Self-checking method of power grid traveling wave fault distance measurement system based on historical data mining

Abstract

The invention provides a power grid traveling wave fault distance measurement system self-checking method based on historical data mining, and relates to the technical field of power grid fault localization, and the method comprises the steps: obtaining historical fault data of a traveling wave fault distance measurement system and historical device related parameters of each transformer substation node; performing supervised learning by using historical fault data and historical device related parameters, and training to obtain a system parameter prediction model; obtaining current device related parameters of the traveling wave fault distance measurement system to be self-detected and the number of main line nodes; generating test data according to the current device related parameters, the main line node number and the historical fault data; inputting each group of test data into the system parameter prediction model to obtain an optimal system parameter; and adjusting system parameters of the power grid traveling wave fault distance measurement system according to the optimal system parameters to realize self-inspection of the power grid traveling wave fault distance measurement system. According to the scheme, manpower and equipment construction cost can be saved, and meanwhile, the accuracy and reliability of the traveling wave fault distance measurement system are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid fault location, and in particular to a power grid traveling wave fault location system self-checking method based on historical data mining. Background Art

[0002] Transmission lines are an important part of the power grid, connecting the power supply side and the power demand side, and are also the main channel for national power transmission. The operation and maintenance of transmission lines is directly related to the safe and stable operation of the power grid, and determines whether electric energy can be delivered economically and reliably. As the amount of data on the lines continues to increase rapidly, the burden on operation and maintenance personnel is becoming heavier. How to use modern information technology to effectively improve the quality and efficiency of transmission line operation and maintenance has become an important issue that needs to be solved urgently.

[0003] In recent years, with the improvement of power grid operation and management level, some power grid companies have gradually begun to deploy centralized power grid traveling wave fault location systems to locate faults in transmission line fault locations. This system has high reliability and accuracy and can efficiently calculate the fault location of the transmission line. Therefore, the operating status of the positioning system will directly affect the reliability of the fault location accuracy of the entire transmission line. In order to ensure the reliability of the operation of the power grid traveling wave fault location system, it is necessary to better understand the operating status of the fault location system on the transmission line.

[0004] At present, when testing the reliability of the traveling wave fault ranging system, a traveling wave fault ranging device tester is usually set up at each substation on both sides of the transmission line, and the GPS satellite time synchronization device is used to trigger and output the pre-recorded or standard traveling wave waveform to the traveling wave fault ranging device, so as to realize the verification of the traveling wave fault ranging device and ensure the accuracy of the fault location result. However, this method not only requires additional testing equipment, but also requires more manpower and equipment construction costs; in addition, when this method uses the standard traveling wave waveform to calibrate the traveling wave fault ranging system, it only analyzes some obvious parameters, which limits the accuracy of the verification of the traveling wave fault ranging system, and thus it is difficult to ensure the accuracy and reliability of the detection results of the traveling wave fault ranging system. Summary of the invention

[0005] In view of this, in order to address the above shortcomings, it is necessary to propose a self-check method for the power grid traveling wave fault location system based on historical data mining to save manpower and equipment construction costs, while improving the accuracy and reliability of the traveling wave fault location system.

[0006] In a first aspect, the present invention provides a self-checking method for a power grid traveling wave fault location system based on historical data mining, comprising:

[0007] Acquire the historical fault data of the transmission lines of the traveling wave fault location system in each area in at least two states and the historical device-related parameters of each substation node at the time of the historical fault; wherein the historical fault data includes: voltage waveform data and current waveform data;

[0008] The historical fault data and historical device-related parameters are used for supervised learning to train a system parameter prediction model; wherein the system parameter prediction model is used to predict the optimal system parameters of the power grid traveling wave fault location system; the system parameters include: initial wave head time, distance between each trunk node in the transmission line and traveling wave propagation speed; wherein the initial wave head time is used to characterize the time when the traveling wave arrives at each substation node;

[0009] Obtain current device-related parameters and trunk line node numbers of each substation of the traveling wave fault location system to be self-checked;

[0010] Generate at least one set of test data according to the current device related parameters, the number of trunk line nodes and the historical fault data; wherein each set of test data corresponds to a fault;

[0011] Input each group of test data into the system parameter prediction model to obtain the optimal system parameters of the current power grid traveling wave fault location system;

[0012] The system parameters of the power grid traveling wave fault location measurement system are adjusted according to the optimal system parameters to achieve self-checking of the power grid traveling wave fault location measurement system.

