Power transmission line fault identification method and device, computer equipment, readable storage medium and program product

Abstract

The invention relates to a power transmission line fault identification method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: acquiring a preset model parameter matrix group and a plurality of line recording information groups collected for a power transmission line at a plurality of moments; according to a time sequence corresponding to each moment, sequentially performing information analysis on each line recording information group based on the model parameter matrix group, and determining full connection layer output information of each line recording information group; for each line recording information group, determining fault probabilities corresponding to the output information of the full connection layer of the line recording information group under different fault categories; and determining a fault identification result of the power transmission line at the moment based on each fault probability. By adopting the method, the accuracy can be improved.

Description

Technical Field

[0001] This application relates to the field of power technologies, and in particular, to a method, device, computer device, computer-readable storage medium, and computer program product for identifying faults in transmission lines. Background Art

[0002] With the continuous growth of social demand for electricity, the safe operation of transmission lines has become crucial. However, transmission channels often pass through complex terrains such as forests, mountains, farmlands, etc. Foreign objects in the external environment (such as fallen trees, building construction, over-dense vegetation, etc.) may pose potential hazards to power facilities, resulting in power outages or even major accidents. Therefore, the fault identification of potential hazards in transmission channels has become a key research direction for power departments.

[0003] In traditional technologies, generally only the processing of single-line detection data by a neural network model is used to achieve the fault identification of transmission lines. However, due to the complex factors affecting the operation of transmission lines, there is a problem of poor accuracy in fault identification only through single-line detection data. Summary of the Invention

[0004] Based on this, it is necessary to provide a method, device, computer device, computer-readable storage medium, and computer program product for identifying faults in transmission lines that can improve accuracy in view of the above technical problems.

[0005] In a first aspect, this application provides a method for identifying faults in transmission lines, including:

[0006] Obtain a preset model parameter matrix group and multiple groups of line recording information collected for the transmission line at multiple moments;

[0007] According to the time sequence corresponding to each of the moments, sequentially perform information analysis on each group of line recording information based on the model parameter matrix group to determine the fully connected layer output information of each group of line recording information;

[0008] For each group of line recording information, determine the fault probabilities corresponding to the fully connected layer output information of the group of line recording information under different fault categories;

[0009] Based on the respective fault probabilities, determine the fault identification result of the transmission line at the moment.

[0010] In one embodiment, the model parameter matrix group includes a first model parameter matrix group and a second model parameter matrix group; the sequentially performing information analysis on each group of line recording information based on the model parameter matrix group according to the time sequence corresponding to each of the moments to determine the fully connected layer output information of each group of line recording information includes:

[0011] According to the chronological order corresponding to each of the said moments, successively perform information analysis on each of the line recording information groups based on the first model parameter matrix group to determine the hidden gate state information of each of the line recording information groups;

[0012] According to the chronological order corresponding to each of the said moments, successively perform information analysis on the hidden gate state information of each of the line recording information groups based on the second model parameter matrix group to determine the fully connected layer output information of each of the line recording information groups.

[0013] In one embodiment, the step of successively performing information analysis on each of the line recording information groups based on the first model parameter matrix group according to the chronological order corresponding to each of the said moments to determine the hidden gate state information of each of the line recording information groups includes:

[0014] Obtain the hidden state information and the memory cell state information;

[0015] Based on the chronological order corresponding to each of the said moments, determine the first moment; the first moment is the earliest moment among all the said moments that has not undergone information analysis.

[0016] Perform information analysis on the line recording information group at the first moment based on the hidden state information, the memory cell state information, and the first model parameter matrix group to obtain the hidden gate state information of the line recording information group and the updated memory cell state information;

[0017] Take the hidden gate state information as the updated memory cell state information, and return to the step of determining the first moment based on the chronological order corresponding to each of the said moments until the hidden gate state information of each of the line recording information groups is determined.

[0018] In one embodiment, the second model parameter matrix group includes a weight matrix and a bias vector; the step of successively performing information analysis on the hidden gate state information of each of the line recording information groups based on the second model parameter matrix group according to the chronological order corresponding to each of the said moments to determine the fully connected layer output information of each of the line recording information groups includes:

[0019] According to the chronological order corresponding to each of the said moments, successively perform information operations on the hidden gate state information of each of the line recording information groups based on the weight matrix to obtain the weighted information of each of the line recording information groups;

[0020] Based on the weighted information and the bias vector, determine the fully connected layer output information of each of the line recording information groups.

[0021] In one embodiment, determining the fault probabilities respectively corresponding to the output information of the fully connected layer of the line recording information group under different fault categories includes:

[0022] Obtaining a preset initial fault probability function and category data respectively corresponding to different fault categories;

[0023] For each of the fault categories, based on the category data, the initial fault probability function, and the output information of the fully connected layer, determining the fault probability corresponding to the output information of the fully connected layer of the line recording information group under the fault category.

