Communication cable fault monitoring system and monitoring method

Abstract

The invention relates to the technical field of cable fault monitoring, in particular to a communication cable fault monitoring system and monitoring method. The method comprises the following steps: injecting a time sequence structure excitation code into a communication cable, obtaining a multi-node excitation record, carrying out full-path phase difference tracking, obtaining phase response trajectory data, constructing a cable response matrix, carrying out structure mapping on the response matrix, positioning a response difference region, and calculating a response distortion feature vector. The method comprises the following steps of: deducing an internal abnormal section of a cable, acquiring geometrical characteristics of the cable, performing radial mapping fitting, reconstructing an internal section of the cable, generating a three-dimensional communication cable model, expanding the three-dimensional model layer by layer, identifying an external damage mode, deducing an external abnormal section, performing interactive authentication by combining the internal and external abnormal sections, and judging a cable fault. And abnormal joint positioning is carried out to form a cable fault section. According to the invention, the accurate identification of the fault source is realized, and the intelligent management efficiency of the communication infrastructure is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of cable fault monitoring, and in particular to a communication cable fault monitoring system and a monitoring method. Background Art

[0002] Traditional communication cable fault monitoring methods usually rely on manual inspections or simple electrical tests, resulting in insufficient timeliness and accuracy in fault detection. Many faults are not quickly identified after they occur, which in turn affects the stability and reliability of communications. Especially in complex network environments, cable damage is difficult to effectively locate, resulting in high maintenance costs and potential service interruption risks. In addition, existing technologies have limitations in phase response analysis and geometric feature acquisition. Many monitoring systems cannot provide real-time and accurate cable status assessments, resulting in a lack of scientific basis for judging internal and external cable anomalies. Especially in multi-layer cables, internal damage is often intertwined with external damage, making it difficult for traditional methods to effectively distinguish and locate it, which in turn affects the efficiency and effectiveness of fault handling. The lack of comprehensive monitoring methods makes the prediction and management of cable faults complicated. Summary of the Invention

[0003] Based on this, it is necessary to provide a communication cable fault monitoring system and monitoring method to solve at least one of the above technical problems.

[0004] To achieve the above object, a communication cable fault monitoring method includes the following steps: Step S1: injecting a timing structure excitation code into the communication cable to obtain a multi-node excitation record; performing full-path phase differential tracking on the communication cable according to the excitation record to obtain phase response trajectory data; and constructing a cable response matrix according to the phase response trajectory data; Step S2: mapping the cable response matrix to a response distribution structure, and locating the response difference area in the communication cable according to the response distribution structure; calculating the response distortion eigenvector of each response difference area, and inferring the abnormal section inside the cable based on the response distortion eigenvector; Step S3: Collecting geometric features of the communication cable; performing radial mapping fitting based on the geometric features of the communication cable, and reconstructing the internal profile of the cable based on the radial mapping fitting data; performing communication cable simulation processing based on the geometric features of the communication cable and the internal profile of the cable, thereby generating a three-dimensional communication cable; Step S4: unfolding the three-dimensional communication cable layer by layer to obtain a multi-layer cable structure unit; identifying external damage patterns based on the structural characteristics of each layer in the multi-layer cable structure unit, and inferring the abnormal section of the cable exterior based on the external damage patterns; Step S5: Interactive authentication is performed on the abnormal section inside the cable and the abnormal section outside the cable. When the abnormal sections overlap, the synchronous overlapping section is determined to be a cable fault. When the abnormal section is unilaterally abnormal, the abnormality is jointly located through the abnormal section inside the cable and the abnormal section outside the cable, thereby generating a cable fault section.

[0005] The present invention realizes the acquisition of multi-node excitation records by injecting timing structure excitation coding into the communication cable, can accurately capture the phase response characteristics of the cable, and obtains phase response trajectory data based on the full-path phase differential tracking technology, which provides a reliable basis for the subsequent construction of the cable response matrix. The generation of the cable response matrix enables the response characteristics of the cable to be systematized, which is convenient for subsequent analysis and processing. The implementation of the response distribution structure mapping helps to locate the response difference area in the cable. The calculated response distortion feature vector provides a quantitative basis for the inference of abnormal sections, which can effectively identify potential problems inside the cable. The collected geometric features of the communication cable provide important data support for the subsequent radial mapping fitting. By reconstructing the internal profile of the cable, an in-depth understanding of the cable structure is achieved. Combining the geometric features with the simulation processing of the internal profile, a three-dimensional communication cable model is generated. It provides an intuitive basis for subsequent analysis and monitoring. The layer-by-layer unfolding three-dimensional model enables the multi-layer cable structure unit to be presented in detail. The ability to identify the structural characteristics of each layer separately enhances the analysis of external damage patterns. The abnormal sections inferred based on external damage patterns can effectively supplement the monitoring of internal abnormalities. Combined with the mechanism of interactive authentication of internal and external abnormal sections, a more comprehensive fault identification solution is provided. When the abnormal sections overlap, the cable fault can be determined quickly and accurately. In the case of unilateral abnormality, the fault source is accurately identified through the joint positioning of the abnormality. The implementation of the overall method has greatly improved the accuracy and efficiency of communication cable fault monitoring, promoted the intelligent management of communication infrastructure, reduced maintenance costs, extended the service life of cables, ensured the stability and reliability of the communication system, and provided strong support and guarantee for the future construction of communication networks.

[0006] The present invention also provides a communication cable fault monitoring system for executing the communication cable fault monitoring method described above, the communication cable fault monitoring system comprising: The excitation modeling module is used to inject timing structure excitation coding into the communication cable to obtain multi-node excitation records; perform full-path phase differential tracking on the communication cable based on the excitation records to obtain phase response trajectory data; and construct a cable response matrix based on the phase response trajectory data; The response location module is used to map the response distribution structure of the cable response matrix and locate the response difference area in the communication cable based on the response distribution structure; calculate the response distortion feature vector of each response difference area and infer the abnormal section inside the cable based on the response distortion feature vector; The simulation and reconstruction module is used to collect the geometric features of the communication cable; perform radial mapping fitting based on the geometric features of the communication cable, and reconstruct the internal profile of the cable based on the radial mapping fitting data; simulate the communication cable through the geometric features of the communication cable and the internal profile of the cable to generate a three-dimensional communication cable; A structure recognition module is used to unfold the three-dimensional communication cable layer by layer to obtain a multi-layer cable structure unit; identify the external damage pattern based on the structural characteristics of each layer in the multi-layer cable structure unit, and infer the abnormal section of the cable outside based on the external damage pattern; The fault judgment module is used to perform interactive authentication by combining the abnormal sections inside the cable and the abnormal sections outside the cable. When the abnormal sections overlap, the synchronous overlapping sections are judged as cable faults. When the abnormal section is unilaterally abnormal, the abnormality is jointly located through the abnormal sections inside the cable and the abnormal sections outside the cable, thereby generating a cable fault section.

[0007] The present invention injects sequential structure excitation coding into the communication cable through the implementation of the excitation modeling module, which can effectively obtain multi-node excitation records and provide basic data for subsequent phase response analysis. The phase response trajectory data obtained by the full-path phase differential tracking technology enables the cable response characteristics to be fully captured, and the constructed cable response matrix provides a reliable basis for subsequent response positioning. The introduction of the response positioning module can accurately locate the response difference area in the communication cable through response distribution structure mapping. The calculated response distortion feature vector provides quantitative support for inferring the abnormal section inside the cable. The simulation reconstruction module collects the geometric features of the communication cable and performs radial mapping fitting, which greatly improves the reconstruction accuracy of the cable internal section. The generated three-dimensional communication cable model provides an intuitive basis for subsequent structural analysis. The layer-by-layer unfolding three-dimensional model enables the multi-layer cable structure unit to be presented in a refined manner, and the structural identification The module can identify external damage patterns based on the structural characteristics of each layer, and the inferred external abnormal sections provide an important reference for troubleshooting. The fault judgment module integrates the abnormal sections inside and outside the cable, and achieves more accurate fault identification through an interactive authentication mechanism. When the abnormal sections overlap, the cable fault can be quickly and accurately determined. In the case of unilateral abnormality, the fault source can be accurately identified through the joint positioning of the abnormality. The implementation of the overall system improves the accuracy and efficiency of communication cable fault monitoring, promotes the intelligent management of communication infrastructure, reduces operation and maintenance costs, extends the service life of cables, ensures the stability and reliability of communication systems, provides a solid technical foundation for the development of future communication networks, enhances the ability to predict potential faults, promotes the establishment of a rapid response mechanism, greatly improves the security and stability of communication networks, and promotes the development of the industry towards intelligence and automation. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 A schematic flow chart of the steps of a communication cable fault monitoring method; Figure 2 Detailed implementation flow chart of step S3; The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0009] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. It is obvious that the embodiments described are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts are within the scope of protection of the present invention.

[0010] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor and / or microcontroller approaches.

[0011] It should be understood that although the terms "first," "second," and the like may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used solely to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the exemplary embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the listed associated items.

