Fault signal detection system and analysis positioning method

Abstract

The invention provides a fault signal detection system and an analysis positioning method, and relates to the technical field of signal detection. An injection module is used for periodically applying narrowband acoustic disturbance capable of adaptively selecting frequency and power to an optical cable sheath; the passive wireless label nodes arranged along the line collect multi-dimensional data such as arrival timestamps, amplitudes, frequencies, phases, local temperatures and the like in real time, and upload the data through near-field coupling. And the convergence module completes node identification association and batch aggregation, the analysis and construction module constructs an acoustic propagation path map, and after temperature difference segmentation compensation, an actual laying path is dynamically recovered by adopting a map search algorithm. And the projection and identification module maps the map to a geographic information system coordinate system and outputs a fault space position, thereby realizing real-time and accurate optical cable fault positioning under a topology-free prior condition, remarkably reducing manual inspection and misjudgment rates, and being suitable for a multi-branch and multi-repair long-distance communication or power optical cable scene.

Description

Technical Field

[0001] This application relates to the field of signal detection technology, and more specifically, to a fault signal detection system and analysis and localization method. Background Technology

[0002] In modern communication infrastructure, optical cables are widely used for interconnection between telecommunications, energy, transportation, and data centers due to their advantages such as large transmission capacity and good stability. However, due to environmental influences, construction errors, natural aging, or accidental damage by third parties, optical cable lines inevitably experience faults such as partial fiber breaks, loose joints, or damaged sheaths. Once such faults occur, if they cannot be located and repaired in a timely manner, it will lead to communication service interruption or severe performance degradation, and may even cause large-scale information silos or economic losses. Currently used optical cable fault detection methods, such as optical time domain reflectometer (OTDR) testing methods, mostly adopt single-end testing mode, relying on pre-drawn and kept-updated line topology diagrams and accurate line length information. However, in actual operating environments, due to the complexity of optical cable network layout, frequent changes and repairs to line topology, and the cumulative differences between the original drawings and the on-site layout, it is often impossible to provide real-time, accurate, and up-to-date topology data. This makes traditional single-end optical testing methods susceptible to multiple reflections, branch scattering, or complex line distribution, resulting in large location errors or blind spots. In addition, such methods usually require multiple manual comparisons and on-site surveys, resulting in slow response times, large positioning errors, strong reliance on manual labor, and high operation and maintenance costs.

[0003] Therefore, under the condition of lacking or having difficulty obtaining accurate line topology information, how to achieve real-time and accurate spatial positioning of optical cable fault points, and significantly reduce reliance on manual methods and the rate of misjudgment, has become an urgent technical problem to be solved. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this application provides a fault signal detection system and analysis and location method.

[0005] In a first aspect, this application provides a fault signal detection system, comprising:

[0006] The injection module is configured to inject a preset narrowband acoustic disturbance signal into the target optical cable;

[0007] Passive wireless tag nodes are spaced along the target optical cable. Each tag node is configured to generate and output multidimensional data characterizing the propagation characteristics of the acoustic disturbance signal when it receives the acoustic disturbance signal.

[0008] The aggregation module is configured to collect the multidimensional data and associate it with the corresponding tag node identifiers;

[0009] The analysis and construction module is configured to calculate the propagation differences between each of the label nodes based on the multidimensional data, and construct an acoustic propagation path map.

[0010] The projection and recognition module is configured to: map the acoustic propagation path map to the geographic information system coordinate system corresponding to the geographic coordinates of the tag nodes; determine whether the propagation difference between each tag node in the multidimensional data exceeds a preset propagation threshold; if it exceeds the threshold, identify the tag node as an abnormal propagation node; and determine the geographic information system coordinates of the abnormal propagation node as the fault spatial location.

[0011] As an optional implementation, injecting a preset narrowband acoustic disturbance signal into the target optical cable includes:

[0012] The injection module is driven to inject at least two sets of test acoustic pulses in ascending frequency order within a preset frequency sweep range, and the arrival amplitude of each set of test acoustic pulses is recorded by a reference tag node set at the deployment point where the injection module is located.

[0013] The target frequency is selected based on the arrival amplitude of each group of test acoustic pulses. The injection frequency corresponding to the test acoustic pulse with the largest amplitude and attenuation rate lower than the preset threshold is determined as the target narrowband injection frequency.

[0014] Based on the arrival amplitude measured in real time by the reference tag node, the injection power is adjusted so that the amplitude of the acoustic disturbance signal of the target narrowband injection frequency at the reference tag node is kept within the preset target amplitude range.

[0015] After the injection power adjustment is completed, the acoustic disturbance signal of the target narrowband injection frequency is cyclically injected into the target optical cable according to the preset injection cycle.

[0016] As an optional implementation, it also includes:

[0017] Each of the passive wireless tag nodes is further configured to report the local temperature data and the identifier of its previous tag node together when outputting the multidimensional data; wherein, the previous tag node refers to the tag node that is adjacent to the current tag node and located upstream of the current tag node in the acoustic propagation path map along the direction of acoustic disturbance signal propagation.

[0018] In response to the absence of other tag nodes between the current tag node and the injection module, the reference tag node is regarded as the previous tag node of the current tag node;

[0019] If the current label node is the farthest label node in the acoustic propagation path, then its previous label node is the nearest upstream label node adjacent to it.

[0020] The analysis and construction module is configured to generate a temperature difference vector between tag pairs based on the local temperature data of adjacent tag nodes;

[0021] Based on the temperature difference vector, the acoustic propagation path map is divided into several isothermal segments, and the temperature difference of each isothermal segment does not exceed a preset threshold. For each isothermal segment, a pre-stored temperature-sound velocity mapping data table corresponding to the optical cable sheath material and tension level is called, and the mapped sound velocity parameters are used to compensate for the propagation difference of the segment.

[0022] As an optional implementation, generating and outputting multidimensional data characterizing the propagation characteristics of the acoustic disturbance signal includes:

[0023] Starting from the moment when the amplitude of the acoustic disturbance signal is detected to exceed the preset trigger threshold for the first time, the acoustic disturbance signal is converted from analog to digital according to the preset sampling rate to obtain a discrete acoustic sampling sequence;

[0024] Within the preset analysis window, the discrete acoustic sampling sequence is processed. When the amplitude of the discrete acoustic sampling sequence first exceeds the rising edge threshold, the moment is recorded as the arrival timestamp. The root mean square amplitude value is calculated as the arrival amplitude. The maximum energy frequency component is extracted as the arrival frequency through fast Fourier transform, and the instantaneous phase of the maximum energy frequency component is obtained simultaneously.

[0025] According to the preset data frame format, the tag node identifier, the previous hop tag node identifier, the local temperature data, and the arrival timestamp, arrival amplitude, arrival frequency and instantaneous phase are encapsulated into an uplink data frame;

[0026] The uplink data frame is sent to the aggregation module using near-field energy coupling.

[0027] As an optional implementation, the step of calculating the propagation differences between each of the tag nodes and constructing an acoustic propagation path map based on the multidimensional data includes:

[0028] Based on the arrival amplitude and the instantaneous phase, label node data with a peak signal power to noise mean square error ratio lower than a preset signal-to-noise ratio threshold and / or a phase difference between adjacent frames greater than 180° are removed to obtain an effective label dataset.

[0029] For each label node in the effective label dataset, select the other label node with the smallest propagation time difference that is greater than zero as the adjacent node pair, and adaptively generate adjacent edges.

[0030] A composite weight is assigned to each adjacent edge, and the composite weight is obtained by linearly combining the propagation time difference and arrival amplitude attenuation value of the adjacent edge according to a preset weighting coefficient;

[0031] Using the composite weights, an acoustic propagation path map is constructed based on a graph search algorithm, and incremental information is updated to the acoustic propagation path map in real time when a new effective label dataset is generated.

[0032] As an optional implementation, constructing an acoustic propagation path map using the composite weights and a graph search algorithm includes:

[0033] Each label node in the effective label dataset is abstracted as a directed graph vertex, the reference label node is used as the source vertex, and the adjacent node pairs determined by the composite weight are mapped as directed edges from the source vertex to the propagation direction.

[0034] During the graph search process, when the composite weight of a candidate directed edge exceeds a preset pruning threshold, the expansion of the corresponding edge is terminated.

[0035] Based on the pruned directed graph, the Dijkstra algorithm is used to find the shortest cumulative composite weight path from the source vertex to the remaining label nodes, and the acoustic propagation path result set is obtained.

[0036] For each path in the acoustic propagation path result set, output a path record consisting of a tag node sequence, cumulative propagation time difference, and cumulative arrival amplitude attenuation, and write the path record into the acoustic propagation path map.

[0037] As an optional implementation, writing the path record into the acoustic propagation path map includes:

[0038] For each path in the acoustic propagation path result set, the path confidence is calculated based on the cumulative composite weight of the path and the number of tag nodes it contains;

[0039] The acoustic propagation path result set is sorted in descending order according to the path confidence, and the top N paths with the highest confidence are selected as the target path set, where N is the preset maximum number of output paths.

[0040] For each path in the target path set, the average temperature difference of the path is calculated based on the local temperature data of the tag nodes that constitute the path.

[0041] The target path set, along with the corresponding tag node sequence, cumulative propagation time difference, cumulative arrival amplitude attenuation, and the average temperature difference value, are written into the acoustic propagation path map.

