A power cable online monitoring fault location system

Abstract

The present invention relates to the technical field of measuring electrical variables, and specifically to a power cable online monitoring fault location system, the system comprising: a signal acquisition module, based on the cable fault site, real-time capture of electromagnetic pulse signals, measurement of the peak amplitude and frequency change of the signal, statistics of the energy of each frequency band signal, and analysis of the frequency distribution characteristics. In the present invention, by sparse processing of the original signal, the data processing speed is accelerated, and the storage requirements are reduced, so that fault detection and location are faster and more accurate. By compressing the data and generating a sparse feature matrix, the efficiency of data analysis is effectively improved. The dictionary training in the signal reconstruction process is optimized to make signal matching more accurate, further enhancing the accuracy of fault diagnosis. The use of deep analysis to accurately classify fault types improves the response speed and accuracy of the fault detection system, thereby improving the overall operating efficiency and reliability of the power system.

Description

Technical Field

[0001] The invention relates to the technical field of measuring electric variables, and in particular to a power cable online monitoring fault location system. Background Art

[0002] The power cable online monitoring fault location system is designed to detect and accurately locate the fault location of the power cable in a timely manner. This system continuously monitors the electrical characteristics of the cable, such as voltage, current, and temperature, through sensors deployed in the power grid. When the system detects abnormal indicators, it can quickly analyze the data and use fault analysis algorithms to determine the specific location of the fault. The main purpose of this technology is to reduce power outages caused by cable faults, optimize maintenance response time, and improve the reliability and efficiency of power supply. In addition, it helps power companies optimize maintenance and operating costs and extend the service life of power facilities.

[0003] However, existing technologies often face problems such as slow processing speed and low storage efficiency when processing large-scale data, especially in high data flow and complex fault scenarios. At the same time, the lack of the ability to flexibly adapt to new signal characteristics limits the accuracy of fault location, affecting the stable operation and timely maintenance of the power system. These technical limitations not only reduce the efficiency of fault response, but may also lead to higher operation and maintenance costs and reduce the overall safety of the system. Summary of the invention

[0004] The purpose of the present invention is to solve the shortcomings in the prior art and to propose a power cable online monitoring fault location system.

[0005] In order to achieve the above object, the present invention adopts the following technical solution: A power cable online monitoring fault location system comprises:

[0006] The signal acquisition module captures electromagnetic pulse signals in real time based on the cable fault site, measures the peak amplitude and frequency change of the signal, counts the energy of the signal in each frequency band, analyzes the frequency distribution characteristics, and obtains the original signal characteristics;

[0007] A sparse representation module performs sparse processing on the signal data based on the original signal features, adjusts the data structure to remove non-critical information, compresses the data using sparse technology to speed up data processing, and generates a sparse feature matrix;

[0008] A dictionary learning module, based on the sparse feature matrix, performs dictionary training, adjusts dictionary elements to match new signal features, optimizes the signal reconstruction process, and obtains dictionary matching results;

[0009] A fault identification module, based on the dictionary matching result, performs in-depth analysis of the fault type, adjusts and sets classification parameters, performs signal pattern matching and fault type classification, and generates a fault identification result;

[0010] The system update module performs system performance identification based on the fault identification result, updates system response parameters and classification efficiency, adjusts system configuration through an online learning mechanism, optimizes the system's response capability to new types of faults, and obtains system performance results.

[0011] As a further solution of the present invention, the original signal characteristics include the measured maximum amplitude, key frequency components and key energy distribution indicators, the sparse feature matrix is ​​the compressed data size, information retention rate and redundant data reduction rate, the dictionary matching results include the optimized number of dictionary elements, matching efficiency and signal reconstruction accuracy, the fault identification results include the identified fault type, classification accuracy and identification response time, and the system performance results include performance evaluation indicators, system stability and adaptive capability update status.

[0012] As a further solution of the present invention, the signal acquisition module includes:

[0013] The signal capture submodule, based on the environment of the cable fault site, accesses the cable port at the fault point, captures the raw data of the electromagnetic pulse signal in real time, collects the timestamp and intensity of the data, and obtains the captured signal data;

[0014] A signal analysis submodule, based on the captured signal data, applies a signal processing tool to measure the peak amplitude of each signal, records the occurrence time of each peak, analyzes the frequency variation of the signal, identifies the main component of the frequency, and generates a signal characteristic analysis;

[0015] The signal statistics submodule calculates the total energy of the signal in each frequency band based on the signal characteristic analysis, compares the energy difference of each frequency band, analyzes the characteristics of frequency distribution, draws an energy distribution diagram, and obtains the original signal characteristics.

