A Double-End Fault Location Method Based on Signal Enhancement

Abstract

The present invention relates to the technical field of data processing, and discloses a double - end fault location method based on signal enhancement, which is used to improve the efficiency of double - end fault location based on signal enhancement. The method includes: collecting and pre - processing the electrical signals at both ends of the transmission line to obtain the pre - processed synchronous electrical signals; performing adaptive enhancement processing on the pre - processed synchronous electrical signals to obtain the enhanced fault feature signals; performing time - frequency analysis and feature extraction on the enhanced fault feature signals to obtain the fault feature vectors and key parameters; performing machine learning analysis on the fault feature vectors and key parameters to obtain the fault type recognition results and preliminary fault location estimates; performing multi - algorithm fusion analysis on the preliminary fault location estimates to obtain the accurate fault location coordinates; performing geographic information mapping and field verification on the accurate fault location coordinates to obtain the fault location results and system optimization data.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular, to a dual - end fault location method based on signal enhancement. Background Art

[0002] In the operation and management of modern distribution networks, inspection, as an important link to ensure the normal operation of power equipment, has received increasing attention. Existing technologies usually adopt regular manual inspections and basic automatic monitoring means to conduct condition assessment and fault detection on the equipment in the distribution network. These methods collect the operation data and image information of the equipment and rely on traditional image - processing technologies and simple feature - extraction algorithms to achieve the detection of equipment abnormalities. However, with the increase in the types and quantities of distribution - network equipment, the traditional manual inspection method is inefficient and prone to missing problems, and cannot reflect the equipment status in a timely and comprehensive manner.

[0003] Although existing technologies have improved the inspection efficiency of distribution networks to a certain extent, there are still many deficiencies. On the one hand, traditional image - processing algorithms are sensitive to image noise in complex environments, resulting in poor image - data processing effects and making it difficult to accurately extract equipment features. On the other hand, existing abnormality - detection means lack intelligence and automation, cannot respond in real time to the dynamic changes of equipment failures, and cannot provide effective decision - making support. This makes the location and assessment of equipment failures have a certain lag, affecting the safe and stable operation of the distribution network. Therefore, there is an urgent need for a more intelligent and efficient inspection management method to solve the many problems existing in the existing technologies. Summary of the Invention

[0004] In view of this, an embodiment of the present invention provides a dual - end fault location method based on signal enhancement, which is used to improve the efficiency of dual - end fault location based on signal enhancement.

[0005] The present invention provides a dual - end fault location method based on signal enhancement, including: collecting and pre - processing the electrical signals at both ends of the transmission line to obtain pre - processed synchronous electrical signals; performing adaptive enhancement processing on the pre - processed synchronous electrical signals to obtain enhanced fault - feature signals; performing time - frequency analysis and feature extraction on the enhanced fault - feature signals to obtain fault - feature vectors and key parameters; performing machine - learning analysis on the fault - feature vectors and key parameters to obtain fault - type recognition results and preliminary fault - location estimates; performing multi - algorithm fusion analysis on the preliminary fault - location estimates to obtain accurate fault - location coordinates; performing geographic - information mapping and on - site verification on the accurate fault - location coordinates to obtain fault - location results and system - optimization data.

[0006] In the technical solution provided by the present invention, electrical signals at both ends of a transmission line are collected and preprocessed to obtain synchronized electrical signals after preprocessing, effectively eliminating noise and interference in the original signals and improving the data quality for subsequent analysis. The synchronized electrical signals after preprocessing are subjected to adaptive enhancement processing to obtain enhanced fault feature signals, highlighting the fault features and suppressing background noise, greatly improving the recognition rate of fault features. Subsequently, time-frequency analysis and feature extraction are performed on the enhanced fault feature signals to obtain fault feature vectors and key parameters, comprehensively capturing the time-domain and frequency-domain features of the fault signals and laying a foundation for accurate positioning. Machine learning analysis is carried out on the fault feature vectors and key parameters to obtain fault type recognition results and preliminary fault location estimates. This innovative application of machine learning technology realizes the automatic recognition and preliminary positioning of fault types, greatly improving the efficiency and accuracy of positioning. Multi-algorithm fusion analysis is performed on the preliminary fault location estimates to obtain accurate fault location coordinates, integrating the advantages of multiple positioning algorithms and significantly improving the positioning accuracy. Finally, geographical information mapping and on-site verification are carried out on the accurate fault location coordinates to obtain fault location results and system optimization data, not only verifying the accuracy of the positioning results but also continuously optimizing the system performance through a feedback mechanism. Overall, this method significantly improves the accuracy, efficiency, and reliability of fault location through the organic combination of technologies such as signal enhancement, feature extraction, machine learning, and multi-algorithm fusion. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for use in the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0008] Figure 1 It is a flowchart of a dual-terminal fault location method based on signal enhancement in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0009] The following will clearly and completely describe the technical solutions of the present invention in conjunction with the drawings. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments 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.

[0010] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be construed as a limitation of the present invention. In addition, the terms "first", "second", "third" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance.

[0011] In addition, the technical features involved in different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0012] For the convenience of understanding, the specific process of the embodiments of the present invention will be described below. Please refer to Figure 1 , Figure 1 which is a flowchart of a dual-terminal fault location method based on signal enhancement according to an embodiment of the present invention. As Figure 1 shown, it includes the following steps:

[0013] S101. Collect and preprocess the electrical signals at both ends of the transmission line to obtain the preprocessed synchronous electrical signals;

[0014] S102. Perform adaptive enhancement processing on the preprocessed synchronous electrical signals to obtain enhanced fault feature signals;

[0015] S103. Perform time-frequency analysis and feature extraction on the enhanced fault feature signals to obtain fault feature vectors and key parameters;

[0016] S104. Perform machine learning analysis on the fault feature vectors and key parameters to obtain fault type recognition results and preliminary fault location estimates;

[0017] S105. Perform multi-algorithm fusion analysis on the preliminary fault location estimates to obtain accurate fault location coordinates;

[0018] S106. Perform geographic information mapping and field verification on the accurate fault location coordinates to obtain fault location results and system optimization data.

[0019] It should be noted that voltage and current signals are simultaneously collected at both ends of the transmission line by high-frequency sampling equipment. The sampling frequency is usually set between several thousand hertz and several tens of thousand hertz to capture transient fault characteristics. After the collected original signals are marked with timestamps, wavelet denoising is performed to eliminate background noise, and then impulse interference is removed through adaptive median filtering. Outlier detection and interpolation processing are performed on the filtered signals to ensure the continuity and reliability of the data. Then, Fourier transform is performed on the processed signals to extract frequency-domain characteristics, and the signals in the target frequency band are retained through a band-pass filter. Finally, Hilbert transform is used for phase correction to ensure the time synchronization of the signals at both ends, and the preprocessed synchronous electrical signals are obtained.