[0042] As an optional implementation, determining whether the propagation difference between the label nodes in the multidimensional data exceeds a preset propagation threshold, and if it does, identifying the label node as an abnormal propagation node; and determining the geographic information system coordinates of the abnormal propagation node as the fault spatial location includes:

[0043] The baseline mean μ and standard deviation σ are calculated for the set of propagation time differences between adjacent tag nodes. When the absolute value of the difference between a certain propagation time difference and the baseline mean is greater than k·σ, the corresponding downstream tag node is marked as a candidate abnormal node; where k is a preset abnormality coefficient.

[0044] According to the direction of propagation, consecutively appearing candidate abnormal nodes are grouped. When the number of nodes in the same group is not less than the preset consecutive abnormal length, the group is identified as an abnormal propagation node cluster.

[0045] For each cluster of anomalous propagation nodes, the center coordinates of the cluster are calculated as the spatial location of the fault based on the geographic information system coordinates of the first and last tag nodes constituting the cluster, through linear interpolation along the optical cable path.

[0046] The spatial location of the fault is written into the acoustic propagation path map, and the corresponding geographic information system coordinates are output to the projection and recognition module.

[0047] Secondly, this application provides an analytical positioning method, including:

[0048] Inject a preset narrowband acoustic disturbance signal into the target optical cable;

[0049] Passive wireless tag nodes are set along the target optical cable at intervals. Each tag node is configured to generate and output multidimensional data characterizing the propagation characteristics of the acoustic disturbance signal when it receives the acoustic disturbance signal.

[0050] Collect the multidimensional data and associate it with the corresponding tag node identifiers;

[0051] Based on the multidimensional data, the propagation differences between each of the label nodes are calculated, and an acoustic propagation path map is constructed.

[0052] The acoustic propagation path map is mapped to the geographic information system coordinate system corresponding to the geographic coordinates of the tag nodes; it is determined whether the propagation difference between each tag node in the multidimensional data exceeds a preset propagation threshold. If it does, the tag node is identified as an abnormal propagation node; and the geographic information system coordinates of the abnormal propagation node are determined as the fault spatial location.

[0053] Compared with existing technologies, this application achieves self-organization detection based on acoustic propagation paths by utilizing passive wireless tag nodes deployed along the optical cable and injected narrowband acoustic disturbance signals when accurate line topology drawings are lacking or unavailable. Compared with traditional methods that rely on complete topology drawings and single-ended optical pulse testing, this application can significantly reduce the need for manual intervention and external drawing maintenance. Attached Figure Description

[0054] Figure 1 This is a schematic diagram of a fault signal detection system provided in an embodiment of this application;

[0055] Figure 2 A flowchart illustrating a method for injecting a preset narrowband acoustic disturbance signal into a target optical cable, as provided in this application embodiment;

[0056] Figure 3 This is a schematic diagram of differential processing provided in an embodiment of this application;

[0057] Figure 4 This is a schematic diagram of a dynamic process processing system based on multi-module collaboration in a smart logistics scenario, provided in an embodiment of this application.

[0058] Figure labeling: 10, Injection module; 20, Passive wireless tag node; 30, Aggregation module; 40, Analysis and construction module; 50, Projection and recognition module. Detailed Implementation

[0059] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0060] See Figure 4 The diagram shown is a schematic of a fault signal detection system provided in an embodiment of this application, including: an injection module 10, a passive wireless tag node 20, a convergence module 30, an analysis and construction module 40, and a projection and recognition module 50, wherein:

[0061] The injection module 10 is configured to inject a preset narrowband acoustic disturbance signal into the target optical cable;

[0062] Passive wireless tag nodes 20 are spaced along the target optical cable. Each tag node is configured to generate and output multidimensional data characterizing the propagation characteristics of the acoustic disturbance signal when it receives the acoustic disturbance signal.

[0063] The aggregation module 30 is configured to collect the multidimensional data and associate it with the corresponding tag node identifier;

[0064] Analysis and construction module 40 is configured to calculate the propagation differences between each of the label nodes based on the multidimensional data, and construct an acoustic propagation path map;

[0065] The projection and recognition module 50 is configured to: map the acoustic propagation path map to the geographic information system coordinate system corresponding to the geographic coordinates of the tag nodes; determine whether the propagation difference between each tag node in the multidimensional data exceeds a preset propagation threshold; if it exceeds the threshold, identify the tag node as an abnormal propagation node; and determine the geographic information system coordinates of the abnormal propagation node as the fault spatial location.

[0066] Among them, narrowband acoustic disturbance signal specifically refers to an acoustic excitation signal with a single center frequency or a very limited bandwidth. This signal is usually injected into the optical cable in the form of pulses to ensure good propagation and signal recognition, thereby effectively reducing noise interference and signal loss, and improving the overall detection accuracy and positioning precision of the system.

[0067] The injection module 10 can consist of a piezoelectric excitation device and a corresponding drive circuit. By controlling the frequency, amplitude, and pulse pattern of the acoustic vibration, it applies a preset narrowband acoustic disturbance signal to the outer sheath of the optical cable. The center frequency of the acoustic disturbance signal can be selected between several kilohertz and over ten kilohertz, which balances propagation within the optical cable sheath material with avoiding severe energy attenuation due to excessively high frequencies. To achieve tunable acoustic output, the control circuit inside the injection module 10 changes the exciter's drive voltage or drive pulse width within a certain range based on a preset sweep frequency range or the echo amplitude measured at the reference tag.

[0068] When the injection module 10 starts working, its piezoelectric transducer is attached or coupled to the optical cable sheath, forming a series of mechanical waves that propagate along the cable direction through short and regular pulses. Compared with traditional fiber optic flaw detection or single-ended optical pulse testing, by directly applying acoustic disturbances to the optical cable sheath, it is possible to penetrate multi-layer sheaths or loose tube structures more flexibly, avoiding signal dead zones at local jumper points.

[0069] In some embodiments, the injection module 10 may also employ a chirp method to smoothly transition from a lower frequency to a higher frequency, thereby obtaining uniform and controllable propagation characteristics across different frequency bands of the sheath for further multidimensional data analysis.

[0070] In terms of mechanical coupling, the injection module 10 is detachably installed near the exposed end of the optical cable or the junction box, using rubber clamps or special coupling agents to reduce acoustic energy loss at the injection interface. If the optical cable is laid in a complex environment or there are interference sources nearby, additional shielding covers or vibration damping components can be installed on the injection module 10 to reduce the impact of environmental noise on acoustic disturbance signals.

[0071] Since the acoustic disturbances generated by the injection module 10 can propagate across fiber breaks and disperse in the branch structure, they can be used in conjunction with passive wireless tag nodes 20 along the route to gradually collect the characteristics of the optical cable transmission path without prior topological information. Compared with the break detection scheme based on optical pulses, this acoustic injection mode is more adaptable to optical fiber splices or branch sections, avoiding the generation of difficult-to-interpret echoes in complex patch cord or junction box areas.

[0072] In addition, narrowband acoustic perturbations can achieve a better signal-to-noise ratio within a controllable frequency band, which not only facilitates the subsequent detection of multi-dimensional parameters and time delay calculation by tag nodes, but also helps to obtain a more uniform power distribution between the near and far ends.

[0073] In this way, regardless of whether the optical cable branches, bends, or is repaired multiple times along the way, the injection module 10 can maintain effective "tapping" of the downstream node, laying an accurate data foundation for the entire fault detection system.

[0074] For example, a rapid fault location is required for an optical cable approximately 10 kilometers long. After multiple relocations and partial repairs, there are significant discrepancies between the original drawings and the site layout, making it difficult to determine the specific location of fiber breaks or loose loops using conventional location methods that rely on optical reflection testing.

[0075] In this embodiment, an exposed sheath section of approximately 1 meter in length is reserved at one end of the optical cable as an injection point, and an injection module 10 equipped with a tunable piezoelectric exciter and control circuit is installed at this point. The exciter is set to an initial operating center frequency of 10kHz and a pulse width of 2ms. Each injection rapidly increases the pulse amplitude to a peak value of 80V through a linearly rising voltage. In this way, a short and concentrated acoustic disturbance is formed on the optical cable.

[0076] To ensure sufficient signal amplitude at the remote tag node while preventing near-end oversaturation, the injection module 10 is also equipped with a reference sensor to monitor the echo signal strength at a distance of approximately 0.5 meters from the injection point in real time. If the amplitude at this location exceeds a preset safety value, the system automatically shortens the pulse width or reduces the peak voltage to avoid excessive impact on the optical cable. Conversely, if the signal amplitude reported from the remote tag node is below the 3dB signal-to-noise ratio threshold, the pulse width is appropriately extended to 3ms, and the voltage is increased to a peak value of 100V to ensure that acoustic disturbances can penetrate sections requiring multiple repairs or junction boxes.

[0077] The acoustic disturbance generated by the injection module 10 can effectively overcome the problems caused by the unknown topology, and provide sufficient signal strength and reliability for fault detection and location of downstream nodes.

[0078] Regarding the aforementioned passive wireless tag node 20:

[0079] Passive wireless tag nodes 20 are fixed at intervals along the optical cable to effectively cover the entire laying section. Each tag node integrates a near-field coupling antenna, a piezoelectric miniature transducer, and low-power control circuitry to capture and analyze acoustic disturbance signals applied by the injection module 10. When acoustic waves propagate along the sheath to the location of the tag node, the miniature transducer converts them into voltage changes, which are then amplified and filtered internally before being sent to the control circuitry. Since the nodes themselves do not require an external power supply, the necessary energy is obtained from external near-field coupling or spontaneous vibration, thus achieving a passive design to simplify deployment and subsequent maintenance.

[0080] After detecting an acoustic disturbance signal, the tag node will discretely collect the signal amplitude and phase information according to the pre-configured sampling rate, and perform preliminary analysis on the sampled waveform to extract multi-dimensional features such as arrival timestamp, peak amplitude, maximum energy frequency component, or instantaneous phase.

