A Fault Identification and Location Method for Distribution Lines Based on DWT and BP Algorithms

Abstract

The present invention relates to the field of discrete signal processing, and specifically relates to a method for fault identification and location of distribution lines based on the DWT and BP algorithms. The method includes the following steps: S1, mounting hardware devices on different phases of nodes on the same line in the distribution network line, and collecting and recording discrete data information of current, voltage and traveling waves in the overhead line; S2, constructing a network topology map of the line to be monitored according to the mounting points of the hardware devices, and combining the use of the particle swarm optimization algorithm and the DWT algorithm to perform noise reduction preprocessing on the original data; S3, performing polynomial fitting and cumulative curve fitting approximation processing on the data after noise reduction processing to obtain the continuous curve data after fitting. The present invention provides a method for fault identification and location of distribution lines based on the DWT and BP algorithms. By combining signal processing technology and intelligent classification algorithms, it realizes the automatic identification and accurate location of distribution line faults, improves the accuracy and efficiency of fault detection, and provides a strong guarantee for the safe and stable operation of the power grid.

Description

Technical Field

[0001] The present invention relates to the field of discrete signal processing, and particularly to a method for identifying and locating faults in a distribution line based on the DWT and BP algorithms. Background Art

[0002] With the rapid development of network technology, network security threats have become increasingly complex and diverse. Network attackers adopt increasingly concealed, multi-level, and multi-source attack means, posing great challenges to network security. Currently, network attack means are becoming increasingly complex and diverse, and traditional security protection methods are difficult to cope with new threats. Existing network kill chain models such as Cyber Kill Chain have been widely applied in security research and defense strategies.

[0003] In the prior art, although there are some fault detection methods based on signal processing, such as Fourier transform, wavelet transform, etc., these methods have limitations in dealing with non-linear and non-stationary signals and are difficult to effectively extract periodic information. At the same time, most traditional fault location methods rely on empirical formulas and simple algorithms, and the location accuracy and reliability need to be improved.

[0004] For example, in the invention patent with the patent publication number CN118584249A and the name "A Fault Detection Method for Distribution Lines Based on State Feature Extraction", it includes: (1): Collect signals; (2): Judge according to the current, if abnormal, alarm; if normal, go to (3); (3): Perform Fourier transform on the current; (4): Calculate the current harmonic distortion rate; (5): Judge according to the current harmonic distortion rate, if abnormal, alarm; if normal, go to (6); (6): Calculate the fundamental wave distortion rate of the current; (7): Judge according to the fundamental wave distortion rate of the current, if abnormal, alarm; if normal, go to (8); (8): Calculate the voltage difference between the head and the end of the distribution line; (9): Judge whether the voltage difference is greater than the reference value, if yes, go to (10); if no, it is normal; (10): Calculate the abnormal current; (11): Judge according to the abnormal current, if abnormal, alarm. The disadvantages are: there are limitations in dealing with non-linear and non-stationary signals and it is difficult to effectively extract periodic information. At the same time, relying on empirical formulas and simple algorithms, the location accuracy and reliability need to be improved. Summary of the Invention

[0005] Aiming at the problems that the fault detection method for distribution lines in the prior art has limitations in dealing with non-linear and non-stationary signals and the location accuracy and reliability are insufficient, the present invention provides a method for identifying and locating faults in a distribution line based on the DWT and BP algorithms. By combining signal processing technology and intelligent classification algorithms, it realizes the automatic identification and precise location of faults in the distribution line, improves the accuracy and efficiency of fault detection, and provides a strong guarantee for the safe and stable operation of the power grid.

[0006] To achieve the above technical objectives, a technical solution provided by the present invention is a method for identifying and locating faults in a distribution line based on the DWT and BP algorithms, including the following steps:

[0007] S1. Mount the hardware devices on different phases of the nodes of the same line in the distribution network line, collect and record the discrete data information of the current, voltage and traveling wave in the overhead line, and upload it to the platform database through the MQTT protocol;

[0008] S2. Construct a network topology diagram of the line to be monitored according to the mounting points of the hardware devices, and use the particle swarm optimization algorithm and the DWT algorithm in combination to perform noise reduction preprocessing on the original data;

[0009] S3. Perform polynomial fitting and cumulative curve fitting approximation processing on the data after noise reduction preprocessing to obtain the continuous curve data after fitting;

[0010] S4. Perform multiple truncation processes on the continuous curve data after fitting, and analyze the frequency components in the truncated interval through fast Fourier transform, and extract the peak information in the frequency domain as the period information;

[0011] S5. Build a BP neural network model, input the preprocessed data into the network for training and classification recognition, judge the fault interval according to the classification result in combination with the line topology diagram, and use the traveling wave fault location algorithm to realize the location of the fault point.