[0020] First, the wavelet packet decomposition is used to decompose the signal into signal components of multiple scales. Then, the energy distribution analysis is carried out on these components, and the key components containing fault characteristics are selected. The selected components are subjected to sparse representation and singular value decomposition to extract the main feature components. Through adaptive gain adjustment, the fault characteristics are highlighted and the background noise is suppressed. Subsequently, the empirical mode decomposition is used to further decompose the signal, extract the intrinsic mode functions, and select the mode function most relevant to the fault through correlation analysis. Finally, an adaptive filter is designed to optimize these key mode functions, and the enhanced fault feature signal is reconstructed. The short-time Fourier transform is used to obtain the time-frequency spectrogram of the signal, and the characteristic frequency points are identified through peak detection. Then, the time-frequency spectrogram is segmented according to these characteristic points, and the energy density distribution of each region is calculated. The principal component analysis is carried out on the energy distribution characteristics, the main feature components are extracted and standardized. At the same time, the wavelet packet transform is used to perform multi-scale decomposition on the signal to extract statistical features. These features are fused to form the final fault feature vector, and key parameters such as the fault occurrence time and duration are calculated. The dimensionality reduction process is carried out on the feature vector to reduce data redundancy. Then, clustering analysis methods such as K-means clustering are used to group the dimensionality-reduced feature data. A decision tree model is constructed based on the clustering results, and the refined classification rules are obtained through pruning optimization. The support vector machine algorithm is used to train the optimized decision rules to obtain the fault type classifier. For location estimation, a neural network model is constructed, and the fault type recognition results and key parameters are used as inputs. The best model structure is selected through cross-validation. Finally, the particle swarm algorithm is used to optimize the parameters of the neural network model to improve the accuracy of location estimation. It combines the advantages of two traditional location methods, the traveling wave method and the impedance method. First, these two methods are respectively used to locate the fault, and then a Bayesian network model is constructed to preliminarily fuse the results of the two methods. The fuzzy inference method is used to consider various uncertain factors to obtain a fuzzy location interval. Subsequently, the genetic algorithm is used to search for the optimal solution within this interval to obtain a series of candidate location points. The Delphi method is used to conduct expert evaluation on these candidate points to give the weight coefficients. The candidate locations are weighted and averaged according to the weight coefficients, and the Kalman filter is used to further optimize the location estimation. Finally, the Monte Carlo simulation method is used to analyze the location error, calculate the confidence interval, and obtain the final accurate fault location coordinates. The fault location coordinates are converted into geographical coordinates and located in the geographic information system. The surrounding terrain features are analyzed according to the location results to plan the optimal inspection route. Field surveys are carried out along the planned route to collect on-site data and images. The computer vision algorithm is used to analyze the fault feature images to verify the accuracy of the fault location. If there is a deviation between the actual fault location and the calculation result, the fault location information is updated, and the cause of the deviation is analyzed. According to the analysis results, the parameters of the location algorithm are adjusted to optimize the system performance.Finally, integrate the corrected fault location and optimized parameters to form the final fault location result and system optimization data.

[0021] By performing the above steps, electrical signals at both ends of the transmission line are collected and preprocessed to obtain preprocessed synchronous electrical signals, effectively eliminating noise and interference in the original signals and improving the data quality for subsequent analysis. Adaptive enhancement processing is performed on the preprocessed synchronous electrical signals to obtain enhanced fault feature signals, highlighting the fault features and suppressing background noise, greatly improving the recognition rate of fault features. Subsequently, time-frequency analysis and feature extraction are performed on the enhanced fault feature signals to obtain fault feature vectors and key parameters, comprehensively capturing the time-domain and frequency-domain features of the fault signals and laying a foundation for accurate location. Machine learning analysis is performed on the fault feature vectors and key parameters to obtain fault type recognition results and preliminary fault location estimates. This innovative application of machine learning technology realizes the automatic recognition and preliminary location of fault types, greatly improving the efficiency and accuracy of location. Multi-algorithm fusion analysis is performed on the preliminary fault location estimates to obtain accurate fault location coordinates, integrating the advantages of multiple location algorithms and significantly improving the accuracy of location. Finally, geographic information mapping and on-site verification are performed on the accurate fault location coordinates to obtain the fault location result and system optimization data, not only verifying the accuracy of the location result but also continuously optimizing the system performance through the feedback mechanism. Overall, this method significantly improves the accuracy, efficiency, and reliability of fault location through the organic combination of technologies such as signal enhancement, feature extraction, machine learning, and multi-algorithm fusion.

[0022] In a specific embodiment, the process of performing step S101 may specifically include the following steps:

[0023] (1) Perform high-frequency sampling on the voltage and current signals at both ends of the transmission line to obtain original electrical signal data, and mark the original electrical signal data with timestamps to obtain timestamped original signal data;

[0024] (2) Perform wavelet denoising on the timestamped original signal data to obtain a preliminary denoised signal, and perform adaptive median filtering on the preliminary denoised signal to obtain a filtered signal;

[0025] (3) Perform outlier detection on the filtered signal to obtain signal data without outliers, and perform interpolation on the signal data without outliers to obtain continuous signal data;

[0026] (4) Perform Fourier transform on the continuous signal data to obtain frequency-domain feature data, and perform band-pass filtering on the frequency-domain feature data to obtain target band signals;

[0027] (5) Perform phase correction on the target frequency band signal through Hilbert transform to obtain the signal after phase correction, and perform time synchronization processing on the signal after phase correction to obtain the electrically synchronized signal at both ends;

[0028] (6) Perform data compression on the electrically synchronized signal at both ends to obtain the compressed synchronized signal, and perform quantization encoding on the compressed synchronized signal to obtain the preprocessed synchronized electrical signal.

[0029] Specifically, a high-precision digital sampling device is used. Usually, the sampling frequency is set between 10 kHz and 100 kHz to ensure that transient fault characteristics can be captured. The original electrical signal data obtained by sampling is immediately marked with timestamps. This step is crucial because it lays the foundation for subsequent signal synchronization processing. The accuracy of the timestamps usually reaches the microsecond level to ensure the precise alignment of the signals at both ends. Next, wavelet denoising processing is performed on the original signal data with timestamps. Wavelet denoising is an effective signal processing technique that can remove background noise while retaining the important features of the signal. The specific operations include selecting an appropriate wavelet basis function, decomposing the signal by wavelet, setting a threshold to remove the noise coefficients, and then reconstructing the signal. After obtaining the preliminary denoised signal, adaptive median filtering is performed. Adaptive median filtering can effectively remove impulse interference while protecting the edge information of the signal. The size of the filtering window will be dynamically adjusted according to the local signal characteristics to achieve the best filtering effect.

