A distributed power transmission line fault accurate positioning system and method

Abstract

This invention relates to the field of distributed fault location technology for transmission lines, and discloses a distributed transmission line fault precise location system and method, including: collecting waveform data at the time of a transmission line fault; constructing graph structure data based on the collected waveform data; inputting the graph structure data into a neural network model, the neural network model outputting a result representing the location of the fault point on the transmission line; calculating the distance between the fault point and the endpoint of the first transient waveform recording of the fault based on the fault distance ratio, extracting the endpoint position of the first transient waveform recording of the fault, and then locating the fault point; this invention uses deep learning to organize the waveform data of the distributed transmission line fault location system, and the fusion of global transient waveform data provides information richness that can remove interference information, thereby improving the accuracy of location to a certain extent.

Description

Technical Field

[0001] This invention relates to the field of distributed fault location technology for power transmission lines, and more specifically, to a distributed power transmission line fault precise location system and method. Background Technology

[0002] Traveling wave location technology is a fault diagnosis method based on measuring the propagation time of transient traveling waves generated by a fault. It features high accuracy and rapid response. This technology has been widely applied to transmission lines at voltage levels of 110kV and above, and is gradually being implemented in distribution networks. Traveling wave location technology includes current traveling wave location and voltage traveling wave location, as well as single-ended measurement methods and double-ended measurement methods, among other classifications.

[0003] Based on their structural approach, they can be divided into: distributed and converged.

[0004] Distributed systems use the same structure as power transmission systems, with sensors installed on the lines. Integrated systems combine traveling wave sensors with pole-mounted switches, either internally or externally located on the pole.

[0005] Although the fault location system of the distributed system architecture has distributed data acquisition terminals, it still uses traditional methods when locating faults and does not make full use of the data collected by the distributed system. Summary of the Invention

[0006] This invention provides a system and method for precise location of faults in distributed transmission lines, solving the technical problems in related technologies.

[0007] This invention provides a method for precise fault location in distributed transmission lines, comprising:

[0008] Step 101: Collect waveform data during power transmission line faults;

[0009] The transmission line is divided into multiple line segments, and waveform data is collected at the endpoints of each data segment;

[0010] The waveform data includes transient waveform data of the endpoints of all line segments during a single transmission line fault.

[0011] Step 102: Construct graph structure data based on the acquired waveform data;

[0012] Graph structure data includes nodes and edges connecting nodes. Nodes include trunk nodes and branch nodes. A trunk node represents a trunk line segment, and a branch node represents a branch line segment.

[0013] The rules for generating edges between nodes are as follows:

[0014] If the line segments represented by two trunk nodes are connected, then an edge is generated between these two trunk nodes;

[0015] If the line segments represented by the trunk node and the branch node are connected, then an edge is generated between the trunk node and the branch node.

[0016] If two branch nodes belong to the same branch, then there is an edge between these two branch nodes;

[0017] Step 103: Input the graph structure data into the neural network model, and the neural network model outputs the result representing the location of the fault point of the transmission line;

[0018] Neural network models include:

[0019] The autoencoder layer is used to take a fault transient waveform as input and output an autoencoder feature. When the input node represents the fault transient waveform of the starting point of the line segment, the output is the autoencoder feature corresponding to that node.

[0020] Tensor layers are used to tensor the autoencoded features of nodes into vector form and then input them into the graph structure learning layer.

[0021] The graph structure learning layer is used to input the autoencoded features of the nodes and the graph structure data, and outputs global features.

[0022] A fully connected layer is used as input for global features and then outputs a result representing the distance ratio to the fault point;

[0023] Step 104: Calculate the distance between the fault point and the endpoint of the transient waveform of the first fault start based on the fault distance ratio, extract the endpoint position of the transient waveform of the first fault start, and then locate the fault point.

[0024] A positive fault distance ratio indicates that the fault point is on the line segment after the endpoint that first initiated the fault transient waveform recording; otherwise, it indicates that the fault point is on the line segment before the endpoint that first initiated the fault transient waveform recording.

[0025] Furthermore, the fault transient waveform is collected synchronously across all line segments. When the endpoint of any line segment triggers the fault transient waveform, the endpoints of all line segments synchronously perform fault transient waveform recording.

[0026] Furthermore, the line segment includes the line segment of the main line and the line segment of the branch line.