[0012] In this technical solution, hardware devices are deployed at key nodes of the distribution network line to comprehensively collect discrete data of current, voltage and traveling wave, and efficiently upload them to the cloud database through the MQTT protocol; subsequently, the particle swarm optimization and DWT technologies are used for deep noise reduction to optimize the quality of the original data; furthermore, through the fitting technologies of polynomials and cumulative curves, the data is transformed into a continuous curve form, laying a foundation for accurate analysis. On this basis, multiple data truncations are implemented, and fast Fourier transform is used for frequency domain analysis to extract the peak information of the fault data, and the fault data is intercepted according to the peak information. Finally, a BP neural network model is built, the fault data is input into the BP network for training to obtain the network model, the new data is classified according to the model result, the fault interval is quickly locked in combination with the line topology diagram, and the accurate location of the fault point is realized relying on the traveling wave fault location algorithm. It significantly improves the accuracy and efficiency of fault detection, builds a solid defense line for the safe and stable operation of the power grid, and effectively guarantees the continuous and reliable operation of the power system.

[0013] The present invention is further configured such that in step S2, the noise reduction preprocessing of the original data by combining the particle swarm optimization algorithm and the DWT algorithm includes: performing denoising processing through the DWT algorithm, where the DWT algorithm decomposes the signal into approximation coefficients and detail coefficients of different scales. The DWT algorithm uses discrete wavelet functions and discrete time scales, and realizes the decomposition of the signal through filtering and downsampling operations, decomposing the original signal into a group of high-frequency sub-signals and low-frequency sub-signals; improving the Heaviside projection function in the DWT algorithm, and optimizing the basic Heaviside function using the sparse regularization Heaviside function.

[0014] It also includes adding a signal-to-noise ratio comparison module based on the PSO algorithm to perform iterative processing on the result after the improved DWT algorithm.

[0015] In this technical solution, by adopting the DWT algorithm, the original signal can be accurately decomposed into approximation coefficients and detail coefficients of different scales. Using discrete wavelet functions and discrete time scales, the separation of high-frequency and low-frequency sub-signals of the signal is realized through filtering and downsampling operations. On the basis of the DWT algorithm, an innovative improvement is made to the Heaviside projection function, and the sparse regularization Heaviside function is introduced. The noise reduction effect is enhanced, providing a more accurate data basis for subsequent fault identification and location. A signal-to-noise ratio comparison module is added within the PSO algorithm framework to perform PSO algorithm iterative processing on the result after the improved DWT algorithm, obtaining the relative optimal solution during data preprocessing and improving the accuracy of fault identification.

[0016] The present invention is further configured such that the Heaviside projection function in the DWT algorithm is optimized by sparse regularization to improve the denoising effect on non-linear and non-smooth data.

[0017] In this technical solution, during the noise reduction processing in step S2, in-depth sparse regularization optimization is performed on the Heaviside projection function in the DWT algorithm. The algorithm shows more excellent performance when dealing with non-linear and non-smooth data, improving the accuracy and stability of the noise reduction processing.

[0018] The present invention is further configured such that in step S3, the cumulative curve fitting approximation processing includes: using a polynomial fitting algorithm to perform preliminary fitting on discrete data, performing secondary fitting through the cumulative curve fitting approximation algorithm, comparing the fNIRS data characteristics with the similar data characteristics collected by the hardware device, using the behavior of the approximate signal curve within the sliding window, obtaining the estimated curve from multiple overlapping windows, and receiving the filtered signal using a weighted logic recombination method to obtain the final fitting curve.

[0019] In this technical solution, a polynomial fitting algorithm is used to perform a preliminary fit on the original discrete data, and then an advanced CCFA algorithm is used for secondary fitting, which further improves the accuracy and stability of the fitting. The fNIRS data features are introduced for comparative analysis with similar data features collected by the hardware device, and the data features are obtained through data comparison and analysis. The behavior of the approximate signal curve in the sliding window is used to cleverly obtain the estimated curve from multiple overlapping windows, and the weighted logical recombination method is used to receive the filtered signal. Through reasonable weight distribution and logical processing, the relevant noise and interference are effectively eliminated, and the final fitting curve is obtained. This fitting curve not only accurately reflects the actual situation of the distribution line fault, but also provides strong data support for subsequent fault location and identification, thereby significantly improving the accuracy and practicality of the entire method.

[0020] The present invention is further configured as follows: between steps S4 and S5, a processing step is also included: the operation in S4 is iterated, and when the peak information in the frequency domain after the fast Fourier transform tends to be consistent, the iterative operation is terminated to obtain a final peak information set.

[0021] In this technical solution, the fast Fourier transform is used to perform frequency domain analysis on the relevant data, and the peak information in the frequency domain is closely monitored. Through continuous iterative operations, the peak information in the frequency domain gradually converges to a consistent state, thereby effectively eliminating accidental errors and noise interference. When the iterative operation reaches the preset termination condition, that is, when the peak information in the frequency domain shows relative consistency, the iteration is terminated to obtain the final peak information set. It not only accurately reflects the characteristics of distribution line faults in the frequency domain, but also provides accurate data support for subsequent fault location and identification, thereby further improving the accuracy and practicality of the entire method.

[0022] The present invention is further configured as follows: in step S4, the multiple truncation processing of the fitted continuous curve data includes: preferentially setting a preliminary truncation threshold, performing fast Fourier transform on the truncated data, analyzing the frequency components of the data in the truncation interval, obtaining peak information in the frequency domain, performing statistical processing on all peak information, obtaining a preliminary peak information set, and iterating the truncation threshold, performing multiple iterations, and obtaining a new peak information set; after multiple iterations, if the peak information obtained after multiple iterations is maintained within a certain range and the frequency of occurrence is greater than a certain threshold, the peak information is defined as periodic information.