[0081] In some embodiments, the node is also equipped with sensors for measuring ambient temperature or vibration noise, enabling the tag node to report not only the transmission characteristics of the acoustic wave itself but also auxiliary information about the external environment. By integrating this multi-dimensional data, propagation path reconstruction and fault location can be performed more accurately in subsequent stages.

[0082] To achieve data backhaul without affecting the normal operation of the optical cable, each tag node packages the extracted multidimensional data into data frames and reports them to the aggregation module 30 via near-field coupling or passive radio frequency. In this mode, the tag node only briefly wakes up to send data after detecting acoustic disturbances and completing data extraction; otherwise, it remains in deep sleep mode, ensuring extremely low overall energy consumption. If no valid acoustic signal is detected within a certain excitation cycle, the tag node will not wake up or send data. This further optimizes the operating efficiency of the passive nodes, enabling long-term online monitoring of the laid lines.

[0083] In actual installation, the passive wireless tag node 20 is typically encapsulated in a waterproof and dustproof housing and securely attached to the surface of the optical cable sheath using adhesive, cable ties, or other fixing methods. Different spacing deployment schemes can be selected according to site conditions, such as every 50 meters or every 100 meters.

[0084] For example, on a fiber optic cable in an urban area to be monitored, a passive wireless tag node 20 can be installed approximately every 50 meters to capture acoustic disturbance signals generated by the injection module 10. In areas with bends, jumpers, or dense junction boxes, the deployment of the tag nodes can be appropriately densified as needed to obtain higher resolution information along the route. After detecting an acoustic disturbance, each tag node extracts multi-dimensional data such as the arrival timestamp, peak amplitude, instantaneous phase, or maximum energy frequency of the disturbance according to a pre-set sampling frequency and analysis window. After a short wake-up, the data is encapsulated into an uplink data frame and transmitted to the aggregation module 30 via near-field coupling or other feasible wireless methods.

[0085] In this exemplary scenario, the tag node itself does not rely on an external power supply, but rather operates through energy coupling with the external environment (such as RF near-field excitation, micro-vibration collection, etc.). To minimize power consumption, the node software logic ensures that it enters a sleep state after each reporting, only reactivating when an acoustic disturbance is detected again. Since each node captures some or all of the features of the same pulse within the same excitation cycle, the system can obtain distributed, multi-dimensional detection data within the same time slice. Compared to traditional single-point measurement or single-end reflection testing methods, this distributed monitoring significantly improves the reconstruction and identification capabilities of the optical cable's condition in the subsequent analysis and construction module 40.

[0086] In the specific implementation of the passive wireless tag node 20, it may include a low-power microcontroller, a set of piezoelectric microphone arrays or miniature transducers, and a small-capacity memory for temporarily storing sampling results. Depending on the design, the tag node may choose to directly perform local digital processing on the acquired acoustic waveforms, or only extract some parameters (such as amplitude and timestamps), and then package them into data frames. If it is necessary to dynamically adjust the detection sensitivity or sampling rate of the tag node during operation, the aggregation module 30 can also remotely configure the tag node through near-field communication to achieve finer-grained monitoring or power management.

[0087] During installation, since each tag node must be tightly coupled to the optical cable sheath to accurately receive acoustic disturbances, waterproof sleeves, fixing clamps, or other stabilizing methods can be considered to attach the nodes to the optical cable surface. If the line needs to be expanded or relocated later, only the corresponding number of tag nodes need to be added to the newly laid optical cable section or the relocation site to continue utilizing the acoustic detection capabilities of this system, without the need to redraw accurate topology maps or conduct extensive manual verification. As the injection module 10 periodically excites acoustic pulses into the optical cable, the tag nodes synchronously monitor the acoustic disturbances arriving at their location and complete signal acquisition and feature extraction locally, laying the foundation for subsequent propagation difference analysis and fault location. Through this scalable distributed deployment method, the overall health status of the optical cable will always be under visualized and detectable management. Even after multiple branches, partial repairs, or jumpers, the system can still maintain high fault location accuracy and real-time performance.

[0088] Regarding the aforementioned aggregation module 30:

[0089] The aggregation module 30 is set up in the fault signal detection system mainly to realize the unified collection and management of multi-dimensional data reported by passive wireless tag nodes 20 scattered along the optical cable. The multi-dimensional data referred to here may include, but is not limited to: arrival timestamps, amplitude information, frequency components, phase characteristics, ambient temperature data, and tag node identification information. Since the same acoustic excitation is collected by multiple nodes simultaneously, the aggregation module 30 needs to be able to distinguish different excitation periods and merge the data corresponding to each node within the same excitation period into the same data set.

[0090] To achieve this function, the aggregation module 30 can be equipped with a near-field communication or passive radio frequency receiving interface in its system architecture, enabling it to interact with all distributed tag nodes. After completing local signal detection and data packaging, each tag node sends an uplink data frame carrying the tag node identifier to the aggregation module 30. When the aggregation module 30 receives the data frame, it determines which specific tag node it comes from and which detection period or acoustic disturbance sequence the data corresponds to by reading the node identifier in the data frame. At the same time, the aggregation module 30 can also determine its relative order in the entire excitation period based on information such as the timestamp in the data frame, thereby performing aggregation processing of the observation results of multiple nodes according to time sequence or disturbance sequence.

[0091] During the collection of multidimensional data, the aggregation module 30 can perform the following operations:

[0092] In the data receiving stage, the data type (such as amplitude, phase, temperature, etc.) is identified according to the data frame header or preset format, and then allocated to the corresponding data buffer area;

[0093] In the data processing stage, a data entry is created for each detection cycle, merging the observations of all nodes under the same acoustic excitation into the same entry to form a multi-dimensional observation matrix. This matrix uses the identifier of the labeled node (e.g., node ID, location code, etc.) as the row index and physical quantities such as timestamp, amplitude, and frequency as the column index.

[0094] During the data association process, the aggregation module 30 ensures that each observation data point is matched with the correct node identifier based on a pre-recorded "node ID-spatial location" mapping table or real-time assigned identification information. For uplink data frames that arrive sequentially or delayed within the same detection period, the aggregation module 30 can insert them into the correct detection period based on the sequence number or excitation event number in the data, maintaining data integrity and continuity.

[0095] After the data is correctly aggregated and associated with node identifiers, the aggregation module 30 can submit the corresponding data batch to the analysis and construction module 40. For example, if the multi-node observations within a detection period contain features such as arrival timestamps, amplitude attenuation, and frequency shifts, the aggregation module 30 will package the data of all nodes in the same batch and mark it with "nth acoustic excitation" or "excitation sequence ID=xxx" so that the analysis and construction module 40 knows that this is a complete set of observations under the same acoustic disturbance event. Based on this, the analysis and construction module 40 can calculate the time delay difference and amplitude difference of each node, construct an acoustic propagation path map, and further locate fault points or determine node anomalies.

[0096] In summary, the purpose of the aggregation module 30 is to: collect multidimensional data reported by each passive wireless tag node 20 using the same or compatible near-field communication protocol; identify and parse the node identifier, detection period or acoustic excitation information carried in each data, and classify and archive them; and after completing the aggregation of multi-node observations, prepare for subsequent analysis, such as packaging the same batch of data and submitting it to the analysis and construction module 40.

[0097] Regarding the analysis and construction module 40 mentioned above:

[0098] In practical implementation, the analysis and construction module 40 is used to reconstruct the actual laying path or connection relationship of the optical cable based on the propagation characteristics between different tag nodes after the system completes the collection and aggregation of multidimensional data, so as to accurately infer the possible fault location.

[0099] Its core process includes: parsing and reading multi-node observation data from the convergence module 30 under the same excitation cycle, calculating the propagation differences between each tag node, and then using these differences to construct or update an acoustic propagation path map.

[0100] After the analysis and construction module 40 obtains multi-dimensional parameters such as the arrival timestamp, amplitude attenuation, frequency shift, and temperature of the tag nodes, it first classifies and indexes these observation data according to the correspondence between node identifiers and excitation events. For example, for the same acoustic pulse excitation, the analysis and construction module 40 reads all tag node records that successfully reported during the excitation period and calculates the time delay difference, amplitude difference, or phase difference between adjacent nodes one by one. By comparing these propagation differences, the degree of time delay increment, energy attenuation, or frequency drift experienced by the acoustic wave as it propagates from the upstream node to the downstream node can be determined. If the time delay difference between two nodes is the smallest and the amplitude attenuation is within a reasonable range, it can be inferred that the physical distance between these two nodes is relatively short and the optical cable connection is relatively direct.

[0101] In some embodiments, the analysis and construction module 40 treats all nodes as vertices in a graph structure and constructs weighted edges based on the propagation differences between adjacent nodes. The weights of the weighted edges are typically obtained by combining factors such as time delay difference and amplitude attenuation, which can reflect both the physical distance of the optical cable and the differences in optical cable quality or splice loss.

[0102] Furthermore, the module can utilize common graph search algorithms (such as shortest path search or minimum spanning tree algorithm) to reconstruct the actual path structure of the optical cable on the weighted graph. If similar propagation difference distributions are obtained under multiple excitations, the module will merge the results and continuously optimize or update the acoustic propagation path map. In this way, even in the absence of original topology drawings, or when pipelines have been repeatedly spliced ​​and have complex branches, the system can still dynamically infer the laying relationship of the optical cable with relatively accurate accuracy.