[0016] As a further solution of the present invention, the sparse representation module includes:

[0017] A data compression submodule, based on the original signal characteristics, uses a mathematical transformation method to trim the data dimension, compress the data, obtain a more compact data representation, and generate a data compression result;

[0018] The information retention submodule adjusts the proportion of retained data based on the data compression result so that key information is not lost, retains key data features by setting an information retention threshold, and generates an information retention index;

[0019] The speed optimization submodule adjusts the data processing flow to improve the processing speed based on the information retention index, implements data structure simplification, and obtains a sparse feature matrix.

[0020] As a further solution of the present invention, the dictionary learning module includes:

[0021] The dictionary training submodule uses the K-SVD algorithm based on the sparse feature matrix to iteratively optimize and compare the data points with the preset dictionary template, adjust the dictionary elements to match the features of the data points, and generate a training dictionary result;

[0022] The element adjustment submodule fine-tunes the dictionary elements based on the training dictionary results, so that each element can match the new data features, refines and adjusts the parameters of each element, and generates element matching results;

[0023] The reconstruction optimization submodule optimizes the signal reconstruction process based on the element matching results, allows the reconstructed signal to retain key features, improves the reconstruction effect by fine-tuning the reconstruction parameters, and obtains the dictionary matching results.

[0024] As a further solution of the present invention, the fault identification module includes:

[0025] A pattern matching submodule compares the patterns in the known fault pattern library based on the dictionary matching result, identifies the matching fault pattern, records the matching degree of the pattern, and generates a pattern analysis result;

[0026] A parameter adjustment submodule, based on the pattern analysis result, sets classification parameters according to the fault type, adjusts the sensitivity in the classification algorithm, and generates parameter setting results;

[0027] The classification execution submodule applies classification logic to determine the type of fault based on the parameter setting result, performs fault classification, and obtains a fault identification result.

[0028] As a further solution of the present invention, the system update module includes:

[0029] The performance evaluation submodule performs a comprehensive evaluation of the system performance based on the fault identification result, records the current performance indicators of the system, and generates a preliminary performance evaluation result;

[0030] A parameter updating submodule adjusts and optimizes the system response parameters based on the preliminary performance evaluation results, updates the classification efficiency, and obtains parameter updating results;

[0031] The adaptive optimization submodule adjusts the system configuration through an online learning mechanism based on the parameter update result to obtain the system performance result.

[0032] As a further solution of the present invention, the K-SVD algorithm uses the formula:

[0033] in, is the regularization parameter, controlling The overall sparsity of is the dictionary regularization parameter, which is used to balance the size of dictionary elements and avoid overly large dictionary elements. is the number of elements in the dictionary, is the dimension of each dictionary element, It is a dictionary Middle The element Quantity, is a random dictionary, is a sparse code, is the data matrix.

[0034] Compared with the prior art, the advantages and positive effects of the present invention are:

[0035] In the present invention, by sparse processing of the original signal, the data processing speed is accelerated, the storage requirement is reduced, and the fault detection and location are made faster and more accurate. By compressing the data and generating a sparse feature matrix, the efficiency of data analysis is effectively improved. The dictionary training in the signal reconstruction process is optimized to make signal matching more accurate, further enhancing the accuracy of fault diagnosis. The fault type is accurately classified using deep analysis, which improves the response speed and accuracy of the fault detection system, thereby improving the overall operating efficiency and reliability of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0038] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are 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 understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0039] See also Figure 1 The present invention provides a technical solution: a power cable online monitoring fault location system comprising:

[0040] The signal acquisition module captures electromagnetic pulse signals in real time based on the cable fault site, measures the peak amplitude and frequency change of the signal, counts the energy of the signal in each frequency band, analyzes the frequency distribution characteristics, and obtains the original signal characteristics;

[0041] The sparse representation module performs sparse processing on the signal data based on the original signal characteristics, adjusts the data structure to remove non-critical information, compresses the data using sparse technology to speed up data processing, and generates a sparse feature matrix;

[0042] The dictionary learning module trains the dictionary based on the sparse feature matrix, adjusts the dictionary elements to match the new signal features, optimizes the signal reconstruction process, and obtains the dictionary matching results;

[0043] The fault identification module performs in-depth analysis of fault types based on dictionary matching results, adjusts and sets classification parameters, performs signal pattern matching and fault type classification, and generates fault identification results;

[0044] The system update module identifies system performance based on the fault identification results, updates system response parameters and classification efficiency, adjusts system configuration through online learning mechanism, optimizes the system's response capability to new types of faults, and obtains system performance results.