[0023] In this technical solution, first, a preliminary truncation threshold is set, and based on this, the continuous curve data is truncated for the first time. The fast Fourier transform is applied to the truncated data to analyze the frequency components of the data in each truncated interval, and the peak information in the frequency domain is accurately captured. The peak information set under this truncation threshold is obtained. By updating the truncation threshold and performing multiple iterations, a new peak information set is obtained. After multiple iterations, if the peak information in the peak information set tends to be relatively stable in multiple iterations and the frequency of occurrence exceeds the preset threshold, then this peak information is defined as periodic information. This not only improves the accuracy of fault identification but also provides strong support for the location and repair of faults.

[0024] The present invention is further configured as: according to the operations in the processing steps, n cycles of continuous curve data are intercepted from the original data in step S3, and sampling processing is performed to obtain discrete waveform data.

[0025] In this technical solution, after step S3, n cycles of continuous curve data are intercepted from the original data. This ensures that the analyzed data segment has complete and representative periodic characteristics, laying a solid foundation for subsequent fault identification. Immediately afterwards, since the BP neural network uses discrete data as input data during training, the continuous curve data is converted into discrete waveform data, that is, sampling processing is performed on the continuous curve data to obtain the final discrete data, and these discrete waveform data not only retain the core characteristics of the original data.

[0026] The present invention is further configured as: in step S5, the building of the BP neural network model includes: creating a training set and a test set, inputting them into the BP neural network for iterative training, obtaining the model parameters of the BP neural network and saving them.

[0027] In this technical solution, first, by creating a training set and a test set, both of these two data sets are derived from the discrete waveform data that has been preprocessed and feature-extracted in the previous steps. These two data sets are input into the constructed BP neural network. The BP neural network adjusts the weight and bias parameters inside the network through multiple iterative trainings in order to achieve the best fitting effect and prediction accuracy. After multiple iterations and adjustments, when the performance of the BP neural network on the training set tends to be stable and meets the termination training conditions, the network model parameters at this time are extracted and saved.

[0028] The present invention is further configured as: in step S5, it further includes: performing actual test verification on the BP neural network model, and optimizing the neuron parameters, loss function, optimizer, initial weights, and network structure according to the test verification results to obtain the final waveform classification model.

[0029] In this technical solution, not only a BP neural network model is established, but also it is further verified through actual tests to ensure the practicability and accuracy of the model. According to the feedback results of these test verifications, comprehensive and systematic optimization processing is carried out on the neuron parameters, loss function, optimization algorithm, initial weights, and network structure in the cycle information extraction.

[0030] The present invention is further set as: in step S5, it further includes: when the classification result of the final waveform classification model is a fault defect, defect interval positioning is performed: the defect interval positioning includes: a short - circuit fault interval judgment algorithm and a single - phase grounding fault interval judgment algorithm.

[0031] In this technical solution, when a fault defect in the distribution line is identified by using the optimized final waveform classification model, the defect interval positioning program is immediately started. This program includes two core algorithms: a short - circuit fault interval judgment algorithm and a single - phase grounding fault interval judgment algorithm. Through the combined use of these two algorithms, the specific location where the fault occurs can be quickly and accurately locked, providing crucial guiding information for subsequent fault repair work, thereby greatly improving the efficiency and accuracy of distribution line fault handling.

[0032] The beneficial effects of the present invention are as follows: (1) By combining signal processing technology and intelligent classification algorithms, automatic identification and precise positioning of distribution line faults are achieved, improving the accuracy and efficiency of fault detection, and providing a strong guarantee for the safe and stable operation of the power grid; (2) Denoising processing of non - linear data, adding the Heaviside projection function in the original discrete wavelet transform algorithm to sparse regularization, improving its processing effect on non - linear and non - smooth data, and combining the particle swarm optimization algorithm to perform secondary denoising processing on the entire discrete data. By nesting the iterative PSO algorithm behind the improved DWT algorithm, that is, by comparing the signal - to - noise ratio of the improved DWT algorithm with the signal - to - noise ratio of the PSO algorithm, it is also judged whether the denoising effect at this time meets the expectation; (3) A BP neural network is created. The classification of discrete point sets does not require the use of complex neural networks. The BP neural network constructed in this solution is based on the triangular waveform analysis network and extended experiments are carried out based on this. It is initially concluded that the above - mentioned network model has a good classification effect on this discrete data. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a schematic diagram of the terminal module structure of the present invention;

[0034] Figure 2 It is a schematic diagram of the terminal working principle of the present invention;

[0035] Figure 3 It is a schematic diagram of the traveling - wave judgment fault interval of the present invention;

[0036] Figure 4 Schematic diagram for judging fault section by short - circuit current of the present invention;

[0037] Figure 5 Schematic diagram of single - ended and double - ended traveling waves of the present invention;

[0038] Figure 6 Schematic flow chart of the present invention. Detailed implementation manners

[0039] To make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific implementation manners described herein are only the best embodiments of the present invention, which are only used to explain the present invention and do not limit the protection scope of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0040] As Figures 1 to 6 shown, as the first embodiment of the present invention, a method for identifying and locating distribution line faults based on DWT and BP algorithms includes the following steps:

[0041] S1, Mount the hardware devices on different phases of nodes on the same line in the distribution network line, collect and record the discrete data information of current, voltage and traveling waves in the overhead line, and upload it to the platform database through the MQTT protocol;

[0042] S2, Construct the network topology diagram of the line to be monitored according to the mounting points of the hardware devices, and use the particle swarm optimization algorithm and DWT algorithm in combination to perform noise reduction pre - processing on the original data;

[0043] S3, Perform polynomial fitting and cumulative curve fitting approximation on the data after noise reduction pre - processing to obtain the continuous curve data after fitting;

[0044] S4, Perform multiple truncation processes on the continuous curve data after fitting, and analyze the frequency components in the truncated interval through fast Fourier transform, and extract the peak information in the frequency domain as the period information;

[0045] S5, Build a BP neural network model, input the pre - processed data into the network for training and classification recognition, judge the fault section according to the classification results in combination with the line topology diagram, and use the traveling wave fault location algorithm to locate the fault point.

[0046] In this technical solution, hardware devices are deployed at key nodes of the distribution network line to comprehensively collect fine data of current, voltage and traveling waves, and efficiently upload them to the cloud database by means of the MQTT protocol; subsequently, the particle swarm optimization and DWT technology are used for deep noise reduction to optimize the quality of the original data; furthermore, through the fitting technology of polynomial and cumulative curve, the data is transformed into a continuous curve form to lay a foundation for accurate analysis. On this basis, multiple data truncations are implemented, and the fast Fourier transform is used for frequency domain analysis to obtain the peak information of the fault data, and the fault data is intercepted according to the peak information. Finally, a BP neural network model is constructed, the preprocessed fault data is input for network training, and the fault interval is quickly locked in combination with the line topology map, and relying on the traveling wave fault location algorithm, the fault point is accurately locked. It significantly improves the accuracy and efficiency of fault detection, builds a solid defense line for the safe and stable operation of the power grid, and effectively guarantees the continuous and reliable operation of the power system.

[0047] It can be understood that the DWT algorithm is the discrete wavelet transform algorithm.

[0048] In an embodiment of the present invention, in step S2, the combined use of the particle swarm optimization algorithm and the DWT algorithm for noise reduction preprocessing of the original data includes: performing denoising processing through the DWT algorithm, and the DWT algorithm decomposes the signal into approximation coefficients and detail coefficients of different scales. The DWT algorithm uses discrete wavelet functions and discrete time scales, and realizes the decomposition of the signal through filtering and downsampling operations, decomposing the original signal into a group of high-frequency sub-signals and low-frequency sub-signals; improving the Heaviside projection function in the DWT algorithm, and optimizing the basic Heaviside function with the sparse regularization Heaviside function;

[0049] It also includes adding a signal-to-noise ratio comparison module based on the PSO algorithm, and performing multiple iterative processes on the result processed by the improved DWT algorithm using the improved PSO algorithm.

[0050] In this technical solution, by adopting the DWT algorithm, the original signal can be accurately decomposed into approximation coefficients and detail coefficients of different scales, that is, using discrete wavelet functions and discrete time scales, through filtering and downsampling operations, the separation of high-frequency sub-signals and low-frequency sub-signals of the signal is realized, and the denoising effect is improved. On the basis of the DWT algorithm, the Heaviside projection function is innovatively improved, and the sparse regularization Heaviside function is introduced. Further enhanced the denoising effect. A signal-to-noise ratio comparison module is added within the PSO algorithm framework, and the result processed by the improved DWT algorithm is iteratively processed by the PSO algorithm to obtain the relative optimal solution during data preprocessing, and the accuracy of fault recognition is improved.

[0051] It is understandable that since the PSO algorithm is population-based, the initial values of the population size m, inertia weight w, acceleration constants c1 and c2, maximum velocity Vmax, maximum number of generations Gmax, etc. of the PSO algorithm are initially set and the algorithm process is executed to obtain an initial result, and its signal-to-noise ratio (SNR) is calculated; and it is compared with the SNR result after denoising by the improved DWT algorithm. If the difference between the SNR of the improved DWT algorithm and the SNR of the PSO algorithm is greater than a certain threshold, then this process is exited; otherwise, the six parameters of the PSO algorithm are iteratively processed using the iterative formula until the SNR meets the above conditions.

[0052] In step S2, it also includes: performing sparse regularization optimization on the Heaviside projection function in the DWT algorithm to improve the processing effect on non-linear and non-smooth data. During the denoising process in step S2, in-depth sparse regularization optimization is performed on the Heaviside projection function in the DWT algorithm. This enables the algorithm to exhibit more excellent performance when processing non-linear and non-smooth data. Through the introduction of sparse regularization, not only the accuracy and stability of the denoising process are improved, but also a clearer and more reliable data basis is provided for subsequent fault identification and location.