[0024] In one embodiment, based on the respective fault probabilities, determining the fault identification result of the transmission line at the moment includes:

[0025] Selecting the one with the maximum fault probability from the respective fault probabilities as the target probability;

[0026] Taking the fault category corresponding to the target probability as the fault identification result of the transmission line at the moment.

[0027] In a second aspect, the present application further provides a transmission line fault identification device, including:

[0028] An information acquisition module, configured to acquire a preset model parameter matrix group and a plurality of line recording information groups collected for a transmission line at multiple moments;

[0029] An information analysis module, configured to sequentially perform information analysis on each of the line recording information groups based on the model parameter matrix group according to the time sequence corresponding to each moment, and determine the output information of the fully connected layer of each line recording information group;

[0030] A fault probability determination module, configured to determine the fault probabilities respectively corresponding to the output information of the fully connected layer of each line recording information group under different fault categories;

[0031] An identification result determination module, configured to determine the fault identification result of the transmission line at the moment based on the respective fault probabilities.

[0032] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the steps of the method described above are implemented.

[0033] In a fourth aspect, the present application further provides a computer-readable storage medium. On the computer-readable storage medium, a computer program is stored, and when the computer program is executed by a processor, the steps of the method described above are implemented.

[0034] In a fifth aspect, the present application also provides a computer program product. The computer program product includes a computer program which, when executed by a processor, implements the steps of the method described above.

[0035] For the above-mentioned transmission line fault identification method, device, computer device, computer-readable storage medium, and computer program product, by obtaining a preset model parameter matrix group and multiple groups of line recording information collected for the transmission line at multiple moments, the line recording situation of the transmission line can be determined. In the time sequence corresponding to each moment, the information analysis of each group of line recording information is sequentially performed based on the model parameter matrix group to determine the output information of the fully connected layer of each group of line recording information. For each group of line recording information, the fault probabilities corresponding to different fault categories of the output information of the fully connected layer of the line recording information group can be determined, and the possible attribution situations of the transmission line for different fault categories at each moment can be determined. Based on the respective fault probabilities, the fault identification result of the transmission line at the moment can be determined. By using the above method, the analysis of the transmission line can be performed for multiple moments and multiple types of information, and then the fault of the transmission line can be identified, which can improve the identification accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments of the present application or related technologies. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0037] Figure 1 It is an application environment diagram of the transmission line fault identification method in an embodiment;

[0038] Figure 2 It is a schematic flowchart of the transmission line fault identification method in an embodiment;

[0039] Figure 3 It is a schematic example diagram of the transmission line fault identification steps in an embodiment;

[0040] Figure 4 It is a schematic example diagram of the transmission line fault identification steps in another embodiment;

[0041] Figure 5 It is a schematic example diagram of the transmission line fault identification result in an embodiment;

[0042] Figure 6 It is a schematic example diagram of the transmission line fault identification result in another embodiment;

[0043] Figure 7 It is a schematic flowchart of a transmission line fault identification method in another embodiment;

[0044] Figure 8 It is a structural block diagram of a transmission line fault identification device in an embodiment;

[0045] Figure 9 It is an internal structure diagram of a computer device in an embodiment. Detailed implementation manners

[0046] In order to make the objectives, technical solutions and advantages of the present application more clear and understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0047] The transmission line fault identification method provided by the embodiments of the present application can be applied to an application environment as shown in Figure 1 . Among them, the terminal 102 communicates with the server 104 through a network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or can be placed in the cloud or other network servers. Among them, the terminal 102 can be, but is not limited to, various personal computers, laptop computers, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart vehicle-mounted devices, projection devices, etc. The portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. The server 104 can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services. Specifically, the server 104 can obtain a pre-set model parameter matrix group and multiple groups of line recording information collected for the transmission line at multiple moments from the terminal 102; in the time sequence corresponding to each moment, sequentially perform information analysis on each group of line recording information based on the model parameter matrix group to determine the fully connected layer output information of each group of line recording information; for each group of line recording information, determine the fault probabilities respectively corresponding to the fully connected layer output information of the group of line recording information under different fault categories; based on the respective fault probabilities, determine the fault identification result of the transmission line at the moment.

[0048] In an exemplary embodiment, as shown in Figure 2 a transmission line fault identification method is provided, and this method is applied to Figure 1Taking the server 104 in it as an example for illustration, including:

[0049] Step S202, obtaining a preset model parameter matrix group and multiple groups of line recording information collected for the transmission line at multiple moments.