[0045] The original signal characteristics include the measured maximum amplitude, key frequency components and key energy distribution indicators. The sparse feature matrix is ​​the compressed data size, information retention rate and redundant data reduction rate. The dictionary matching results include the optimized number of dictionary elements, matching efficiency and signal reconstruction accuracy. The fault identification results include the identified fault type, classification accuracy and identification response time. The system performance results include performance evaluation indicators, system stability and adaptive capability updates.

[0046] In this embodiment, the signal acquisition module includes:

[0047] The signal capture submodule, based on the environment of the cable fault site, accesses the cable port at the fault point, captures the raw data of the electromagnetic pulse signal in real time, collects the timestamp and intensity of the data, and obtains the captured signal data;

[0048] The signal analysis submodule, based on the captured signal data, applies signal processing tools to measure the peak amplitude of each signal, records the occurrence time of each peak, analyzes the frequency changes of the signal, identifies the main components of the frequency, and generates signal characteristic analysis;

[0049] The signal statistics submodule calculates the total energy of the signal in each frequency band based on signal characteristic analysis, compares the energy difference of each frequency band, analyzes the characteristics of frequency distribution, draws the energy distribution diagram, and obtains the original signal characteristics.

[0050] The signal capture submodule is designed to access the cable port at the fault point. High-frequency broadband sensors dedicated to capturing raw electromagnetic pulse signals are used to achieve data acquisition. These sensors are carefully selected to adapt to industrial environments with high electromagnetic interference to ensure complete signal capture and high signal-to-noise ratio. The specific process includes: installing high-frequency broadband sensors at the cable fault point to monitor and collect raw data of electromagnetic pulses, which includes the timestamp and intensity of the signal. These data are first filtered and amplified by the sensor's built-in microprocessor to filter out environmental noise and enhance signal readability. Subsequently, the data is transmitted to the central processing unit for further analysis. This step is critical to ensuring the accuracy and real-time nature of the data. The selection of sensors takes into account their response speed and accuracy to improve the accuracy and efficiency of data acquisition, thereby effectively improving the accuracy and efficiency of fault detection. Finally, the system uses these preliminarily processed data for fault analysis to obtain a clear capture of the original signal characteristics.

[0051] In the above content, the signal analysis submodule uses the captured signal data and applies signal processing tools to measure the peak amplitude of each signal according to the formula , calculate the peak amplitude of the signal.

[0052] In the formula, represents the peak amplitude, Represents a point in time, represents the observation time window, Represents at a point in time signal strength.

[0053] Detailed explanation of the formula and the process of formula calculation and derivation:

[0054] In signal processing, peak amplitude is obtained by finding Signal strength within For example, in the time window In the data, the signal strength data is , then the peak amplitude This calculation is obtained by iterating over each element in the signal strength array and continuously updating the maximum value.

[0055] The design of the signal statistics submodule is based on signal characteristic analysis. By calculating the total energy of the signal in each frequency band, the detailed execution process includes using the algorithm to perform frequency analysis on the signal, dividing the signal into multiple frequency bands, calculating the energy of each frequency band, and converting the time series signal into a frequency series signal using Fourier transform, and then calculating the energy of the signal in each frequency band. This process includes the summation of energy and comparison of differences. The energy calculation formula is: ,in represents a frequency band set, Indicates the frequency The system draws an energy distribution diagram according to the energy values ​​of these frequency bands. The energy distribution diagram obtained by analysis can be used to evaluate the signal characteristics. This method calculates the energy of each frequency band in detail to provide basic data for further analysis of the signal, thereby identifying the characteristics of the signal frequency distribution and obtaining a detailed description of the original signal characteristics.