[0103] In some optional designs, the analysis and construction module 40 can also perform hierarchical processing on the branch structure: if certain nodes are detected to form sub-paths that differ significantly from the main path, it indicates that there may be optical cable branches with independent directions; the module can extract the branch to generate a local path map and maintain its correlation with the trunk map. For observation data accumulated over multiple detection cycles, the module can also use a time-series overlay method to evaluate whether the optical cable shows a continuous increase in attenuation or the disappearance of a certain tag node in a specific section by comparing the changes in propagation differences at different times (which may indicate cable breakage or worsening sheath damage).

[0104] After completing the acoustic propagation path map construction or update, the analysis and construction module 40 sends the key results of the acoustic propagation path map to the projection and identification module 50. These results may include inter-node distances or loss estimates, as well as markers for certain high-risk sections. Upon receiving the map, the projection and identification module 50 maps it to an actual geographic coordinate system, providing maintenance personnel with a visualized view of the fiber optic cable distribution and potential fault locations.

[0105] Regarding the aforementioned projection and recognition module 50:

[0106] This system is responsible for converting the acoustic propagation path maps generated or updated by the analysis and construction module 40 into visual coordinates corresponding to the actual geographical locations of the labeled nodes, and determining whether a node is abnormal based on a preset propagation threshold. Its implementation logic can be roughly divided into two levels: spatial coordinate mapping and threshold detection and fault labeling.

[0107] Regarding spatial coordinate mapping, the projection and recognition module 50 first reads the path map provided by the analysis and construction module 40. This map contains metrics on the associations and propagation differences between labeled nodes (such as latency differences and amplitude attenuation). Simultaneously, to implement the abstract graph structure into a Geographic Information System (GIS), the module retrieves the geographic coordinates corresponding to each labeled node from a previously established "node identifier-spatial coordinate" mapping table or reference coordinate information maintained on-site. These coordinates can be in latitude and longitude form, or they can be planar coordinates or projected coordinates, depending on the on-site environment and the usage habits of the upper-level GIS platform.

[0108] After extracting or transforming the node coordinates, the projection and recognition module 50 will draw the propagation edges between each node in the map onto the GIS coordinate system, thus forming a schematic diagram of one or more line distributions. For node sets that are determined to be independent branches or tributary structures during the analysis and construction phase, the module can also distinguish them on the GIS interface with different colors or line types, making it easy for maintenance personnel to clearly understand the current actual route and connection relationship of the optical cable.

[0109] After completing spatial mapping, the projection and recognition module 50 will compare the propagation difference characteristics of each node with a set threshold based on the data accumulated from each node or multiple detections within the same detection cycle. For example, if the delay difference, amplitude attenuation, or frequency offset between node A and its neighboring nodes is found to significantly exceed the threshold in a certain detection, the module will mark node A as an "abnormal propagation node." This abnormality may indicate that the optical cable has experienced problems such as sheath damage, fiber breakage, or loose joints near node A. When the same out-of-limit condition occurs in multiple consecutive detection cycles, it further indicates that the reliability of the fault at that location is relatively high.

[0110] Once the projection and recognition module 50 determines that a certain tag node exceeds the preset propagation threshold range, it will present the node as a fault mark in the GIS coordinate system and generate an alarm prompt on the maintenance personnel's display terminal or management backend.

[0111] Furthermore, to better assist on-site handling, the module can further combine optical cable linear interpolation or coordinates of other surrounding nodes to spatially refine the fault location or estimate its center point, thereby providing a more accurate "fault spatial location" coordinate. If integrated with external databases or geographic pipeline information systems in the future, the module can also add the administrative district, street name, or location of ground landmarks where the fault point is located, enabling the repair team to quickly and accurately locate the fault site.

[0112] Through the aforementioned technical means, the projection and recognition module 50 combines the topology information and propagation characteristic results output by the analysis and construction module 40 with the spatial coordinates of the tag nodes to achieve visual mapping and fault node identification in the GIS coordinate system. Furthermore, by utilizing the set propagation threshold and anomaly judgment strategy, the module can automatically identify suspected problem sections and output their exact location in the real physical space, thereby helping maintenance and repair teams to efficiently and accurately locate faults. Working in conjunction with other modules of this invention, even in environments lacking original topology drawings or with multiple relocations, it can still ensure accurate detection and rapid handling of optical cable fault points.

[0113] As an optional implementation, please refer to Figure 2 , Figure 2 A flowchart of a method for injecting a preset narrowband acoustic disturbance signal into a target optical cable, provided in an embodiment of this application, includes steps S201 to S204, wherein:

[0114] S201: Drive the injection module 10 to inject at least two sets of test acoustic pulses in ascending frequency order within a preset sweep frequency range, and record the arrival amplitude of each set of test acoustic pulses by the reference tag node set at the deployment point of the injection module 10.

[0115] S202: Select the target frequency based on the arrival amplitude of each group of test acoustic pulses, and determine the injection frequency corresponding to the test acoustic pulse with the largest amplitude and attenuation rate lower than the preset threshold as the target narrowband injection frequency.

[0116] S203: Based on the arrival amplitude measured in real time by the reference tag node, adjust the injection power so that the amplitude of the acoustic disturbance signal of the target narrowband injection frequency at the reference tag node is kept within the preset target amplitude range.

[0117] S204: After completing the injection power adjustment, the acoustic disturbance signal of the target narrowband injection frequency is cyclically injected into the target optical cable according to the preset injection cycle.

[0118] In some scenarios, to quickly find the optimal or near-optimal injection frequency and power during initial excitation, enabling remote tag nodes to reliably detect acoustic disturbance signals, the injection module 10 can optionally be configured to operate in a frequency sweep mode. This involves injecting several sets of probe acoustic pulses in ascending frequency order according to a pre-defined frequency sweep range, with a reference tag node located at or near the injection module 10 recording the arrival amplitude of each set of probe pulses. The reference tag node can employ a transducer and control circuit similar to the downstream passive wireless tag node 20, but is deployed near the injection module 10, physically adjacent to the injection point, to allow for real-time measurement of the acoustic energy amplitude actually applied to the optical cable sheath.

[0119] During this frequency sweep, each set of probe pulses is applied to the optical cable at a slightly different injection frequency. The reference tag node then determines the propagation effect of that frequency under the current sheath and environmental conditions based on the arrival amplitude. If the recorded amplitude of a certain set of pulses is much higher than that of other frequency points, it indicates that the frequency is easier to propagate in the optical cable with lower loss. If the recorded amplitude of certain frequency points is significantly reduced or the attenuation rate is too high, it indicates that there may be sheath resonance attenuation, air cavity effects, or other reasons in that frequency band, making it difficult to form a stable and detectable sound wave.

[0120] Based on the arrival amplitude of each set of test pulses, the internal or upper-level control program of the injection module 10 will select the injection frequency corresponding to the test pulse with the largest amplitude and an attenuation rate lower than a certain preset threshold, and determine it as the subsequent target narrowband injection frequency. This target narrowband injection frequency can remain unchanged for a certain period of time, or it can be periodically rescanned to adapt to changes in environmental conditions (such as temperature, burial depth, and sheath aging degree) over time.

[0121] After selecting the target frequency, the system also needs to adjust the injected power accordingly. At this time, the reference tag node will measure the arrival amplitude of the injected pulse at its location in real time. If the amplitude exceeds or falls below the preset target amplitude range, the injection module 10 will correspondingly increase or decrease the exciter's drive voltage, or adjust the pulse width, to ensure that the amplitude at the reference tag node remains stably within a suitable range that guarantees detection by subsequent downstream nodes. This closed-loop power control avoids sheath damage caused by near-end oversaturation and also prevents insufficient detection signals at the far end.

[0122] After the frequency sweeping, frequency selection, and power adjustment steps are completed, the injection module 10 enters the periodic injection mode. According to a pre-set injection cycle (e.g., every 1 second or every 10 seconds), the module repeatedly applies acoustic disturbance signals to the optical cable based on a determined target narrowband injection frequency and appropriate power parameters. All downstream distributed passive wireless tag nodes 20 can monitor and capture this disturbance within the corresponding cycle and report the multi-dimensional data to the aggregation module 30 for subsequent analysis.

[0123] For example, the preset frequency sweep range is within 8kHz to 12kHz. Upon startup, the module injects multiple sets of probe acoustic pulses in increments of 0.5kHz. The reference tag node records the amplitude of each pulse set and determines the probe frequency with the largest amplitude and attenuation rate <10% as the optimal frequency for this round. If a frequency consistently achieves a high amplitude value in multiple probes, the system selects this frequency as the target narrowband injection frequency and performs closed-loop adjustment of the injection power to maintain it at 80dB within the target amplitude range at the reference tag node. Finally, acoustic pulses at the target frequency are injected into the optical cable at a 1Hz injection cycle, providing long-term, stable fault detection support to the downstream passive wireless tag node 20.

[0124] In this way, the system can adaptively select the optimal injection signal parameters under different optical cable materials, temperatures, or burial depths, thereby improving the detection signal-to-noise ratio and availability of remote tag nodes.

[0125] As an optional implementation, it also includes:

[0126] Each of the passive wireless tag nodes 20 is further configured to report the local temperature data and the identifier of its previous tag node together when outputting the multidimensional data; wherein, the previous tag node refers to the tag node that is adjacent to the current tag node and located upstream of the current tag node in the acoustic propagation path map along the direction of acoustic disturbance signal propagation.

[0127] In response to the absence of other tag nodes between the current tag node and the injection module 10, the reference tag node is regarded as the previous tag node of the current tag node;

[0128] If the current label node is the farthest label node in the acoustic propagation path, then its previous label node is the nearest upstream label node adjacent to it.