[0013] Preferably, the historical device related parameters include: line type, line length, fault waveform file, and timing, startup settings and line parameters of the local device.

[0014] Preferably, the neural network model in the supervised learning includes a classification network and a regression network connected in series therewith;

[0015] The classification network is used to classify the historical fault data according to factors causing fault location errors;

[0016] The regression network is used to determine the optimal values ​​of system parameters under the known fault location errors under the current system parameters.

[0017] Preferably,

[0018] The classification network includes a data input layer, eight fully connected layers and a classification output layer; wherein the data input layer uses the historical fault data and historical device related parameters as input data, and uses the number of trunk line nodes as the number of input layer nodes;

[0019] The regression network includes a data input layer, eight fully connected layers and a regression output layer; wherein the data input layer of the regression network uses the data output by the classification output layer of the classification network as input data.

[0020] Preferably, the factors causing the fault location error include: an error caused by a time synchronization quantity in the system parameters, an error caused by a line parameter in the system parameters, and an error caused by a traveling wave velocity in the system parameters.

[0021] Preferably, the adjusting of the system parameters of the power grid traveling wave fault location system according to the optimal system parameters includes:

[0022] If the deviation between the optimal system parameters and the current system parameters of the traveling wave fault location system is greater than a preset deviation threshold, an alarm prompt is issued so that the staff can manually confirm the self-test result.

[0023] Preferably, after each self-check is completed, the system parameter prediction model is further updated using data obtained after the self-check.

[0024] In a second aspect, the present invention provides a self-checking device for a power grid traveling wave fault location system based on historical data mining, comprising: a historical data acquisition module, a model training module, a current data acquisition module, a test data generation module, a system parameter prediction module, and a system parameter adjustment module;

[0025] The historical data acquisition module is configured to acquire the historical fault data of the transmission line of the traveling wave fault location system in each area in at least two states and the historical device-related parameters of each substation node at the time of the historical fault; wherein the historical fault data includes: voltage waveform data and current waveform data;

[0026] The model training module is configured to use the historical fault data and historical device related parameters for supervised learning to train a system parameter prediction model; wherein the system parameter prediction model is used to predict the optimal system parameters of the power grid traveling wave fault location system; the system parameters include: initial wave head time, the distance between each trunk node in the transmission line and the traveling wave propagation speed; wherein the initial wave head time is used to characterize the time when the traveling wave arrives at each substation node;

[0027] The current data acquisition module is configured to obtain current device-related parameters and the number of trunk line nodes of each substation of the traveling wave fault location measurement system to be self-checked;

[0028] The test data generation module is configured to generate at least one set of test data according to the current device related parameters, the number of trunk line nodes and the historical fault data; wherein each set of test data corresponds to a fault;

[0029] The system parameter prediction module is configured to input each set of test data into the system parameter prediction model to obtain the optimal system parameters of the current power grid traveling wave fault location system;

[0030] The system parameter adjustment module is configured to adjust the system parameters of the power grid traveling wave fault location system according to the optimal system parameters to achieve self-checking of the power grid traveling wave fault location system.

[0031] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to execute any of the methods described in the first aspect.

[0032] In a fourth aspect, the present invention provides a computing device, including a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, it implements any method described in the first aspect.