[0030] Outlier detection is also required for the filtered signal. Outlier detection aims to identify and remove abnormal data points, which may be caused by equipment failures or transient interferences. Common detection methods include statistical-based methods (such as the 3σ criterion) or density-based methods (such as the DBSCAN algorithm). After detecting and removing the outliers, interpolation processing is performed on the data to ensure the continuity of the signal. The interpolation method can be linear interpolation, spline interpolation, or more complex algorithms. The specific choice depends on the characteristics and accuracy requirements of the signal. Performing Fourier transform on the continuous signal data is to obtain the frequency domain characteristics of the signal. Fast Fourier transform (FFT) is usually used for this purpose, which can efficiently convert the time domain signal to the frequency domain. In the frequency domain, different types of faults will exhibit different characteristic frequencies. After obtaining the frequency domain characteristic data, a band-pass filter is applied to retain the target frequency band signal. The parameter settings of the band-pass filter are based on prior knowledge of the fault characteristic frequencies and usually retain the frequency bands containing fault information (such as power frequency and its harmonics, fault transient frequencies, etc.) while filtering out high-frequency noise and low-frequency interference.

[0031] The Hilbert transform is used to perform phase correction on the signals in the target frequency band. The Hilbert transform can obtain the analytic representation of the signal, thereby separating the amplitude and phase information of the signal. The purpose of phase correction is to eliminate the phase deviation caused by factors such as transmission line characteristics and measurement equipment, and ensure the phase consistency of the signals at both ends. After phase correction, time synchronization processing is performed on the signals. Time synchronization is the key to the two-terminal positioning method. It aligns the signals by comparing the timestamps of the signals at both ends and combining the line parameters (such as line length, propagation speed, etc.). Finally, data compression and quantization coding are performed on the synchronized electrical signals at both ends. Data compression aims to reduce the amount of data and improve the transmission and storage efficiency. Commonly used compression methods include wavelet compression, compressive sensing, etc. The compressed signal is subjected to quantization coding, which maps the continuous signal values to discrete digital codes. The quantization process needs to balance data accuracy and storage space, and usually adopts non-uniform quantization methods to reduce the amount of data while ensuring the accuracy of key information.

[0032] For example: A single-phase grounding fault occurs on a 330 kV transmission line, and the fault location devices installed at both ends of the line are started simultaneously. The sampling frequency is set to 20 kHz, and voltage and current data for 1 second are collected, obtaining 20,000 sampling points. Each sampling point is attached with a timestamp accurate to the microsecond. The original data is decomposed by db4 wavelet for 5 layers for denoising, and the noise threshold is set to 0.05, removing approximately 90% of the background noise. The adaptive median filter window size is dynamically adjusted between 3 and 7, successfully filtering out 10 pulse interferences caused by switch operations. The 3σ criterion is used for outlier detection, identifying and removing 5 abnormal data points, and then cubic spline interpolation is used to fill the data gaps. The processed 19,995 data points are subjected to a 2048-point FFT to obtain the spectrum from 0 to 10 kHz. A band-pass filter is designed with passbands of 45 - 55 Hz and 400 - 600 Hz, retaining the power frequency signal and fault transient characteristics. After the Hilbert transform, a 2-degree phase difference is found between the signals at both ends and corrected. Finally, the data volume is compressed to 30% of the original using the wavelet compression algorithm, and 8-bit non-uniform quantization is used for coding. The entire preprocessing process takes approximately 50 milliseconds.

[0033] In a specific embodiment, the process of performing step S102 may specifically include the following steps:

[0034] (1) Perform wavelet packet decomposition on the preprocessed synchronized electrical signals to obtain multi-scale signal components, and perform energy distribution analysis on the multi-scale signal components to obtain an energy feature vector;

[0035] (2) Perform adaptive threshold selection on the multi-scale signal components according to the energy feature vector to obtain candidate feature components, and perform sparse representation on the candidate feature components to obtain a sparse coefficient matrix;

[0036] (3) Perform singular value decomposition on the sparse coefficient matrix to obtain the main eigencomponents, and perform adaptive gain adjustment on the main eigencomponents to obtain the signal after gain adjustment;

[0037] (4) Decompose the signal after gain adjustment through empirical mode decomposition to obtain an ensemble of intrinsic mode functions, and perform correlation analysis on the ensemble of intrinsic mode functions to obtain the key mode functions;

[0038] (5) Design an adaptive filter for the key mode functions to obtain optimized filter parameters, and perform filtering on the key mode functions using the optimized filter parameters to obtain the filtered mode functions;

[0039] (6) Reconstruct the filtered mode functions to obtain the enhanced fault feature signal.

[0040] Specifically, wavelet packet decomposition is an extension of wavelet transform. It also subdivides the high-frequency part of the signal, providing a finer frequency division. In specific implementation, an appropriate wavelet basis function (such as Daubechies wavelet) and the decomposition level are selected to decompose the signal into sub-signals of different frequency bands. Perform energy distribution analysis on these multi-scale signal components, calculate the energy value of each component, and form an energy feature vector. The energy feature vector reflects the energy distribution of the signal in different frequency bands, which helps to identify fault features. According to the energy feature vector, an adaptive threshold selection method is used to screen out the candidate feature components containing fault information. The setting of the adaptive threshold takes into account the overall energy distribution and local characteristics of the signal, and usually uses methods based on statistical characteristics or information entropy to determine. The selected candidate feature components are then subjected to sparse representation, which is a method of representing a signal as a linear combination of a small number of basic elements. Sparse representation solves an optimization problem to find a small number of atoms in a predefined dictionary that can best represent the signal features, obtaining a sparse coefficient matrix. Sparse representation helps to highlight the essential features of the signal and suppress noise and interference.

[0041] Performing singular value decomposition on a sparse coefficient matrix is an effective method for extracting the main characteristic components. Singular value decomposition decomposes the matrix into the product of three matrices, where the singular values reflect the importance of the characteristics. By retaining the larger singular values and their corresponding singular vectors, the main characteristic components are reconstructed. Adaptive gain adjustment is performed on these main characteristic components, aiming to highlight the fault characteristics and suppress the background noise. The amplitude of the gain adjustment is dynamically adjusted according to the local characteristics of the signal to avoid over-amplifying the noise or suppressing useful information. Empirical mode decomposition is an adaptive signal processing method, especially suitable for the analysis of non-linear and non-stationary signals. It decomposes the signal into a series of intrinsic mode functions (IMFs), and each IMF represents an inherent oscillation mode of the signal. Empirical mode decomposition is performed on the signal after gain adjustment to obtain a set of IMFs. Subsequently, correlation analysis is performed on these IMFs, the correlation coefficient between each IMF and the original signal is calculated, and the IMFs with higher correlation are selected as the key mode functions. These key mode functions usually contain the most important fault characteristic information.