[0027] Furthermore, the distance ratio to the fault point is expressed as follows:

[0028]

[0029] Among them, L 0，+L represents the length of the line segment preceding the endpoint where the fault transient recording is first initiated. 0，- L represents the length of the line segment following the endpoint where the fault transient recording was first initiated. 0，b This indicates the distance between the fault point and the endpoint that first initiated the fault transient waveform recording.

[0030] Furthermore, the autoencoder layer takes the waveform data of segments A, B, C, and D of the fault transient waveform data as input, and then outputs the weighted features of segment A, segment B, segment C, and segment D.

[0031] Each waveform segment A, B, C, and D is itself a time series, containing waveform sample values ​​from multiple time steps. They can be considered as four independent input sequences, denoted as X. A X B X C X D .

[0032] For each sequence, add the corresponding segment type embedding. For example, for X A The embedding vector of segment A is copied and expanded to be the same as that of segment X. A A matrix of equal length is denoted as E. A Then X A and E A By concatenating along the feature dimensions, we obtain X′. A .

[0033] The four enhanced input sequences X′ A , X′ B , X′ C , X′ D Concatenate them chronologically to form a long sequence X′=[X′ A , X′ B , X′ C , X′ D ].

[0034] Finally, X′ is input into the attention layer;

[0035] The formula for calculating the attention layer is as follows:

[0036] Perform a linear transformation on X′ to obtain the Query matrix Q, Key matrix K, and Value matrix V.

[0037] Q = X′W Q

[0038] K = X′W K

[0039] V = X′W V

[0040] in Let d be a learnable weight matrix. k For the dimension of attention head.

[0041] Then calculate the attention weight matrix.

[0042]

[0043] Finally, the attention weight matrix is ​​multiplied by the value matrix and linearly transformed to obtain the output autoencoded features.

[0044] Z = AVW O

[0045] in This is the output layer weight matrix.

[0046] Furthermore, the graph structure learning layer includes three hidden layers and one linear layer. The calculation formula for the first hidden layer is as follows:

[0047]

[0048] in Let be the attention weight of node u to node v, and let N(v) represent the set of nodes that have an edge with node v. The first weight parameter of the first layer is calculated using the following formula:

[0049]

[0050] 'a' is a learnable attention vector, and || denotes the vector concatenation operation. This represents the second weight parameter of the first layer, and LeakyReLU represents the LeakyReLU function;

[0051]

[0052] where x v and x u Let v and u represent the feature representations of the v-th and u-th nodes, respectively.

[0053] The formula for calculating the linear layer is as follows:

[0054] Q t Represents global features. Let N(all) represent the feature representation of node c in the third layer, N(all) represent the set of all nodes, and Concat represents vector concatenation.

[0055] Furthermore, tensor quantization in tensor layers involves expanding the matrix row by row into vectors and then concatenating them.

[0056] Furthermore, the neural network model also includes a second fully connected layer, which takes global features as input and outputs a classification result representing the fault type.

[0057] This invention provides a distributed transmission line fault precise location system, comprising:

[0058] Signal detection terminals installed on the main lines and branches of transmission lines;

[0059] A communication terminal used to receive data from a signal detection terminal; the communication terminal is used to send data to the system master station.

[0060] The system master station is used to execute steps 102-104 of the aforementioned method for precise location of faults in distributed transmission lines.

[0061] Furthermore, the signal detection terminal synchronizes the recording time based on BeiDou or GPS timing. The signal detection terminal can record waveforms according to the recording command. The recording start time included in the recording command can obtain synchronized fault transient waveform recording when the command is executed.

[0062] The present invention provides a storage medium storing computer-readable instructions that, when read, can execute the aforementioned method for precise location of faults in distributed transmission lines.