[0033] It can be seen from the above technical scheme that when the traveling wave fault ranging system of the power grid is self-checked, the scheme first obtains the historical fault data of the transmission line of the traveling wave fault ranging system in each area in at least two states and the historical device-related parameters of each substation node at the time of the historical fault, and then uses the historical fault data and the historical fault-related parameters for supervised learning to train and obtain a system parameter prediction model for predicting the optimal system parameters of the traveling wave fault ranging system of the power grid. Further, the current device-related parameters and the number of trunk line nodes of each substation of the traveling wave fault ranging system to be self-checked are obtained, and several groups of test data are generated based on this and the historical fault data, and then each group of test data is input into the system parameter prediction model to obtain the optimal system parameters of the current traveling wave fault ranging system of the power grid, and then the system parameters of the traveling wave fault ranging system of the power grid are adjusted according to the optimal system parameters. It can be seen that the scheme can realize the self-check of the traveling wave fault ranging system, and the implementation logic can be directly configured in the traveling wave fault ranging system, and there is no need to develop additional test equipment for the traveling wave fault ranging system, which not only requires excessive human participation, but also can greatly save the construction cost of the equipment. In addition, this scheme can mine deeper features from historical data through machine learning, so that the optimal system parameters of the traveling wave fault ranging system to be self-inspected can be more accurately predicted based on the historical fault data and device parameter training model. Therefore, the traveling wave fault ranging system can be configured with the optimal system parameters to obtain more accurate fault location results, making the traveling wave fault ranging system more accurate and reliable. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 A flow chart of a self-checking method for a power grid traveling wave fault location system based on historical data mining is provided in an embodiment of the present invention.

[0035] Figure 2 This is a schematic diagram of a regional fault location principle.

[0036] Figure 3 The figure is a schematic diagram of the traveling wave fault location principle of a transmission line.

[0037] Figure 4 A schematic diagram of a self-checking device for a power grid traveling wave fault location system based on historical data mining provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0039] like Figure 1 As shown, the present invention provides a self-checking method for a power grid traveling wave fault location system based on historical data mining, which may include the following steps:

[0040] Step 101: Acquire historical fault data of transmission lines in at least two states of the traveling wave fault location system in each area and historical device-related parameters of each substation node at the time of historical faults; wherein the historical fault data includes: voltage waveform data and current waveform data;

[0041] Step 102: supervised learning is performed using historical fault data and historical device-related parameters to train a system parameter prediction model; wherein the system parameter prediction model is used to predict the optimal system parameters of the power grid traveling wave fault location system; the system parameters include: initial wave head time, distance between each trunk node in the transmission line and traveling wave propagation speed; wherein the initial wave head time is used to characterize the time when the traveling wave arrives at each substation node;

[0042] Step 103: obtaining current device-related parameters and the number of trunk line nodes of each substation of the traveling wave fault location measurement system to be self-checked;

[0043] Step 104: Generate at least one set of test data according to current device related parameters, the number of trunk line nodes and historical fault data; wherein each set of test data corresponds to a fault;

[0044] Step 105: input each group of test data into the system parameter prediction model to obtain the optimal system parameters of the current power grid traveling wave fault location system;

[0045] Step 106: adjusting the system parameters of the power grid traveling wave fault location system according to the optimal system parameters to achieve self-checking of the power grid traveling wave fault location system.

[0046] In this embodiment, the implementation logic can be directly configured into the traveling wave fault ranging system, and there is no need to develop additional test equipment for the traveling wave fault ranging system, which not only requires excessive human involvement, but also can greatly save the equipment construction cost. In addition, this solution can mine deeper features from historical data through machine learning, so that the optimal system parameters of the traveling wave fault ranging system to be self-tested can be more accurately predicted based on the historical fault data and device parameter training model, so that the traveling wave fault ranging system can configure the optimal system parameters to obtain more accurate fault location results, making the traveling wave fault ranging system more accurate and reliable.

[0047] Moreover, this scheme uses existing fault waveform data to verify and test the accuracy of the results of the traveling wave ranging device. This scheme uses a large amount of data for deduction, so that some self-test parameters that were not obvious before can be learned by the model, thereby improving the reliability of the system and the accuracy of fault ranging, while reducing the false operation rate and refusal rate. In addition, the scheme can automatically identify the operation mode, effectively reducing the workload of the staff. The deep learning network does not require manual feature extraction, and can learn deeper data feature connections through data samples. It has strong nonlinear fitting capabilities, and its learning ability is stronger as the network depth increases. This self-test method inputs historical fault waveforms and other data into the self-test model to test the response of the fault location system and the accuracy of the ranging results, which can significantly improve its accuracy and stability, thereby providing a more solid guarantee for the safe operation of the smart grid.