[0042] Designing an adaptive filter for the key mode functions is an important step in further optimizing the signal. The adaptive filter can automatically adjust its parameters according to the characteristics of the input signal to achieve the best filtering effect. Commonly used adaptive filtering algorithms include the least mean square error (LMS) algorithm and the recursive least squares (RLS) algorithm. Through these algorithms, the optimal filter parameters are determined, and then these parameters are used to filter the key mode functions to further improve the signal-to-noise ratio and obtain the filtered mode functions. Finally, the filtered mode functions are reconstructed to obtain the enhanced fault characteristic signal. The reconstruction process usually directly adds the selected filtered mode functions or uses weighted summation, and the weights can be determined based on the energy or correlation of each mode function. The reconstructed signal retains the key fault characteristics in the original signal while significantly reducing the influence of noise and interference.

[0043] For example: A single-phase ground fault occurs on a 500 kV transmission line, and the length of the preprocessed synchronous electrical signal collected is 10,000 points. First, use the db4 wavelet basis to perform 5-layer wavelet packet decomposition on the signal to obtain signal components in 32 frequency bands. Calculate the energy of each component to form a 32-dimensional energy feature vector. Set the energy threshold to 1% of the total energy, and select 10 candidate feature components with energy exceeding the threshold. Perform sparse representation on these 10 components, use an overcomplete dictionary with 100 atoms, and solve to obtain the sparse coefficient matrix. Perform singular value decomposition on this matrix, and retain the feature components corresponding to the first 5 largest singular values. Adaptive gain adjustment uniformly adjusts the peaks of these 5 components to 80% of the peak of the original signal. Empirical mode decomposition yields 8 IMFs, and 3 key IMFs with a correlation coefficient greater than 0.6 are selected through correlation analysis. Use the LMS algorithm to design a 50th-order adaptive FIR filter to filter these 3 IMFs. Finally, reconstruct the weighted sum of the 3 filtered IMFs, and the weights are determined according to the energy ratio of each IMF. The entire signal enhancement process takes about 100 milliseconds. The fault features in the enhanced signal are more obvious, and the signal-to-noise ratio is increased by about 15 dB, laying a solid foundation for subsequent fault location analysis.

[0044] In a specific embodiment, the process of executing step S103 may specifically include the following steps:

[0045] (1) Perform short-time Fourier transform on the enhanced fault feature signal to obtain a time-frequency spectrogram, and perform peak detection on the time-frequency spectrogram to obtain a set of characteristic frequency points;

[0046] (2) Perform region segmentation on the time-frequency spectrogram according to the set of characteristic frequency points to obtain time-frequency feature regions, and calculate the energy density of the time-frequency feature regions to obtain energy distribution characteristics;

[0047] (3) Perform principal component analysis on the energy distribution characteristics to obtain main feature components, and perform normalization processing on the main feature components to obtain a normalized feature vector;

[0048] (4) Perform multi-scale decomposition on the enhanced fault feature signal through wavelet packet transform to obtain multi-scale coefficients, and extract statistical features from the multi-scale coefficients to obtain a set of statistical features;

[0049] (5) Perform correlation analysis on the set of statistical features to obtain key statistical features, and combine the key statistical features to obtain a statistical feature vector;

[0050] (6) Fuse the normalized feature vector and the statistical feature vector to obtain a fault feature vector, and calculate the fault occurrence time and duration according to the fault feature vector to obtain a fault feature vector and key parameters.

[0051] Specifically, the short-time Fourier transform divides the signal into small time windows and performs Fourier transform on each window, thereby obtaining the frequency information of the signal varying with time. A time-frequency spectrogram is generated, which intuitively shows the frequency components of the signal at different time points. Peak detection is performed on the time-frequency spectrogram to identify the characteristic frequency points where the energy is concentrated, and these points usually correspond to fault characteristics. Peak detection uses a local maximum search algorithm and sets an appropriate threshold to filter out weak peaks, obtaining a set of characteristic frequency points. Based on the set of characteristic frequency points, the time-frequency spectrogram is regionally segmented, dividing the time-frequency plane into multiple characteristic regions. The segmentation method can adopt threshold-based segmentation or clustering algorithms such as K-means clustering. The segmented time-frequency characteristic regions reflect the distribution characteristics of the fault signal in the time-frequency domain. The energy density is calculated for each time-frequency characteristic region to obtain the energy distribution characteristics. The energy density calculation considers the ratio of the sum of the squares of the signal amplitudes in the region to the region area, reflecting the concentration degree of the fault energy on the time-frequency plane.

[0052] Performing principal component analysis on the energy distribution characteristics is an effective method for dimensionality reduction and extracting main features. Principal component analysis transforms the possibly correlated variables into linearly independent variables, namely principal components, through orthogonal transformation. The first few principal components with larger contribution rates are selected as the main feature components, which not only retain the main information of the original data but also reduce the data dimension. The main feature components are standardized to convert the features with different dimensions into a unified scale, obtaining a normalized feature vector. The z-score method is usually adopted for standardization to ensure that different features have the same weight in subsequent analysis. The enhanced fault characteristic signal is subjected to multi-scale decomposition through wavelet packet transform, providing a detailed representation of the signal in the time and frequency domains. Wavelet packet transform is an extension of wavelet transform, which also subdivides the high-frequency part of the signal, providing a finer frequency division. The coefficients obtained from multi-scale decomposition reflect the characteristics of the signal at different scales and frequency bands. Statistical feature extraction is performed on these multi-scale coefficients, including calculating statistics such as mean, variance, skewness, and kurtosis, obtaining a set of statistical features. These statistical features capture the distribution characteristics of the signal at different scales.

[0053] Perform a correlation analysis on the statistical feature set, screen out the features highly correlated with the fault type and location, and obtain the key statistical features. The correlation analysis can adopt methods such as Pearson correlation coefficient or mutual information. Select features with high correlation and low correlation with each other to avoid information redundancy. Combine the selected key statistical features to form a statistical feature vector. The combination method can be simple splicing or weighted summation, and the weights can be determined according to the importance of each feature. Fuse the normalized feature vector and the statistical feature vector to obtain the final fault feature vector. The fusion method can adopt feature-level fusion or decision-level fusion. Feature-level fusion directly splices the two vectors, and decision-level fusion classifies the two vectors separately and then synthesizes the results. The fused feature vector contains the comprehensive information of the time-frequency domain and the statistical domain. According to the fused fault feature vector, by setting thresholds or using machine learning models, the occurrence time and duration of the fault can be accurately judged, and these, as key parameters, are crucial for fault location.