[0056] In this embodiment, the sparse representation module includes:

[0057] The data compression submodule uses mathematical transformation methods to prune data dimensions based on the original signal characteristics, compress the data, obtain a more compact data representation, and generate data compression results;

[0058] The information retention submodule adjusts the proportion of retained data based on the data compression results to prevent key information from being lost. It retains key data features by setting the information retention threshold and generates information retention indicators.

[0059] The speed optimization submodule adjusts the data processing flow to improve the processing speed based on the information retention index, implements data structure simplification, and obtains a sparse feature matrix.

[0060] The data compression submodule realizes the pruning and compression of data dimensions. The specific process adopts principal component analysis technology. This technology selects the main components of the data to reduce the dimension of the data set based on the original signal characteristics while retaining the most critical information. This technology calculates the covariance matrix of the data set, then solves the eigenvalues ​​and eigenvectors, and selects the eigenvectors corresponding to the largest eigenvalues ​​to form a new data expression. In this process, the original data is converted into a set of new orthogonal features. These new features are variables that can best represent the characteristics of the data. This method effectively reduces the redundancy and noise of the data. The final data compression result includes a set of concise and expressive features, which not only improves the efficiency of data processing, but also facilitates subsequent data storage and processing.

[0061] In the above content, the information retention submodule adjusts the proportion of retained data according to the formula , calculate the proportion of information retained.

[0062] In the formula, Represents the proportion of information retained, Representative eigenvalues, represents the number of eigenvalues ​​selected, Represents the total number of eigenvalues.

[0063] Detailed explanation of the formula and the process of formula calculation and derivation:

[0064] When compressing data, the retention of key information is achieved by setting a threshold for information retention. For example, the characteristic values ​​of the original data set are from large to small. , select before The largest eigenvalue is used to retain information. For example, if the eigenvalue is , select the first two eigenvalues ​​to retain, and the proportion of retained information for , which means that about 83.33% of the information is retained.

[0065] The speed optimization submodule improves the processing speed by adjusting the data processing flow. The execution process includes implementing data structure simplification and algorithm optimization, and using sparse matrix technology to process sparse feature matrices. This technology significantly reduces storage space and computing time by storing only the non-zero elements in the matrix. For each non-zero element, its row position, column position, and value are recorded. This data structure optimization is particularly suitable for matrices in which most elements are zero, achieving faster data access and processing speeds. In addition, the parallel processing of the algorithm is realized through a parallel computing framework, which further improves the efficiency of data processing. The optimized data processing flow not only speeds up the calculation speed, but also improves the overall performance of the system, making the processing of sparse feature matrices more efficient and scalable.

[0066] In this embodiment, the dictionary learning module includes:

[0067] The dictionary training submodule uses the K-SVD algorithm based on the sparse feature matrix to iteratively optimize and compare the data points with the preset dictionary template, adjust the dictionary elements to match the characteristics of the data points, and generate the training dictionary results;

[0068] The element adjustment submodule fine-tunes the dictionary elements based on the training dictionary results, so that each element can match the new data features, refines and adjusts the parameters of each element, and generates element matching results;

[0069] The reconstruction optimization submodule optimizes the signal reconstruction process based on the element matching results, allowing the reconstructed signal to retain key features, improves the reconstruction effect by fine-tuning the reconstruction parameters, and obtains the dictionary matching results.

[0070] The dictionary training submodule adjusts the dictionary elements to match the data point features through the K-SVD algorithm. This process involves sparse representation of the original data and iterative update of the dictionary. The K-SVD algorithm first initializes a dictionary, and then in each iteration, fixes the dictionary and updates the sparse code, then fixes the sparse code and updates the dictionary columns. This iterative process is repeated until convergence. The update of each column involves solving a small-scale least squares problem, which can be effectively handled by singular value decomposition. In this way, the algorithm can gradually adjust the dictionary elements to better represent the features in the data set, thereby generating a trained dictionary, which is used in practical applications to effectively represent new data samples.

[0071] In the above content, the element adjustment submodule fine-tunes the dictionary elements according to the formula , adjust the dictionary elements.

[0072] In the formula, represents the adjusted dictionary, Represents the original dictionary, Represents the adjustment amount of the dictionary, Represents the learning rate.

[0073] Detailed explanation of the formula and the process of formula calculation and derivation:

[0074] During the element adjustment process, each dictionary element is fine-tuned according to the data characteristics. for , adjustment amount for , learning rate Set to 0.1. According to the formula, the adjusted dictionary The calculation is as follows:

[0075] .