[0129] The analysis and construction module 40 is configured to generate a temperature difference vector between tag pairs based on the local temperature data of adjacent tag nodes;

[0130] Based on the temperature difference vector, the acoustic propagation path map is divided into several isothermal segments, and the temperature difference of each isothermal segment does not exceed a preset threshold. For each isothermal segment, a pre-stored temperature-sound velocity mapping data table corresponding to the optical cable sheath material and tension level is called, and the mapped sound velocity parameters are used to compensate for the propagation difference of the segment.

[0131] In certain high-temperature gradient or complex environment scenarios for optical cables, there may be significant differences in soil temperature, burial depth, or duct ventilation along the route, which directly affects the propagation speed and attenuation coefficient of acoustic waves within the sheath. To address this, this application further configures each passive wireless tag node 20 with an additional temperature measurement function, allowing the node to include local temperature data and the "previous hop tag node identifier" when reporting multidimensional data. In this way, the analysis and construction module 40 can incorporate the temperature difference between adjacent nodes into the calculation of propagation differences, performing temperature compensation or segmented correction when necessary, thereby significantly improving the matching accuracy between the actual propagation path and latency.

[0132] In practical implementation, each passive wireless tag node 20 has a built-in or external low-power temperature sensor (such as a digital semiconductor thermometer or thermistor). After detecting acoustic disturbances, it packages the local real-time temperature data, along with arrival timestamps, peak amplitude, and other parameters, into an uplink data frame. If the system also needs to determine the relative order or topological relationship between nodes, it can use a chained tagging system based on the "previous hop tag node identifier": when a node reports data, it tells the system which upstream node it received the acoustic wave from, i.e., who the "previous hop tag node" is.

[0133] To facilitate the formation of a clear path chain, each passive wireless tag node 20 is assigned a globally unique ID. When node A detects an acoustic disturbance during the current excitation cycle, it checks which node is "closer to injection module 10 in the propagation direction". If there are no other nodes within that distance, the reference tag node or injection module 10 itself is considered A's previous hop node; if A itself is the farthest node, the previous hop node can be traced back to its nearest upstream node. Through this identification mechanism, after receiving data at the aggregation module 30, the nodes can be automatically linked together in the order of "previous hop - next hop" to form an approximate acoustic propagation chain structure.

[0134] To facilitate temperature compensation in the analysis and construction module 40, the system records the temperature difference between each pair of "adjacent nodes (A, B)" to form a "label-to-temperature difference vector." If the temperature difference between two adjacent nodes exceeds a certain preset threshold, it indicates that there may be a significant temperature gradient within that segment, and the acoustic wave velocity deviates considerably from the normal temperature value. Based on this, the analysis and construction module 40 divides the path map into several "isothermal segments," ensuring that the temperature difference between nodes within the same segment does not exceed the threshold. This allows it to be approximated as a uniform temperature range for unified calculation of wave velocity or attenuation coefficient.

[0135] To obtain more accurate acoustic propagation parameters, the module can access a pre-established "temperature-sound velocity" or "temperature-attenuation coefficient" mapping data table, which lists the sound velocity values ​​under different temperatures, sheath materials, and tension levels. The system will select or interpolate the corresponding sound velocity based on the average temperature of the segment, and then correct the time delay of all node pairs within that segment, thereby more accurately inferring the distance or attenuation between nodes. If there are tension differences in a segment (such as pipe subsidence or stretching), additional data can be used for composite queries.

[0136] For example, the temperature difference between optical cable segments in a certain urban area can reach 5-8℃, especially the temperature difference between the surface line and the deeply buried pipe gallery. Each passive wireless tag node 20 includes the local temperature when reporting data. and the previous hop label node identifier Upon receiving the data, the analysis and construction module 40 will first... By performing an interpolation calculation, if the difference is greater than 3°C, that segment of the link is considered the segment boundary, forming a new isothermal segment. The temperature-velocity mapping table v=f(T) is then used to correct the velocity of this segment. In this way, the propagation delay error caused by temperature differences can be significantly reduced, and the positioning accuracy is improved accordingly.

[0137] in, This refers to the local temperature, i.e., the ambient temperature detected by the current node (e.g., node A), which can be in degrees Celsius (°C) or other temperature scales. The temperature is the temperature of the previous hop tag node, i.e., the local temperature reported by the immediate upstream node (e.g., node B) along the direction of acoustic disturbance propagation; v=f(T) is the function / table of the relationship between temperature T and sound velocity v, representing the possible changes in the sound velocity value in the optical cable sheath under different temperature environments. This function can be in interpolation form or discrete lookup table form. For example, when the sheath material, tension level, and other conditions are fixed, several temperature points and corresponding measured or calibrated sound velocity values ​​can be stored. This serves as the identifier for the previous hop label node, used for chain tracking of which upstream node A received the acoustic disturbance from during the current excitation cycle.

[0138] For example, when performing propagation difference analysis on adjacent nodes (A, B), if it is found that... Exceeding a preset threshold This indicates a significant temperature difference between points A and B, suggesting a possible substantial change in the velocity of sound within this section of the optical cable sheath. Therefore, the analysis and construction module 40 divides this section into a new "isothermal segment" or "thermal segment" and calls the aforementioned v=f(T) to adjust the time delay or attenuation calculation for this segment. A common mapping relationship can be written as: ;

[0139] The corrected speed of sound within the segment;

[0140] Reference temperature The calibrated value of the speed of sound below;

[0141] α is the temperature coefficient of sound speed, used to quantify the linear effect of temperature on sound speed;

[0142] ,in The average temperature of this segment (which can be obtained from...) and (Calculated by combining the average or multiple node temperatures).

[0143] It is understandable that those skilled in the art can also use other more complex functions or lookup table methods to reflect the effects of nonlinearity or multi-parameter coupling.

[0144] In some cases, if abnormal temperature data occurs at a certain node (such as large jumps or distortions), the system can mark it as a suspicious data point and trace back along the previous hop to determine whether there is a sensor malfunction or missing data reporting. Through such upstream and downstream cooperation and segmented compensation, the analysis and construction module 40 can generate an acoustic propagation path map that better fits the actual physical environment, providing the backend projection and recognition module 50 with a segmented temperature-compensated path result, further improving the accuracy and robustness of fault location.

[0145] In this way, it can not only cope with the cumulative errors that traditional optical cable ranging is prone to in temperature difference environments, but also be compatible with the complex laying conditions of multiple jumpers, messy branches, and huge differences in burial depth, providing repair personnel with a more reliable and intuitive basis for fault location.

[0146] As an optional implementation, generating and outputting multidimensional data characterizing the propagation characteristics of the acoustic disturbance signal includes:

[0147] Starting from the moment when the amplitude of the acoustic disturbance signal is detected to exceed the preset trigger threshold for the first time, the acoustic disturbance signal is converted from analog to digital according to the preset sampling rate to obtain a discrete acoustic sampling sequence;

[0148] Within the preset analysis window, the discrete acoustic sampling sequence is processed. When the amplitude of the discrete acoustic sampling sequence first exceeds the rising edge threshold, the moment is recorded as the arrival timestamp. The root mean square amplitude value is calculated as the arrival amplitude. The maximum energy frequency component is extracted as the arrival frequency through fast Fourier transform, and the instantaneous phase of the maximum energy frequency component is obtained simultaneously.

[0149] According to the preset data frame format, the tag node identifier, the previous hop tag node identifier, the local temperature data, and the arrival timestamp, arrival amplitude, arrival frequency and instantaneous phase are encapsulated into an uplink data frame;

[0150] The uplink data frame is sent to the aggregation module 30 using near-field energy coupling.

[0151] When a node detects an acoustic disturbance signal and determines that its amplitude exceeds a preset trigger threshold for the first time, the tag node switches from standby or shallow sleep mode to sampling mode. In this mode, the node continuously converts the analog sound waves transmitted from the fiber optic cable sheath into digital sampled data according to a pre-configured sampling rate (e.g., 10kHz or 20kHz).

[0152] After storing the sampled sequence, the node uses the basic computing power of its local microcontroller to perform preliminary data processing. The processing logic may include a time-domain scan of amplitude changes: when the amplitude of the sampled sequence first exceeds the rising edge threshold at a certain point, it is considered that a pulse has been detected, and the node records the timestamp corresponding to that sampling point as the arrival timestamp.

[0153] In this way, the system can compare the relative timing of pulse arrivals at different nodes to obtain the reference time required for subsequent propagation difference analysis.

[0154] To balance computational complexity and robustness to transient noise when calculating the arrival amplitude, a root mean square (RMS) amplitude calculation can be performed on the data within the analysis window. The node squares a small segment of the waveform after the rising edge of the sampled sequence, sums the squares, and then takes the square root to obtain an amplitude index representing the overall energy level. If this amplitude is significantly higher than the node's background noise, it indicates that the signal still retains a considerable amount of energy upon pulse arrival; if the amplitude is equal to the noise or does not exceed the threshold, it may indicate that the pulse has severely attenuated during propagation.

[0155] When frequency information needs to be extracted, nodes can also perform simplified frequency domain analysis on the sampled sequence, such as performing a Fast Fourier Transform (FFT). To avoid consuming too many computational resources in passive mode, some implementations only perform low-resolution or truncated FFTs, focusing on searching for the frequency component with the highest energy and recording the instantaneous phase of that component. This helps in subsequent determination of whether acoustic disturbances in the optical cable cause additional distortion due to sheath resonance or environmental noise, and also provides more basis for analyzing and constructing module 40 to identify abnormal segments.