[0054] For example: A two-phase short-circuit fault occurs in a 220 kV transmission line, and the length of the enhanced fault feature signal collected is 20,000 points, with a sampling frequency of 10 kHz. First, perform a short-time Fourier transform using a Hanning window, with a window length of 512 points and an overlap rate of 50% to obtain a time-frequency spectrogram. Perform peak detection on the time-frequency spectrogram, set the energy threshold to 3 times the average energy, and detect 15 characteristic frequency points. Based on these points, use the K-means algorithm (K = 5) to segment the time-frequency spectrogram. Calculate the energy density of each region to obtain a 5-dimensional energy distribution feature. Perform principal component analysis on the energy distribution feature, select the first 3 principal components with a cumulative contribution rate reaching 95% as the main feature components, and perform z-score normalization to obtain a normalized feature vector. At the same time, use the db4 wavelet basis to perform 4-layer wavelet packet decomposition on the signal to obtain the coefficients of 16 frequency bands. Calculate the mean, variance, skewness, and kurtosis for each frequency band to form a 64-dimensional statistical feature set. By calculating the Pearson correlation coefficient, select 10 features with an absolute value of the correlation coefficient greater than 0.7 as the key statistical features. Directly splice the normalized feature vector (3-dimensional) and the key statistical features (10-dimensional) to obtain a 13-dimensional final fault feature vector. According to the changes in energy mutation and frequency features in the feature vector, judge that the fault occurs at the 8732nd sampling point (corresponding to the actual time of 0.8732 seconds) and the duration is 0.15 seconds. The entire feature extraction and analysis process takes about 200 milliseconds, laying a foundation for subsequent accurate fault location.

[0055] In a specific embodiment, the process of executing step S104 may specifically include the following steps:

[0056] (1) Perform dimensionality reduction on the fault feature vector to obtain the dimensionality-reduced feature data, and perform clustering analysis on the dimensionality-reduced feature data to obtain feature clusters;

[0057] (2) Construct a decision tree model based on the feature clusters to obtain preliminary classification rules, and perform pruning optimization on the preliminary classification rules to obtain optimized decision rules;

[0058] (3) Train the optimized decision rules through the support vector machine algorithm to obtain a fault type classifier, and use the fault type classifier to classify the new fault feature vector to obtain the fault type recognition result;

[0059] (4) Construct a neural network model based on the fault type recognition result and key parameters to obtain a position estimation model, and perform cross-validation on the position estimation model to obtain a candidate position estimation model;

[0060] (5) Optimize the parameters of the candidate position estimation model using the particle swarm algorithm to obtain an optimized position estimation model, and process the fault feature vector through the optimized position estimation model to obtain a preliminary fault position estimation.

[0061] Specifically, common dimensionality reduction methods include principal component analysis (PCA) and linear discriminant analysis (LDA). PCA transforms potentially correlated variables into linearly independent variables through orthogonal transformation, retaining the principal components with larger variance contribution rates. LDA, on the other hand, considers class information and seeks the projection direction that maximizes the between-class variance and minimizes the within-class variance. The dimensionality-reduced feature data retains the main information of the original data while greatly reducing the data dimension. For clustering analysis of the dimensionality-reduced feature data, the K-means or hierarchical clustering algorithm is usually adopted to group similar fault features into one category to form feature clusters. Clustering analysis helps to discover the internal structure in the data and provides a basis for subsequent fault type recognition. Constructing a decision tree model based on feature clusters is an intuitive and efficient classification method. The decision tree classifies data through a series of if-then rules, where each internal node represents a feature test and each leaf node represents a category. Common decision tree algorithms include ID3, C4.5, and CART. During the construction process, the algorithm selects the best splitting feature according to the information gain or Gini coefficient and recursively partitions the data set. The obtained preliminary classification rules may be too complex and there is a risk of overfitting. Therefore, pruning optimization is required, including pre-pruning and post-pruning. Pre-pruning restricts the tree growth process, while post-pruning simplifies the fully grown tree. The goal of pruning is to reduce the complexity of the tree while maintaining the classification accuracy and improving the generalization ability of the model.

[0062] By training the optimized decision rules using the Support Vector Machine (SVM) algorithm, a more robust fault type classifier can be obtained. The core idea of SVM is to construct an optimal separating hyperplane in a high-dimensional space to maximize the margin between samples of different classes. For the case of non-linearly separable data, SVM maps the data into a high-dimensional space through the kernel trick. Commonly used kernel functions include linear kernel, polynomial kernel, and Radial Basis Function (RBF) kernel. The training process of SVM involves solving a quadratic programming problem to obtain the support vectors and the decision boundary. After training, the fault type classifier is used to classify the new fault feature vectors to obtain the fault type recognition results. Based on the fault type recognition results and key parameters, a neural network model is constructed for fault location estimation. The neural network model has a powerful non-linear mapping ability and is suitable for dealing with complex fault location estimation problems. Commonly used neural network structures include Multi-Layer Perceptron (MLP) and Long Short-Term Memory network (LSTM). During the construction process, it is necessary to determine the network structure (such as the number of layers, the number of neurons in each layer), activation functions (such as ReLU, sigmoid), and loss functions (such as mean squared error). The network is trained using the backpropagation algorithm, and the network parameters are optimized by the gradient descent method. To evaluate the generalization ability of the model, cross-validation is performed on the location estimation model. A commonly used method is K-fold cross-validation, where the dataset is divided into K parts, and K - 1 parts are used as the training set and 1 part as the validation set in turn. Through cross-validation, multiple candidate location estimation models can be obtained.

[0063] The Particle Swarm Optimization (PSO) algorithm is used to optimize the parameters of the candidate location estimation model to further improve the model performance. PSO is a swarm intelligence optimization algorithm that searches for the optimal solution by simulating the foraging behavior of bird flocks. In this solution, PSO is used to optimize the hyperparameters of the neural network, such as the learning rate, the number of hidden layers, the number of neurons in each layer, etc. Each particle represents a combination of hyperparameters. By iteratively updating the position and velocity of the particles, the parameter combination that makes the model performance optimal is searched. During the optimization process, an appropriate fitness function (such as the mean squared error on the validation set) is defined to evaluate the quality of the parameter combination. Finally, the optimized location estimation model is obtained, which has better generalization ability and prediction accuracy.

[0064] For example: A single-phase ground fault occurs on a 330 kV transmission line, and the collected fault feature vector is 50-dimensional. First, PCA is used for dimensionality reduction, retaining the principal components with a cumulative variance contribution rate of 95%, and reducing the feature vector to 15 dimensions. The K-means algorithm is used to cluster the data after dimensionality reduction, with K = 5 set, resulting in 5 feature clusters. Based on these feature clusters, a decision tree is constructed using the CART algorithm, with an initial tree depth of 10. The optimal pruning parameter is determined through cross-validation, reducing the tree depth to 6 to obtain the optimized decision rule. The SVM with an RBF kernel is used to train the decision rule, with the kernel parameter γ = 0.1 and the penalty parameter C = 10, achieving a classification accuracy of 98% on the validation set. Based on the SVM classification results and key parameters such as the fault occurrence time and duration, a 3-layer MLP neural network is constructed, with the number of neurons in the hidden layers being 20 and 10 respectively, and the ReLU activation function is used. Five-fold cross-validation is adopted to obtain 5 candidate location estimation models. The PSO is used to optimize the neural network hyperparameters, with the number of particles set to 30 and the number of iterations to 100. The search range includes the learning rate (0.001 - 0.1) and the number of neurons in the hidden layer (10 - 50). The average absolute error of the optimized best model on the test set is 0.5% of the total line length. The new fault feature vector is input into the optimized model, and the fault location is estimated to be 68.3 km from the starting point of the line. The entire fault type recognition and location estimation process takes about 300 milliseconds, providing strong support for rapid and accurate fault location.