[0076] This adjustment makes the dictionary elements match the new data features more accurately, thereby improving the accuracy of the reconstructed data.

[0077] The reconstruction optimization submodule improves the expression quality of the signal by optimizing the signal reconstruction process. In this process, fine-tuning the reconstruction parameters is a key step, which involves adjusting the regularization parameters and learning rate in the reconstruction algorithm. These parameters directly affect the quality and speed of the reconstructed signal. Specifically, the regularization parameter controls the intensity of sparsity in the reconstruction process, while the learning rate determines the speed of parameter adjustment. By gradually testing different parameter configurations, the system can find the optimal reconstruction setting to ensure that the reconstructed signal can maintain key features while minimizing the reconstruction error. The optimized reconstruction parameters make the system more efficient when processing large-scale data, while ensuring the accuracy and stability of signal reconstruction, and achieving overall optimization and improvement of dictionary matching results.

[0078] In this embodiment, the fault identification module includes:

[0079] The pattern matching submodule compares the patterns in the known fault pattern library based on the dictionary matching results, identifies the matching fault patterns, records the matching degree of the patterns, and generates pattern analysis results;

[0080] The parameter adjustment submodule sets the classification parameters according to the fault type based on the pattern analysis results, adjusts the sensitivity in the classification algorithm, and generates parameter setting results;

[0081] The classification execution submodule applies classification logic to determine the type of fault based on the parameter setting results, performs fault classification, and obtains the fault identification result.

[0082] The pattern matching submodule compares the known fault pattern library based on the dictionary matching results. This process uses the cosine similarity method to calculate the matching degree. By performing a dot product on each fault pattern vector and the dictionary matching result vector, and then dividing it by the product of the modulus lengths of the two vectors, a value between -1 and 1 is obtained. The closer this value is to 1, the more similar the two vectors are, thereby determining the matching fault pattern. The key to this method is to quickly and accurately identify patterns similar to the current fault data, and then generate pattern analysis results. These results provide a basis for determining the subsequent fault type, ensuring the accuracy and efficiency of fault diagnosis.

[0083] In the above content, the parameter adjustment submodule sets the classification parameters according to the pattern analysis results, according to the formula , adjust the classification parameters.

[0084] In the formula, Represents the new parameter value, Represents the original parameter value, represents the adjustment factor, Represents the adjustment amount of the parameter.

[0085] Detailed explanation of the formula and the process of formula calculation and derivation:

[0086] In the classification algorithm, the sensitivity is adjusted by introducing an adjustment factor to achieve, for example, the original classification parameters is 0.5, and the classification sensitivity needs to be increased according to the fault type, and the adjustment factor Set to 0.05, the adjustment amount of the parameter is 0.1. According to the formula, the new parameter value Calculated as:

[0087] .

[0088] This adjustment method allows the system to flexibly adjust the sensitivity of the classification algorithm according to different fault types, ensuring the accuracy and adaptability of the classification results.

[0089] The classification execution submodule applies classification logic to determine the type of fault. In this process, according to the parameter setting results, the classification algorithm uses decision tree logic to determine the fault type. Each decision node determines which child node the data flows to according to the set parameter value until it reaches the leaf node. The label of the leaf node is the judgment result of the fault type. The key to this method is to quickly and effectively perform classification according to fault characteristics and ultimately generate fault identification results. The accuracy of classification execution directly affects the efficiency of fault response and processing, ensuring the stability and reliability of the system in actual operation.

[0090] In this embodiment, the system update module includes:

[0091] The performance evaluation submodule performs a comprehensive evaluation of system performance based on the fault identification results, records the current performance indicators of the system, and generates preliminary performance evaluation results;

[0092] The parameter update submodule adjusts and optimizes the system response parameters based on the initial performance evaluation results, updates the classification efficiency, and obtains the parameter update results;

[0093] The adaptive optimization submodule adjusts the system configuration based on the parameter update results through the online learning mechanism to obtain the system performance results.

[0094] The performance evaluation submodule uses the fault identification results to conduct a comprehensive evaluation of system performance. This process includes collecting system operation data, such as response time, error rate and system load, as well as the efficiency of fault handling. These data are summarized and used to calculate system performance indicators, such as average response time and system availability. These indicators reflect the current performance status of the system. Based on these indicators, the performance evaluation submodule generates a preliminary performance evaluation result. This result provides system administrators with a quantitative performance overview, allowing them to understand the performance of the system in actual operation and provide a basis for possible performance improvement measures.