[0012] To achieve this, please refer to Figures 1 to 2 , a communication cable fault monitoring method, comprising the following steps: Step S1: injecting a timing structure excitation code into the communication cable to obtain a multi-node excitation record; performing full-path phase differential tracking on the communication cable according to the excitation record to obtain phase response trajectory data; and constructing a cable response matrix according to the phase response trajectory data; Step S2: mapping the cable response matrix to a response distribution structure, and locating the response difference area in the communication cable according to the response distribution structure; calculating the response distortion eigenvector of each response difference area, and inferring the abnormal section inside the cable based on the response distortion eigenvector; Step S3: Collecting geometric features of the communication cable; performing radial mapping fitting based on the geometric features of the communication cable, and reconstructing the internal profile of the cable based on the radial mapping fitting data; performing communication cable simulation processing based on the geometric features of the communication cable and the internal profile of the cable, thereby generating a three-dimensional communication cable; Step S4: unfolding the three-dimensional communication cable layer by layer to obtain a multi-layer cable structure unit; identifying external damage patterns based on the structural characteristics of each layer in the multi-layer cable structure unit, and inferring the abnormal section of the cable exterior based on the external damage patterns; Step S5: Interactive authentication is performed on the abnormal section inside the cable and the abnormal section outside the cable. When the abnormal sections overlap, the synchronous overlapping section is determined to be a cable fault. When the abnormal section is unilaterally abnormal, the abnormality is jointly located through the abnormal section inside the cable and the abnormal section outside the cable, thereby generating a cable fault section.

[0013] The present invention realizes the acquisition of multi-node excitation records by injecting timing structure excitation coding into the communication cable, can accurately capture the phase response characteristics of the cable, and obtains phase response trajectory data based on the full-path phase differential tracking technology, which provides a reliable basis for the subsequent construction of the cable response matrix. The generation of the cable response matrix enables the response characteristics of the cable to be systematized, which is convenient for subsequent analysis and processing. The implementation of the response distribution structure mapping helps to locate the response difference area in the cable. The calculated response distortion feature vector provides a quantitative basis for the inference of abnormal sections, which can effectively identify potential problems inside the cable. The collected geometric features of the communication cable provide important data support for the subsequent radial mapping fitting. By reconstructing the internal profile of the cable, an in-depth understanding of the cable structure is achieved. Combining the geometric features with the simulation processing of the internal profile, a three-dimensional communication cable model is generated. It provides an intuitive basis for subsequent analysis and monitoring. The layer-by-layer unfolding three-dimensional model enables the multi-layer cable structure unit to be presented in detail. The ability to identify the structural characteristics of each layer separately enhances the analysis of external damage patterns. The abnormal sections inferred based on external damage patterns can effectively supplement the monitoring of internal abnormalities. Combined with the mechanism of interactive authentication of internal and external abnormal sections, a more comprehensive fault identification solution is provided. When the abnormal sections overlap, the cable fault can be determined quickly and accurately. In the case of unilateral abnormality, the fault source is accurately identified through the joint positioning of the abnormality. The implementation of the overall method has greatly improved the accuracy and efficiency of communication cable fault monitoring, promoted the intelligent management of communication infrastructure, reduced maintenance costs, extended the service life of cables, ensured the stability and reliability of the communication system, and provided strong support and guarantee for the future construction of communication networks.

[0014] In an embodiment of the present invention, the communication cable fault monitoring method includes the following steps: Step S1: injecting a timing structure excitation code into the communication cable to obtain a multi-node excitation record; performing full-path phase differential tracking on the communication cable according to the excitation record to obtain phase response trajectory data; and constructing a cable response matrix according to the phase response trajectory data; In this embodiment, the communication cable to be tested is excited by the timing structure excitation coding signal, and the frequency domain characteristic uniformly distributed spread spectrum excitation source signal is used for modulation. The excitation signal carrier frequency is set to 80 MHz, and the signal duration is 25 , with a repetition period of 100 The modulated signal is evenly injected into multiple preset endpoints of the communication cable through a 16-channel multi-point injection device. At the same time, a bidirectional synchronous acquisition module is connected to each injection endpoint and the corresponding excitation response data is synchronously recorded to form a multi-node excitation record. The acquisition process adopts a sampling frequency of 1GS / s (billion times per second) and performs timing alignment with a resolution of 300ps. The phase differential locking of the acquired multi-node excitation record is performed. By fine-tuning and calibrating the synchronous reference point of the excitation signal, the phase difference tracking data under the complete path is constructed. The phase difference data is then differentially mapped according to the node order to extract the phase response trajectory data. The trajectory data is constructed into a complete time series structure diagram with each node as the starting point. Based on the trajectory data, a two-dimensional cable response matrix is ​​constructed according to the bidirectional propagation relationship between the nodes. Each element in the response matrix represents the propagation phase change difference between two nodes. Its unit is radian and recorded with a resolution accuracy of 0.01rad.

[0015] Step S2: mapping the cable response matrix to a response distribution structure, and locating the response difference area in the communication cable according to the response distribution structure; calculating the response distortion eigenvector of each response difference area, and inferring the abnormal section inside the cable based on the response distortion eigenvector; In this embodiment, the constructed cable response matrix is ​​subjected to response distribution structure mapping. Matrix slicing is used to divide local response windows with a step size of 2×2. The structural offset coefficient and extreme value distribution eigenvector of each window are extracted, and a local response reconstruction map is constructed. The response pattern differences between the structural mutation region and the preset stable segment in the map are then compared using normalized convolution to obtain response difference regions. These difference regions identify the locations of structural mutations where propagation disturbances exist in the cable. Furthermore, the disturbance slope of each response difference region in the joint time-frequency space is extracted to construct a response distortion eigenvector for that region. The vector dimension is set to 128 and includes indicators such as phase inflection rate, time drift rate, and structural splitting rate. Multidimensional vector matching is used to compare the tagged distortion templates in the historical response model database. The matching result is determined by the minimum Euclidean distance. The matched distortion eigenvectors are then reverse mapped to infer the location of the corresponding internal abnormal section in the actual physical cable structure. The location range of the internal abnormal section is divided and labeled in units of 20 cm.

[0016] Step S3: Collecting geometric features of the communication cable; performing radial mapping fitting based on the geometric features of the communication cable, and reconstructing the internal profile of the cable based on the radial mapping fitting data; performing communication cable simulation processing based on the geometric features of the communication cable and the internal profile of the cable, thereby generating a three-dimensional communication cable; In this embodiment, a structured laser scanning device is used to perform geometric structure acquisition on the target communication cable, and the scanning resolution is set to The obtained external contour point cloud data is subjected to radial differential fitting processing, and the radial curvature change fitting function in the cylindrical coordinate system is used to fit the radius curve of each contour segment. All radial fitting results are then linearly reconstructed to obtain the complete cable outer contour fitting data. Based on this fitting data and the established material stacking structure rules, the internal cross-sectional structure model of the cable is reconstructed. Each layer in the model includes the outer sheath, shielding layer, insulation layer, conductor layer, etc. Each layer has its material density, electromagnetic constant and thickness parameters set. The cross-sectional model is imported through a multi-physics field simulation platform (such as COMSOL Multiphysics). The unit current pulse with an interval of 0.5 mA from 1 mA to 10 mA is applied in simulation and a frequency domain impedance model is constructed in combination with the cable material characteristics. The full-frequency response data is obtained through simulation calculation and mapped into a spatial structure response image to construct a three-dimensional communication cable model. The model resolution is set to 50 points per millimeter to support subsequent structural layered expansion analysis.

[0017] Step S4: unfolding the three-dimensional communication cable layer by layer to obtain a multi-layer cable structure unit; identifying external damage patterns based on the structural characteristics of each layer in the multi-layer cable structure unit, and inferring the abnormal section of the cable exterior based on the external damage patterns; In this embodiment, the constructed three-dimensional communication cable model is structurally unfolded in a longitudinal unfolding manner. During the unfolding process, the cross-sectional profile along the longitudinal axis is extracted in a stepping manner of 0.5 mm, and each layer of structural units is extracted separately, including the outer sheath layer, braided shielding layer, aluminum foil shielding layer, insulation layer and conductor layer. The structural features of each layer of structural unit are extracted using local thickness offset recognition technology. Combined with the distorted structure boundary extraction method based on the density gradient field, the main characteristic boundary contours and main axis curves of each layer of structure are extracted, and then the external damage mode is identified through boundary deformation and local depression model. The damage mode annotation adopts a three-classification structure: indentation, erosion, and extrusion penetration. The feature area with a convolutional atlas comparison and matching similarity index greater than 0.85 is used as the external abnormal reference segment. The projection interval of the corresponding abnormal segment along the longitudinal axis is further inferred and mapped as the external abnormal segment of the cable. The boundary of each segment is annotated in units of 10 cm.

[0018] Step S5: Interactive authentication is performed on the abnormal section inside the cable and the abnormal section outside the cable. When the abnormal sections overlap, the synchronous overlapping section is determined to be a cable fault. When the abnormal section is unilaterally abnormal, the abnormality is jointly located through the abnormal section inside the cable and the abnormal section outside the cable, thereby generating a cable fault section.

[0019] In this embodiment, the internal abnormal sections and external abnormal sections obtained above are matched and compared one by one. In the matching process, the overlapping length between the start and end positions of the internal abnormal sections and the external abnormal sections is used as a criterion. If any pair of abnormal sections has an overlapping length greater than or equal to 3 cm in the longitudinal position, the overlapping interval is determined to be a synchronous overlapping section, and the section is marked as a cable fault section. If there is no synchronous overlap between all abnormal sections with an overlapping length reaching a preset threshold, the type is defined as a unilateral abnormality. By calculating the relative position offset, structural hierarchy correspondence and physical property comparison between the internal and external abnormal sections corresponding to all unilateral abnormalities, a joint abnormality vector matrix is ​​constructed. For each group The joint anomaly vector performs spatial consistency measurement, and the structural mismatch score (MisfitScore) is used to determine the degree of correlation between the two. Areas above the set correlation threshold are mapped as potential linkage anomaly sections. These potential linkage anomaly sections are used to construct a cable fault space model. Each potential fault section in the model is identified with a length unit of 30 cm. Based on this model, the cable fault section is located, and the specific longitudinal section start and end positions are uploaded to the specified data interface module in the monitoring system. The position format uses the cable starting mark point as the benchmark, and the relative starting distance and length data of the section are recorded. Finally, a complete cable fault section information data packet is generated and stored in the monitoring system database for subsequent maintenance scheduling.