[0065] In a specific embodiment, the process of executing step S105 may specifically include the following steps:

[0066] (1) Perform traveling wave method analysis on the preliminary fault location estimation to obtain the traveling wave method location estimation, and perform impedance method analysis on the preliminary fault location estimation to obtain the impedance method location estimation;

[0067] (2) Construct a Bayesian network model based on the traveling wave method location estimation and the impedance method location estimation to obtain the preliminary fusion result, and perform fuzzy reasoning on the preliminary fusion result to obtain the fuzzy location interval;

[0068] (3) Optimize the fuzzy location interval using the genetic algorithm to obtain the candidate location point set, and evaluate the candidate location point set using the Delphi method to obtain the weight coefficients;

[0069] (4) Perform weighted averaging on the candidate location point set according to the weight coefficients to obtain the weighted location estimation, and perform Kalman filtering on the weighted location estimation to obtain the filtered location estimation;

[0070] (5) Perform error analysis on the filtered location estimation through Monte Carlo simulation to obtain the error distribution, and calculate the confidence interval according to the error distribution to obtain the accurate fault location coordinates.

[0071] Specifically, in the double - end fault location method based on signal enhancement, analyzing the initial fault location estimate by the traveling - wave method and the impedance method is a key step to improve the location accuracy. The traveling - wave method locates faults by using the propagation characteristics of high - frequency electromagnetic waves generated by faults on the line. In specific implementation, first, wavelet transform is performed on the high - frequency signals collected at both ends to identify the arrival time of the initial fault traveling wave. Then, according to the traveling - wave propagation speed and the time difference of the traveling - wave arrival at both ends, the distances from the fault point to both ends of the line are calculated. The impedance method locates faults based on the relationship between the fault - point impedance and the impedance per unit length of the line. During the implementation process, by measuring the voltage and current at the fault moment, the equivalent impedance of the fault point is calculated, and then the fault distance is deduced in combination with the line parameters. A Bayesian network model is constructed based on the position estimation results of the traveling - wave method and the impedance method to achieve the probability fusion of the two methods. The Bayesian network is a reasoning method based on the probabilistic graphical model, which can handle uncertainties and complex dependencies. In this solution, the nodes of the Bayesian network include the results of the traveling - wave method, the results of the impedance method, and the actual fault location, and the relationships between the nodes are described by the conditional probability table. The network parameters are trained using historical data to obtain the preliminary fusion result. Subsequently, fuzzy reasoning is performed on the preliminary fusion result, introducing expert knowledge and experience rules. The fuzzy reasoning process includes three steps: fuzzification, fuzzy rule reasoning, and defuzzification, and finally a fuzzy position interval is obtained. Optimizing the fuzzy position interval by the genetic algorithm is to find the optimal set of candidate position points within the fuzzy interval. The genetic algorithm simulates the biological evolution process and continuously optimizes the population of solutions through selection, crossover, and mutation operations. In this solution, each chromosome represents a possible fault location, and the fitness function is designed based on the errors of the traveling - wave method and the impedance method. Through multiple generations of evolution, a set of high - quality candidate position points is obtained. These candidate position points are evaluated by the Delphi method, integrating the opinions of multiple experts. The Delphi method is a structured group communication technique, and through multiple rounds of anonymous questionnaires, expert consensus is gradually formed. Each expert scores the candidate positions according to their own experience and the provided data, and after multiple rounds of iteration, the weight coefficients of each candidate position are finally obtained.

[0072] The candidate location point set is weighted and averaged according to the weight coefficients to obtain a comprehensive location estimate. The credibility of each candidate location is fully considered, making the final result more reliable. Kalman filtering is performed on the weighted location estimate to further improve the positioning accuracy. Kalman filtering is a recursive state estimation algorithm that can optimally estimate the system state in the presence of measurement noise. In this solution, Kalman filtering uses historical data and current measurement values to continuously update the estimate of the fault location and obtain the filtered location estimate. Error analysis of the filtered location estimate is carried out through Monte Carlo simulation, which can comprehensively evaluate the uncertainty of the positioning result. The Monte Carlo method simulates the behavior of complex systems through a large number of random samplings. In this solution, according to the statistical characteristics of measurement errors and model parameters, a large number of samples are generated, and the fault location of each sample is calculated. By analyzing these simulation results, the error distribution of the fault location estimate is obtained. Based on the error distribution, the confidence interval is calculated, and a statistically significant fault location range is given. This confidence interval not only provides the exact fault location coordinates but also quantifies the reliability of the positioning result.

[0073] For example: A 500 kV transmission line is 300 km long and a single-phase ground fault occurs. The preliminary fault location estimate is at 180 km from the starting point. Traveling wave method analysis is performed on this location. By performing wavelet transform on the high-frequency signals at both ends, the arrival times of the initial traveling waves of the fault are identified as t1 = 0.6 ms and t2 = 0.8 ms respectively. Considering the traveling wave propagation speed v = 290 m / μs, the traveling wave method location estimate is calculated to be 175.8 km. At the same time, using the impedance method, the equivalent impedance of the fault point is measured to be 45 + j60 Ω. Combining with the line parameters (impedance per kilometer is 0.25 + j0.32 Ω / km), the impedance method location estimate is calculated to be 183.2 km.

[0074] Build a Bayesian network model to fuse the results of the traveling wave method and the impedance method. After training, the network gives a preliminary fusion result of 179.5 km. Fuzzy inference is carried out on this result, considering line parameter errors and measurement errors, and a fuzzy position interval of [177 km, 182 km] is obtained. The genetic algorithm is used to optimize within this interval, with the population size set to 100 and the number of evolutionary generations to 50, resulting in 10 candidate position points. Through the Delphi method, 5 experts evaluate these 10 points in 3 rounds, and finally obtain the weight coefficients. The weighted average is performed according to the weight coefficients, and the weighted position estimate is 179.8 km. Kalman filtering is carried out on this estimate, considering historical fault data and current measurement values, and the filtered position estimate is 179.6 km. Finally, 10,000 Monte Carlo simulations are carried out to generate the error distribution. The analysis results show that the 95% confidence interval is [179.2 km, 180.0 km]. Therefore, the final accurate fault location coordinate is determined to be 179.6 km, with a confidence interval of ±0.4 km. The entire multi-algorithm fusion analysis process takes about 500 ms, significantly improving the accuracy and reliability of fault location.