[0050] Among them, the model parameters refer to the model parameters in the recurrent neural network model. For example, the model parameters may include the number of input features, the dimension of the hidden layer, the weight matrix or the bias vector, etc. The model parameter matrix group refers to the model parameter matrix group in the recurrent neural network model. For example, in this application, the model parameter matrix group may include the weight matrix and the bias vector, etc. The model structure of the recurrent neural network (RNN) is mainly based on its ability to process sequential data, and its structural feature lies in the cyclic flow of information in the time dimension. The basic structure of the RNN includes an input layer, a hidden layer, and an output layer. At each time step t, the RNN receives the current input x(t) and the hidden state h(t - 1) of the previous time step. Then, through the calculation of the hidden layer, the RNN generates the output o(t) of the current time step and the new hidden state h(t).

[0051] The transmission line refers to the power transmission system connecting power plants, substations and users, mainly used for long-distance and large-capacity power transmission. It is a crucial part of the power system, carrying the tasks of safe and efficient transmission of electric energy. The transmission line is usually composed of metal wires and is suspended in the air through high towers or brackets to reduce power loss and improve transmission efficiency.

[0052] The line recording information refers to the information obtained by continuously or periodically sampling and recording the electrical quantities of the transmission line during the operation of the transmission line. For example, the line recording information may include three-phase voltages, three-phase currents, zero-sequence voltages or zero-sequence current sequences, etc., which are respectively expressed as: three-phase voltages: (Ua(t), Ub(t), Uc(t)); zero-sequence voltage (U0(t)); three-phase currents: (Ia(t), Ib(t), Ic(t)); zero-sequence current: (I0(t)); where n represents the time series length of the line recording information.

[0053] Specifically, in order to identify faults in the transmission line, multiple groups of line recording information collected for the transmission line at multiple moments can be obtained first, and then the change situation of the electrical quantities in the transmission line can be determined, and a preset model parameter matrix group can be obtained, so as to perform information analysis on the line recording information group through the model parameter matrix group subsequently. Exemplarily, the operation of obtaining a preset model parameter matrix group and multiple groups of line recording information collected for the transmission line at multiple moments can be an active acquisition or a passive reception.

[0054] Exemplarily, although there are various types of line recording information, when determining the line recording information group, generally, some line recording information is selected from each line recording information to form the line recording information group. For example, in this embodiment, 4 channels can be arbitrarily selected from the line recording information. Generally, voltage (three-phase voltage + 1 zero-sequence voltage) or current (three-phase current + 1 zero-sequence current) is selected, or it can also be selected according to the actual situation, which is not limited in this embodiment.

[0055] Step S204: According to the time sequence corresponding to each moment, sequentially perform information analysis on each line recording information group based on the model parameter matrix group to determine the output information of the fully connected layer of each line recording information group.

[0056] Among them, the output information of the fully connected layer refers to the output information corresponding to the fully connected layer in the recurrent neural network model. Essentially, it is the result after linear transformation (the product of the input features and the weight matrix, plus the bias term) and possible non-linear activation function processing. These output information can be regarded as the feature representation or classification prediction result of the input data.

[0057] Specifically, after obtaining the model parameter matrix group and the line recording information group, since the corresponding moments of each line recording information group are different, therefore, in order to ensure the orderliness of information processing, it is also necessary to sequentially perform information analysis on each line recording information group based on the model parameter matrix group according to the time sequence corresponding to each moment to determine the output information of the fully connected layer of each line recording information group. Exemplarily, the process of performing information analysis on each line recording information group based on the model parameter matrix group specifically refers to performing information operation on the information of the line recording information group based on the model parameter matrix group to obtain the input gate state information, output gate state information, memory cell information, and hidden gate state information corresponding to the line recording information group, etc. Then, the hidden gate state information is further subjected to information operation through the model parameter matrix group to obtain the output information of the fully connected layer.

[0058] Step S206: For each line recording information group, determine the fault probabilities corresponding to the output information of the fully connected layer of the line recording information group under different fault categories.

[0059] Among them, the fault category refers to various fault types that may occur during the operation of the transmission line, and these faults may have a serious impact on the stable operation of the power system. The fault probability is the probability that a fault may exist in the transmission line at the moment when the line recording information group is located. For example, the fault category may include B-phase short circuit, AC-phase short circuit, A-phase short circuit, C-phase short circuit, B-phase lightning strike, etc. Among them, A-phase, B-phase, and C-phase are respectively one of the three phases in the three-phase AC power system.

[0060] Specifically, the output information of the fully connected layer for each line wave recording information group is determined, that is, the line condition characteristics of the transmission line at each moment are determined. Therefore, for each line wave recording information group at each moment, the server can determine different fault probabilities corresponding to different fault categories of the transmission line according to the output information of the fully connected layer of this line wave recording information group.