[0020] Preferably, step S1 includes the following steps: Step S11: Performing timing structure excitation coding injection on the communication cable, using a pseudo-random modulation signal with a pulse width of 50-150ns, and setting the excitation point spacing to a physical interval of 20-50cm to obtain multi-node excitation records; Step S12: performing phase differential locking on the multi-node excitation records to generate full-path phase differential tracking data; Step S13: extracting phase response trajectory data based on the full-path phase differential tracking data, with the extraction path segmentation granularity set to 10 cm per segment; Step S14: Integrate the response of each node through the phase response trajectory data to obtain node response integration data, where the number of response nodes is limited to 16-128; and construct a cable response matrix through the node response integration data.

[0021] In this embodiment, a high-precision signal source generator is used to construct a pseudo-random modulation excitation signal. The signal adopts a broadband time domain coding form of fixed frequency modulation. The pulse period of the excitation signal is set between 2 microseconds and 10 microseconds. Among them, the modulation structure with 5 microseconds as the base period has the highest adoption rate in the experiment. The pulse width of the modulation signal is set between 50 nanoseconds and 150 nanoseconds. Among them, the bandwidth modulation signal with an 80 nanosecond pulse width can effectively cover the main communication frequency band within 100 MHz. The signal modulation adopts a maximum length sequence (M sequence) modulation structure. The signal frequency is kept stable by the phase-locked loop control module. The excitation signal is transmitted to the input end of the communication cable through a pulse injector with an impedance matching network. The physical interval of the excitation point is set to 100 nanoseconds along the length of the cable. The distance between 20 cm and 50 cm is between 20 cm and 50 cm. 32 excitation nodes are set up in the experiment. Each node is connected through a coaxial open-circuit coupling interface, and finally a multi-node excitation record consisting of the excitation response voltage and time series of each node is obtained. The excitation record is synchronously collected in a 1GS / s (billion times per second) high-speed sampling module. A high-pass filter is used to suppress low-frequency interference. When the multi-node excitation record is phase-differentially locked, the collected excitation response signal of each node is sent to the phase-controlled analysis module. The module used is based on a fast Fourier transform (FFT) array integrated in a high-speed FPGA to perform frequency domain transformation on the response of each excitation node in the entire time axis, and obtain the main frequency phase value of the corresponding node in each time window. A sliding window with a time window width of 1 microsecond is used. The structure is constructed with a sliding step of 250 nanoseconds. The main frequency phase evolution trajectory in the stimulus response of each node is extracted, the main frequency phase difference between adjacent nodes is calculated and the change interval is locked. The full path phase difference data set is constructed. The phase difference is normalized with the reference node as the starting point, and the cumulative phase change from the source node to the end node on each propagation path is finally output. The path phase difference sequence obtained above is segmented, and each path is divided into interval segments with a granularity of 10 cm. The median of the phase difference value in each segment is taken as the phase response value of the segment. The segment response sequence of each path is expressed as a path trajectory. The path numbers are arranged in order of physical position. Each segment data contains a time index, path number, segment length number and Corresponding to the four categories of phase values, a complete phase response trajectory data set is formed. The format of the data set is matrix, and the dimension is the number of paths multiplied by the number of segments. Each data point is recorded in a high-precision double floating-point variable to prevent precision loss. The path segment phase response trajectory data is clustered by node, and the phase response values ​​of all path segments passing through a node are averaged and the position of the maximum phase mutation point is recorded to construct the integrated feature vector of the response node. For each node, the length of the response vector corresponds to the number of path segments it is physically connected to. The number of response nodes is set between 16 and 128. The total number of nodes is determined by the number of excitation points and the layout spacing. The node response integration data finally forms a response matrix with the dimension of the number of nodes multiplied by the number of path segments.This response matrix serves as the input data source for subsequent response distribution structure mapping and response distortion analysis. When constructing the cable response matrix, each row of the matrix represents the integrated sequence of path responses for an excitation node, and each column represents the cumulative response status corresponding to the path segment number. The matrix data is written to an external NVMe SSD storage device in a lossless compression format for subsequent use, completing the cable response matrix construction process. Once the response matrix is ​​constructed, it is called by the external structural analysis module to enter the response distribution analysis processing flow. All raw and processed data collected up to this stage are retained for subsequent traceability and comparison.

[0022] Preferably, performing response distribution structure mapping on the cable response matrix in step S2, and locating the response difference area in the communication cable according to the response distribution structure includes: Performing normalized amplitude modulation transformation on the cable response matrix to obtain a normalized response matrix; The normalized response matrix is ​​segmented and aggregated to generate segmented response features; Mapping the response distribution structure based on segmented response features; Analyze local response trends based on response distribution structure; Determine the response difference blocks through local response trends; A response difference area in the communication cable is located based on the response difference block.

[0023] In this embodiment, each row in the cable response matrix is ​​input as a node response sequence into the normalization processing unit. The processing unit performs a linear scaling operation on each response sequence based on the minimum-maximum normalization algorithm, where the minimum value is defined as the lower limit of all response values ​​in the current row, and the maximum value is defined as the upper limit of all response values ​​in the current row. The normalization formula is each response value minus the minimum value of the row and divided by the difference between the maximum and the minimum values. All calculation processes are completed in parallel by the FPGA array, and each node sequence is assigned an independent multiplier-adder unit to perform the normalization operation. After normalization, each response value in the matrix is ​​scaled to a real value between 0 and 1. The original amplitude information is backed up by retaining the normalization coefficient vector. The normalized response matrix is ​​stored in the internal DMA cache in double-precision floating-point format and waits for the next step of processing. According to the physical path segment of each row of response data, The statistical operation is completed in the ARM control unit. The segmented aggregation content includes six parameters: maximum value, minimum value, mean value, variance, number of continuous mutation points and maximum mutation value in each aggregation unit. The number of mutation points is counted by taking the point count with an absolute value greater than 0.3 after the difference of adjacent response values. The maximum mutation value is the maximum absolute value of the adjacent differential values ​​in the group. Finally, the six features of each node after aggregation in each group of path segments constitute a feature vector. The aggregation result constitutes a three-dimensional feature matrix, whose dimension is the number of nodes multiplied by the number of aggregation segments multiplied by the six statistical parameters. The matrix adopts column-first arrangement and compression storage for rapid call of subsequent operations. The segmented response feature matrix obtained in the previous step is input into the structural mapping model, which is based on the self-organizing map network (Self-Organizing Map Network). The SOM network performs unsupervised learning. Each input feature vector is mapped to a node position in a two-dimensional grid. The input layer of the SOM network has a dimension of 6, and the output layer has a dimension of 12 by 12 neurons in a two-dimensional plane. The number of training iterations is set to 500, and the learning rate is initialized to 0.01 and gradually decay, and finally obtain the mapping position index of the aggregation segment corresponding to each node. Combined with the original node number and the path segment number, the mapping structure is inversely transformed to generate a complete response distribution structure diagram. The distribution structure diagram is expressed in the form of a two-dimensional image, where the gray value of each pixel represents the comprehensive distance weight of the response feature of the aggregation segment. The image format is 256 times 256 pixels, and the grayscale distribution range is 0 to 255. The distribution diagram is divided into several non-overlapping local windows, and the size of each window is set to 16 times 16 pixels. The local mean filter is used to extract the response change trend within each window. The filter window adopts the Gaussian weighting system. The response trend is expressed as the grayscale average value and the maximum gradient direction of each window. The trend information of each window is represented by a two-dimensional vector. The horizontal axis is the grayscale change direction angle, and the vertical axis is the grayscale increase value. The trend vectors of all windows are aggregated to obtain the response trend distribution vector field of the entire image. The vector field structure uses the vector length greater than the threshold of 0.5 as the local trend significance judgment condition, thereby screening out local areas with obvious response feature changes for the next step of difference block extraction. The boundary extraction operation is performed on all significant local areas in the response trend distribution vector field, using a combination method based on threshold binarization and connected domain analysis. All areas with a vector length greater than 0.5 are set as foreground areas, and other areas are set as backgrounds. The boundary point set of each foreground area is extracted using eight-neighborhood connectivity analysis, and the area and average gradient direction consistency of each boundary area are calculated. Pseudo-difference blocks with an area less than 9 pixels or a direction consistency less than 0.8 are excluded, and only significant and stable difference blocks are retained as the final response difference block results. After all response difference blocks are numbered, their starting path segment, ending path segment, corresponding node number and regional average gradient direction are recorded in array form. The path segment numbers of the above-mentioned difference blocks are mapped one-to-one with the physical cable length calibration data, and calibration is performed. The data is derived from actual scale records during the installation of communication cables. Using point 0 as the starting reference point at the cable input, each path segment is converted to physical coordinates using a spatial index accuracy of 100 points / m. The start and end indexes of the path segments in the difference blocks are multiplied by the segment length of 10 cm to obtain the corresponding physical distance range. The physical location intervals of all difference areas are uniformly numbered and sorted, and a final physical coordinate table of the corresponding difference areas is output. Each entry contains the distance from the cable start end to the area, the area length, the area number, the corresponding node number, and the area feature number. This result is stored in the fault location module for subsequent annotation display and response compensation calculations.