[0063] The beneficial effects of this invention are as follows: This invention uses deep learning to organize the waveform data of a distributed transmission line fault location system, and utilizes the transmission characteristics of transient waves on transmission lines to fuse transient waveform data from multiple line segments. This solves the problem that it is difficult to remove interference information from a single transient waveform data when performing fault location. The fusion of global transient waveform data provides information richness that can remove interference information, thereby improving the accuracy of location to a certain extent. Attached Figure Description

[0064] Figure 1 This is a flowchart of the method for accurate location of distributed transmission line faults according to the present invention;

[0065] Figure 2 This is a schematic diagram of the neural network model. Figure 1 ;

[0066] Figure 3 This is a schematic diagram of the neural network model. Figure 2 ;

[0067] Figure 4-1 Schematic diagram of single-line configuration for overhead power lines;

[0068] Figure 4-2 Site view of a 220kV overhead power line;

[0069] Figure 4-3 Schematic diagram of single-line cable configuration;

[0070] Figure 4-4 Schematic diagram of a mixed overhead and cable line configuration;

[0071] Figure 4-5 Schematic diagram of overhead T-junction line configuration;

[0072] Figure 4-6 Schematic diagram of cable line branch box configuration;

[0073] Figure 5-1 Typical waveform diagram of lightning strike fault;

[0074] Figure 5-2 Typical waveform diagram of lightning backflashover fault;

[0075] Figure 5-3 Typical waveform diagram of tree obstruction fault;

[0076] Figure 5-4 Typical waveform diagram of wildfire failure;

[0077] Figure 5-5 Typical waveform diagram of a hanging object failure;

[0078] Figure 5-6 Typical waveform diagram of icing fault;

[0079] Figure 5-7 Typical waveform diagram of wind deflection fault;

[0080] Figure 5-8 Typical waveform diagram of construction equipment touching the line failure. Detailed Implementation

[0081] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, some features described in the examples may be combined in other examples.

[0082] At least one embodiment of the present invention discloses a method for precise fault location in distributed transmission lines, such as... Figure 1 As shown, including:

[0083] Step 101: Collect waveform data during power transmission line faults;

[0084] The transmission line is divided into multiple segments, and waveform data is collected at the endpoints of each segment.

[0085] The fault transient waveform is collected synchronously across all line segments. When the endpoint of any line segment triggers the fault transient waveform, the endpoints of all line segments synchronously perform fault transient waveform recording.

[0086] Therefore, the waveform data includes fault transient waveform data of all line segments at the endpoints during a single transmission line fault.

[0087] Of course, in order to ensure the time synchronization of the start point of waveform recording, it is necessary to synchronize the waveform recording time according to Beidou or GPS time synchronization. Even if there is a time delay in the transmission of waveform recording command, the waveform recording start time contained in the command can be synchronized with the fault transient waveform recording when the command is executed.

[0088] Fault transient waveform triggering is generally initiated based on abnormal fluctuations in current and voltage. Typically, a threshold value for current or voltage is set as the triggering condition. This threshold value can be the magnitude or phase of the current or voltage.

[0089] A line segment includes the main line segment and branch line segments;

[0090] Step 102: Construct graph structure data based on the acquired waveform data;

[0091] Graph structure data includes nodes and edges connecting nodes. Nodes include trunk nodes and branch nodes. A trunk node represents a trunk line segment, and a branch node represents a branch line segment.

[0092] The rules for generating edges between nodes are as follows:

[0093] If the line segments represented by two trunk nodes are connected, then an edge is generated between these two trunk nodes;

[0094] If the line segments represented by the trunk node and the branch node are connected, then an edge is generated between the trunk node and the branch node.

[0095] If two branch nodes belong to the same branch, then there is an edge between these two branch nodes;

[0096] This is because fluctuations caused by branch failures have a smaller impact on the trunk line, but a larger impact on the same branch line. Therefore, full connection edge operations are performed on the nodes inside the branch line. As for the trunk line itself, the faults in the trunk line are more regularly propagated in the transient waves of the trunk line, so edge operations are only performed on the nodes of adjacent line segments.

[0097] Step 103: Input the graph structure data into the neural network model, and the neural network model outputs the result representing the location of the fault point of the transmission line;

[0098] Neural network models include:

[0099] The autoencoder layer is used to take a fault transient waveform as input and output an autoencoder feature. When the input node represents the fault transient waveform of the starting point of the line segment, the output is the autoencoder feature corresponding to that node.

[0100] Tensor layers are used to tensor the autoencoded features of nodes into vector form and then input them into the graph structure learning layer.

[0101] The graph structure learning layer is used to input the autoencoded features of the nodes and the graph structure data, and outputs global features.

[0102] A fully connected layer is used as input for global features and then outputs a result representing the distance ratio to the fault point;

[0103] The distance ratio to the fault point is expressed as follows:

[0104]

[0105] Among them, L 0，+ L represents the length of the line segment preceding the endpoint where the fault transient recording is first initiated. 0，- L represents the length of the line segment following the endpoint where the fault transient recording was first initiated. 0，b This indicates the distance between the fault point and the endpoint of the first fault transient waveform recording.