[0061] In some specific embodiments, the server can obtain a preset initial fault probability function and category data corresponding to different fault categories respectively. For each fault category, based on the category data, the initial fault probability function, and the output information of the fully connected layer, determine the fault probability corresponding to the output information of the fully connected layer of the line wave recording information group under the fault category. In some other specific embodiments, a fault probability determination model can also be established, and based on this fault probability determination model, determine the fault probabilities corresponding to the output information of the fully connected layer of the line wave recording information group under different fault categories respectively.

[0062] Step S208: Based on each fault probability, determine the fault identification result of the transmission line at the moment.

[0063] Among them, the fault identification result refers to the result after fault identification for the transmission line, which corresponds one-to-one with the fault category.

[0064] Specifically, after determining the fault probabilities corresponding to each line wave recording information group under different fault categories respectively, the fault identification result of the transmission line at this moment can be determined based on each fault probability. Exemplarily, the server can determine a target probability from each fault probability, and use the fault category corresponding to this target probability as the fault identification result of the transmission line at this moment.

[0065] For the above transmission line fault identification method, by obtaining the preset model parameter matrix group and multiple line wave recording information groups collected for the transmission line at multiple moments, the line wave recording situation of the transmission line can be determined. According to the time sequence corresponding to each moment, sequentially perform information analysis on each line wave recording information group based on the model parameter matrix group to determine the output information of the fully connected layer of each line wave recording information group. For each line wave recording information group, determine the fault probabilities corresponding to the output information of the fully connected layer of the line wave recording information group under different fault categories respectively, and the possible attribution situations of the transmission line for different fault categories at each moment can be determined. Based on each fault probability, determine the fault identification result of the transmission line at the moment. By using the above method, multi-moment and multi-type information analysis can be performed on the transmission line, and then the fault of the transmission line can be identified, which can improve the identification accuracy.

[0066] In an exemplary embodiment, the model parameter matrix group includes a first model parameter matrix group and a second model parameter matrix group. In the context of this embodiment, in the chronological order corresponding to each moment, information analysis is sequentially performed on each line recording information group based on the model parameter matrix group to determine the output information of the fully connected layer of each line recording information group, including: in the chronological order corresponding to each moment, information analysis is sequentially performed on each line recording information group based on the first model parameter matrix group to determine the hidden gate state information of each line recording information group; in the chronological order corresponding to each moment, information analysis is sequentially performed on the hidden gate state information of each line recording information group based on the second model parameter matrix group to determine the output information of the fully connected layer of each line recording information group.

[0067] Among them, both the first model parameter matrix group and the second model parameter matrix group contain a weight matrix and a bias vector. Only the weight matrix and bias vector in the first model parameter matrix group are used for calculating the hidden gate state information, while the weight matrix and bias vector in the second model parameter matrix group are used for calculating the output information of the fully connected layer. The hidden gate state information is the state information corresponding to the hidden gate in the recurrent neural network model. The weight matrix represents the connection strength between neurons. The bias vector (usually simply referred to as bias) is used in a neural network to adjust the output of a neuron. In an RNN, each neuron (including neurons from the input layer to the hidden layer, from the hidden layer to the output layer, and within the hidden layer) may have a bias vector associated with it.

[0068] Specifically, the core of an RNN lies in the state of its hidden layer, which can capture the temporal dependence in sequential data. At each moment, the RNN receives an input vector (line recording information group) and the hidden gate state information from the previous moment as inputs. Through a non-linear function (such as tanh or ReLU), combining the current input and the previous hidden state, the hidden gate state information at the current moment is calculated. After obtaining the hidden gate state information, the RNN usually calculates the output through a linear transformation and an activation function (such as softmax). This linear transformation can be regarded as an operation of a fully connected layer, which maps the hidden gate state information to the output space. Therefore, during the operation of the recurrent neural network model, the server can first perform information analysis on each line recording information group based on the first model parameter matrix group in the chronological order corresponding to each moment to determine the hidden gate state information of each line recording information group, and then perform information analysis on the hidden gate state information of each line recording information group based on the second model parameter matrix group in the chronological order corresponding to each moment to determine the output information of the fully connected layer of each line recording information group.

[0069] In this embodiment, the server sequentially determines the hidden gate status information and the fully connected layer output information of each line wave recording information group according to the time sequence, which can ensure the accuracy of time capture and feature prediction in the status information of the transmission line.

[0070] In an exemplary embodiment, according to the time sequence corresponding to each moment, information analysis is sequentially performed on each line wave recording information group based on the first model parameter matrix group to determine the hidden gate status information of each line wave recording information group, including: obtaining the hidden state information and the memory cell state information; determining the first moment based on the time sequence corresponding to each moment; the first moment is the earliest moment among all moments that has not undergone information analysis; based on the hidden state information, the memory cell state information, and the first model parameter matrix group, information analysis is performed on the line wave recording information group at the first moment to obtain the hidden gate status information of the line wave recording information group and the updated memory cell state information; taking the hidden gate status information as the updated memory cell state information, and returning to the step of determining the first moment based on the time sequence corresponding to each moment until the hidden gate status information of each line wave recording information group is determined.