[0024] Preferably, in step S2, calculating the response distortion feature vector of each response difference region, and inferring the abnormal section inside the cable based on the response distortion feature vector includes: Extract response contour data for response difference regions; Separate different types of distortion patterns in response contour data; Calculate the response distortion feature vector of each response difference region according to different types of distortion patterns; performing abnormal pattern aggregation on the response distortion feature vector data to obtain abnormal aggregated data; Relocating the abnormal segment boundary based on the abnormal aggregation data to generate a relocated abnormal segment boundary; The abnormal section inside the cable is determined by relocating the abnormal section boundary.

[0025] In this embodiment, a fiber optic reflective distributed sensor (such as a distributed sensing system based on Brillouin scattering or Rayleigh scattering) is used to perform spatial resolution sampling of the positions along the cable. Sampling points are set at intervals of 1 meter, and the sampling frequency is set to 100 Hz. At each sampling point, a time series of changes in the intensity of the corresponding reflected light is obtained. Then, a sliding window segmentation method is used to extract the response contour curve of the continuous time period in each difference area. The window size is set to 100 sampling points, and the step size is set to 20 sampling points. The extracted curve is recorded in the form of a vector, each vector containing 100 continuous normalized light intensity values, and a profile vector set is formed in position order. The set is used as the response contour data for subsequent processing, and the empirical mode decomposition (EMD) algorithm is used to decompose each response contour vector into components, and each contour vector is decomposed into multiple intrinsic mode functions (Intrinsic Mode Functions). Function, abbreviated as IMF) sequence, and calculate the frequency center value and root mean square amplitude of each IMF. According to the components with frequency center values ​​lower than 5Hz, they are classified as low-frequency deformation variables, those with frequency center values ​​in the range of 5Hz to 20Hz are classified as medium-frequency mutation type distortion, and those with frequency center values ​​higher than 20Hz are classified as high-frequency spike type distortion. At the same time, the energy distribution of each type of IMF subset is calculated, and the energy spectrum distribution curve of each distortion mode is constructed. Then, each response contour is separated into three types of distortion mode data according to its corresponding IMF distribution, and they are stored as independent distortion subsets. The first 10 main frequency components of the energy spectrum data of each type of distortion subset are calculated. The energy is divided and the energy feature components are constructed separately. For low-frequency deformation patterns, the average trend change rate feature is added. For medium-frequency mutation patterns, the position offset (calculated as the change in the peak center relative to the initial position within the subwindow) and the edge steepness coefficient (calculated as the mean of the slopes on both sides) are added. For high-frequency spike patterns, the spike density (the number of spikes per unit time) and spike amplitude variance features are added. Finally, the 6-dimensional features of each distortion pattern are spliced ​​into an 18-dimensional response distortion feature vector. A corresponding feature vector is generated for each difference region and stored in the response distortion feature vector matrix. All feature vectors are clustered using the Gaussian Mixture Model (GMM), with the number of cluster categories K set to 3. After the model is trained and converged using the Expectation Maximum (EM) algorithm, the category label and corresponding cluster probability of each feature vector are output. Clustering is performed for features with a cluster probability less than 0.The feature vectors of 6 are marked as fuzzy samples and removed. The number and mean vector of each category are counted for the remaining clustering results. The clusters with a number greater than 40% are retained as abnormal aggregation data. The difference block numbers and their center coordinates corresponding to this type of abnormal aggregation data are further extracted to form a rough positioning set of abnormal segments. The contour response data are re-extracted from the area 10 meters outward before and after the segment, and the EMD decomposition and distortion feature extraction operations are performed again. A new distortion feature vector is constructed for the extended edge area, and the Euclidean distance is calculated with the original abnormal aggregation feature mean vector. The edge points with a distance less than the preset threshold of 0.3 are selected as boundary update points, and the start and end points of the segment are updated respectively. The relocated abnormal section boundary set is formed by retrieving the starting and ending positions, and the sampling index of the boundary points is recorded in the log file for subsequent traceability. Based on the relocated abnormal section boundary set, the corresponding spatial range between each pair of boundary points is calculated in sequence and mapped to the actual physical location coordinates in the original cable routing path. By matching the position with the cable number and path length in the construction drawings, each abnormal section is associated with the actual physical structure of the communication cable. Ultimately, a complete structured description of the abnormal section within the cable, including the number, starting and ending positions, the trunk line number to which it belongs, and the corresponding difference type, is generated and output as a CSV format text file for reading and use by the fault warning system.

[0026] Preferably, step S3 includes the following steps: Step S31: Collecting geometric features of the communication cable; dividing the geometric radial sectors based on the geometric features of the communication cable; Step S32: matching inner and outer layer features of the geometric radial sector to obtain matching inner and outer multi-layer structural features; splicing the internal cross-section of the cable based on the matching inner and outer multi-layer structural features; Step S33: reconstructing a geometric shape according to the geometric features of the communication cable; fusing the geometric shape and the structure of the internal cross-section of the cable to obtain structural fusion simulation data; Step S34: reorganizing the three-dimensional structure based on the structure fusion simulation data to generate a three-dimensional communication cable.

[0027] In this example, the geometric features of the communication cable were collected using a FARO Quantum Max FaroArm laser 3D scanner to perform high-resolution scanning of the cable cross-section. The device model was ±0.02 mm. The point cloud sampling accuracy was ±0.02 mm. The point cloud data obtained from the scans was filtered using the MeshLab point cloud processing system to remove noise. Isolated point clouds were removed using a radius filter with a radius set to 0.5 mm, the minimum number of neighborhood points is set to 12, and the processed point cloud data is extracted through the RANSAC (random sampling consensus) algorithm to extract the cylindrical shape structure, and the parameters such as the cross-sectional radius, the number of core wires, the angular position of each core wire, the cladding thickness, and the outer diameter of the sheath are vectorized and encoded. A set of geometric parameter matrices including the outer diameter, sheath thickness, core wire distribution angle, and the number of layers are constructed for each cable. The parameter matrix is ​​then input into the polar coordinate sector division module, and 12 geometric radial sectors are divided from 0° to 360° according to the circumferential angle. Each sector corresponds to an angle range of 30°. Each core line and internal structure is mapped to the corresponding sector to form a partition structure feature mapping table and stored in the data cache for subsequent processing. When performing inner and outer layer feature matching for each geometric radial sector, a graph structure registration algorithm based on the mutual information entropy maximization criterion is adopted to establish the inner and outer layer structural features of each sector as undirected graph structures, in which each graph node represents the coordinates of the core line center point, and each edge represents the Euclidean distance between the two core lines. The outer graph is used as the reference structure and the inner graph is used as the structure to be matched. The position of the inner graph nodes is adjusted iteratively to make the mutual information value between the overall graphs The corresponding node pairs after matching are maximized, and the connection vectors between the node pairs are used as the structural mapping primitives. All node pair data are written into the structural mapping matrix. After completing the 12 sector structure matching, the cross-sectional structure path is reconstructed according to the coordinate data of each node pair. The 12 sector structures are spliced ​​in sequence in a clockwise direction. The curve fitting algorithm is used to perform C2 continuity interpolation on the structural lines at the boundaries of adjacent sectors. The interpolation algorithm uses B-spline interpolation, and the number of control points is set to 5. After the fitting is completed, the internal cross-sectional data of the complete communication cable spliced ​​is output, and the three-dimensional construction is used. The API interface program of the modeling tool SolidWorks is used for batch modeling operations. Each set of parameter matrices is converted into a feature sketch and the stretch command is called to build a three-dimensional model. In the specific operation, the outer diameter parameter is set as the boundary radius of the solid stretching. The sheath thickness is constructed by the difference between the inside and outside to construct the offset shell. The core wire distribution is arranged on the circumference using polar coordinates. Each dot matrix is ​​used as the starting point of the primitive stretching body to generate a core wire cylinder. Then, the complete communication cable solid model is constructed through Boolean operations. After the geometric shape is constructed, the model is exported to STL file format and loaded into the ANSYS SpaceClaim module for geometric fusion operation with the previously spliced ​​internal section of the communication cable. The specific fusion operation is to construct the internal section volume using the STL point-surface reconstruction method, and the section data is embedded into the three-dimensional shape through the Boolean intersection operation. After the fusion is completed, the model is stored as Parasolid (.The data was then converted to the BRep (Boundary Representation) format for subsequent structural simulations. The OpenCASCADE geometric modeling library was used to construct a 3D BRep (Boundary Representation) structure. The Parasolid file was first parsed, all boundary surface data extracted, and the corresponding topology constructed. Each cross-section segment was then linearly stretched at 2 mm intervals along the Z axis to construct the segment volume. Finally, all the cross-section segments were combined into a complete 3D communication cable model, with each segment spaced 2 mm apart. This structure includes the coordinate path of each core, the location of each insulation layer, and the 3D geometry of the outer sheath. After this construction, the entire structure was re-meshed using tetrahedral elements with an average element edge length of 0.8 mm. Convergence analysis was used to adjust the element density to ensure mesh continuity at the interfaces between layers. After re-meshing, the complete 3D communication cable structure was generated and exported to the VTK (Visualization Toolkit) format for visualization analysis and subsequent damage simulation modules.

[0028] Preferably, in step S4, unfolding the three-dimensional communication cable layer by layer includes: Confirm the number of three-dimensional communication cable layers based on the geometric characteristics of the communication cable; Deconstruct the three-dimensional communication cable by layers according to the number of layers, thereby generating cable layer analysis data; Confirm the layer boundary outline based on the cable layer analysis data; The global-local structural units are reorganized through the hierarchical boundary contours to obtain multi-layer cable structural units.