[0048] For step 101, historical fault data of the transmission line in at least two states of the traveling wave fault location system in each area and historical device-related parameters of each substation node at the time of historical faults are obtained; wherein the historical fault data includes: voltage waveform data and current waveform data;

[0049] The existing two-terminal traveling wave distance measurement system for transmission lines uses the initial traveling wave head generated by the fault to determine the fault point location by calculating the time difference between the transient traveling wave reaching the two ends of the line, as shown in the following formula:

[0050]

[0051] In the formula, l 1 is the distance from the fault point to the busbar, t 1 and t 2 are the time it takes for the traveling wave to reach both ends of the line, v is the speed of the traveling wave, and L is the total length of the line.

[0052] The wave velocity and the time when the initial traveling wave arrives at the busbar are key issues in fault distance calculation. If the arrival time of the traveling wave is 1μs, it may cause a sub-measurement error of 150m. The influence of wave velocity on fault distance accuracy is related to the location of the fault point and the line length. The longer the line, the greater the influence of wave velocity. When the fault point is located near the two ends of the line, the influence of wave velocity on distance accuracy is particularly large.

[0053] Regional fault location is developed from the existing two-terminal traveling wave ranging. Figure 2 As shown in the figure, when the fault point is located in the MN section of the line, it is assumed that the device at the M end is abnormal and the initial wave head time t at the M end cannot be obtained. 1 , only the initial wave head moment t at the N end 2 , then it is impossible to complete the dual-end fault location. Regional fault location can be achieved by the initial wave head time (t 1 ′,t′ 2 ,…t′ n ) and the initial wave head moment t at the N end 2 Pairing, forming a dual-end ranging, the initial wave head time t of the S end 1 ′,t′ 2 ,…t′ n With t 1 Satisfies the following formula:

[0054]

[0055] Where, L n is the line length of the adjacent line n, t′ n For the traveling wave to reach S n At the end moment, v n is the propagation speed of the traveling wave on line n. Theoretically, t is calculated based on the traveling waves on each line. 1 The results should be consistent, that is, satisfy the following formula:

[0056]

[0057] Therefore, the initial wave head moment, the distance between the trunk nodes in the transmission line and the traveling wave propagation speed are considered as the system parameters of the traveling wave fault location system, and then these parameters are used to perform self-inspection on the traveling wave fault location system. The sample data of the training model can be constructed by combining the historical fault data with the historical device-related parameters of the substation node.

[0058] For historical fault data, it may include voltage waveform data, current waveform data, etc. under different states. Different states may include the traveling wave fault ranging system being in a normal state, a state requiring attention, a state with an abnormality, and a state with a serious abnormality.

[0059] The relevant parameters of historical devices may include: line type, line length, fault waveform file, and timing, startup settings and line parameters of local devices.

[0060] When constructing sample training data, it is further possible to simulate the actual situation after the transmission line is grounded or short-circuited. Figure 3 As shown, the transient traveling wave generated by the fault will propagate along the line in the power grid, and each terminal on the transmission line can detect the transient traveling wave. At this time, multiple groups of sample fault parameter generation data of the target access point are obtained.

[0061] Depend on Figure 3 It can be seen that t 1 ,t 2 ,t 3 ,t 4 are the time when the transient traveling wave reaches each actual substation terminal measurement node (called the initial wave head moment in this paper); t 1 ′、t 2 ′、t 3 ′、t 4 ′ are respectively based on the initial wave head time t of the actual substation terminal node n The calculated initial wave head time of the virtual node on the trunk line (hereinafter referred to as the trunk node); 1 , l 2 , l 3 , l 4 are the distances between the backbone nodes, and b 1 , b 2 、b 3 、b 4 are the lengths of each branch line respectively. Assuming the line wave speed is v, the initial wave head time of each trunk node is expressed as follows:

[0062]

[0063] Assuming that each trunk node of the substation is the starting point of the traveling wave, the initial wave head time of each node of the substation along the line is calculated segment by segment. n , the distance between each backbone node is l n and the traveling wave propagation speed v n is the system parameter. From the historical fault data and historical device related parameters, a large number of fault condition data can be obtained, that is, t n =(t 1 ,t 2 ,t 3 …t i ) T , l n =(l 1 ,l 2 ,l 3 …li ) T , v n =(v 1 ,v 2 ,v 3 …v i ) T The fault conditions covered by the sample data may include: single-phase grounding, two-phase grounding, two-phase short circuit, three-phase short circuit, line break, overload, insulation degradation, equipment failure, impact of natural disasters, external force damage, bird damage, and flashover, etc.; the specific fault data may include information such as different fault locations, different fault types, different fault sections, and different grounding impedances.