[0071] Among them, the first moment is the earliest moment among all moments that has not undergone information analysis. In an RNN, the hidden state is an internal state used to store and transmit information when processing the input sequence. It is updated at each time step to capture the information and context relationship in the input sequence. The hidden gate status information can be regarded as the encoding of the input sequence at the current time step, containing the information of all previous inputs, thereby helping the model to consider the context when processing the current input. This enables the RNN to process and understand the complex structures and patterns in the sequence data. The memory cell state information is more related to the memory mechanism or the state representation in certain specific algorithms.

[0072] Specifically, the server can calculate the information corresponding to the states of the input gate, forget gate, candidate memory cell, output gate, hidden gate, etc. in sequence according to the preset first model parameter matrix group at time t, and obtain the hidden gate state information of each line recording information group. Optionally, in the above process, it is necessary to first obtain the hidden state information and memory cell state information, determine the first moment based on the time sequence corresponding to each moment, and then perform information analysis on the line recording information group at the first moment based on the hidden state information, memory cell state information, and the first model parameter matrix group to obtain the hidden gate state information of the line recording information group at the first moment and the updated memory cell state information. Take the hidden gate state information of the line recording information group at the first moment as the updated memory cell state information, and return to the step of determining the first moment based on the time sequence corresponding to each moment until the hidden gate state information of each line recording information group is determined. It can be understood that the first moment is successively the earliest moment among each moment that has not undergone information analysis. In a specific embodiment, the value of time t is t = 1, 2, 3,... n. Where n represents the actual time length of the line recording information. Suppose initially, the hidden state information h0 = 0, the memory cell state information c0 = 0, t = 1, and xt = X_Input(t = 1).

[0073] The calculation of the state information of each gate is as follows,

[0074] The state information of the input gate is,

[0075] The state information of the forget gate is,

[0076] The candidate memory cell information is,

[0077] The state information of the output gate is,

[0078] The memory cell state information is,

[0079] The hidden gate state information is,

[0080] The final output is the hidden gate state information of the line recording information group X_Input.

[0081] Among them,

[0082] Wi is a 200×4 single-precision floating-point number matrix:

[0083]

[0084] Wf is a 200×4 single-precision floating-point number matrix:

[0085]

[0086] Wg is a 200×4 single-precision floating-point number matrix:

[0087]

[0088] Wo is a 200×4 single-precision floating-point number matrix:

[0089]

[0090] Ri is a 200×200 single-precision floating-point number matrix:

[0091]

[0092] Rf is a 200×200 single-precision floating-point number matrix:

[0093]

[0094] Rg is a 200×200 single-precision floating-point number matrix:

[0095]

[0096] Ro is a 200×200 single-precision floating-point number matrix:

[0097]

[0098] Bi is a 200×1 single-precision floating-point number column vector:

[0099]

[0100] Bf is a 200×1 single-precision floating-point number column vector:

[0101]

[0102] Bg is a 200×1 single-precision floating-point number column vector:

[0103]

[0104] Bo is a 200×1 single-precision floating-point number column vector:

[0105]

[0106] In this embodiment, through multiple determinations at the first moment, it is ensured that the hidden gate state information and memory element information are calculated for each set of line recording wave information at each moment, which can ensure the accuracy of information operation. Moreover, since the hidden gate can capture the information and context relationship in the input sequence, therefore, by adopting the method of this embodiment, the accuracy of transmission line fault prediction can also be improved.

[0107] In an exemplary embodiment, the second model parameter matrix group includes a weight matrix and a bias vector; in the time order corresponding to each moment, based on the second model parameter matrix group in sequence, information analysis is performed on the hidden gate state information of each set of line recording wave information to determine the fully connected layer output information of each set of line recording wave information, including: in the time order corresponding to each moment, based on the weight matrix in sequence, information operation is performed on the hidden gate state information of each set of line recording wave information to obtain the weighted information of each set of line recording wave information; based on the weighted information and the bias vector, the fully connected layer output information of each set of line recording wave information is determined.

[0108] Among them, the weight matrix: maps the input vector or the hidden state of the previous moment to a new space through a linear transformation (i.e., matrix multiplication), and this new space usually corresponds to the output of the activation function or gating unit in the hidden layer. The bias vector: adds a fixed offset to the output of the activation function or gating unit, thereby increasing the flexibility of the model and allowing the learned representation to be offset from the origin.