[0029] In this embodiment, the structure model of the communication cable is imported into the structure recognition module developed based on Python. The module performs modeling operations based on the OpenCASCADE geometry core library and adopts a hierarchical region growing algorithm to identify the hierarchical structure inside the cable. First, the outer sheath is used as the starting surface, and the packaging layer, coating layer, insulation layer and core layer are identified from the outside to the inside. The boundary line extraction operation is performed on each layer according to the gray density difference corresponding to its material interface. The Sobel operator based on the gradient amplitude is used to scan the gray boundary on the cross-sectional slice. When the boundary spacing is greater than 0.3 mm, it is determined to be a layer of structural boundary. Each group of grayscale mutation intervals is recorded as a layer of structural area, and the corresponding contour point set and thickness parameters are recorded. All the identified layers and corresponding boundary point data are output as a structural level index table. The format of the structural level index table is that each row contains five parameters: level number, starting Z coordinate, ending Z coordinate, average thickness and material label. All data are written to the database for unified scheduling, and VTK (Visualization The Threshold Filter module in the 3D Cable Toolkit performs Z-axis segment extraction on the 3D structure model. For each layer, the corresponding voxel region is extracted using the threshold interval between the starting and ending Z coordinates recorded in the index table. A Marching Cubes surface reconstruction operation is performed during the extraction process to ensure the topological closure of each layer. Each layer is stored in the PolyData format and named Layer_01 to Layer_N, where N is the total number of layers. Each layer is then topologically labeled and volume calculated using a triangular mesh volume segmentation method with a ±1% error. The 3D facet information, surface normal, center point location, and material label of each layer are summarized to form the cable layer analysis data. All data is stored in JSON format, where each entry contains six fields: layer number, number of 3D facets, surface area, volume, mean boundary normal, and material type. A contour tracing algorithm based on OpenCV is used to perform 2D projection of each layer's cross-section. Each layer is first orthogonally projected along the Z axis with a projection accuracy of 0.1 mm, after generating a two-dimensional grayscale image, the contour boundary is extracted by the Canny edge detection algorithm, and the edge detection threshold is set to an upper limit of 150 and a lower limit of 100. Then, the Suzuki-ABE contour search method is used to extract the main contour path and perform curvature fitting operation. The fitting method is five-order Bezier curve fitting. Each curve segment is used to construct the fitting path with 20 control points. After the fitting is completed, the path coordinate sequence of each boundary contour is output and the boundary contour map is constructed. The contour map format is SVG (Scalable Vector Graphics). All contour map files are stored separately according to the structure layer number and summarized into the structure boundary library for subsequent reorganization operation call. All SVG boundary maps are converted into DXF format and imported into AutoCAD Mechanical Global alignment was performed in 2024. An automatic benchmark alignment script was used to align the geometric centers using the center of each layer's boundary as an anchor point. The boundary outline was then extended outward by 2 mm to construct a redundant envelope for structural nesting determination. Based on this nesting determination, local feature regions of each layer were extracted and broken into independent, closed segments, numbered LocalUnit_1 to LocalUnit_M, where M is the number of local structural units. Thickness, area, and spatial coordinates were calculated for each local unit to form a local structural unit data table. Finally, all local structural unit data was merged with the original hierarchical global information. A global-local bidirectional index structure was constructed using hierarchical pointers. This structural unit was stored in JSON format, with each item containing five fields: global layer number, local number, geometric feature sequence, position index, and material property. After completion, the multi-layer cable structural unit data was imported into the structural visualization module and bound to the subsequent fault simulation module for use.

[0030] Preferably, in step S4, identifying the external damage pattern based on the structural features of each layer in the multi-layer cable structure unit, and inferring the abnormal section outside the cable based on the external damage pattern includes: Extracting structural features of each layer in the multi-layer cable deconstruction unit; Extract the main axis of structural features of each level; Identify external damage patterns through structural feature principal axes; Spatial penetration mapping is performed through external damage patterns to generate sheath-braid interface penetration response data; Determine the braided shield layer penetration based on the sheath-braid interface penetration response data, thereby obtaining the braided shield layer abnormality data; Detecting abnormal distribution sections connected in three-dimensional communication cables based on abnormal data of braided shielding layers; Based on the connected abnormal distribution segments, the start and end points and contours of each abnormal segment are located to obtain the external abnormal segment boundary; The outer anomaly section is determined by the outer anomaly section boundary.

[0031] In this embodiment, the hierarchical profile data in the aforementioned structural boundary library is called, and the VTK (Visualization Toolkit) and OpenCASCADE geometric modeling library are combined in the Python environment to extract the feature vectors of the three-dimensional patch data of each layer of the structure. The specific steps include using the principal component analysis (PCA) algorithm to reduce the dimensionality of the hierarchical structure point cloud data, converting the three-dimensional point cloud into a two-dimensional principal plane coordinate system, and then calculating the spatial distribution of the structural feature points contained in each layer, extracting the direction with the largest variance as the structural feature axis of the layer, and the feature axis reflects the main direction distribution of the layer structure. The singular value decomposition (Singular Value Decomposition) is used in the extraction process. The eigenvectors and eigenvalues ​​are obtained using the decomposition method. The eigenvalue represents the variance of the distribution. All eigenvectors are sorted by eigenvalue, and the eigenvector corresponding to the largest eigenvalue is taken as the principal axis direction. The extracted results are stored as three-dimensional vectors with a layer index and the corresponding structural point set. The data format is JSON text, containing the layer number, the characteristic principal axis vector, and the statistical parameters of the corresponding coordinate point set. This allows the extraction and storage of the principal axis of the structural features. Using the calculated principal axis vector as the reference direction, a projection transformation is performed on the structural point cloud at each level. The structural points in three-dimensional space are projected onto the corresponding principal axis direction and its perpendicular plane. Morphological filtering is performed on the projected two-dimensional point cloud, using opening and closing operations to remove noise and enhance edge features. The clustering-based DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm is then used to identify outlier point clusters. The minimum number of points is set to 30, and the neighborhood radius is set to 0.5 mm, after determining the abnormal point cluster, extract its geometric features, including length, width, height and volume parameters, and judge the abnormal pattern type by comparing the volume ratio of the abnormal cluster with the overall volume of the hierarchical structure. At the same time, combined with the relative angle of the main axis of the multi-layer structure and the distribution position of the abnormal cluster, build an external damage pattern model, and finally generate a structural characterization data set describing various damage patterns. The data format adopts the XML standard, including pattern number, spatial coordinate range, volume characteristics and adjacent hierarchical relationship information, and establishes a three-dimensional spatial coordinate mapping relationship between the sheath and the braided layer. Using the above-mentioned hierarchical layered analysis results, the cubic spline interpolation method (Cubic Spline Interpolation) is used based on the spatial interpolation algorithm to realize the spatial mapping grid construction of the sheath layer and the braided shielding layer. The grid density is controlled at 100 units per cubic millimeter. The finite element analysis software ANSYS is used. The Meshing module in Workbench generates volumetric mesh data for the corresponding structure. This mesh data is exported in VTK format for subsequent calculations. For each mesh node, resistivity change data and microwave reflection intensity data collected by surface sensors are collected. Data fusion techniques are used to associate the sensor data with spatial mesh nodes. A weighted average method is used to calculate the penetration response value for each mesh node. The weight is calculated using an exponential decay function based on the spatial distance between the node and the actual sensor location. This ultimately generates a penetration response data matrix for the sheath-braid interface. This matrix is ​​formatted as a two-dimensional array, with each row corresponding to a mesh node number and columns containing the node's spatial coordinates and penetration response intensity. This achieves a tightly coupled mapping between physical sensing information and geometric space. The penetration response matrix data is normalized, using a minimum-maximum normalization method to linearly map the response values ​​to the range of 0 to 1. The threshold is then set to 0 based on the damage threshold in the historical sample library.65. Nodes exceeding the threshold are marked as infiltration anomalies. Connectivity analysis of the anomalies is performed using a three-dimensional region growing algorithm based on six-neighborhood connections. The connected domain volume threshold is set to 5 cubic millimeters. Small isolated point groups are removed and the abnormal region point cloud set processed by region growing is output. The volume morphology of the abnormal point cloud is analyzed, the bounding box and volume parameters of the abnormal region are calculated, and the woven shielding layer abnormal data file is generated. The file is stored in JSON format and contains the abnormal region number, spatial range, volume information and spatial relationship with the adjacent layer, providing basic data for subsequent abnormal segment detection. The abnormal point cloud set in the abnormal data file is read and density-based spatial aggregation analysis is performed on the point cloud. The clustering algorithm DBSCAN is used to divide the abnormal points with a spatial distance less than 1 mm into the same abnormal segment. The minimum point number threshold is set to 50 to eliminate low-density noise. After clustering, the spatial topological relationship of each abnormal segment is calculated, and the KD tree structure is used to index the adjacent abnormal segments to achieve efficient spatial connection judgment. The abnormal segment topology map is constructed based on the adjacency relationship. The nodes in the map represent abnormal segments, and the edges represent spatial adjacency relationships. All connected subgraphs are detected by the depth-first search algorithm and marked as connected abnormal segments. An abnormal segment list file is generated, which includes the abnormal segment number, the number of points, the spatial coordinate range and the adjacent abnormal segment number, to achieve spatial connectivity judgment and number archiving of abnormal segments. Based on the spatial information of the connected abnormal segments, the boundary extraction algorithm is used to locate the start and end positions and contours of the abnormal segments. The voxel slicing technology is used to slice the abnormal segment point cloud data at equal intervals. The slice spacing is set to 0.5 mm. Cut along the Z axis to obtain a series of two-dimensional cross-sectional point clouds. The convex hull algorithm (Convex Hull Algorithm) is used for each cross section. Hull) calculates the boundary point set of the abnormal section, and the boundary points are stored in the form of a two-dimensional coordinate array. The three-dimensional boundary is reconstructed along the slice direction by combining all the cross-section boundary points. The boundary surface model is constructed using the polygon mesh generation algorithm. The surface smoothing process uses the Laplace smoothing algorithm. The number of iterations is set to 10 times and the smoothing coefficient is 0.15. Finally, a three-dimensional geometry file of the abnormal section boundary model is generated and saved in OBJ format. The file contains the start and end Z coordinates of the abnormal section, the coordinates of the boundary surface vertices and the topological connection relationship, so as to achieve high-precision spatial positioning and three-dimensional reconstruction of the abnormal section boundary. The abnormal section boundary model is connected to the The three-dimensional models of the outer sheath of the communication cable were spatially compared, and a nearest neighbor search algorithm based on point clouds was used to calculate the minimum distance between the two models. The distance threshold was set to 0.2 mm. All boundary model points whose distance from the outer sheath surface was less than the threshold were marked as external abnormal regions. These abnormal regions were further merged through Boolean operations to form continuous external abnormal segments. The final boundary description file for the external abnormal segment was output in STL (triangular patch mesh) format. This file contains geometric information such as the spatial position of the abnormal segment, surface normal, and boundary curve, providing accurate spatial data for subsequent maintenance positioning and fault diagnosis.