[0064] It is easy to understand that after obtaining historical fault data, we should consider preprocessing the data, such as normalizing and standardizing the data. The preprocessing algorithm can be, for example, Z-score, Z-mean. The preprocessing algorithm can also include multi-source data dimensionality reduction such as PCA and LDA to improve the speed of subsequent model training.

[0065] When generating sample data, the data used to train the model should contain a large number of fault condition data sets. Usually, a suitable data set should contain at least 500 fault condition data sets. Generally speaking, the more data, the better the model training effect, but the training time will also increase accordingly. The specific selection should be based on the requirements for model accuracy and model training time.

[0066] For step 102, supervised learning is performed using historical fault data and historical device-related parameters to train a system parameter prediction model; wherein the system parameter prediction model is used to predict the optimal system parameters of the power grid traveling wave fault location system;

[0067] A neural network is a computational model inspired by the structure of the human brain. It processes and analyzes data by imitating the connection mode of brain neurons. It can mainly include processes such as data input, forward propagation, loss calculation, back propagation, weight update, model evaluation and testing. The neural network learning process mainly consists of input forward propagation, error back propagation, cyclic memory training and judgment of learning results. Through back propagation, the model can be continuously corrected so that the network output can be more in line with the expected prediction results. Therefore, this step considers training the system parameter prediction model through a neural network to achieve accurate prediction of the optimal system parameters.

[0068] In this step, the neural network model in supervised learning can be composed of a classification network and a regression network connected in series. The classification network is used to classify historical fault data according to the factors that cause fault distance measurement errors; the regression network is used to determine the optimal values ​​of system parameters under the known fault location errors under the current system parameters. Specifically, the training model used is a serial ten-layer neural network model. The classification network includes a data input layer, which selects historical fault data and historical device-related parameters as input data. The number of input layer nodes is i, that is, the number of trunk line nodes; eight fully connected layers and a classification output layer. The regression network includes a data input layer, which uses the data output by the classification output layer of the classification network as input data.

[0069] For the classified network, the classification is based on the factors causing the fault location error, including the error caused by the time synchronization in the system parameters, the error caused by the line parameters in the system parameters, and the error caused by the traveling wave speed in the system parameters.

[0070] For example, models built using neural networks have a significant impact on both recognition rate and convergence speed. In neural network models, activation functions such as ReLU, Sigmoid, or Tanh are used to capture the nonlinear characteristics of these errors. Through training data, neural networks can learn the relationship between these errors and input quantities, which can include the following three situations:

[0071] (1) The error changes with time. The error at the measurement time depends on the GPS timing error. When the time when the transient wave reaches each actual substation terminal measurement node changes compared with the historical data time, the problem usually manifests as a GPS synchronization clock problem. If the GPS time synchronization system encounters a fault, such as unstable receiver signal, satellite signal loss of lock, or clock jump, it may cause time synchronization error, which in turn affects the accuracy of fault distance measurement. During model training, the activation function is used to simulate this error, reflecting a nonlinear relationship related to the time point, in which the error changes with time. Specifically, the error is a time synchronization error for each self-check.

[0072] (2) The error is a constant value. When the distance between the transient traveling wave and the actual substation terminal measurement nodes changes compared with the historical data, it indicates that there is a problem of line parameter changes. Transmission line parameters (such as resistance, inductance, capacitance, etc.) and line structures (such as chain grids, triangular ring networks, cable hybrid lines, etc.) have a direct impact on the propagation of traveling waves. Changes in these parameters and structures may lead to reduced accuracy in fault distance measurement. The errors caused by line parameters may be caused by line aging, environmental changes, or measurement equipment errors. The activation function reflects a constant nonlinear relationship, that is, no matter how the input changes, the error is expressed as a fixed value, which is specifically manifested in each self-test, the error is constant.