[0106] Step 104: Calculate the distance between the fault point and the endpoint of the transient waveform of the first fault start based on the fault distance ratio, extract the endpoint position of the transient waveform of the first fault start, and then locate the fault point.

[0107] A positive value for B0 indicates that the fault point is on the line segment after the endpoint of the first fault transient waveform recording; otherwise, it indicates that the fault point is on the line segment before the endpoint of the first fault transient waveform recording.

[0108] The autoencoder layer takes the waveform data of segments A, B, C, and D of the fault transient waveform data as input, and then outputs the weighted features of segment A, segment B, segment C, and segment D.

[0109] Each waveform segment A, B, C, and D is itself a time series, containing waveform sample values ​​from multiple time steps. They can be considered as four independent input sequences, denoted as X. A X B X C X D .

[0110] For each sequence, add the corresponding segment type embedding. For example, for X A The embedding vector of segment A is copied and expanded to be the same as that of segment X. A A matrix of equal length is denoted as E. A Then XA and E A By concatenating along the feature dimensions, we obtain X′. A .

[0111] The four enhanced input sequences X′ A , X′ B , X′ C , X′ D Concatenate them chronologically to form a long sequence X′=[X′ A , X′ B , X′ C , X′ D ].

[0112] Finally, X′ is input into the attention layer;

[0113] The formula for calculating the attention layer is as follows:

[0114] Perform a linear transformation on X′ to obtain the Query matrix Q, Key matrix K, and Value matrix V.

[0115] Q = X′W Q

[0116] K = X′W K

[0117] V = X′W V

[0118] in Let d be a learnable weight matrix. k For the dimension of attention head.

[0119] Then calculate the attention weight matrix.

[0120]

[0121] Finally, the attention weight matrix is ​​multiplied by the value matrix and linearly transformed to obtain the output autoencoded features.

[0122] Z = AVW O

[0123] in This is the output layer weight matrix.

[0124] The graph structure learning layer consists of three hidden layers and one linear layer. The calculation formula for the l-th hidden layer is as follows:

[0125]

[0126] in Let be the attention weight of node u to node v, and let N(v) represent the set of nodes that have an edge with node v. The first weight parameter of the l-th layer is calculated using the following formula:

[0127]

[0128] 'a' is a learnable attention vector, and '∥' represents a vector concatenation operation. This represents the second weight parameter of the l-th layer, and LeakyReLU represents the LeakyReLU function;

[0129]

[0130] where x v and x u Let v and u represent the feature representations of the v-th and u-th nodes, respectively.

[0131] The formula for calculating the linear layer is as follows:

[0132] Q t Represents global features. Let N(all) represent the feature representation of node c in the third layer, N(all) represent the set of all nodes, and Concat represents vector concatenation.

[0133] In one embodiment of the present invention, tensor quantization of the tensor layer is performed by expanding the matrix row by row into vectors and then concatenating them.

[0134] The fully connected layer uses a general linear mapping to output the value of the fault point distance ratio.

[0135] In existing technologies, the waveform of fault recording is compared with... Figures 5-1 to 5-8 The fault type is identified by comparing typical waveforms of the faults shown. Although these typical waveforms are representative, they are generally obtained through tests under power-off conditions. The fault waveforms monitored online are much more complex than typical waveforms, making them difficult to identify and easily leading to errors in fault type identification.

[0136] In one embodiment of the invention, the neural network model further includes a second fully connected layer, which takes global features as input and outputs a classification result representing the fault type.

[0137] Furthermore, the cross-entropy loss function of the classification result output by the second fully connected layer can be weighted and summed with the prediction loss function of the fault point distance ratio to calculate the comprehensive loss for training the neural network model, which helps to improve the convergence speed of the neural network model and the recognition accuracy of the fault point distance ratio.

[0138] At least one embodiment of the present invention provides a distributed transmission line fault precise location system, comprising:

[0139] Signal detection terminals installed on the main lines and branches of transmission lines;

[0140] A communication terminal used to receive data from a signal detection terminal; the communication terminal is used to send data to the system master station.

[0141] The system master station is used to execute steps 102-104 of the aforementioned method for precise location of faults in distributed transmission lines.