[0075] In a specific embodiment, the process of executing step S106 may specifically include the following steps:

[0076] (1) Perform geocoding conversion on the accurate fault location coordinate to obtain the geographical coordinate, and perform positioning in the geographic information system according to the geographical coordinate to obtain the geographical location of the fault point;

[0077] (2) Perform terrain analysis on the geographical location of the fault point to obtain terrain feature data, and perform path planning according to the terrain feature data to obtain the target inspection route;

[0078] (3) Conduct on-site inspection along the target inspection route to obtain on-site data, and perform image processing on the on-site data to obtain the fault feature image;

[0079] (4) Analyze the fault feature image through computer vision algorithms to obtain the fault verification result, and update the fault location information according to the fault verification result to obtain the corrected fault location;

[0080] (5) Conduct deviation analysis on the corrected fault location and the initial accurate fault location coordinate to obtain the positioning error data, and adjust the parameters of the positioning algorithm according to the positioning error data to obtain the optimized parameters;

[0081] (6) Integrate the optimized parameters and the corrected fault location to obtain the fault location result and the system optimization data.

[0082] Specifically, a dedicated coordinate conversion algorithm is usually used to convert the distance points on the line into longitude and latitude coordinates. When converting, the earth's curvature and projection method need to be considered to ensure the conversion accuracy. After obtaining the geographical coordinates, they are located in a Geographic Information System (GIS), mapping the abstract coordinate points to the actual geographical environment to obtain the specific geographical location of the fault point, including information such as the terrain and surrounding environment. Conducting terrain analysis on the geographical location of the fault point is an important link in preparing for on-site investigation. Terrain analysis usually includes the analysis of factors such as slope, aspect, elevation, and vegetation cover. These analyses are carried out using Digital Elevation Model (DEM) and remote sensing image data, and terrain feature data are extracted through GIS spatial analysis tools. Based on these terrain feature data and combined with the actual trend of the transmission line, the shortest path algorithm (such as Dijkstra algorithm or A* algorithm) is used for path planning. When planning, factors such as terrain difficulty and traffic accessibility need to be considered, and finally an optimal target inspection route is obtained. Conducting on-site investigation along the target inspection route is a direct means to verify the fault location. During the investigation process, a high-precision GPS device is used to record the path, and at the same time, multimedia data such as on-site images and videos are collected. Image processing is performed on the collected on-site data, including steps such as image enhancement, noise removal, and feature extraction. Image enhancement can use methods such as histogram equalization or adaptive contrast adjustment to improve the image quality. Noise removal can adopt techniques such as median filtering or wavelet transform. Feature extraction is the key to identifying fault features and can use methods such as edge detection and texture analysis, and finally an image highlighting the fault features is obtained. Analyzing the fault feature image through computer vision algorithms is the core step of automatic fault recognition. Multiple algorithms may be involved, such as object detection algorithms based on deep learning (such as YOLO or Faster R-CNN) for locating fault components, and Convolutional Neural Network (CNN) for fault type classification. These algorithms require a large amount of labeled data for training to identify various possible fault features. The analysis results give the fault verification results, including information such as fault type and severity. Based on these results and combined with the GPS positioning data, the fault location information is updated to obtain the corrected fault location.

[0083] Perform deviation analysis on the corrected fault location and the initial accurate fault location coordinates, and calculate the distance difference and direction offset between the two. This analysis obtains the positioning error data, which reflects the deviation between the algorithm positioning result and the actual situation. Based on these error data, adjust the parameters of the positioning algorithm. The adjustment process may involve multiple aspects, such as the correction of the traveling wave velocity, the calibration of impedance parameters, the update of the weights of the Bayesian network, etc. These adjustments aim to reduce the errors of future positioning and improve the accuracy of the algorithm. Finally, a set of optimized parameters are obtained, and these parameters will be used for subsequent fault location work. Integrate the optimized parameters and the corrected fault location to form the final fault location result and system optimization data. These data not only include the accurate fault location information, but also include the relevant parameters of algorithm optimization. These results will be used to update the database of the fault location system, further improving the positioning accuracy and reliability of the system.

[0084] For example: A single-phase grounding fault occurs on a 330 kV transmission line. The initial accurate fault location coordinates are 150.5 km from the starting point. Through the coordinate conversion algorithm, this location is converted into longitude and latitude coordinates (120.5678° E, 30.1234° N). After positioning in the GIS, it is found that this point is located in a hilly area. The terrain analysis shows that the average slope within 5 km around the fault point is 15°, the highest altitude is 500 m, and the vegetation coverage rate is 60%. Based on these data, use the A* algorithm to plan a 12.3 km long inspection route, avoiding steep areas and dense vegetation. Conduct on-site surveys along the planned route and collect 200 high-definition images and 10 minutes of 4K video. Preprocess these data, including enhancing the image contrast using histogram equalization and removing noise through Gaussian filtering. Then use the Canny edge detection algorithm to extract the line and tower profiles in the images. Use the YOLOv5 object detection algorithm to analyze the processed images and identify a damaged insulator at a location 151.2 km from the starting point.

[0085] Compare the corrected fault location (151.2 km) with the initial location (150.5 km), and calculate a positioning error of 0.7 km. Based on this error, correct the traveling wave velocity, adjusting it from the original 298 m / μs to 297.3 m / μs. At the same time, update the line parameters in the impedance method, correcting the resistance per unit length from 0.028 Ω / km to 0.0282 Ω / km. These optimized parameters are integrated into the system, and it is expected that the positioning error in future similar situations will be reduced by approximately 20%. The final fault location result is determined to be a damaged insulator at 151.2 km from the starting point, with a positioning error less than 100 m. The entire verification and optimization process takes about 4 hours, significantly improving the accuracy of fault location and the reliability of the system.

[0086] The above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the embodiments, those of ordinary skill in the art should understand that: it is still possible to modify the specific implementation manners of the present invention or make equivalent replacements, and any modification or equivalent replacement that does not depart from the spirit and scope of the present invention shall be covered by the scope of the claims of the present invention.