[0073] (3) The error changes with the change of wave speed. The uncertainty of wave speed measurement depends on the selection of wave speed parameters from historical data. When the speed of the transient wave reaching the actual substation terminal measurement node changes compared with the historical data, it indicates that there is a problem with the change of wave speed parameters. The error caused by the wave speed value may be caused by changes in the transmission medium, temperature effects or other environmental factors. For example: ① Line inductance, capacitance, earth resistivity, etc. are affected by weather and are not stable constants; ② Engineering parameters such as the spacing between different ground lines are often consistent and often have errors; ③ If the tower data is missing, there are generally errors in the line length. The activation function reflects a nonlinear relationship that fluctuates with the input, in which the error is not fixed, but changes with the change of wave speed. Specifically, the error is drifting in each self-test.

[0074] In this way, data classification can be achieved according to the above classification standards. During the training process, the learning rate of the training error function is variable. After multiple experiments, the content and location of the specific problematic system parameters are determined through classifier post-processing. Each round of classifier output layer post-processing is a verification of the current system parameters.

[0075] For the regression network, it performs regression training by receiving data from the classification network. Regression training includes collecting and preprocessing data, selecting and training models, tuning hyperparameters, evaluating and diagnosing models, optimizing models, testing model performance, and deploying and monitoring models. After obtaining the output value of the regression network, if the error between it and the historical system parameter data is greater than the set error, the weights of each layer need to be corrected. That is, the weights and deviations of each layer of the network are adjusted to reduce the error function along the negative gradient direction until the error reaches the minimum value or meets the set error requirement.

[0076] For step 103, the current device-related parameters and the number of trunk line nodes of each substation of the traveling wave fault location system to be self-checked are obtained;

[0077] In this step, for the traveling wave fault location system that currently needs self-checking, the corresponding device-related parameters and the number of trunk line nodes are obtained.

[0078] For step 104, at least one set of test data is generated according to current device-related parameters, the number of trunk line nodes and historical fault data; wherein each set of test data corresponds to a fault;

[0079] In this step, according to the device-related parameters of the traveling wave fault location system to be self-checked and the number of trunk line nodes, the test data under the corresponding fault conditions are extracted from the historical fault data. For example, the test data under one condition may include [t n =(t 1 ,t 2 ,t 3 …t i ) T ,l n =(l 1 ,l 2 ,l 3 …l i ) T ,v n =(v 1 ,v 2 ,v 3 …v i ) T ].

[0080] Step 105, input each group of test data into the system parameter prediction model to obtain the optimal system parameters of the current power grid traveling wave fault location system;

[0081] In this step, after the model training is completed, the test data is input into the trained system parameter prediction model to obtain the optimal system parameters predicted by the model.

[0082] Step 106, adjusting the system parameters of the power grid traveling wave fault location system according to the optimal system parameters, so as to realize the self-check of the power grid traveling wave fault location system.

[0083] In this step, the system parameters of the traveling wave fault location system are adjusted according to the optimal system parameters predicted and output by the system parameter prediction model, thereby ensuring that the traveling wave fault location system can use the configured optimal system parameters to locate faults, thereby ensuring the accuracy of fault location.

[0084] In one embodiment, in order to further improve the reliability of system parameter adjustment, if the deviation between the optimal system parameters predicted by the system prediction model and the current system parameters of the traveling wave fault ranging system is greater than a preset deviation threshold, an alarm can be issued so that the staff can manually confirm the self-test results and then adjust the system parameters, thereby further ensuring the reliability of the self-test results.

[0085] In addition, in order to continuously optimize the system parameter prediction model, the system parameter prediction model can be further updated using the data obtained after each self-test. That is, after the self-test is completed, the updated data after each self-test is added to the training data set for training, expanding the database to ensure that the model becomes more accurate before the line fault occurs, completing the self-update, and further improving the accuracy of fault location.