[0095] In the above content, the parameter update submodule adjusts and optimizes the system response parameters according to the formula , update the classification efficiency.

[0096] In the formula, represents the updated parameter value, Represents the original parameter value, Represents the adjustment factor.

[0097] Detailed explanation of the formula and the process of formula calculation and derivation:

[0098] The key to parameter adjustment is to improve classification efficiency. For example, the original parameter value is 50, adjustment factor Set to 0.02, indicating that we want to improve the efficiency by 2%. According to the formula, the new parameter value Calculated as: .

[0099] This formula reflects the ability to fine-tune parameters to improve the system's responsiveness and processing capabilities to adapt to changing operating conditions and demands.

[0100] The adaptive optimization submodule adjusts the system configuration through an online learning mechanism based on the parameter update results. This process is achieved by real-time monitoring of system performance and dynamic adjustment of configuration parameters. It involves using machine learning algorithms to analyze historical performance data and current operating status and automatically recommend the optimal configuration. This method enables the system to adapt to different workloads and environmental changes and maintain efficient operation. The optimized parameters include cache size, number of threads, and resource allocation strategy. These adjustments help the system maintain or improve performance levels. The final system performance results show the improvement of the adjusted system and provide support for future operational decisions.

[0101] In this embodiment, the K-SVD algorithm uses the formula: in, is the regularization parameter, controlling The overall sparsity of is the dictionary regularization parameter, which is used to balance the size of dictionary elements and avoid overly large dictionary elements. is the number of elements in the dictionary, is the dimension of each dictionary element, It is a dictionary Middle The element A quantity.

[0102] Execution process:

[0103] First, set up the dictionary and sparse codes Initialization value, based on the original data , the iterative process starts with the application of the improved K-SVD algorithm. In each iteration, update To minimize the reconstruction error and sparsity penalty , then update , while introducing To control the size of the dictionary elements. By adjusting and , which not only reduces the reconstruction error, but also maintains a reasonable size of the dictionary and the sparsity of the sparse code. The optimal and , in order to achieve the minimum verification error.

[0104] Calculation process:

[0105] First get the parameters: regularization parameters : The regularization parameter controlling the sparsity is set to . Dictionary regularization parameter : The parameter used to control the size of the dictionary elements is set to . Number of dictionary elements :The number of elements in the dictionary is The dimensions of each dictionary element are : The dimension of each dictionary element is . Generate a random dictionary and a sparse code , and then use these to generate data .dictionary and sparse codes All generated by random numbers: dictionary : A matrix with random elements. Sparse code : A matrix with random elements. pass Calculated to ensure that the reconstruction error is zero, because Directly from and Generated: The reconstruction error is calculated as and through and The square of the Frobenius norm between the reconstructed data: because Is Generated directly, so the error here is zero.

[0106] Calculate the sparsity penalty:

[0107] The sparse penalty is calculated as The L1 norm of : The specific value is , which indicates that the sparse code The sum of the elements of is multiplied by the regularization parameter.

[0108] Compute dictionary regularization:

[0109] The dictionary regularization is calculated as the sum of the square roots of the Euclidean norm of each dictionary element multiplied by : The specific value is , which means that the size of all dictionary elements is controlled within a reasonable range, avoiding dictionary elements that are too large.

[0110] Adding all the parts together gives the total cost:

[0111] The results show that the optimization process successfully balances the reconstruction error, the sparsity of the sparse code, and the size of the dictionary.