[0156] After processing the discrete acoustic sampling sequence, the node packages key parameters such as arrival timestamp, amplitude value (or RMS amplitude), maximum energy frequency, and phase information, along with its own identifier, the identifier of the previous hop tag node, and local temperature data (if the node is equipped with a temperature sensor), according to a preset data frame format. To adapt to short-range communication or near-field coupling in passive mode, the passive wireless tag node 20 often adopts a highly compact frame structure, recording each field in a fixed order or as little-endian key-value pairs. After packaging, the node activates the near-field coupling transmission module and sends the uplink data frame to the aggregation module 30. If the node completes reporting or confirms packet reception within a limited number of times, it can quickly switch back to sleep mode, keeping overall power consumption at an extremely low level.

[0157] In this way, distributed, multi-dimensional detection data can be obtained, providing accurate and rich observational data for subsequent fault location and path recovery in the aggregation module 30 and the analysis and construction module 40. When used in conjunction with the aforementioned temperature and previous hop identifier, it can also significantly improve the accuracy of physical laying conditions in multi-branch, multi-temperature zone, or variable wave speed scenarios, ensuring high positioning accuracy and feasibility in large-scale optical cable networks.

[0158] As an optional implementation, please refer to Figure 3 The flowchart of a method for constructing an acoustic propagation path map provided in this application includes steps S301 to S304, wherein:

[0159] S301: Based on the arrival amplitude and the instantaneous phase, remove tag node data that are lower than the peak signal power to noise mean square error ratio below a preset signal-to-noise ratio threshold and / or have a phase difference greater than 180° between adjacent frames, to obtain an effective tag dataset;

[0160] S302: For each label node in the effective label dataset, select the other label node with the smallest propagation time difference and greater than zero as the adjacent node pair, and adaptively generate adjacent edges.

[0161] S303: Assign a composite weight to each of the adjacent edges, wherein the composite weight is obtained by linearly combining the propagation time difference and arrival amplitude attenuation value of the adjacent edge according to a preset weighting coefficient;

[0162] S304: Using the composite weights, construct an acoustic propagation path map based on a graph search algorithm, and update the acoustic propagation path map with incremental information as soon as a new valid label dataset is generated.

[0163] In practical implementation, relying solely on basic parameters such as arrival timestamps and amplitudes is insufficient to completely eliminate data anomalies caused by sporadic noise or multipath interference. Therefore, this application adds a secondary screening and weighted graph construction process for multidimensional data to the analysis and construction module 40. This process effectively removes node observations with excessive noise or conflicting phases during the construction of the acoustic propagation path map. Furthermore, it performs composite weight calculations on the propagation differences between nodes while retaining valid node data, resulting in more robust path map update results.

[0164] In practice, after the aggregation module 30 integrates the multidimensional data of the tag nodes within the current detection cycle (or the cumulative data from multiple detections), the analysis and construction module 40 first performs data cleaning based on arrival amplitude and instantaneous phase. To measure the signal-to-noise ratio, the system can calculate the ratio of peak signal power to noise mean square error. If this value is lower than a certain preset signal-to-noise ratio threshold, it indicates that the observation data of the tag node in the current cycle may be almost submerged in noise and is not suitable for subsequent path inference. At the same time, if the phase difference between adjacent frames (such as the phase difference between the main energy peaks of consecutive sampled frames of the same node within the same cycle) exceeds 180°, it can also be determined that there is a phase information jump or mismatch, and it is also marked as an anomaly, thereby removing the valid tag dataset.

[0165] After this initial screening, the analysis and construction module 40 selects another labeled node with the smallest and greater than zero propagation time difference from the remaining valid node data as an adjacent node pair for each labeled node. This adaptively generates connections between the upstream node and the current node, or between the current node and the downstream node. In some environments, if a node and multiple adjacent nodes satisfy the condition of similar propagation time differences, further criteria such as amplitude matching and phase similarity can be used to determine the final adjacent pair. The node pairs established in this way can be considered as directed edges in a graph structure, representing the direction of the most likely and shortest path of acoustic wave propagation.

[0166] To integrate latency and attenuation or other features, the system assigns a composite weight to each adjacent edge. A common approach is to linearly combine the propagation time difference and arrival amplitude attenuation value using a preset weighting coefficient. Alternatively, a nonlinear function mapping can be used to enhance sensitivity to a specific feature. For example, the composite weight could be calculated as: γ × (propagation time difference) + β × (amplitude attenuation), where γ and β are configurable parameters. Amplitude attenuation can be estimated based on the amplitude ratio between node A and node B or the RMS energy difference. Excessive amplitude attenuation may indicate poor path quality; excessive time difference suggests significant physical distance or varying medium conditions between the two nodes.

[0167] Once the analysis and construction module 40 has built such a weighted graph, common graph search algorithms (such as shortest path search or minimum spanning tree algorithm) can be called to generate the overall acoustic propagation path map and determine the positional relationship of each node in the chain propagation. For the already constructed map, incremental information can be immediately updated to the graph structure when a new detection generates a valid label dataset. In this way, the system can maintain the dynamism and real-time nature of the path map even when the optical cable environment continues to change or nodes are added or removed in batches, thus reflecting the most realistic laying status and propagation relationships at present.

[0168] During deployment, if the system detects frequent changes in the adjacency relationships of certain nodes or a significant and continuous increase in the composite weight of a certain edge during multiple consecutive updates, it may indicate that the optical cable segment has experienced dynamic degradation (e.g., stretching, sheath immersion in water, or loose joints). In this case, the analysis and construction module 40 can further output alarm information to the projection and recognition module 50, highlighting the segment in the GIS coordinate system to prompt maintenance personnel to conduct preliminary checks.

[0169] As an optional implementation, constructing an acoustic propagation path map using the composite weights and a graph search algorithm includes:

[0170] Each label node in the effective label dataset is abstracted as a directed graph vertex, the reference label node is used as the source vertex, and the adjacent node pairs determined by the composite weight are mapped as directed edges from the source vertex to the propagation direction.

[0171] During the graph search process, when the composite weight of a candidate directed edge exceeds a preset pruning threshold, the expansion of the corresponding edge is terminated.

[0172] Based on the pruned directed graph, the Dijkstra algorithm is used to find the shortest cumulative composite weight path from the source vertex to the remaining label nodes, and the acoustic propagation path result set is obtained.

[0173] For each path in the acoustic propagation path result set, output a path record consisting of a tag node sequence, cumulative propagation time difference, and cumulative arrival amplitude attenuation, and write the path record into the acoustic propagation path map.

[0174] Furthermore, in some complex optical fiber networks, relying solely on adjacency pairs and composite weighting is insufficient to obtain a complete and stable propagation path distribution across the entire network. Therefore, this application can further introduce a graph search algorithm in the analysis and construction module 40 to perform global-level search and pruning of these weighted edges.

[0175] Specifically, each label node in the valid label dataset is first abstracted as a directed graph vertex. If a reference label node is identified within the detection period, it can be considered as the source vertex of the entire graph, allowing a series of directed edges to be generated downstream along the propagation direction. After assigning composite weights to the nodes, the system terminates candidate edges with excessively large weights in advance during the graph search process based on a "preset pruning threshold," avoiding overexpansion or the generation of a large number of invalid branches.

[0176] When the composite weight of a candidate edge exceeds a preset pruning threshold, the analysis and construction module 40 determines that the node pair corresponding to that edge does not constitute a practically feasible propagation path in the current scenario, possibly due to excessive attenuation, excessive latency, or data anomalies. By terminating the expansion of such edges in a timely manner during the search, the system can quickly converge to paths with lower weights and stable transmission with less computation. The graph search algorithm itself can employ various methods, such as Dijkstra's algorithm, to find the shortest cumulative composite weight path from the source vertex to the remaining nodes on the pruned directed graph. Each shortest path actually corresponds to a feasible and relatively low-loss, low-latency optical fiber propagation chain, thereby helping the system reconstruct or iteratively update the acoustic propagation path map.

[0177] When filling the path graph with the result set obtained from the shortest path search, the analysis and construction module 40 records data such as the output node sequence, cumulative propagation time difference, and cumulative arrival amplitude attenuation for each path. These elements reflect how acoustic disturbances are attenuated or delayed along the link from the source to the downstream node. If some paths seem feasible but have abnormally large cumulative attenuation, the system can also prompt for manual verification in the next step. Through this graph search and pruning strategy, the analysis and construction module 40 can integrate a large number of scattered node observation results into an interpretable and visualized propagation link, and immediately perform incremental updates as soon as new valid label data appears in subsequent detections, ensuring that the graph's reflection of the optical cable laying status and fault risk remains real-time or near real-time.

[0178] For example, if the composite weight is a linear combination of delay difference and amplitude attenuation, directed edge weights exceeding a certain threshold (e.g., 10, or a specific upper limit after normalization) will be filtered out during the search. Then, Dijkstra's algorithm is used to find the minimum weight path for each other node starting from the reference label node. If a path with a relatively small cumulative delay difference and relatively small amplitude attenuation from the source to the target node is found, it indicates that this group of nodes is more likely to be physically adjacent or branched. After the search is complete, each shortest cumulative weight path can be recorded in the acoustic propagation path graph. The system saves the node order and accumulates the relative attenuation for visualization and subsequent fault identification in the projection and recognition module 50. If new detection period data subsequently causes some edge weights to be recalculated, the module can perform incremental corrections on existing paths in the graph without having to rebuild from scratch.

[0179] This allows the entire system to maintain good adaptive reconfiguration capabilities even in the absence of a clear topology and with multiple points of attenuation, and provides a high-quality path link foundation for subsequent localization analysis.