[0142] The main functions of the signal detection terminal provided in this embodiment include:

[0143] Real-time monitoring function

[0144] The signal detection terminal collects the current, altitude, latitude and longitude of the transmission line conductors in real time and monitors their changes.

[0145] Scheduled data collection and uploading

[0146] The signal detection terminal periodically uploads measurement data such as conductor current, altitude, latitude and longitude, conductor temperature, and humidity to the communication terminal;

[0147] The communication terminal periodically uploads measurement data such as conductor current, altitude, latitude and longitude, conductor temperature, and humidity to the system master station.

[0148] Fault signal acquisition and uploading

[0149] After the condition is triggered, the signal detection terminal completes high-speed sampling of the current, records the fault current waveform and the current traveling wave waveform, and uploads them to the communication terminal.

[0150] After receiving the fault waveform data from the signal detection terminal, the communication terminal filters the data and uploads the valid data to the backend main station.

[0151] Data storage and historical data upload

[0152] The communication terminal supports up to 2G Nand Flash and 4G DDRII storage, with a standard configuration of 128M Nand Flash and 2G DDRII storage. The standard configuration can achieve 100,000*12 SOE event records and 333*12 complete waveform data storage (each waveform file is 32*12KB).

[0153] The communication terminal can not only actively upload monitoring data for line faults, but also record waveforms and perform measurements in real time, and supports the function of remotely and manually retrieving data from the back-end main station.

[0154] Equipment self-test

[0155] The signal detection terminal can periodically collect and upload data such as induced current, GPS online status, and battery / capacitor voltage.

[0156] The signal detection terminal has alarm functions for low battery / capacitor voltage, GPS not online, and low short-range wireless communication signal strength, as well as information uploading functions.

[0157] The communication terminal can periodically collect and upload parameters such as solar power supply current (optional), battery voltage, GPS online status, short-range wireless communication signal strength, and GPRS received signal strength (quality);

[0158] The communication terminal has alarm functions for low battery voltage, GPS not online, low short-range wireless communication signal strength, and low GPRS received signal strength (poor quality), as well as information uploading functions.

[0159] The communication terminal has software and hardware self-testing functions to prevent the program from 'running away' or 'deadlocking'.

[0160] Single-line configuration scheme for overhead lines

[0161] For overhead lines, one monitoring point is set at each power source, and the location and number of monitoring points can be flexibly configured along the line. The distance between two adjacent monitoring points should not be less than 5km and not more than 30km. Each monitoring point is equipped with a set of overhead signal detection terminals, with each set consisting of 3 signal detection terminals, installed on the three-phase conductors respectively.

[0162] Single-line configuration scheme for cable lines

[0163] For cable lines, one monitoring point is set at each end of the cable line, with a distance of no less than 5 km between the two monitoring points. Each monitoring point is equipped with one cable-type signal detection terminal, which can be installed on a pole or in a branch box. Each signal detection terminal can connect to 12 Rogowski coil current transformers, and can simultaneously measure 3 cable lines.

[0164] Overhead and cable hybrid line configuration scheme

[0165] For mixed overhead and cable lines, configurations shall be made in accordance with 4.1 and 4.2 respectively.

[0166] Overhead T-junction configuration scheme

[0167] For T-connected overhead lines, both the main line and each branch are configured as single-line lines according to the configuration scheme in 4.1.

[0168] Cable line branch box configuration scheme

[0169] Install one cable-type signal detection terminal in the cable line branch box, which can be configured with up to 12 Rogowski coil current transformers.

[0170] In at least one embodiment of the present invention, a storage medium stores computer-readable instructions that, when read, can execute the aforementioned method for precise location of distributed transmission line faults.

[0171] The embodiments of the present invention have been described above. However, the embodiments are not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make more equivalent embodiments under the guidance of the present embodiments, and all of them are within the protection scope of the present embodiments.