[0109] For example, Wfull (weight matrix) is a 6×200 single-precision floating-point number matrix:

[0110]

[0111] Bfull is a 6×1 single-precision floating-point number column vector (bias vector):

[0112]

[0113] Specifically, as shown in the following formula, the server can, in the time order corresponding to each moment, based on the weight matrix Wfull in sequence, perform information operation on the hidden gate state information of each set of line recording wave information to obtain the weighted information of each set of line recording wave information, and then based on the weighted information and the Bfull bias vector, determine the fully connected layer output information Full_Outt of each set of line recording wave information, which can improve the accuracy of fault identification.

[0114]

[0115] In an exemplary embodiment, determining the fault probabilities respectively corresponding to the output information of the fully connected layer of the line recording information group under different fault categories includes: obtaining a preset initial fault probability function and category data respectively corresponding to different fault categories; for each fault category, determining the fault probability corresponding to the output information of the fully connected layer of the line recording information group under the fault category based on the category data, the initial fault probability function, and the output information of the fully connected layer.

[0116] Among them, the category data refers to the category data corresponding to the fault category. It can be understood that the category data corresponds one-to-one with the fault category. The following formula is the initial fault probability function. A short circuit in phase B, a short circuit between phases A and C, a short circuit in phase A, a short circuit in phase C, and a lightning strike in phase B respectively correspond to the numbers 1, 2, 3, 4, 5, and 6.

[0117]

[0118] Specifically, after obtaining the preset initial fault probability function and the category data respectively corresponding to different fault categories, the server can, for each fault category, determine the fault probability corresponding to the output information of the fully connected layer of the line recording information group under the fault category based on the category data, the initial fault probability function, and the output information of the fully connected layer. As shown in the above formula, for each fault category, inputting the category data and the output information of the fully connected layer into the initial fault probability function can determine the fault probability corresponding to the output information of the fully connected layer of the line recording information group under the fault category.

[0119] In this embodiment, through the above initial fault probability function, the fault probabilities of the line recording information group corresponding to different fault categories at each moment can be determined specifically, which can improve the accuracy of subsequent fault identification.

[0120] In an exemplary embodiment, determining the fault identification result of the transmission line at a moment based on the respective fault probabilities includes: selecting the one with the largest fault probability from the respective fault probabilities as the target probability; taking the fault category corresponding to the target probability as the fault identification result of the transmission line at that moment.

[0121] Specifically, for each line recording information group, select the one with the largest fault probability from the respective fault probabilities as the target probability, denoted as 1, and the others as 0, as shown in the following formula:

[0122]

[0123] Figure 3 This is a process of calculating Locate_x(t). The one in the circle is an example of "assuming that X2 is marked as 1", that is, for X2 corresponding to the moment t = 2, among which the fault probability of fault category 2 is the largest, being 0.9997. Therefore, in Figure 4Among them, the fault category 2 of X2 is recorded as 1. It can be understood that Figure 4 the horizontal rows are moments, and the vertical columns correspond to fault categories. For example Figure 5 as shown, since the probability at the fault category 1 at each moment is not the maximum, they are all "normal", representing normal. In an example of an actual scenario, as follows Figure 6 , the circle represents t = 33, normal. The rectangular box represents t = 38, phase A short circuit.

[0124] In a specific embodiment, as shown in the following formula, according to the predefined transmission line fault type classification array Categorical_X, the classification value Locate_x(i) of the line recording information group can be classified to obtain the classification Ca_X_Output of the line recording information group X_Input.

[0125]

[0126] Among them, Categorical_X is represented as follows:

[0127]

[0128] In this embodiment, the fault probability with the largest value is selected from each fault probability as the target probability, and the fault category corresponding to the target probability is used as the fault identification result of the transmission line at this moment, which can improve the accuracy of determining the fault identification result.

[0129] In a specific embodiment, as Figure 7 shown, a transmission line fault identification method is also provided, including:

[0130] Step S701, obtain a preset model parameter matrix group and multiple line recording information groups collected for the transmission line at multiple moments;

[0131] Among them, the model parameter matrix group includes a first model parameter matrix group and a second model parameter matrix group; the second model parameter matrix group includes a weight matrix and a bias vector;

[0132] Step S702, obtain hidden state information and memory cell state information, and determine the first moment based on the time sequence corresponding to each moment;

[0133] Among them, the first moment is the earliest moment among all moments that has not undergone information analysis;

[0134] Step S703, perform information analysis on the line recording information group at the first moment based on the hidden state information, memory cell state information, and the first model parameter matrix group to obtain the hidden gate state information of the line recording information group and the updated memory cell state information;

[0135] Step S704: Take the hidden door status information as the updated memory cell status information, and return to the step of determining the first moment based on the time sequence corresponding to each moment until the hidden door status information of each line recording information group is determined.

[0136] Step S705: According to the time sequence corresponding to each moment, perform information operations on the hidden door status information of each line recording information group based on the weight matrix in turn to obtain the weighted information of each line recording information group.