[0053] In step S3, the cumulative curve fitting approximation process includes: using a polynomial fitting algorithm to initially fit the discrete data, performing a secondary fitting through the CCFA algorithm, comparing the fNIRS data characteristics with the similar data characteristics collected by the hardware device, using the behavior of the approximate signal curve within the sliding window, obtaining the estimated curve from multiple overlapping windows, and adopting a weighted logic recombination method to receive the filtered signal to obtain the final fitting curve. The polynomial fitting algorithm is used to initially fit the original discrete data. Immediately afterwards, the advanced CCFA algorithm is used for secondary fitting, further improving the accuracy and stability of the fitting. The fNIRS data characteristics are introduced and compared with the similar data characteristics collected by the hardware device. Through data comparison and analysis, the data characteristics are obtained. Using the behavior of the approximate signal curve within the sliding window, the estimated curve is cleverly obtained from multiple overlapping windows, enriching the information content and reliability of the fitting curve. The weighted logic recombination method is adopted to receive the filtered signal. Through reasonable weight allocation and logical processing, the denoising effect of the data is improved, and the final fitting curve is obtained. This fitting curve not only accurately reflects the real situation of the distribution line fault, but also provides strong data support for subsequent fault location and identification, thus significantly improving the accuracy and practicality of the entire method.

[0054] It is understandable that the CCFA algorithm is a special curve fitting technique mainly used to reduce the distortion effect caused by the non-stationarity of fNIRS data.

[0055] Between steps S4 and S5, referring to Figure 2 , further includes: a processing step: iteratively operating on the operations in S4, and terminating the iterative operation when the peak information in the frequency domain after fast Fourier transform tends to be consistent, to obtain a final set of peak information.

[0056] It can be understood that the specific processing steps include: Processing step one: Set the initial truncation threshold to the termination truncation threshold, gradually decrease the initial truncation threshold (a fixed value), repeat the truncation processing operation, and store the period information in a certain period information set until the termination truncation threshold is reached, and terminate this operation; Processing step two: Process the obtained period information set, that is, perform k-means clustering operation on it to obtain its clustering centers. If the clustering centers are close, it is defaulted that this group of data is a periodic function, and any value is taken as the period of this group of data. On the contrary, if there are multiple clustering centers that are far apart, it is considered that this group of data may contain non-periodic components, but mainly sort by the size of the clustering centers, and take the largest three clustering centers as possible periods (in a sense, this may indicate that there are faults in the data). Use fast Fourier transform to perform frequency domain analysis on the relevant data. Through continuous iterative operations, the method gradually adjusts and optimizes the processing parameters, making the peak information in the frequency domain gradually tend to be consistent, thus effectively eliminating accidental errors and noise interference, and improving the reliability and stability of the data. When the iterative operation reaches the preset termination condition, that is, the peak information in the frequency domain shows a high degree of consistency, the method stops the iteration and outputs the final set of peak information. It not only accurately reflects the characteristics of the distribution line fault in the frequency domain, but also provides accurate data support for subsequent fault location and identification, thus further improving the accuracy and practicality of the entire method.

[0057] In step S4, the multiple truncation processing of the fitted continuous curve data includes: First, set a preliminary truncation threshold, perform fast Fourier transform on the truncated data, analyze the frequency components of the data in the truncated interval, obtain the peak information in the frequency domain, and obtain a set of peak information under this truncation threshold. By updating the truncation threshold and iterating the truncation threshold, through multiple iterative operations, a new set of peak information is obtained; After multiple iterative operations, if the peak information obtained after multiple iterations tends to be relatively stable and the frequency of occurrence is greater than the preset threshold, then these peak information are defined as period information. It not only improves the accuracy of fault identification, but also provides strong support for fault location and repair.

[0058] According to the operations in the processing steps, n - period continuous curve data is intercepted from the original data in step S3, and sampling processing is performed to obtain discrete waveform data. This ensures that the analyzed data segment has complete and representative periodic characteristics, laying a solid foundation for subsequent fault identification. Immediately afterwards, in order to convert these continuous curve data into a form more suitable for digital processing, further sampling processing is performed on these data. Through a scientific sampling strategy, the original continuous curve is accurately converted into a series of discrete waveform data points. These discrete waveform data not only retain the core characteristics of the original data, but also greatly simplify the subsequent data processing flow and improve the calculation efficiency.

[0059] In step S5, the construction of the BP neural network model includes: creating a training set and a test set, inputting them into the BP neural network for iterative training, obtaining the model parameters of the BP neural network and saving them. First, by creating a training set and a test set, both of these two data sets are derived from the discrete waveform data that has been pre - processed and feature - extracted in the previous steps. These two data sets are input into the constructed BP neural network. The BP neural network adjusts the internal weights and bias parameters of the network through multiple iterative trainings in order to achieve the best fitting effect and prediction accuracy. After multiple iterations and adjustments, when the performance of the BP neural network on the training set tends to be stable and meets the termination training conditions, the network model parameters at this time are extracted and saved. These parameters not only contain the structural information of the neural network, but also contain the knowledge of fault identification and location learned through a large amount of data, which is the key to subsequent rapid and accurate fault identification and location.

[0060] In step S5, it also includes: conducting actual test verification on the BP neural network model, and optimizing the neuron parameters, loss function, optimizer, initial weights, and network structure in the periodic information according to the test verification results to obtain the final waveform classification model. Not only is the BP neural network model established, but it is also further subjected to actual test verification to ensure the practicality and accuracy of the model. According to the feedback results of these test verifications, comprehensive and systematic optimization processing is carried out on the neuron parameters, loss function, optimization algorithm, initial weights, and network structure in the periodic information extraction.