[0032] It is particularly important to detect abnormal distribution sections in three-dimensional communication cables based on abnormal data of the braided shield layer, including: Perform 3D node mapping fusion on the abnormal data of the braided shield layer to generate 3D node abnormality labels; The three-dimensional node abnormal label data is spatially clustered and topologically reorganized to obtain adjacent abnormal point clusters; Based on the adjacent outlier cluster data, boundary connectivity reduction processing is performed to obtain the outlier cluster distribution area; The contour lines of the abnormal cluster distribution area are extracted, and the regional trajectory is constructed based on the contour lines, so as to obtain the connected abnormal distribution sections in the three-dimensional communication cable.

[0033] In this embodiment, the coordinate data of all abnormal points in the space are collected, and a high-precision laser scanner is used to collect point cloud data with an accuracy of less than 0.1 mm. The point cloud data is imported into the three-dimensional space processing system, and the point cloud is efficiently organized and indexed by constructing an octree structure. Combined with the resistance change value measured by the time series sensor, the abnormal index of each node is weighted according to the distance from the sensor and the sensor signal strength. The weight gradually decreases exponentially with the increase of distance. The weight is calculated based on the fact that when the node distance from the sensor increases, the weight decreases at a fixed attenuation ratio. After the comprehensive weight is combined, the spatial coordinates are bound to the abnormal index, and the nodes with abnormal index values ​​greater than 0.7 are marked as abnormal and assigned an abnormal label of 1. The other nodes are Assign the label 0 to form a three-dimensional node data table containing node number, spatial coordinates and abnormal label to ensure the accurate positioning of abnormal nodes in space. Filter all node data with abnormal label 1 and use the spatial index structure to search for neighboring points of these nodes. The search range is limited to a radius of 1 mm. The adjacent abnormal nodes are classified into the same cluster unit. The density-based clustering algorithm is used to group the abnormal nodes. Each cluster is set to contain at least 50 nodes and the neighborhood range is 0.8 mm. Isolated abnormal nodes are eliminated to generate multiple abnormal point clusters. The spatial adjacency relationship between each cluster is then calculated. Only when the distance between the closest points between two clusters does not exceed 1.2 mm, they are considered adjacent to build an adjacency relationship network between clusters. The network is constructed by using a depth-first traversal algorithm to identify all connected cluster combinations, and a structured data file containing the cluster number, number of nodes, and spatial range is output to ensure that the clustering results are spatially coherent and complete. When the abnormal cluster distribution area is obtained by performing boundary connectivity reduction processing based on the adjacent abnormal point cluster data, a three-dimensional point cloud boundary construction technology is used for each abnormal point cluster to generate a spatial polyhedron boundary that tightly surrounds the point set. The tightness of the boundary is controlled by setting the boundary parameter to 2 mm. This parameter ensures that the boundary can appropriately envelop the shape details of the point set. Subsequently, the spatial connectivity of the boundary polyhedrons of adjacent abnormal point clusters is judged by calculating the overlapping area and adjacent distance between boundaries. Only when the overlapping area is greater than 10 square millimeters and the distance between boundaries is less than When the error is 1 mm, a spatial merging operation of polyhedrons is performed. During the merging, three-dimensional geometric Boolean operation technology is used to merge multiple boundary polyhedrons into a larger continuous area, and small isolated areas are eliminated. The merged abnormal clusters are stored in the form of triangular meshes. Each abnormal cluster is assigned a number and a spatial range identifier to ensure that the abnormal distribution area is continuous and has a complete spatial representation. The contour lines of the abnormal cluster distribution area are extracted, and the regional trajectory is constructed based on the contour lines. When obtaining the connected abnormal distribution sections in the three-dimensional communication cable, the three-dimensional mesh model of the abnormal cluster is first uniformly sliced ​​along the cable axial direction. The slice spacing is set to 0.5 mm. The slices obtain a series of two-dimensional closed curve point sets. The curve simplification algorithm is used to reduce the number of contour points. The error during simplification is controlled within 0.0.5 mm, maintaining the shape accuracy of the contour line. The simplified point set is connected into a continuous contour line, with the contour line data points evenly distributed and the point spacing controlled at 0.3 mm. The curvature is calculated on the contour line, and the inflection points with significant curvature changes are identified. The trajectory line is smoothed using the cubic spline interpolation method, with the smoothing level set to a medium level. The interpolation result is a smooth and continuous spatial trajectory line, which includes the spatial position and curvature information of each point. The trajectory data is saved in a structured text format. The trajectory shows the spatial representation of the abnormal distribution section in the communication cable, which is used for subsequent fault detection and location analysis.

[0034] Preferably, in step S5, performing interactive authentication in combination with the abnormal section inside the cable and the abnormal section outside the cable, and when the abnormal sections overlap, determining the synchronously overlapping section as a cable fault includes: Combine the abnormal sections inside the cable and the abnormal sections outside the cable to obtain the section mapping data, where the maximum allowable section start and end difference is 20cm; Filter the synchronous overlapping segments in the segment mapping data, where 85% is set as the minimum overlap judgment threshold; When a synchronous overlapping section exists, a cable fault determination is performed based on the synchronous overlapping section, and the location of the faulty cable section is uploaded.

[0035] In this embodiment, when combining the abnormal section inside the cable and the abnormal section outside the cable to obtain the section mapping data, the abnormal section data collected by the internal sensor of the cable and the abnormal section data obtained by three-dimensional scanning and abnormality detection on the outside of the cable are imported into a unified coordinate system. The internal abnormal section data includes the start and end position coordinates of the cable length direction, and the external abnormal section data obtains the corresponding length direction range through the three-dimensional space positioning algorithm. The start and end coordinates of the internal and external abnormal sections are aligned using the spatial registration algorithm, and the error of the start and end positions of the cable length direction is allowed to be no more than 20 cm. The upper limit of the error is determined by the equipment accuracy and data acquisition delay, and a weighted method is used. Overlap detection algorithm calculates the ratio of the overlapping part of the length of two abnormal segments to the length of the shorter segment, and uses it as the segment mapping data to generate a mapping data table including the start and end points of the internal segment, the start and end points of the external segment, the overlapping length and the overlapping ratio to ensure the accuracy of the data correspondence and facilitate subsequent analysis. When screening the synchronous overlapping segments in the segment mapping data, the overlap ratio field in the mapping data table is used for screening, and the overlap ratio value is compared with the set minimum overlap judgment threshold of 85%. All records with an overlap ratio greater than or equal to 85% are identified as synchronous overlapping segments. The screening operation is implemented using a data filtering algorithm. The algorithm first sorts the mapping data table and sorts them by overlap ratio. The examples are arranged from large to small, and then the list is traversed. For each record, the overlapping ratio is compared with the 85% threshold, and the data that meets the conditions is retained. In addition, for each synchronous overlapping segment record, the corresponding start and end position data are combined to calculate the specific spatial range of the overlapping segment, and output a synchronous overlapping segment list. The list contains the starting position, end position and overlapping ratio for the fault judgment module to call. When a synchronous overlapping segment exists, the cable fault is judged based on the synchronous overlapping segment, and the faulty cable segment position is uploaded. The judgment process is implemented by the fault judgment unit. First, the synchronous overlapping segment list is received, and the fault confirmation operation is performed on each segment in turn. The fault confirmation is confirmed by comparing with the preset threshold. The threshold is 85% overlap ratio to ensure the accuracy of fault section determination. After confirmation, the positioning module is called to determine the spatial position of the fault section in the entire cable through the start and end length coordinates of the section. The spatial position data is calibrated in meters, and the positioning accuracy reaches 0.1 meters. The fault section location information is then uploaded to the monitoring center in a structured data format. The upload process uses a wired communication interface based on the TCP / IP protocol to ensure the real-time and stability of data transmission. The transmission content includes the start and end point coordinates of the fault section, the overlap ratio and the fault level identification, ensuring that the monitoring center can obtain accurate fault cable location data in a timely manner and complete remote monitoring and management of the fault section.