[0137] Step S706: Based on the weighted information and the bias vector, determine the fully connected layer output information of each line recording information group.

[0138] Step S707: For each line recording information group, obtain the preset initial fault probability function and the category data corresponding to different fault categories respectively.

[0139] Step S708: For each fault category, based on the category data, the initial fault probability function, and the fully connected layer output information, determine the fault probability corresponding to the fully connected layer output information of the line recording information group under the fault category.

[0140] Step S709: Select the one with the largest fault probability from each fault probability as the target probability, and take the fault category corresponding to the target probability as the fault recognition result of the transmission line at the moment.

[0141] It should be understood that although the steps in the flowcharts involved in the above embodiments are shown in sequence according to the arrows, these steps do not necessarily need to be executed in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages. These steps or stages do not necessarily need to be executed at the same moment, but can be executed at different moments. The execution order of these steps or stages is not necessarily sequential either, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0142] Based on the same inventive concept, the embodiments of the present application also provide a transmission line fault recognition device for implementing the above-mentioned transmission line fault recognition method. The implementation solutions provided by this device to solve problems are similar to those described in the above method. Therefore, the specific limitations in one or more embodiments of the following transmission line fault recognition device can refer to the limitations on the transmission line fault recognition method in the above text, and will not be repeated here.

[0143] In an exemplary embodiment, asFigure 8 As shown in the figure, a transmission line fault identification device 800 is provided, including: an information acquisition module 802, an information analysis module 804, a fault probability determination module 806, and an identification result determination module 808, where:

[0144] The information acquisition module 802 is configured to acquire a pre-set model parameter matrix group and a plurality of line wave recording information groups collected for the transmission line at multiple moments;

[0145] The information analysis module 804 is configured to sequentially perform information analysis on each line wave recording information group based on the model parameter matrix group in the time sequence corresponding to each moment, and determine the fully connected layer output information of each line wave recording information group;

[0146] The fault probability determination module 806 is configured to determine, for each line wave recording information group, the fault probabilities respectively corresponding to the fully connected layer output information of the line wave recording information group under different fault categories;

[0147] The identification result determination module 808 is configured to determine the fault identification result of the transmission line at the moment based on the respective fault probabilities.

[0148] In an exemplary embodiment, the model parameter matrix group includes a first model parameter matrix group and a second model parameter matrix group. In the case of this embodiment, the information analysis module 804 includes:

[0149] A hidden gate information determination unit, configured to sequentially perform information analysis on each line wave recording information group based on the first model parameter matrix group in the time sequence corresponding to each moment, and determine the hidden gate state information of each line wave recording information group;

[0150] A fully connected layer information determination unit, configured to sequentially perform information analysis on the hidden gate state information of each line wave recording information group based on the second model parameter matrix group in the time sequence corresponding to each moment, and determine the fully connected layer output information of each line wave recording information group.

[0151] In an exemplary embodiment, the hidden gate information determination unit is specifically configured to:

[0152] Acquire hidden state information and memory cell state information;

[0153] Based on the time sequence corresponding to each moment, determine the first moment; the first moment is the earliest moment among all moments that has not undergone information analysis;

[0154] Perform information analysis on the line wave recording information group at the first moment based on the hidden state information, the memory cell state information, and the first model parameter matrix group, to obtain the hidden gate state information of the line wave recording information group and the updated memory cell state information;

[0155] Take the hidden door status information as the updated memory cell status information, and return the step of determining the first moment based on the time sequence corresponding to each moment until the hidden door status information of each line recording information group is determined.

[0156] In an exemplary embodiment, the second model parameter matrix group includes a weight matrix and a bias vector. In the case of this embodiment, the fully connected layer information determination unit is specifically configured to:

[0157] According to the time sequence corresponding to each moment, sequentially perform information operations on the hidden door status information of each line recording information group based on the weight matrix to obtain the weighted information of each line recording information group;

[0158] Based on the weighted information and the bias vector, determine the fully connected layer output information of each line recording information group.

[0159] In an exemplary embodiment, the fault probability determination module 806 is specifically configured to:

[0160] Obtain a preset initial fault probability function and category data corresponding to different fault categories respectively;

[0161] For each fault category, based on the category data, the initial fault probability function, and the fully connected layer output information, determine the fault probability corresponding to the fully connected layer output information of the line recording information group under the fault category.

[0162] In an exemplary embodiment, the recognition result determination module 808 is specifically configured to:

[0163] Select the one with the maximum fault probability from the fault probabilities as the target probability;

[0164] Take the fault category corresponding to the target probability as the fault recognition result of the transmission line at the moment.

[0165] Each module in the above transmission line fault recognition device can be implemented in whole or in part by software, hardware, and their combination. Each of the above modules can be embedded in the processor of the computer device in hardware form or be independent of it, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each of the above modules.