[0086] like Figure 4 As shown, the present invention also provides a self-checking device for a power grid traveling wave fault location system based on historical data mining, comprising: a historical data acquisition module 401, a model training module 402, a current data acquisition module 403, a test data generation module 404, a system parameter prediction module 405 and a system parameter adjustment module 406;

[0087] The historical data acquisition module 401 is configured to acquire the historical fault data of the transmission line of the traveling wave fault location system in each area in at least two states and the historical device-related parameters of each substation node at the time of the historical fault; wherein the historical fault data includes: voltage waveform data and current waveform data;

[0088] The model training module 402 is configured to use historical fault data and historical device related parameters for supervised learning, and train to obtain a system parameter prediction model; wherein the system parameter prediction model is used to predict the optimal system parameters of the power grid traveling wave fault location system; the system parameters include: the initial wave head time, the distance between each trunk node in the transmission line and the traveling wave propagation speed; wherein the initial wave head time is used to characterize the time when the traveling wave arrives at each substation node;

[0089] The current data acquisition module 403 is configured to acquire current device-related parameters and the number of trunk line nodes of each substation of the traveling wave fault location system to be self-checked;

[0090] The test data generation module 404 is configured to generate at least one set of test data according to current device related parameters, the number of trunk line nodes and historical fault data; wherein each set of test data corresponds to a fault;

[0091] The system parameter prediction module 405 is configured to input each set of test data into the system parameter prediction model to obtain the optimal system parameters of the current power grid traveling wave fault location system;

[0092] The system parameter adjustment module 406 is configured to adjust the system parameters of the power grid traveling wave fault location system according to the optimal system parameters to achieve self-checking of the power grid traveling wave fault location system.

[0093] The present specification also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed in a computer, the computer is caused to execute a method in any one of the embodiments in the specification.

[0094] The present specification also provides a computing device, including a memory and a processor, wherein executable codes are stored in the memory, and when the processor executes the executable codes, a method in any embodiment of the present specification is implemented.

[0095] Since the device embodiments provided by the present invention are based on the same inventive concept as the method embodiments of this specification, the specific contents can be found in the description of the method embodiments of this specification and will not be repeated here.

[0096] In summary, the power grid traveling wave fault location system self-checking method based on historical data mining provided by the present invention has at least the following beneficial effects:

[0097] 1) The self-check method of the power grid traveling wave fault location system based on historical data mining can effectively make up for the defects and blind spots of the existing online monitoring system in terms of operating status reliability, and has a significant positive effect on improving the reliability of the transmission line fault monitoring system.

[0098] 2) The accuracy of the self-test results can more truly reflect the risk factors of the transmission line fault monitoring system, such as the system parameter configuration, synchronization time of field equipment, start-up threshold and line parameters, which helps to manage and control the accuracy of transmission line fault results in a more targeted manner.

[0099] 3) Establish a self-checking model of the neural network, and obtain more accurate parameter values ​​of the current system through each machine learning and model training.

[0100] 4) This solution is directly configured into the traveling wave fault location system, and does not require additional development of test equipment for the traveling wave fault location system, which not only requires excessive manpower participation, but also can greatly save the equipment construction cost.

[0101] The modules or units in the device of the embodiment of the present invention can be combined, divided and deleted according to actual needs. The above disclosure is only the preferred embodiment of the present invention, and of course it cannot be used to limit the scope of the rights of the present invention. Those skilled in the art can understand that all or part of the processes of the above embodiment are implemented, and the equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims

1. A self-checking method for a power grid traveling wave fault location system based on historical data mining, characterized in that: include: Acquire the historical fault data of the transmission lines of the traveling wave fault location system in each area in at least two states and the historical device-related parameters of each substation node at the time of the historical fault; wherein the historical fault data includes: voltage waveform data and current waveform data; The historical fault data and historical device-related parameters are used for supervised learning to train a system parameter prediction model; wherein the system parameter prediction model is used to predict the optimal system parameters of the power grid traveling wave fault location system; the system parameters include: initial wave head time, distance between each trunk node in the transmission line and traveling wave propagation speed; wherein the initial wave head time is used to characterize the time when the traveling wave arrives at each substation node; Obtain current device-related parameters and trunk line node numbers of each substation of the traveling wave fault location system to be self-checked; Generate at least one set of test data according to the current device related parameters, the number of trunk line nodes and the historical fault data; wherein each set of test data corresponds to a fault; Input each group of test data into the system parameter prediction model to obtain the optimal system parameters of the current power grid traveling wave fault location system; The system parameters of the power grid traveling wave fault location measurement system are adjusted according to the optimal system parameters to achieve self-checking of the power grid traveling wave fault location measurement system.