[0061] In step S5, it further includes: when the classification result of the final waveform classification model is a fault defect, defect interval positioning is performed. The defect interval positioning includes: a short-circuit fault interval judgment algorithm and a single-phase ground fault interval judgment algorithm. When a fault defect in the distribution line is identified using the optimized final waveform classification model, the defect interval positioning program is immediately started. This program includes two core algorithms: a short-circuit fault interval judgment algorithm and a single-phase ground fault interval judgment algorithm. Through the comprehensive application of these two algorithms, the specific location where the fault occurs can be quickly and accurately locked, providing crucial guiding information for subsequent fault repair work, thereby greatly improving the efficiency and accuracy of distribution line fault handling.

[0062] It can be understood that the short-circuit fault interval judgment algorithm mainly uses the short-circuit current to judge the interval and supplements it with the change of the traveling wave head direction; the single-phase ground fault interval judgment algorithm mainly uses the change of the traveling wave polarity and the judgment of the grounding current. The network topology map constructed in step S2 will be supplemented in this stage.

[0063] It can be understood that for short-circuit fault interval judgment: when a short circuit occurs at a certain point, a large short-circuit current will appear in the direction of the power supply from this node, and there will be no short-circuit current flowing through the load direction of this node. Early fault indicators operate based on this principle to judge the short-circuit fault interval; as Figure 4 shown: When a short-circuit fault occurs in phases AB at point F1, then in the direction of the power supply from point F1, short-circuit currents appear in both phases AB, there is no short-circuit current in the load direction, and there is no short-circuit current in the non-fault phases. It should be noted that the interval refers to the line from the monitoring device to the power supply direction. If there is a short-circuit current in a certain device on the main line, there must be a short-circuit point in the load direction (both the main line and branches behind are possible); for single-phase ground faults, there are transmission and reflection of the single traveling wave in the line, and the following situations may exist for the same device:

[0064] 1. Receive the first traveling wave signal and report it;

[0065] 2. Only receive the reflected traveling wave of the first traveling wave signal and report it;

[0066] 3. Receive the first traveling wave signal, report it, and receive the first reflected traveling wave signal and report it;

[0067] For the above situations, the following features should be analyzed in combination:

[0068] As Figure 3As shown in the figure, for single-phase grounding fault section judgment, the traveling waves generated at a certain moment at the same fault point propagate towards both ends, and the signal amplitude attenuates with distance. For the equipment on the line from near to far, the traveling wave signal amplitude should gradually attenuate; otherwise, there may be a situation where a reflected traveling wave is received. When selecting the traveling wave signal, two traveling waves with opposite polarities and the largest signal amplitudes should be selected for section judgment (generally the equipment near and around the fault point), and according to the time difference at the "inflection point" position of the traveling wave, it is calculated whether it meets the time requirement for signal propagation; otherwise, it cannot be used as a condition for fault section judgment.

[0069] It can be understood that in step S1, the power frequency data (normal operation data) is sent once every certain period of time, while the high-frequency data (line fault data) is uploaded in real time; the data formats of both are the same, and the format is as follows:

[0070] Traveling wave data (this format data packet is continuously sent 40 times, and the total number of data points is (6400 * 40));

[0071] Voltage data (this format data packet is continuously sent 10 times, and the total number of data points is (6400 * 10));

[0072] Current data (this format data packet is continuously sent 10 times, and the total number of data points is (6400 * 10)).

[0073] It can be understood that in step S5, in the construction of the BP neural network, the BP neural network mainly consists of 1 input layer, 5 hidden layers, and 1 output layer, which specifically includes the following steps:

[0074] Input layer: 1280 neurons; the first hidden layer: 1536 neurons; the second hidden layer: 1792 neurons; the third hidden layer: 1920 neurons; the fourth hidden layer: 1792 neurons; the fifth hidden layer: 1536 neurons; output layer: n neurons, and the value of n is determined by the types of fault types.

[0075] It can be understood that the creation of the training set adopts the backpropagation algorithm. Backpropagation is an optimization algorithm that enables the neural network to learn the mapping relationship between input and output by continuously adjusting the connection weights between neurons in the network. Specifically, the backpropagation algorithm calculates the state and activation value of each layer, calculates the error from the last layer forward, and updates the parameters to minimize the error between the predicted output and the actual output of the network. This process will be iterated continuously until the stopping criterion is met.

[0076] It can be understood that in step S5, the determination of the fault interval by combining the classification result with the line topology diagram includes: after determining the fault interval, the processing of the fault point distance is carried out, which is mainly calculated according to some physical characteristics. Among them, the traveling wave velocity of the overhead line is about 290 m / us, and the traveling wave velocity of the power cable is about 170 m / us. There are two calculation methods. One is the single-ended traveling wave method: the fault is located by the time difference of the traveling wave traveling back and forth to the fault point. The other is the double-ended traveling wave method: the fault is located by the time difference of the traveling wave reaching both sides of the line.