[0180] As an optional implementation, writing the path record into the acoustic propagation path map includes:

[0181] For each path in the acoustic propagation path result set, the path confidence is calculated based on the cumulative composite weight of the path and the number of tag nodes it contains;

[0182] The acoustic propagation path result set is sorted in descending order according to the path confidence, and the top N paths with the highest confidence are selected as the target path set, where N is the preset maximum number of output paths.

[0183] For each path in the target path set, the average temperature difference of the path is calculated based on the local temperature data of the tag nodes that constitute the path.

[0184] The target path set, along with the corresponding tag node sequence, cumulative propagation time difference, cumulative arrival amplitude attenuation, and the average temperature difference value, are written into the acoustic propagation path map.

[0185] In some fault detection tasks, a single Dijkstra algorithm run may generate a large number of shortest cumulative composite weighted paths. For example, in multi-branch structures or densely distributed optical fiber networks, there may be several feasible paths with similar weights between each terminal node and the source vertex. To facilitate subsequent interpretation and focus, this application introduces a path confidence calculation mechanism in the analysis and construction module 40. All paths are sorted in descending order, and then the most representative or highest-quality paths are selected according to the preset maximum number of output paths N. Additional information such as their average temperature difference is recorded and written into the final acoustic propagation path map.

[0186] After performing a Dijkstra search on the pruned directed graph, the system obtains one or more shortest cumulative composite weight paths from the source vertex (usually the reference label node or the location of the injection module 10) to the remaining label nodes. The analysis and construction module 40 organizes these paths into an acoustic propagation path result set, and calculates a path confidence for each path based on its cumulative composite weight and the number of label nodes it contains.

[0187] The path confidence score can be defined as a floating-point number between 0 and 1 or other fractional values, or it can be assigned a higher confidence score to paths with smaller weights and a more moderate number of nodes through linear / non-linear functions. For example, the confidence score can be inversely proportional to 1 / (cumulative weight × number of nodes), but other correction coefficients can also be added to this.

[0188] After calculating the confidence scores for all paths, the system sorts the entire path result set in descending order according to the confidence scores and generates a target path set from the top N paths (e.g., the highest-ranked paths). Here, N can be a constant pre-set in the system configuration or dynamically adjusted according to the actual scenario to balance the need to "retain multiple candidate paths to avoid missed detections" with the need to "avoid excessive results." A larger N value allows the system to retain a sufficient number of candidate paths even when multiple branches or paths with similar weights are present; a smaller N value is more conducive to focusing on high-confidence paths in the interface or downstream processing stages.

[0189] After selecting the target path set, the system can traverse all tag nodes within each path, extract the local temperature data reported by these nodes, and then calculate the average temperature difference value of the path. If the number of nodes in the path is 'a', the system can average the temperature difference between adjacent nodes, or take the arithmetic average of the temperatures of a single node and compare it with a reference temperature. It can also perform more refined segmented analysis to obtain a comprehensive value that represents the temperature gradient experienced by the path as needed. When the average temperature difference value of the same path is confirmed to be consistent or shows a regular change after multiple detection cycles, it indicates that the temperature conditions of the path in physical space are relatively stable. If the temperature difference fluctuates drastically in a short period of time, it may indicate a change in environmental conditions (such as pipeline damage, or a section of the road being changed from ground level to elevated).

[0190] The system will ultimately write the target path set, along with the corresponding tag node sequence, cumulative propagation time difference, cumulative arrival amplitude attenuation, and the aforementioned average temperature difference, into the acoustic propagation path map. This not only records the specific node sequence traversed by each high-confidence path but also provides additional temperature and attenuation data, facilitating precise location of potential fault points in the subsequent projection and recognition module 50. If maintenance personnel wish to view the full attenuation and temperature curves of a specific path from the source vertex to the end node, the system can also generate visual reports or trend charts based on these records, further improving the efficiency of fault diagnosis and maintenance decision-making.

[0191] In this way, for urban optical cable networks with varying burial depths, significant ground heat island effects, or frequent seasonal temperature fluctuations, the system's adaptability and insight into complex environments can be improved, reducing the subsequent processing burden caused by missed detections, false alarms, or redundant paths.

[0192] As an optional implementation, determining whether the propagation difference between the label nodes in the multidimensional data exceeds a preset propagation threshold, and if it does, identifying the label node as an abnormal propagation node; and determining the geographic information system coordinates of the abnormal propagation node as the fault spatial location includes:

[0193] The baseline mean μ and standard deviation σ are calculated for the set of propagation time differences between adjacent tag nodes. When the absolute value of the difference between a certain propagation time difference and the baseline mean is greater than k·σ, the corresponding downstream tag node is marked as a candidate abnormal node; where k is a preset abnormality coefficient.

[0194] According to the direction of propagation, consecutively appearing candidate abnormal nodes are grouped. When the number of nodes in the same group is not less than the preset consecutive abnormal length, the group is identified as an abnormal propagation node cluster.

[0195] For each cluster of anomalous propagation nodes, the center coordinates of the cluster are calculated as the spatial location of the fault based on the geographic information system coordinates of the first and last tag nodes constituting the cluster, through linear interpolation along the optical cable path.

[0196] The spatial location of the fault is written into the acoustic propagation path map, and the corresponding geographic information system coordinates are output to the projection and recognition module 50.

[0197] In some application scenarios with extremely complex pipeline networks or geographical environments, simply applying threshold judgments to a shortest path or adjacent node pairs may not be able to identify some hidden fault areas that are scattered along the route in a timely manner. To address this, this application adds a baseline mean-standard deviation statistical method to the projection and recognition module 50 or the analysis and construction module 40 to comprehensively evaluate the propagation time difference between adjacent labeled nodes. This method is supplemented by node clustering and coordinate interpolation to determine the specific fault center location on the GIS coordinate system.

[0198] Once the system obtains the set of propagation time differences between tag node pairs within a certain detection period, it connects the adjacent relationships of nodes according to the previously determined acoustic propagation path map, extracts a list of propagation time differences between node pairs, and calculates the baseline mean μ and standard deviation σ of this list. The baseline mean μ can be understood as the overall central trend of the delay of the entire network under normal transmission conditions, while the standard deviation σ reflects the random fluctuation or dispersion of the delay in different segments. If the absolute value of the deviation between the propagation time difference of a pair of nodes and μ exceeds k·σ, where k is a pre-configured anomaly coefficient (usually between 2 and 3), it can be determined that this adjacent relationship has extremely abnormal delay signs in the current period. The system will mark the corresponding downstream tag node as a candidate abnormal node to indicate that there may be local sheath damage, loose connectors, or other physical damage.

[0199] By employing this statistical distribution-based anomaly identification method, the system can not only capture outliers that are relatively large compared to the average latency, but also take into account the overall network dispersion. If certain latency values ​​themselves have relatively high fluctuations within the pipeline as a whole, the range of the mean plus a multiple of σ will be more lenient, avoiding false alarms for nodes that are already prone to bends or severe scattering. Conversely, for pipelines with relatively small overall fluctuations and relatively stable distributions, even a large deviation can be quickly detected by the k·σ threshold.

[0200] After identifying candidate abnormal nodes, the system further performs continuous detection on these nodes according to the direction of acoustic disturbance propagation. If several nodes are found to have consecutive abnormal markers in adjacent positions (meaning they are adjacent in physical propagation order, not adjacent in ID), and maintain this abnormal state in the same or multiple detection cycles, then these consecutively appearing candidate abnormal nodes can be grouped. When the number of nodes in the same group is not less than the "preset consecutive abnormal length" (e.g., 2 to 3 or more), it indicates that the optical cable segment is highly likely to have a concentrated fault risk at the branch or section level, and the system can further define this group of nodes as an "abnormal propagation node cluster".

[0201] Each cluster of anomalous propagation nodes contains a head node and a tail node. Based on the positions of these two extreme nodes in the GIS coordinate system, the system can calculate a central coordinate as a representative value of the fault's spatial location through linear interpolation of the fiber optic cable path or more refined curve interpolation. If it is known that the nodes in the cluster are not strictly linear in physical space, curve approximation or subdivided segmental interpolation can be applied to fit the arrangement of multiple nodes in the GIS coordinate system, and the curve center or centroid can be selected as the final interpolation result. Since the fault is usually patchy or covers a large area, this aggregation positioning can help maintenance personnel quickly locate a suspicious section, rather than just a single point.

[0202] After locating the cluster of abnormal propagation nodes, the system writes the interpolated fault spatial location information into the acoustic propagation path map and simultaneously outputs the corresponding GIS coordinates to the projection and recognition module 50. The projection and recognition module 50 can highlight the segment or the fault center location on the geographic information system with a highlight, warning icon, or different color to alert on-site maintenance personnel that a clustered or multi-point fault may occur there. After the fault is investigated or repaired, the system can also perform verification in the next detection cycle. Once the propagation delay of the segment returns to normal, the system automatically removes the area from the abnormal cluster or updates its status to maintain the real-time accuracy of the network-wide path and fault distribution information.

[0203] In this way, it can filter out scattered and isolated jump point interference, and make rapid judgments and fixed-point interpolation for potential large-scale or continuous faults in the network, thereby significantly shortening the maintenance decision and troubleshooting time, and providing more adaptable intelligent detection capabilities for complex laying environments of large-scale, multi-branch pipelines.

[0204] Based on the same inventive concept, this application also provides an analysis and location method corresponding to the fault signal detection system. Since the principle of the method in this application is similar to that of the fault signal detection system described above in this application, the implementation of the method can refer to the implementation of the system, and the repeated parts will not be described again.