2. The self-checking method of power grid traveling wave fault location system based on historical data mining according to claim 1 is characterized in that: The historical device related parameters include: line type, line length, fault waveform file, and the timing, startup setting and line parameters of the local device.

3. The self-checking method of power grid traveling wave fault location system based on historical data mining according to claim 1 is characterized in that: The neural network model in the supervised learning includes a classification network and a regression network connected in series; The classification network is used to classify the historical fault data according to factors causing fault location errors; The regression network is used to determine the optimal values ​​of system parameters under the known fault location errors under the current system parameters.

4. The self-checking method of power grid traveling wave fault location system based on historical data mining according to claim 3 is characterized in that: The classification network includes a data input layer, eight fully connected layers and a classification output layer; wherein the data input layer uses the historical fault data and historical device related parameters as input data, and uses the number of trunk line nodes as the number of input layer nodes; The regression network includes a data input layer, eight fully connected layers and a regression output layer; wherein the data input layer of the regression network uses the data output by the classification output layer of the classification network as input data.

5. The self-checking method of power grid traveling wave fault location system based on historical data mining according to claim 3 is characterized in that: The factors causing the fault location error include: an error caused by a time synchronization amount in the system parameters, an error caused by a line parameter in the system parameters, and an error caused by a traveling wave velocity in the system parameters.

6. The self-checking method of power grid traveling wave fault location system based on historical data mining according to claim 1 is characterized in that: The adjusting of the system parameters of the power grid traveling wave fault location system according to the optimal system parameters includes: If the deviation between the optimal system parameters and the current system parameters of the traveling wave fault location system is greater than a preset deviation threshold, an alarm prompt is issued so that the staff can manually confirm the self-test result.

7. The self-checking method of power grid traveling wave fault location system based on historical data mining according to claim 1 is characterized in that: After each self-check is completed, the system parameter prediction model is further updated using the data obtained after the self-check.

8. A self-checking device for a power grid traveling wave fault location system based on historical data mining, characterized in that: include: Historical data acquisition module, model training module, current data acquisition module, test data generation module, system parameter prediction module and system parameter adjustment module; The historical data acquisition module is configured to acquire the historical fault data of the transmission line of the traveling wave fault location system in each area in at least two states and the historical device-related parameters of each substation node at the time of the historical fault; wherein the historical fault data includes: voltage waveform data and current waveform data; The model training module is configured to use the historical fault data and historical device related parameters for supervised learning to train a system parameter prediction model; wherein the system parameter prediction model is used to predict the optimal system parameters of the power grid traveling wave fault location system; the system parameters include: initial wave head time, the distance between each trunk node in the transmission line and the traveling wave propagation speed; wherein the initial wave head time is used to characterize the time when the traveling wave arrives at each substation node; The current data acquisition module is configured to obtain current device-related parameters and the number of trunk line nodes of each substation of the traveling wave fault location measurement system to be self-checked; The test data generation module is configured to generate at least one set of test data according to the current device related parameters, the number of trunk line nodes and the historical fault data; wherein each set of test data corresponds to a fault; The system parameter prediction module is configured to input each set of test data into the system parameter prediction model to obtain the optimal system parameters of the current power grid traveling wave fault location system; The system parameter adjustment module is configured to adjust the system parameters of the power grid traveling wave fault location system according to the optimal system parameters to achieve self-checking of the power grid traveling wave fault location system.

9. A computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to execute the method according to any one of claims 1 to 8.

10. A computing device comprising a memory and a processor, wherein the memory stores executable codes, and when the processor executes the executable codes, the method according to any one of claims 1 to 8 is implemented.

Images