[0077] In this step, it is necessary to accurately locate the wavehead positions of the traveling waves uploaded by two different devices. Here, the slope judgment method is used for calculation, and its judgment method is divided into two steps: The first step is to determine the range of the interval where the traveling wave changes violently, and intercept 1000 points before and after it. This step will mainly use a low-pass filter to process it and obtain the rough range of the violent fluctuation of the traveling wave. The second step is to start calculating the point where the wavehead is located after determining the rough range of the fluctuation, which is further as follows:

[0078] 1. Denoise and curve fit the data in this interval;

[0079] 2. Sample at the reporting frequency of the original reporting point of the data;

[0080] 3. Calculate the slope values of the previous point and the next point of the sampled data and judge them with a threshold. If it is greater than the threshold, it is a candidate point and recorded in the candidate set; if it is less than the threshold, no processing is done;

[0081] When all candidate sets in the two devices are determined, by comparing the data in the candidate sets, the earliest and the smallest difference data are obtained as the final time difference, and then the relevant fault point positions are calculated.

[0082] It can be understood that step S1 specifically includes the following steps:

[0083] Step 1.1, equipment installation. Before installation, it is necessary to fully understand the specific situation of the line to be installed, including but not limited to the following content:

[0084] 1. Line topology diagram. The topology diagram marks the line orientation, length, geographical location, switch position and type, main line and branch topology relationship, conductor type, cross-sectional area, main line and branch design load size, and whether it is a mixed connection line. The detailed topology diagram is an important basis for the line configuration plan.

[0085] 2. Span table. The span table marks the front and rear spans of the line poles. Not all units can provide it. Accurate spans can reduce the ranging error. When the span table cannot be provided, calculate according to an average span of 50 m.

[0086] 3. Line fault operating status. For example: the normal operating current of the main line and important branches of the line (determining whether the terminal can be online for a long time), fault frequency, fault type, fault-prone areas, etc.

[0087] Step 1.2: The terminal configuration principles are as follows:

[0088] 1. At the outlet of the substation line;

[0089] 2. At the connection of the cable line and the overhead line. In principle, both sides are installed if there is a cable line in the middle;

[0090] 3. If the main overhead line exceeds 5 km, an additional set of configurations is added;

[0091] 4. For T-connected lines, if the branch line is greater than 1.5 km, generally a set is installed at the first pole of the branch line;

[0092] 5. For lines with more users or higher load density, the installation quantity can be appropriately increased;

[0093] 6. Appropriately increase the configuration quantity for important loads on the line;

[0094] 7. Cross-region demarcation points;

[0095] 8. When installed on the same pole as the switch, it can be appropriately adjusted to the next pole;

[0096] Step 1.3: Determine the data upload protocol, the topic address for device upload, and data such as XOL / IMEI number / WAVEFORM_DATA.

[0097] It can be understood that the specific steps of step S2 are as follows: Step 2.1, perform denoising processing through an improved discrete wavelet transform algorithm. The improved DWT algorithm decomposes the signal into approximate coefficients and detail coefficients of different scales, that is, the improved DWT uses discrete wavelet functions and discrete time scales to achieve signal decomposition through filtering and downsampling operations. Decompose the discrete data signal into a group of high-frequency and low-frequency sub-signals. And calculate the corresponding signal-to-noise ratio data.

[0098] As Figure 1 shown, a terminal module is provided on the hardware device, and a timing module, a processor, a 4G module, an electric field module, a traveling wave / current module, and a lithium battery pack are provided on the terminal module.

[0099] Step 2.2, perform processing using the improved PSO algorithm, perform secondary denoising processing, and set the iteration termination condition. Perform multiple iterative processes, and always compare with the signal-to-noise ratio result in step 3.1. If the difference meets the termination condition, it means that the denoising processing is completed. Otherwise, modify the PSO initial parameters and continue to perform iterative processing on it.

[0100] It can be understood that step S3 specifically includes the following steps: Step 3.1, use the polynomial fitting algorithm to perform an initial fit on the results obtained in step 3. Here, an nth-degree polynomial is used, and the desired result is to approximately connect all point sets;

[0101] Step 3.2, use the Cumulative Curve Fitting Approximation (CCFA) algorithm to perform a fitting process on the results of step 3.1, and further perform denoising processing, and perform supplementary processing on the results of step S2.

[0102] It can be understood that the precise positioning of the fault point using the traveling wave fault location algorithm specifically includes the following steps:

[0103] Determine the range of the interval where the traveling wave changes violently, and intercept the 1000 points before and after it. This step will use a low-pass filter to process it and obtain the rough range of the traveling wave fluctuation;

[0104] Perform denoising and curve fitting on the data in this interval;

[0105] Sample at the reporting frequency of the original reporting points of the data;

[0106] By calculating the slope values of the previous point and the next point of the sampled data and comparing them with a threshold value, if it is greater than the threshold value, it is a candidate point and is recorded in the candidate set; if it is less than the threshold value, no processing is performed;

[0107] When all candidate sets in the two devices are determined, by comparing the data in the candidate sets, the earliest and smallest-difference data is obtained as the final time difference, and then the relevant fault point position is calculated. The calculation method is as follows:

[0108] Single-ended traveling wave method: Fault location is performed by the time difference of the traveling wave traveling back and forth to the fault point:

[0109] X = 1 / 2 (tmf - tm) * V;

[0110] Double-ended traveling wave method: Fault location is performed by the time difference of the traveling wave reaching both sides of the line:

[0111] 2X = L + Vtm - Vtn;

[0112] X = 1 / 2 + (tm - tn) V / 2;

[0113] Furthermore, the final fault location is obtained.

