After the feature extraction of the N-th layer, the feature matrix extracted by GCN is obtained.

[0222] Finally, input the extracted feature matrix into the classifier. The classifier includes a fully connected layer and an activation function such as softmax to obtain the probability of whether each node has a fault. Regard the transmission line with a fault probability higher than the threshold as the line that may have a fault.

[0223] When training GCN and the classifier, pre-mark the tripped nodes in the sample data. According to the tripping probability of each node output by the classifier and the actual tripped nodes, calculate the loss value to update GCN and the classifier.

[0224] Construct an adjacency matrix based on the electrical diagram of the overhead line, use the node information of the adjacency matrix as the state matrix, and perform fault prediction and location based on the adjacency matrix, the state matrix and the total fault prediction model, thereby improving the accuracy of fault prediction. Adopt the method of transfer learning, save the convolution layer parameters in the fault prediction sub-model obtained by training with 200 km of training data, use the training data of other different lines to train the fault prediction sub-models applicable to different line lengths respectively, thereby improving the pertinence and accuracy of fault prediction.

[0225] The following continues to describe the exemplary structure of the fault prediction device 255 provided in the embodiment of the present application as a software module. In some embodiments, refer to Figure 2 , the software module stored in the fault prediction device 255 in the memory 250 may include:

[0226] The data acquisition module 2551 is used to acquire the real-time line data of each first transmission line in the overhead line system.

[0227] The construction module 2552 is used to construct a volt-ampere characteristic curve graph for each first transmission line based on the real-time line data of the first transmission line.

[0228] The first fault prediction module 2553 is used to use the fault prediction sub-model corresponding to the first transmission line to perform prediction based on the volt-ampere characteristic curve graph to obtain the first fault prediction result of the first transmission line.

[0229] The second fault prediction module 2554 is used to obtain the first adjacency matrix corresponding to the overhead line system, and use the total fault prediction model to perform prediction based on the first adjacency matrix, the first fault prediction result corresponding to the first transmission line and the real-time line data to obtain the second fault prediction result of the overhead line system.

[0230] In some embodiments, the second fault prediction module 2554 is further used to construct a state matrix based on the first fault prediction result and the real-time line data corresponding to the first transmission line; determine the first adjacency matrix as the convolution kernel of the total fault prediction model, use the total fault prediction model to extract features from the state matrix to obtain a feature matrix; perform prediction based on the feature matrix to obtain the second fault prediction result of the overhead line system.

[0231] The fault prediction device 255 further includes a first adjacency matrix construction module, configured to obtain an electrical diagram of the overhead line system, where the electrical diagram includes: at least one line segment included in each of the first transmission lines, the connection relationships between the line segments, and the deployment positions of the sensors for collecting real-time line data of the first transmission lines; for each of the line segments, based on the connection relationships between the line segments and the deployment positions of the sensors, construct a second adjacency matrix corresponding to the line segment; and based on the second adjacency matrices corresponding to the line segments, determine a first adjacency matrix corresponding to the overhead line system.

[0232] The fault prediction device 255 further includes a second adjacency matrix construction module, configured to, for each of the line segments, in the case where a fault occurs in the line segment, if there is an available line segment between the i-th sensor and the j-th sensor, determine the element in the i-th row and j-th column of the second adjacency matrix as a first value, where i = 1, 2,..., N, j = 1, 2,..., N, and N is the total number of sensors in the overhead line system; if there is no available line segment between the i-th sensor and the j-th sensor, determine the element in the i-th row and j-th column of the second adjacency matrix as a second value.

[0233] The fault prediction device 255 further includes a third adjacency matrix construction module, configured to perform bitwise summation on the second adjacency matrices corresponding to the line segments to obtain a third adjacency matrix; and set the elements in the third adjacency matrix that are less than a preset threshold to a third value to obtain the first adjacency matrix.

[0234] The fault prediction device 255 further includes a first fault prediction sub-model training module, configured to obtain a fault prediction sub-model to be trained and fault waveform data corresponding to a plurality of second transmission lines with different candidate lengths; for the fault waveform data corresponding to each of the second transmission lines with the candidate lengths, preprocess the fault waveform data to obtain training data corresponding to the second transmission lines with the candidate lengths; based on the average length among the plurality of different candidate lengths, determine a first length; use the training data corresponding to the second transmission lines with the first length to train the fault prediction sub-model to be trained to obtain a trained fault prediction sub-model corresponding to the second transmission lines with the first length; and use the training data corresponding to the second transmission lines with other lengths to adjust the parameters of the trained fault prediction sub-model corresponding to the second transmission lines with the first length to obtain a trained fault prediction sub-model corresponding to the second transmission lines with the other lengths, where the other lengths are any of the candidate lengths other than the first length among the plurality of different candidate lengths.