[0166] In an exemplary embodiment, a computer device is provided. The computer device can be a terminal, and its internal structure diagram can be as Figure 9As shown in the figure. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit, and an input device. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface, the display unit, and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a mobile cellular network, near field communication (NFC), or other technologies. When the computer program is executed by the processor, it implements a method for identifying transmission line faults. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the housing of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0167] Those skilled in the art can understand that Figure 9 the structure shown in the figure is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0168] In one embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps of the above method are implemented.

[0169] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0170] In one embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0171] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0172] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in this application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in this application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, artificial intelligence (AI) processors, etc., without limitation.

[0173] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this application.

[0174] The above-described embodiments merely represent several implementation manners of this application. The description is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of this application. It should be noted that for those of ordinary skill in the art, without departing from the concept of this application, several variations and improvements can still be made, and these all belong to the protection scope of this application. Therefore, the protection scope of this application should be subject to the appended claims.

Claims

1. A method for identifying a power transmission line fault, characterized in that: The method comprises: Obtaining a preset model parameter matrix group and a plurality of line recording information groups collected for the transmission line at a plurality of time instants; According to the time sequence corresponding to each of the moments, information analysis is performed on each of the line recording information groups based on the model parameter matrix group in turn to determine the fully connected layer output information of each of the line recording information groups; For each of the line recording information groups, determining the fault probabilities corresponding to the fully connected layer output information of the line recording information group under different fault categories; Based on each of the fault probabilities, a fault identification result of the power transmission line at the moment is determined.

2. The method according to claim 1, characterized in that The model parameter matrix group includes a first model parameter matrix group and a second model parameter matrix group; the information analysis is performed on each line recording information group based on the model parameter matrix group in turn according to the time sequence corresponding to each moment, and the fully connected layer output information of each line recording information group is determined, including: According to the time sequence corresponding to each of the moments, information analysis is performed on each of the line recording information groups based on the first model parameter matrix group in turn to determine the hidden door state information of each of the line recording information groups; According to the time sequence corresponding to each of the moments, information analysis is performed on the hidden door state information of each of the line recording information groups based on the second model parameter matrix group to determine the fully connected layer output information of each of the line recording information groups.

3. The method according to claim 2, characterized in that The step of analyzing each line recording information group based on the first model parameter matrix group in accordance with the time sequence corresponding to each of the moments to determine the hidden door state information of each line recording information group includes: Get hidden state information and memory meta-state information; Determine a first moment based on the time sequence corresponding to each of the moments; the first moment is the earliest moment among the moments for which information analysis has not been performed; Performing information analysis on the line recording information group at the first moment based on the hidden state information, the memory element state information and the first model parameter matrix group to obtain hidden door state information of the line recording information group and updated memory element state information; The hidden door status information is used as the updated memory element status information, and the step of determining the first moment based on the time sequence corresponding to each moment is returned to, until the hidden door status information of each line recording information group is determined.

4. The method according to claim 2, characterized in that: The second model parameter matrix group includes a weight matrix and a bias vector; the hidden door state information of each line recording information group is analyzed in sequence based on the second model parameter matrix group according to the time sequence corresponding to each moment, and the fully connected layer output information of each line recording information group is determined, including: According to the time sequence corresponding to each of the moments, information calculation is performed on the hidden door state information of each of the line recording information groups based on the weight matrix in turn to obtain weighted information of each of the line recording information groups; Based on the weighted information and the bias vector, the fully connected layer output information of each line recording information group is determined.

5. The method according to claim 1, characterized in that The determining of the fault probabilities corresponding to the fully connected layer output information of the line recording information group under different fault categories includes: Obtaining a preset initial function of fault probability and category data corresponding to different fault categories; For each of the fault categories, based on the category data, the initial function of the fault probability, and the fully connected layer output information, the fault probability corresponding to the fault category of the fully connected layer output information of the line recording information group is determined.

6. The method according to claim 1, characterized in that The determining, based on each of the fault probabilities, a fault identification result of the power transmission line at the moment includes: Selecting the maximum fault probability from among the fault probabilities as the target probability; The fault category corresponding to the target probability is used as the fault identification result of the transmission line at the moment.

7. A transmission line fault identification device, characterized in that: The device comprises: An information acquisition module, used to acquire a preset model parameter matrix group and a plurality of line recording information groups collected for the transmission line at a plurality of time periods; An information analysis module, used to analyze each of the line recording information groups based on the model parameter matrix group in sequence according to the time sequence corresponding to each of the moments, and determine the fully connected layer output information of each of the line recording information groups; A fault probability determination module is used to determine, for each of the line recording information groups, the fault probabilities corresponding to the fully connected layer output information of the line recording information group under different fault categories; The identification result determination module is used to determine the fault identification result of the transmission line at the moment based on each of the fault probabilities.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images