[0205] Reference Figure 1 The diagram shown is a flowchart of an analysis and localization method provided in an embodiment of this application, including steps S101 to S105, wherein:

[0206] S101: Inject a preset narrowband acoustic disturbance signal into the target optical cable;

[0207] S102: Passive wireless tag nodes are set along the target optical cable intervals. Each tag node is configured to generate and output multi-dimensional data characterizing the propagation characteristics of the acoustic disturbance signal when it receives the acoustic disturbance signal.

[0208] S103: Collect the multidimensional data and associate it with the corresponding tag node identifier;

[0209] S104: Based on the multidimensional data, calculate the propagation differences between each of the label nodes and construct an acoustic propagation path map;

[0210] S105: Map the acoustic propagation path map to the geographic information system coordinate system corresponding to the geographic coordinates of the tag node; determine whether the propagation difference between each tag node in the multidimensional data exceeds a preset propagation threshold; if it does, identify the tag node as an abnormal propagation node; and determine the geographic information system coordinates of the abnormal propagation node as the fault spatial location.

[0211] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this application can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

Claims

1. A fault signal detection system, characterized in that, include: The injection module is configured to inject a preset narrowband acoustic disturbance signal into the target optical cable; Passive wireless tag nodes are spaced along the target optical cable. Each tag node is configured to generate and output multidimensional data characterizing the propagation characteristics of the acoustic disturbance signal when it receives the acoustic disturbance signal. The aggregation module is configured to collect the multidimensional data and associate it with the corresponding tag node identifiers; The analysis and construction module is configured to calculate the propagation differences between each of the label nodes based on the multidimensional data, and construct an acoustic propagation path map. The projection and recognition module is configured to: map the acoustic propagation path map to the geographic information system coordinate system corresponding to the geographic coordinates of the tag nodes; determine whether the propagation difference between each tag node in the multidimensional data exceeds a preset propagation threshold; if it exceeds the threshold, identify the tag node as an abnormal propagation node; and determine the geographic information system coordinates of the abnormal propagation node as the fault spatial location.

2. The fault signal detection system according to claim 1, characterized in that, The injection of a preset narrowband acoustic disturbance signal into the target optical cable includes: The injection module is driven to inject at least two sets of test acoustic pulses in ascending frequency order within a preset frequency sweep range, and the arrival amplitude of each set of test acoustic pulses is recorded by a reference tag node set at the deployment point where the injection module is located. The target frequency is selected based on the arrival amplitude of each group of test acoustic pulses. The injection frequency corresponding to the test acoustic pulse with the largest amplitude and attenuation rate lower than the preset threshold is determined as the target narrowband injection frequency. Based on the arrival amplitude measured in real time by the reference tag node, the injection power is adjusted so that the amplitude of the acoustic disturbance signal of the target narrowband injection frequency at the reference tag node is kept within the preset target amplitude range. After the injection power adjustment is completed, the acoustic disturbance signal of the target narrowband injection frequency is cyclically injected into the target optical cable according to the preset injection cycle.

3. The fault signal detection system according to claim 2, characterized in that, Also includes: Each of the passive wireless tag nodes is further configured to report the local temperature data and the identifier of its previous tag node together when outputting the multidimensional data; wherein, the previous tag node refers to the tag node that is adjacent to the current tag node and located upstream of the current tag node in the acoustic propagation path map along the direction of acoustic disturbance signal propagation. In response to the absence of other tag nodes between the current tag node and the injection module, the reference tag node is regarded as the previous tag node of the current tag node; If the current label node is the farthest label node in the acoustic propagation path, then its previous label node is the nearest upstream label node adjacent to it. The analysis and construction module is configured to generate a temperature difference vector between tag pairs based on the local temperature data of adjacent tag nodes; Based on the temperature difference vector, the acoustic propagation path map is divided into several isothermal segments, and the temperature difference of each isothermal segment does not exceed a preset threshold. For each isothermal segment, a pre-stored temperature-sound velocity mapping data table corresponding to the optical cable sheath material and tension level is called, and the mapped sound velocity parameters are used to compensate for the propagation difference of the segment.

4. The fault signal detection system according to claim 3, characterized in that, The generated and outputted multidimensional data characterizing the propagation properties of the acoustic disturbance signal includes: Starting from the moment when the amplitude of the acoustic disturbance signal is detected to exceed the preset trigger threshold for the first time, the acoustic disturbance signal is converted from analog to digital according to the preset sampling rate to obtain a discrete acoustic sampling sequence; Within the preset analysis window, the discrete acoustic sampling sequence is processed. When the amplitude of the discrete acoustic sampling sequence first exceeds the rising edge threshold, the moment is recorded as the arrival timestamp. The root mean square amplitude value is calculated as the arrival amplitude. The maximum energy frequency component is extracted as the arrival frequency through fast Fourier transform, and the instantaneous phase of the maximum energy frequency component is obtained simultaneously. According to the preset data frame format, the tag node identifier, the previous hop tag node identifier, the local temperature data, and the arrival timestamp, arrival amplitude, arrival frequency and instantaneous phase are encapsulated into an uplink data frame; The uplink data frame is sent to the aggregation module using near-field energy coupling.

5. The fault signal detection system according to claim 4, characterized in that, The step of calculating the propagation differences between each of the labeled nodes and constructing an acoustic propagation path map based on the multidimensional data includes: Based on the arrival amplitude and the instantaneous phase, label node data with a peak signal power to noise mean square error ratio lower than a preset signal-to-noise ratio threshold and / or a phase difference between adjacent frames greater than 180° are removed to obtain an effective label dataset. For each label node in the effective label dataset, select the other label node with the smallest propagation time difference that is greater than zero as the adjacent node pair, and adaptively generate adjacent edges. A composite weight is assigned to each adjacent edge, and the composite weight is obtained by linearly combining the propagation time difference and arrival amplitude attenuation value of the adjacent edge according to a preset weighting coefficient; Using the composite weights, an acoustic propagation path map is constructed based on a graph search algorithm, and incremental information is updated to the acoustic propagation path map in real time when a new effective label dataset is generated.

6. The fault signal detection system according to claim 5, characterized in that, Using the aforementioned composite weights, and based on a graph search algorithm, the acoustic propagation path map is constructed as follows: Each label node in the effective label dataset is abstracted as a directed graph vertex, the reference label node is used as the source vertex, and the adjacent node pairs determined by the composite weight are mapped as directed edges from the source vertex to the propagation direction. During the graph search process, when the composite weight of a candidate directed edge exceeds a preset pruning threshold, the expansion of the corresponding edge is terminated. Based on the pruned directed graph, the Dijkstra algorithm is used to find the shortest cumulative composite weight path from the source vertex to the remaining label nodes, and the acoustic propagation path result set is obtained. For each path in the acoustic propagation path result set, output a path record consisting of a tag node sequence, cumulative propagation time difference, and cumulative arrival amplitude attenuation, and write the path record into the acoustic propagation path map.

7. The fault signal detection system according to claim 6, characterized in that, Writing the path record into the acoustic propagation path map includes: For each path in the acoustic propagation path result set, the path confidence is calculated based on the cumulative composite weight of the path and the number of tag nodes it contains; The acoustic propagation path result set is sorted in descending order according to the path confidence, and the top N paths with the highest confidence are selected as the target path set, where N is the preset maximum number of output paths. For each path in the target path set, the average temperature difference of the path is calculated based on the local temperature data of the tag nodes that constitute the path. The target path set, along with the corresponding tag node sequence, cumulative propagation time difference, cumulative arrival amplitude attenuation, and the average temperature difference value, are written into the acoustic propagation path map.

8. The fault signal detection system according to claim 7, characterized in that, Determine whether the propagation difference between each label node in the multidimensional data exceeds a preset propagation threshold; if it does, the label node is identified as an abnormal propagation node. Determining the geographic information system coordinates of the abnormal propagation node as the spatial location of the fault includes: The baseline mean μ and standard deviation σ are calculated for the set of propagation time differences between adjacent tag nodes. When the absolute value of the difference between a certain propagation time difference and the baseline mean is greater than k·σ, the corresponding downstream tag node is marked as a candidate abnormal node; where k is a preset abnormality coefficient. According to the direction of propagation, consecutively appearing candidate abnormal nodes are grouped. When the number of nodes in the same group is not less than the preset consecutive abnormal length, the group is identified as an abnormal propagation node cluster. For each cluster of anomalous propagation nodes, the center coordinates of the cluster are calculated as the spatial location of the fault based on the geographic information system coordinates of the first and last tag nodes constituting the cluster, through linear interpolation along the optical cable path. The spatial location of the fault is written into the acoustic propagation path map, and the corresponding geographic information system coordinates are output to the projection and recognition module.

9. An analytical localization method, characterized in that, include: Inject a preset narrowband acoustic disturbance signal into the target optical cable; Passive wireless tag nodes are set along the target optical cable at intervals. Each tag node is configured to generate and output multidimensional data characterizing the propagation characteristics of the acoustic disturbance signal when it receives the acoustic disturbance signal. Collect the multidimensional data and associate it with the corresponding tag node identifiers; Based on the multidimensional data, the propagation differences between each of the label nodes are calculated, and an acoustic propagation path map is constructed. The acoustic propagation path map is mapped to the geographic information system coordinate system corresponding to the geographic coordinates of the tag nodes; it is determined whether the propagation difference between each tag node in the multidimensional data exceeds a preset propagation threshold. If it does, the tag node is identified as an abnormal propagation node; and the geographic information system coordinates of the abnormal propagation node are determined as the fault spatial location.

Images