[0112] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0113] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0114] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A power cable online monitoring fault location system, characterized in that: The system comprises: The signal acquisition module captures electromagnetic pulse signals in real time based on the cable fault site, measures the peak amplitude and frequency change of the signal, counts the energy of the signal in each frequency band, analyzes the frequency distribution characteristics, and obtains the original signal characteristics; A sparse representation module performs sparse processing on the signal data based on the original signal features, adjusts the data structure to remove non-critical information, compresses the data using sparse technology to speed up data processing, and generates a sparse feature matrix; A dictionary learning module, based on the sparse feature matrix, performs dictionary training, adjusts dictionary elements to match new signal features, optimizes the signal reconstruction process, and obtains dictionary matching results; The dictionary learning module includes: The dictionary training submodule uses the K-SVD algorithm based on the sparse feature matrix to iteratively optimize and compare the data points with the preset dictionary template, adjust the dictionary elements to match the features of the data points, and generate a training dictionary result; The element adjustment submodule fine-tunes the dictionary elements based on the training dictionary results, so that each element can match the new data features, refines and adjusts the parameters of each element, and generates element matching results; The reconstruction optimization submodule optimizes the signal reconstruction process based on the element matching results, so that the reconstructed signal retains key features, improves the reconstruction effect by fine-tuning the reconstruction parameters, and obtains the dictionary matching results; The K-SVD algorithm uses the formula: in, is the regularization parameter, controlling The overall sparsity of is the dictionary regularization parameter, which is used to balance the size of dictionary elements and avoid dictionary elements that are too large. is the number of elements in the dictionary, is the dimension of each dictionary element, It is a dictionary Middle The element Quantity, is a random dictionary, is a sparse code, is the data matrix; A fault identification module, based on the dictionary matching result, performs in-depth analysis of the fault type, adjusts and sets classification parameters, performs signal pattern matching and fault type classification, and generates a fault identification result; The system update module performs system performance identification based on the fault identification result, updates system response parameters and classification efficiency, adjusts system configuration through an online learning mechanism, optimizes the system's response capability to new types of faults, and obtains system performance results.

2. The power cable online monitoring fault location system according to claim 1 is characterized in that: The original signal characteristics include the measured maximum amplitude, key frequency components and key energy distribution indicators; the sparse feature matrix is ​​the compressed data size, information retention rate and redundant data reduction rate; the dictionary matching results include the optimized number of dictionary elements, matching efficiency and signal reconstruction accuracy; the fault identification results include the identified fault type, classification accuracy and identification response time; the system performance results include performance evaluation indicators, system stability and adaptive capability update status.

3. The power cable online monitoring fault location system according to claim 1, characterized in that: The signal acquisition module comprises: The signal capture submodule, based on the environment of the cable fault site, accesses the cable port at the fault point, captures the raw data of the electromagnetic pulse signal in real time, collects the timestamp and intensity of the data, and obtains the captured signal data; A signal analysis submodule, based on the captured signal data, applies a signal processing tool to measure the peak amplitude of each signal, records the occurrence time of each peak, analyzes the frequency variation of the signal, identifies the main component of the frequency, and generates a signal characteristic analysis; The signal statistics submodule calculates the total energy of the signal in each frequency band based on the signal characteristic analysis, compares the energy difference of each frequency band, analyzes the characteristics of frequency distribution, draws an energy distribution diagram, and obtains the original signal characteristics.

4. The power cable online monitoring fault location system according to claim 1, characterized in that: The sparse representation module comprises: A data compression submodule, based on the original signal characteristics, uses a mathematical transformation method to trim the data dimension, compress the data, obtain a more compact data representation, and generate a data compression result; The information retention submodule adjusts the proportion of retained data based on the data compression result so that key information is not lost, retains key data features by setting an information retention threshold, and generates an information retention index; The speed optimization submodule adjusts the data processing flow to improve the processing speed based on the information retention index, implements data structure simplification, and obtains a sparse feature matrix.

5. The power cable online monitoring fault location system according to claim 1, characterized in that: The fault identification module comprises: A pattern matching submodule compares the patterns in the known fault pattern library based on the dictionary matching result, identifies the matching fault pattern, records the matching degree of the pattern, and generates a pattern analysis result; A parameter adjustment submodule, based on the pattern analysis result, sets classification parameters according to the fault type, adjusts the sensitivity in the classification algorithm, and generates parameter setting results; The classification execution submodule applies classification logic to determine the type of fault based on the parameter setting result, performs fault classification, and obtains a fault identification result.

6. The power cable online monitoring fault location system according to claim 1, characterized in that: The system update module includes: The performance evaluation submodule performs a comprehensive evaluation of the system performance based on the fault identification result, records the current performance indicators of the system, and generates a preliminary performance evaluation result; A parameter updating submodule adjusts and optimizes the system response parameters based on the preliminary performance evaluation results, updates the classification efficiency, and obtains parameter updating results; The adaptive optimization submodule adjusts the system configuration through an online learning mechanism based on the parameter update result to obtain the system performance result.

Images