[0235] The fault prediction device 255 further includes a second fault prediction sub-model training module, which is configured to keep the parameters of the convolutional layer in the trained fault prediction sub-model corresponding to the second transmission line of the first length unchanged, set the parameters of the fully connected layer in the trained fault prediction sub-model corresponding to the second transmission line of the first length to a fourth value, and obtain an intermediate fault prediction sub-model; and use the training data corresponding to the second transmission lines of other lengths to adjust the parameters of the fully connected layer in the intermediate fault prediction sub-model to obtain the trained fault prediction sub-models corresponding to the second transmission lines of other lengths.

[0236] The fault prediction device 255 further includes an actual fault prediction sub-model determination module, which is configured to obtain the second length of each of the first transmission lines; for each of the first transmission lines, determine the first length with the smallest difference from the second length among multiple candidate lengths; and determine the trained fault prediction sub-model corresponding to the first length as the fault prediction sub-model corresponding to the first transmission line.

[0237] The fault prediction device 255 further includes a fault prediction total model training module, which is configured to obtain a fault prediction total model to be trained, the historical fault prediction results and historical line data corresponding to each of the first transmission lines, and the actual fault lines; construct a historical state matrix based on the historical fault prediction results and the historical line data; perform prediction based on the historical state matrix by using the fault prediction total model to be trained to obtain a third prediction result; determine a loss value based on the third prediction result and the actual fault lines; and perform backpropagation training on the fault prediction total model to be trained by using the loss value to obtain a trained fault prediction total model.

[0238] An embodiment of the present application provides a computer program product, which includes a computer program or computer executable instructions, and the computer program or computer executable instructions are stored in a computer-readable storage medium. A processor of an electronic device reads the computer executable instructions from the computer-readable storage medium, and the processor executes the computer executable instructions, so that the electronic device executes the fault prediction method described above in the embodiments of the present application.

[0239] An embodiment of the present application provides a computer-readable storage medium, in which computer executable instructions or a computer program are stored, and when the computer executable instructions or the computer program are executed by a processor, the processor will be caused to execute the fault prediction method provided by the embodiment of the present application.

[0240] In some embodiments, the computer-readable storage medium may be a memory such as RAM, ROM, flash memory, magnetic surface memory, optical disc, or CD-ROM; or may be various devices including one or any combination of the above memories.

[0241] In some embodiments, the computer-executable instructions may be in the form of a program, software, software module, script, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including being deployed as a stand-alone program or being deployed as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0242] As an example, the computer-executable instructions may or may not correspond to a file in a file system, may be stored as part of a file that holds other programs or data, for example, in one or more scripts in a Hyper Text Markup Language (HTML) document, stored in a single file dedicated to the program being discussed, or, stored in multiple cooperating files (such as files that store one or more modules, subroutines, or portions of code).

[0243] As an example, the computer-executable instructions may be deployed to execute on one electronic device, or on multiple electronic devices located at one location, or, on multiple electronic devices distributed at multiple locations and interconnected by a communication network.

[0244] In summary, through the embodiments of the present application, a fault prediction sub-model applicable to different line lengths can be trained, and the fault prediction and location can be performed in combination with the overall fault prediction model, improving the accuracy and pertinence of fault prediction and location.

[0245] The above is only the embodiments of the present application and is not intended to limit the protection scope of the present application. Any modifications, equivalent replacements, and improvements made within the spirit and scope of the present application are all included in the protection scope of the present application.

Claims

1. A fault prediction method, characterized in that: The method comprises: Acquiring real-time line data of each first transmission line in the overhead line system; For each first power transmission line, construct a volt-ampere characteristic curve diagram based on the real-time line data of the first power transmission line; Using the fault prediction sub-model corresponding to the first transmission line, and based on the volt-ampere characteristic curve diagram, a prediction is performed to obtain a first fault prediction result of the first transmission line; The fault prediction sub-model corresponding to the first transmission line is determined based on a second length of the first transmission line, and the second length represents an actual length of the first transmission line; Acquire a first adjacency matrix corresponding to the overhead line system, and construct a state matrix based on a first fault prediction result corresponding to the first transmission line and real-time line data; Determine the first adjacency matrix as the convolution kernel of the overall fault prediction model, and use the overall fault prediction model to perform feature extraction on the state matrix to obtain a feature matrix; The characteristic matrix is ​​input into the classifier in the overall fault prediction model to obtain a second fault prediction result of the overhead line system.

2. The method according to claim 1, characterized in that The method further comprises: Acquire an electrical diagram of the overhead line system, the electrical diagram comprising: at least one line section included in each of the first transmission lines, a connection relationship between each of the line sections, and a deployment position of each sensor, the sensor being used to collect real-time line data of the first transmission line; For each of the route sections, based on the connection relationship between the route sections and the deployment position of each sensor, construct a second adjacency matrix corresponding to the route section; Based on the second adjacency matrix corresponding to each of the line sections, a first adjacency matrix corresponding to the overhead line system is determined.