[0036] Preferably, in step S5, when the abnormal section is a unilateral abnormality, performing joint abnormality location by using the abnormal section inside the cable and the abnormal section outside the cable includes: Combining the abnormal section inside the cable and the abnormal section outside the cable, when the synchronous overlapping section does not exist, the abnormal section is determined to be a unilateral abnormality; Analyze the abnormal correlation between the abnormal section inside the cable and the abnormal section outside the cable based on the single-side abnormality; Determine cable abnormalities through abnormal correlation; The cable fault section location is predicted based on the cable abnormality, and the cable fault section is queried based on the cable fault section location.

[0037] In this embodiment, the abnormal section inside the cable and the abnormal section outside the cable are combined. When the synchronous overlapping section does not exist, the abnormal section is determined to be a unilateral abnormal section. The specific operation is to first import the spatial position data of the abnormal section inside the cable and the abnormal section outside the cable into a unified three-dimensional coordinate system. The internal abnormal section data comes from the resistance change and temperature abnormality indicators collected by the built-in sensor. The external abnormal section data obtains the cable surface abnormality information through a high-resolution three-dimensional laser scanner. The two sets of data are compared using a spatial matching algorithm, and the overlap threshold is set to 85%. If the overlap ratio of the two sets of abnormal sections in the length direction does not reach the threshold, it is determined to be a non-synchronous overlapping section, and then marked as a unilateral abnormality. The data processing flow The Pandas library in the Python environment was used to filter and annotate the data table. Spatial comparison was achieved using a point cloud alignment algorithm, with a registration error of less than 0.2 mm. The spatial coordinate range, amplitude, and timestamp data of the abnormal section within the cable, as well as the corresponding information of the abnormal section outside the cable, were extracted. A multidimensional correlation matrix was constructed using a spatiotemporal correlation analysis method. The matrix dimensions included abnormal position deviation, amplitude difference, and time interval. Potential correlations were determined by calculating pairs of nodes with an abnormal position deviation of less than 30 cm and a time interval of less than 5 minutes. A correlation coefficient was used to measure the similarity of the abnormal amplitudes, with a correlation coefficient threshold of 0. 7. Node pairs that meet the conditions are included in the association network, and graph theory algorithms are used to construct an abnormal association graph. Nodes represent abnormal sections, and edges represent associations. The degree centrality of nodes in the graph is calculated to identify key abnormal sections. The abnormal association graph data is input. Combined with cable design parameters and historical fault databases, an abnormal discrimination model based on machine learning is used for evaluation. The model input features include the spatial distribution density of abnormal nodes, the strength of associations, and the fluctuation of abnormal amplitudes. The discrimination model is constructed using the random forest algorithm. The model training data set contains 1,000 cable abnormality samples of different fault types. After feature normalization, the input data is mapped to the model. The model outputs the probability distribution of abnormal types and sets the threshold. The probability is 0.8, and a probability higher than this value is considered a definite anomaly. The model runs in the TensorFlow framework on a GPU-accelerated workstation. The judgment results are further verified by the rule engine. The rules include that the continuous length of abnormal nodes must not be less than 10 cm, and the abnormal amplitude change must not be less than 0.2 ohms. Finally, a comprehensive assessment report of the cable abnormality is output. The report format includes the abnormality level, the spatial location of the abnormal section, and the fault type code. The determined abnormality is input into the fault prediction module. This module uses a spatiotemporal regression model and utilizes the mapping relationship between abnormal indicators and fault locations in historical fault data. The regression model parameters are determined by least squares fitting, and the spatial error is controlled at 0.Within 15 meters, with a time error limit of one hour, the model outputs the specific start and end locations of the section where the fault may occur, with coordinates accurate to the millimeter level. The cable maintenance database interface is then called to query fault records based on the predicted location. Matching conditions include an overlap of greater than 75% in the start and end coordinates of the fault section and a consistent fault type. Query results include historical maintenance records for the fault section, fault type, and handling recommendations. All data is accessed through the database using Structured Query Language (SQL) statements, and the returned data is formatted as JSON, ensuring efficient and accurate fault prediction and query operations.

[0038] It is particularly important to identify cable anomalies through abnormal correlations, including: Deconstruct the abnormal correlation data based on structural correspondence to obtain abnormal pattern matching data; Perform regional coverage overlap simulation based on anomaly pattern matching data to generate anomaly coverage coincidence data; According to the abnormal coverage coincidence data, abnormal path connectivity is tracked to obtain the abnormal mapping path; Determine cable anomalies through abnormal mapping paths.

[0039] In this embodiment, the input abnormal association graph data is imported into the graph data processing module. The abnormal association graph is composed of nodes and edges. Nodes represent abnormal detection points, and edges represent the association relationship between abnormalities. The graph data is stored in the form of an adjacency matrix. The element value in the adjacency matrix indicates the strength of the association between two nodes. The spectral clustering algorithm in graph theory is used to Clustering) performs feature decomposition on the adjacency matrix, extracts the main eigenvectors, divides the nodes into several abnormal pattern clusters, and uses the K-means clustering algorithm to complete the node grouping in the feature space. The number of clusters is determined based on the silhouette coefficient index during clustering. The clustering process adopts iterative optimization until the cluster center is stable. The deconstructed abnormal pattern cluster is mapped into abnormal pattern matching data through graph structure. The abnormal pattern matching data contains a list of nodes in the cluster and its correlation strength distribution. A three-dimensional covering sphere is defined according to the node space coordinates and node radius of each abnormal pattern cluster in the abnormal pattern matching data. The node radius is set according to the sensor detection radius parameter, which is generally 5 cm. The covering sphere is described by the spherical equation. A three-dimensional spatial index structure such as octree is used to perform spatial management on all covering spheres to improve query efficiency and to search for all abnormal patterns. The common pattern cluster covers the spheres for pairwise spatial intersection detection. The intersection calculation is based on the sphere distance formula. When the distance between the centers of the two spheres is less than the sum of the radii of the two spheres, it is determined to be an overlapping area. The overlapping area data includes the overlapping volume size and overlapping node pairs. The overlapping simulation process is implemented in C++, and CGAL (computational geometry library) is called for precise spatial geometry calculation. The output of abnormal coverage coincidence data contains the three-dimensional spatial coordinates, volume size and node number list of each overlapping area. The overlapping area information in the abnormal coverage coincidence data is imported to construct a path tracking graph. The nodes in the graph represent overlapping areas, and the edges represent the spatial connectivity between overlapping areas. The spatial connectivity is completed by calculating the shortest distance between the boundaries of overlapping areas. The distance threshold is set to 1 cm. If the distance is less than the threshold, an edge connection is established. The path tracking adopts the depth-first search algorithm (Depth-First The system traverses all nodes in the graph to find connected paths between covered areas. During the traversal process, access marks and path stacks are maintained to ensure that nodes are not visited repeatedly. When searching for paths, edges with larger association strengths are traversed first to ensure that the paths represent the most significant abnormal connectivity. The final path set contains multiple abnormal mapping paths. Each path records the starting point, end point, and path length. The path data is stored in the graph database format. Path query and management are implemented through the Neo4j graph database. The database index strategy is based on node attributes and edge weights. The abnormal mapping paths are imported into the fault judgment unit. The fault judgment unit combines cable design parameters and historical abnormal case libraries, and uses a rule engine to match path features. Path features include path length, number of path nodes, mean path association strength, and path spatial direction.Paths with a length exceeding 15 centimeters and at least five nodes are marked as key abnormal paths. The mean correlation strength is calculated by averaging the weights of all path edges. The spatial orientation of the path is calculated by fitting the three-dimensional coordinate curves of the path nodes to calculate the direction of the tangent vector, which is used to determine the physical orientation of the abnormal path. The fault determination unit uses a weighted scoring mechanism to comprehensively score each path. The scoring rules are developed based on expert experience, and the scoring results are mapped to the abnormality level. The final output is structured data including the abnormal path identifier, the starting and ending spatial coordinates, and the abnormality level.

[0040] The present invention also provides a communication cable fault monitoring system for executing the communication cable fault monitoring method described above, the communication cable fault monitoring system comprising: The excitation modeling module is used to inject timing structure excitation coding into the communication cable to obtain multi-node excitation records; perform full-path phase differential tracking on the communication cable based on the excitation records to obtain phase response trajectory data; and construct a cable response matrix based on the phase response trajectory data; The response location module is used to map the response distribution structure of the cable response matrix and locate the response difference area in the communication cable based on the response distribution structure; calculate the response distortion feature vector of each response difference area and infer the abnormal section inside the cable based on the response distortion feature vector; The simulation and reconstruction module is used to collect the geometric features of the communication cable; perform radial mapping fitting based on the geometric features of the communication cable, and reconstruct the internal profile of the cable based on the radial mapping fitting data; simulate the communication cable through the geometric features of the communication cable and the internal profile of the cable to generate a three-dimensional communication cable; A structure recognition module is used to unfold the three-dimensional communication cable layer by layer to obtain a multi-layer cable structure unit; identify the external damage pattern based on the structural characteristics of each layer in the multi-layer cable structure unit, and infer the abnormal section of the cable outside based on the external damage pattern; The fault judgment module is used to perform interactive authentication by combining the abnormal sections inside the cable and the abnormal sections outside the cable. When the abnormal sections overlap, the synchronous overlapping sections are judged as cable faults. When the abnormal section is unilaterally abnormal, the abnormality is jointly located through the abnormal sections inside the cable and the abnormal sections outside the cable, thereby generating a cable fault section.