Claims

1. A two-terminal fault location method based on signal enhancement, characterized in that: include: Collect and preprocess the electrical signals at both ends of the transmission line to obtain preprocessed synchronous electrical signals; Adaptively enhance the pre-processed synchronous electrical signal to obtain an enhanced fault characteristic signal; Perform time-frequency analysis and feature extraction on the enhanced fault feature signal to obtain the fault feature vector and key parameters; Perform machine learning analysis on fault feature vectors and key parameters to obtain fault type identification results and preliminary fault location estimation; Perform multi-algorithm fusion analysis on the preliminary fault location estimation to obtain the precise fault location coordinates; The step of performing multi-algorithm fusion analysis on the preliminary fault location estimation to obtain accurate fault location coordinates includes: Performing a traveling wave method analysis on the preliminary fault location estimate to obtain a traveling wave method location estimate, and performing an impedance method analysis on the preliminary fault location estimate to obtain an impedance method location estimate; constructing a Bayesian network model based on the traveling wave method location estimate and the impedance method location estimate to obtain a preliminary fusion result, and performing fuzzy reasoning on the preliminary fusion result to obtain a fuzzy location interval; performing genetic algorithm optimization on the fuzzy location interval to obtain a candidate location point set, and performing a Delphi method evaluation on the candidate location point set to obtain a weight coefficient; performing a weighted average on the candidate location point set according to the weight coefficient to obtain a weighted location estimate, and performing a Kalman filter on the weighted location estimate to obtain a filtered location estimate; performing an error analysis on the filtered location estimate through Monte Carlo simulation to obtain an error distribution, and calculating a confidence interval based on the error distribution to obtain precise fault location coordinates; The precise fault location coordinates are mapped and verified on the spot to obtain fault location results and system optimization data.

2. The double-terminal fault location method based on signal enhancement according to claim 1, characterized in that: The step of collecting and preprocessing the electrical signals at both ends of the transmission line to obtain the preprocessed synchronous electrical signals includes: Performing high-frequency sampling on voltage and current signals at both ends of the transmission line to obtain original electrical signal data, and performing time stamp marking on the original electrical signal data to obtain original signal data with time stamp; Performing wavelet denoising on the original signal data with the timestamp to obtain a preliminary denoised signal, and performing adaptive median filtering on the preliminary denoised signal to obtain a filtered signal; Performing outlier detection on the filtered signal to obtain signal data without outliers, and performing interpolation processing on the signal data without outliers to obtain continuous signal data; Performing Fourier transform on the continuous signal data to obtain frequency domain feature data, and performing bandpass filtering on the frequency domain feature data to obtain a target frequency band signal; Performing phase correction on the target frequency band signal by Hilbert transform to obtain a phase-corrected signal, and performing time synchronization processing on the phase-corrected signal to obtain an electrical signal synchronized at both ends; The electrical signals synchronized at both ends are data compressed to obtain a compressed synchronization signal, and the compressed synchronization signal is quantized and encoded to obtain a pre-processed synchronization electrical signal.

3. The double-terminal fault location method based on signal enhancement according to claim 1, characterized in that: The step of performing adaptive enhancement processing on the pre-processed synchronous electrical signal to obtain an enhanced fault characteristic signal comprises: Performing wavelet packet decomposition on the preprocessed synchronous electrical signal to obtain multi-scale signal components, and performing energy distribution analysis on the multi-scale signal components to obtain energy eigenvectors; Adaptively selecting a threshold value for a multi-scale signal component according to the energy eigenvector to obtain a candidate eigencomponent, and sparsely representing the candidate eigencomponent to obtain a sparse coefficient matrix; Performing singular value decomposition on the sparse coefficient matrix to obtain main characteristic components, and performing adaptive gain adjustment on the main characteristic components to obtain a gain-adjusted signal; Decomposing the gain-adjusted signal by empirical mode decomposition to obtain an intrinsic mode function set, and performing correlation analysis on the intrinsic mode function set to obtain a key mode function; Performing adaptive filter design on the key modal function to obtain optimized filter parameters, and filtering the key modal function using the optimized filter parameters to obtain a filtered modal function; The filtered modal function is reconstructed to obtain an enhanced fault characteristic signal.

4. The double-terminal fault location method based on signal enhancement according to claim 1, characterized in that: The step of performing time-frequency analysis and feature extraction on the enhanced fault feature signal to obtain the fault feature vector and key parameters includes: Performing short-time Fourier transform on the enhanced fault characteristic signal to obtain a time-frequency spectrum, and performing peak detection on the time-frequency spectrum to obtain a characteristic frequency point set; Performing regional segmentation on the time-frequency spectrum according to the characteristic frequency point set to obtain time-frequency characteristic regions, and performing energy density calculation on the time-frequency characteristic regions to obtain energy distribution characteristics; Performing principal component analysis on the energy distribution characteristics to obtain main characteristic components, and performing standardization processing on the main characteristic components to obtain normalized characteristic vectors; Performing multi-scale decomposition on the enhanced fault characteristic signal by wavelet packet transform to obtain multi-scale coefficients, and performing statistical feature extraction on the multi-scale coefficients to obtain a statistical feature set; Performing correlation analysis on the statistical feature set to obtain key statistical features, and combining the key statistical features to obtain a statistical feature vector; The normalized feature vector and the statistical feature vector are fused to obtain a fault feature vector, and the fault occurrence time and duration are calculated according to the fault feature vector to obtain the fault feature vector and key parameters.

5. The double-terminal fault location method based on signal enhancement according to claim 1, characterized in that: The step of performing machine learning analysis on the fault feature vector and key parameters to obtain a fault type identification result and a preliminary fault location estimation step includes: Performing dimensionality reduction processing on the fault feature vector to obtain feature data after dimensionality reduction, and performing cluster analysis on the feature data after dimensionality reduction to obtain feature clusters; Constructing a decision tree model according to the feature cluster to obtain a preliminary classification rule, and performing pruning optimization on the preliminary classification rule to obtain an optimized decision rule; The optimized decision rule is trained by a support vector machine algorithm to obtain a fault type classifier, and the fault type classifier is used to classify a new fault feature vector to obtain a fault type identification result; Building a neural network model according to the fault type identification result and key parameters to obtain a position estimation model, and cross-validating the position estimation model to obtain a candidate position estimation model; The candidate location estimation model is optimized by using a particle swarm algorithm to obtain an optimized location estimation model, and the fault feature vector is processed by using the optimized location estimation model to obtain a preliminary fault location estimation.

6. The double-terminal fault location method based on signal enhancement according to claim 1, characterized in that: The steps of mapping and verifying the precise fault location coordinates on the ground to obtain the fault location results and system optimization data include: Performing geocoding conversion on the precise fault location coordinates to obtain geographic coordinates, and locating the fault point in a geographic information system according to the geographic coordinates to obtain the geographic location of the fault point; Performing terrain analysis on the geographical location of the fault point to obtain terrain feature data, and performing path planning based on the terrain feature data to obtain a target inspection route; Conducting field survey along the target inspection route to obtain field data, and performing image processing on the field data to obtain a fault feature image; Analyzing the fault feature image by a computer vision algorithm to obtain a fault verification result, and updating the fault location information according to the fault verification result to obtain a corrected fault location; Performing deviation analysis on the corrected fault location and the initial accurate fault location coordinates to obtain positioning error data, and adjusting parameters of the positioning algorithm according to the positioning error data to obtain optimized parameters; The optimization parameters and the corrected fault location are integrated to obtain the fault location result and system optimization data.

Images