3. The method according to claim 2, characterized in that The step of constructing, for each of the route sections, a second adjacency matrix corresponding to the route section based on the connection relationship between the route sections and the deployment positions of the sensors, includes: For each of the line sections, when a fault occurs in the line section, if there is an available line section between the i-th sensor and the j-th sensor, determine the element of the i-th row and j-th column of the second adjacency matrix as a first value, where i=1,2,…,N, j=1,2,…,N, and N is the total number of sensors in the overhead line system; If there is no available route section between the i-th sensor and the j-th sensor, the element in the i-th row and j-th column of the second adjacency matrix is ​​determined as a second value.

4. The method according to claim 2, characterized in that: The determining, based on the second adjacency matrix corresponding to each of the line sections, a first adjacency matrix corresponding to the overhead line system comprises: Performing bitwise summation on the second adjacency matrices corresponding to the route sections to obtain a third adjacency matrix; The elements in the third adjacency matrix that are smaller than a preset threshold are set to a third value to obtain the first adjacency matrix.

5. The method according to any one of claims 1 to 4, characterized in that: The method further comprises: Acquire a fault prediction sub-model to be trained and fault waveform data corresponding to a plurality of second transmission lines of different candidate lengths; For each of the fault waveform data corresponding to the second transmission line of the candidate length, preprocessing the fault waveform data to obtain training data corresponding to the second transmission line of the candidate length; Determining a first length based on an average length among the plurality of different candidate lengths; Using the training data corresponding to the second transmission line of the first length, training the fault prediction sub-model to be trained to obtain a trained fault prediction sub-model corresponding to the second transmission line of the first length; By using the training data corresponding to the second transmission lines of other lengths, the parameters of the trained fault prediction sub-model corresponding to the second transmission lines of the first length are adjusted to obtain the trained fault prediction sub-model corresponding to the second transmission lines of the other lengths, wherein the other lengths are any candidate lengths among the multiple different candidate lengths except the first length.

6. The method according to claim 5, characterized in that The method of using the training data corresponding to the second transmission line of other lengths to adjust the parameters of the trained fault prediction sub-model corresponding to the second transmission line of the first length to obtain the trained fault prediction sub-model corresponding to the second transmission line of other lengths includes: Keeping unchanged the parameters of the convolutional layer in the trained fault prediction sub-model corresponding to the second transmission line of the first length, setting the parameters of the fully connected layer in the trained fault prediction sub-model corresponding to the second transmission line of the first length to a fourth value, to obtain an intermediate fault prediction sub-model; The training data corresponding to the second transmission lines of other lengths are used to adjust the parameters of the fully connected layer in the intermediate fault prediction sub-model to obtain the trained fault prediction sub-model corresponding to the second transmission lines of other lengths.

7. The method according to claim 6, characterized in that The method further comprises: Obtaining a second length of each of the first transmission lines; For each of the first transmission lines, determine a first length having the smallest difference with the second length from a plurality of the candidate lengths; The trained fault prediction sub-model corresponding to the first length is determined as the fault prediction sub-model corresponding to the first transmission line.

8. The method according to any one of claims 1 to 4, characterized in that: The method further comprises: Obtaining the overall fault prediction model to be trained, historical fault prediction results and historical line data corresponding to each of the first transmission lines, and actual fault lines; Based on the historical fault prediction results and the historical line data, construct a historical state matrix; Using the overall fault prediction model to be trained, prediction is performed based on the historical state matrix to obtain a third prediction result; Determining a loss value based on the third prediction result and the actual fault line; The loss value is used to perform back-propagation training on the fault prediction overall model to be trained to obtain a trained fault prediction overall model.

9. A fault prediction device, characterized in that: The device comprises: A data acquisition module, used to acquire real-time line data of each first transmission line in the overhead line system; A construction module, configured to construct a volt-ampere characteristic curve diagram for each first power transmission line based on real-time line data of the first power transmission line; a first fault prediction module, configured to use a fault prediction sub-model corresponding to the first transmission line to perform prediction based on the volt-ampere characteristic curve diagram to obtain a first fault prediction result of the first transmission line; wherein the fault prediction sub-model corresponding to the first transmission line is determined based on a second length of the first transmission line, and the second length represents an actual length of the first transmission line; The second fault prediction module is used to obtain a first adjacency matrix corresponding to the overhead line system, and construct a state matrix based on a first fault prediction result corresponding to the first transmission line and real-time line data; determine the first adjacency matrix as a convolution kernel of a total fault prediction model, and use the total fault prediction model to perform feature extraction on the state matrix to obtain a feature matrix; input the feature matrix into a classifier in the total fault prediction model to obtain a second fault prediction result of the overhead line system.

10. An electronic device, characterized in that: The electronic device comprises: A memory for storing computer executable instructions or computer programs; A processor, configured to implement the method according to any one of claims 1 to 8 when executing computer executable instructions or computer programs stored in the memory.

11. A computer-readable storage medium storing computer-executable instructions or a computer program, characterized in that: When the computer executable instructions or computer programs are executed by a processor, the method according to any one of claims 1 to 8 is implemented.

12. A computer program product comprising computer executable instructions or a computer program, characterized in that: When the computer executable instructions or computer programs are executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images