Claims

1. A method for precise fault location in distributed transmission lines, characterized in that, include: Step 101: Collect waveform data during power transmission line faults; The transmission line is divided into multiple line segments, and waveform data is collected at the endpoints of each data segment; The waveform data includes transient waveform data of the endpoints of all line segments during a single transmission line fault. Step 102: Construct graph structure data based on the acquired waveform data; Graph structure data includes nodes and edges connecting nodes. Nodes include trunk nodes and branch nodes. A trunk node represents a trunk line segment, and a branch node represents a branch line segment. Step 103: Input the graph structure data into the neural network model, and the neural network model outputs the result representing the location of the fault point of the transmission line; Neural network models include: The autoencoder layer is used to take a fault transient waveform as input and output an autoencoder feature. When the input node represents the fault transient waveform of the starting point of the line segment, the output is the autoencoder feature corresponding to that node. The autoencoder layer takes the waveform data of segments A, B, C, and D of the fault transient waveform data as input, and then outputs the weighted features of segment A, segment B, segment C, and segment D. Each waveform segment A, B, C, and D is itself a time series, denoted as follows: ; For each sequence, add the corresponding segment type embedding; Four enhanced input sequences Concatenate them into a long sequence according to time sequence ; Finally Input is fed into the attention layer; The formula for calculating the attention layer is as follows: right Perform a linear transformation to obtain the Query matrix Q, Key matrix K, and Value matrix V; ； in The weight matrix is ​​a learnable matrix. The dimension of the attention head; Calculate the attention weight matrix : ； The attention weight matrix is ​​multiplied by the value matrix and then linearly transformed to obtain the output autoencoded features. : ； in This is the output layer weight matrix; Tensor layers are used to tensor the autoencoded features of nodes into vector form and then input them into the graph structure learning layer. The graph structure learning layer is used to input the autoencoded features of the nodes and the graph structure data, and outputs global features. A fully connected layer is used as input for global features and then outputs a result representing the distance ratio to the fault point; Step 104: Calculate the distance between the fault point and the endpoint of the transient waveform of the first fault start based on the fault distance ratio, extract the endpoint position of the transient waveform of the first fault start, and then locate the fault point.

2. The method for precise fault location in distributed transmission lines according to claim 1, characterized in that, The fault transient waveform is collected synchronously across all line segments. When the endpoint of any line segment triggers the fault transient waveform, the endpoints of all line segments synchronously perform fault transient waveform recording.

3. The method for precise fault location in distributed transmission lines according to claim 1, characterized in that, The rules for generating edges between nodes are as follows: If the line segments represented by two trunk nodes are connected, then an edge is generated between these two trunk nodes; If the line segments represented by the trunk node and the branch node are connected, then an edge is generated between the trunk node and the branch node. If two branch nodes belong to the same branch, then there is an edge between these two branch nodes.

4. The method for precise fault location in a distributed transmission line according to claim 1, characterized in that, The distance ratio to the fault point is expressed as follows: ； in, This indicates the length of the line segment preceding the endpoint where the fault transient recording is first initiated. This indicates the length of the line segment following the endpoint where the fault transient recording was first initiated. This indicates the distance between the fault point and the endpoint that first initiated the fault transient waveform recording.

5. The method for precise fault location in a distributed transmission line according to claim 1, characterized in that, The graph structure learning layer consists of three hidden layers and one linear layer. The calculation formula for the l-th hidden layer is as follows: ； in It is the attention weight of node u to node v. Let v represent the set of nodes that have edges with node v. The first weight parameter of the l-th layer is calculated using the following formula: ； 'a' is a learnable attention vector. This represents a vector concatenation operation. This represents the second weight parameter of the l-th layer. express function; ； in and Let v and u represent the feature representations of the v-th and u-th nodes, respectively. The formula for calculating the linear layer is as follows: ; in Represents global features. This represents the feature representation of node c in the third layer. Represents the set of all nodes. This indicates vector concatenation.

6. The method for precise fault location in a distributed transmission line according to claim 1, characterized in that, Tensor quantization in a tensor layer involves expanding a matrix row by row into vectors and then concatenating them.

7. The method for precise fault location in a distributed transmission line according to claim 1, characterized in that, The neural network model also includes a second fully connected layer, which takes global features as input and outputs a classification result representing the type of fault.

8. A distributed transmission line fault precise location system, characterized in that, include: Signal detection terminals installed on the main lines and branches of transmission lines; A communication terminal used to receive data from a signal detection terminal; the communication terminal is used to send data to the system master station. The system master station is used to execute steps 102-104 of the distributed transmission line fault accurate location method as described in any one of claims 1-7.

9. A storage medium storing computer-readable instructions, characterized in that, When the computer-readable instructions are read, they can execute a method for precise location of faults in a distributed transmission line as described in any one of claims 1-7.

Images