[0114] The above embodiments' specific description of the present invention is only used to further illustrate the present invention and cannot be understood as a limitation on the protection scope of the present invention. Those skilled in the art's non-essential improvements and adjustments made based on the content of the above invention all fall within the protection scope of the present invention.

Claims

1. A method for identifying and locating distribution line faults based on DWT and BP algorithm, characterized in that: The following steps are involved: S1, install the hardware equipment on different phases of the nodes of the same line in the distribution network line, collect and record the discrete data information of current, voltage and traveling wave in the overhead line, and upload it to the platform database through the MQTT protocol; S2, build the network topology of the line to be monitored according to the hardware device mounting points, and use the particle swarm optimization algorithm and DWT algorithm to pre-process the original data for noise reduction; S3, performing polynomial fitting and cumulative curve fitting approximation processing on the data after noise reduction preprocessing to obtain continuous curve data after fitting; S4, performing multiple truncation processing on the fitted continuous curve data, and analyzing the frequency components in the truncation interval by fast Fourier transform, and extracting the peak information in the frequency domain as the period information; S5, build a BP neural network model, input the preprocessed data into the network for training and classification recognition, judge the fault interval based on the classification results combined with the line topology map, and use the traveling wave fault distance measurement algorithm to locate the fault point; Improve the Heaviside projection function in the DWT algorithm and optimize the basic Heaviside function using the sparse regularized Heaviside function; It also includes adding a signal-to-noise ratio comparison module based on the PSO algorithm, and iterating the results processed by the improved DWT algorithm again; In step S3, the cumulative curve fitting approximation processing includes: using a polynomial fitting algorithm to perform preliminary fitting on discrete data, performing secondary fitting through a cumulative curve fitting approximation algorithm, comparing fNIRS data features with similar data features collected by the hardware device, using the behavior of the approximate signal curve in the sliding window, obtaining an estimated curve from multiple overlapping windows, and using a weighted logical recombination method to receive the filtered signal to obtain the final fitting curve.

2. A method for identifying and locating distribution line faults based on DWT and BP algorithm according to claim 1, characterized in that: In step S2, the combination of the particle swarm optimization algorithm and the DWT algorithm to perform noise reduction preprocessing on the original data includes: performing denoising processing through the DWT algorithm, the DWT algorithm decomposing the signal into approximate coefficients and detail coefficients of different scales, the DWT algorithm using discrete wavelet functions and discrete time scales, and realizing signal decomposition through filtering and downsampling operations, decomposing the original signal into a group of high-frequency sub-signals and low-frequency sub-signals.

3. A method for identifying and locating distribution line faults based on DWT and BP algorithm according to claim 2, characterized in that: In step S2, it also includes: performing sparse regularization optimization on the Heaviside projection function in the DWT algorithm to improve the denoising effect on nonlinear and non-smooth data.

4. A method for identifying and locating distribution line faults based on DWT and BP algorithm according to claim 1, 2 or 3, characterized in that: Between step S4 and step S5, a processing step is also included: performing an iterative operation on the operation in S4, and when the peak information in the frequency domain after the fast Fourier transform tends to be consistent, terminating the iterative operation to obtain a final peak information set.

5. A method for identifying and locating distribution line faults based on DWT and BP algorithm according to claim 4, characterized in that: In step S4, the multiple truncation processing of the fitted continuous curve data includes: preferentially setting a preliminary truncation threshold, performing fast Fourier transform on the truncated data, analyzing the frequency components of the data in the truncation interval, obtaining the peak information in the frequency domain, performing statistical processing on all the peak information, obtaining a preliminary peak information set, and iterating the truncation threshold, performing multiple iterations, and obtaining a new peak information set; after multiple iterations, if the peak information obtained after multiple iterations is maintained within a certain range and the frequency of occurrence is greater than a certain threshold, the peak information is defined as periodic information.

6. A method for identifying and locating distribution line faults based on DWT and BP algorithm according to claim 5, characterized in that: According to the operation in the processing step, the continuous curve data of n cycles are intercepted from the original data in step S3, and sampling processing is performed to obtain discrete waveform data.

7. A method for identifying and locating distribution line faults based on DWT and BP algorithm according to claim 6, characterized in that: In step S5, the building of the BP neural network model includes: creating a training set and a test set, inputting them into the BP neural network for iterative training, obtaining the model parameters of the BP neural network and saving them.

8. A method for identifying and locating distribution line faults based on DWT and BP algorithm according to claim 7, characterized in that: In step S5, it also includes: performing actual test verification on the BP neural network model, and optimizing the neuron parameters, loss function, optimizer, initialization weights and network structure according to the test verification results to obtain the final waveform classification model.

9. A method for identifying and locating distribution line faults based on DWT and BP algorithm according to claim 8, characterized in that: In step S5, it also includes: when the classification result of the final waveform classification model is a fault defect, defect interval positioning is performed: the defect interval positioning includes: a short circuit fault interval judgment algorithm and a single-phase grounding fault interval judgment algorithm.

Images