[0041] The present invention injects sequential structure excitation coding into the communication cable through the implementation of the excitation modeling module, which can effectively obtain multi-node excitation records and provide basic data for subsequent phase response analysis. The phase response trajectory data obtained by the full-path phase differential tracking technology enables the cable response characteristics to be fully captured, and the constructed cable response matrix provides a reliable basis for subsequent response positioning. The introduction of the response positioning module can accurately locate the response difference area in the communication cable through response distribution structure mapping. The calculated response distortion feature vector provides quantitative support for inferring the abnormal section inside the cable. The simulation reconstruction module collects the geometric features of the communication cable and performs radial mapping fitting, which greatly improves the reconstruction accuracy of the cable internal section. The generated three-dimensional communication cable model provides an intuitive basis for subsequent structural analysis. The layer-by-layer unfolding three-dimensional model enables the multi-layer cable structure unit to be presented in a refined manner, and the structural identification The module can identify external damage patterns based on the structural characteristics of each layer, and the inferred external abnormal sections provide an important reference for troubleshooting. The fault judgment module integrates the abnormal sections inside and outside the cable, and achieves more accurate fault identification through an interactive authentication mechanism. When the abnormal sections overlap, the cable fault can be quickly and accurately determined. In the case of unilateral abnormality, the fault source can be accurately identified through the joint positioning of the abnormality. The implementation of the overall system improves the accuracy and efficiency of communication cable fault monitoring, promotes the intelligent management of communication infrastructure, reduces operation and maintenance costs, extends the service life of cables, ensures the stability and reliability of communication systems, provides a solid technical foundation for the development of future communication networks, enhances the ability to predict potential faults, promotes the establishment of a rapid response mechanism, greatly improves the security and stability of communication networks, and promotes the development of the industry towards intelligence and automation.

[0042] The present invention is therefore intended to be illustrative and non-restrictive in all respects, with the scope of the invention being defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the application documents are intended to be embraced therein.

[0043] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A communication cable fault monitoring method, characterized in that: The following steps are involved: Step S1: injecting a timing structure excitation code into the communication cable to obtain a multi-node excitation record; Perform full-path phase differential tracking on the communication cable based on the excitation record to obtain phase response trajectory data; Constructing a cable response matrix based on the phase response trace data; Step S2: mapping the cable response matrix to a response distribution structure, and locating the response difference area in the communication cable according to the response distribution structure; Calculating the response distortion feature vector of each response difference area, and inferring the abnormal section inside the cable based on the response distortion feature vector; Step S3: Collecting geometric features of the communication cable; performing radial mapping fitting based on the geometric features of the communication cable, and reconstructing the internal profile of the cable based on the radial mapping fitting data; performing communication cable simulation processing based on the geometric features of the communication cable and the internal profile of the cable, thereby generating a three-dimensional communication cable; Step S4: unfolding the three-dimensional communication cable layer by layer to obtain a multi-layer cable structure unit; Identify external damage patterns based on the structural features of each layer in the multi-layer cable structure unit, and infer the abnormal section outside the cable based on the external damage pattern; Step S5: Interactive authentication is performed on the abnormal section inside the cable and the abnormal section outside the cable. When the abnormal sections overlap, the synchronous overlapping section is determined to be a cable fault. When the abnormal section is unilaterally abnormal, the abnormality is jointly located through the abnormal section inside the cable and the abnormal section outside the cable, thereby generating a cable fault section.

2. The communication cable fault monitoring method according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: Performing timing structure excitation coding injection on the communication cable, using a pseudo-random modulation signal with a pulse width of 50-150ns, and setting the excitation point spacing to a physical interval of 20-50cm to obtain multi-node excitation records; Step S12: performing phase differential locking on the multi-node excitation records to generate full-path phase differential tracking data; Step S13: extracting phase response trajectory data based on the full-path phase differential tracking data, with the extraction path segmentation granularity set to 10 cm per segment; Step S14: Integrate the response of each node through the phase response trajectory data to obtain node response integration data, where the number of response nodes is limited to 16-128; and construct a cable response matrix through the node response integration data.

3. The communication cable fault monitoring method according to claim 1, characterized in that: In step S2, mapping the cable response matrix to a response distribution structure and locating the response difference area in the communication cable according to the response distribution structure includes: Performing normalized amplitude modulation transformation on the cable response matrix to obtain a normalized response matrix; The normalized response matrix is ​​segmented and aggregated to generate segmented response features; Mapping the response distribution structure based on segmented response features; Analyze local response trends based on response distribution structure; Determine the response difference blocks through local response trends; A response difference area in the communication cable is located based on the response difference block.

4. The communication cable fault monitoring method according to claim 1, characterized in that: In step S2, the response distortion feature vector of each response difference region is calculated, and based on the response distortion feature vector, the abnormal section inside the cable is inferred to include: Extract response contour data for response difference regions; Separate different types of distortion patterns in response contour data; Calculate the response distortion feature vector of each response difference region according to different types of distortion patterns; performing abnormal pattern aggregation on the response distortion feature vector data to obtain abnormal aggregated data; Relocating the abnormal segment boundary based on the abnormal aggregation data to generate a relocated abnormal segment boundary; The abnormal section inside the cable is determined by relocating the abnormal section boundary.

5. The communication cable fault monitoring method according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: Collecting geometric features of the communication cable; dividing the geometric radial sectors based on the geometric features of the communication cable; Step S32: matching inner and outer layer features of the geometric radial sector to obtain matching inner and outer multi-layer structural features; splicing the internal cross-section of the cable based on the matching inner and outer multi-layer structural features; Step S33: reconstructing a geometric shape according to the geometric features of the communication cable; fusing the geometric shape and the structure of the internal cross-section of the cable to obtain structural fusion simulation data; Step S34: reorganizing the three-dimensional structure based on the structure fusion simulation data to generate a three-dimensional communication cable.

6. The communication cable fault monitoring method according to claim 1, characterized in that: Step S4 of unfolding the three-dimensional communication cable layer by layer includes: Confirm the number of three-dimensional communication cable layers based on the geometric characteristics of the communication cable; Deconstruct the three-dimensional communication cable by layers according to the number of layers, thereby generating cable layer analysis data; Confirm the layer boundary outline based on the cable layer analysis data; The global-local structural units are reorganized through the hierarchical boundary contours to obtain multi-layer cable structural units.

7. The communication cable fault monitoring method according to claim 1, characterized in that: In step S4, the external damage pattern is identified based on the structural features of each layer in the multi-layer cable structure unit, and the abnormal section outside the cable is inferred based on the external damage pattern, including: Extracting structural features of each layer in the multi-layer cable deconstruction unit; Extract the main axis of structural features of each level; Identify external damage patterns through structural feature principal axes; Spatial penetration mapping is performed using external damage patterns to generate sheath-braid interface penetration response data; Determine the braided shield layer penetration based on the sheath-braid interface penetration response data, thereby obtaining the braided shield layer abnormality data; Detecting abnormal distribution sections connected in three-dimensional communication cables based on abnormal data of braided shielding layers; Based on the connected abnormal distribution segments, the start and end points and contours of each abnormal segment are located to obtain the external abnormal segment boundary; The outer anomaly section is determined by the outer anomaly section boundary.

8. The communication cable fault monitoring method according to claim 1, characterized in that: In step S5, performing interactive authentication on the abnormal section inside the cable and the abnormal section outside the cable, and determining the synchronously overlapping section as a cable fault when the abnormal sections overlap, includes: Combine the abnormal sections inside the cable and the abnormal sections outside the cable to obtain the section mapping data, where the maximum allowable section start and end difference is 20cm; Filter the synchronous overlapping segments in the segment mapping data, where 85% is set as the minimum overlap judgment threshold; When a synchronous overlapping section exists, a cable fault determination is performed based on the synchronous overlapping section, and the location of the faulty cable section is uploaded.

9. The communication cable fault monitoring method according to claim 1, characterized in that: In step S5, when the abnormal section is a unilateral abnormality, performing joint abnormal location by using the abnormal section inside the cable and the abnormal section outside the cable includes: Combining the abnormal section inside the cable and the abnormal section outside the cable, when the synchronous overlapping section does not exist, the abnormal section is determined to be a unilateral abnormality; Analyze the abnormal correlation between the abnormal section inside the cable and the abnormal section outside the cable based on the single-side abnormality; Determine cable abnormalities through abnormal correlation; The cable fault section location is predicted based on the cable abnormality, and the cable fault section is queried based on the cable fault section location.

10. A communication cable fault monitoring system, characterized in that: For executing the communication cable fault monitoring method according to claim 1, the communication cable fault monitoring system comprises: The excitation modeling module is used to inject timing structure excitation coding into the communication cable to obtain multi-node excitation records; perform full-path phase differential tracking on the communication cable based on the excitation records to obtain phase response trajectory data; and construct a cable response matrix based on the phase response trajectory data; The response location module is used to map the response distribution structure of the cable response matrix and locate the response difference area in the communication cable based on the response distribution structure; calculate the response distortion feature vector of each response difference area and infer the abnormal section inside the cable based on the response distortion feature vector; The simulation and reconstruction module is used to collect the geometric features of the communication cable; perform radial mapping fitting based on the geometric features of the communication cable, and reconstruct the internal profile of the cable based on the radial mapping fitting data; simulate the communication cable through the geometric features of the communication cable and the internal profile of the cable to generate a three-dimensional communication cable; A structure recognition module is used to unfold the three-dimensional communication cable layer by layer to obtain a multi-layer cable structure unit; identify the external damage pattern based on the structural characteristics of each layer in the multi-layer cable structure unit, and infer the abnormal section of the cable outside based on the external damage pattern; The fault judgment module is used to perform interactive authentication by combining the abnormal sections inside the cable and the abnormal sections outside the cable. When the abnormal sections overlap, the synchronous overlapping sections are judged as cable faults. When the abnormal section is unilaterally abnormal, the abnormality is jointly located through the abnormal sections inside the cable and the abnormal sections outside the cable, thereby generating a cable fault section.

Images