A highway DAS catastrophe evolution monitoring and grading early warning method and system

Abstract

The application provides a highway DAS disaster evolution monitoring and grading early warning method and system, and belongs to the technical field of highway infrastructure safety monitoring. In view of the problems of the existing DAS monitoring, such as the fragmentation of single-window detection results, the misguiding of abnormal propagation determination by fiber coordinate proximity, and frequent jumping of the early warning level, the application organizes the event regions belonging to the same abnormal source into an event chain through cross-window association, establishes a structure section graph with highway structure sections as nodes and engineering connection relationships as edges for propagation consistency verification, and adopts a warning state machine with differentiated setting of upgrade conditions and rollback conditions to control the early warning level. The application can track the complete evolution process from the inception to the subsidence of the disaster, effectively suppress false positives caused by non-disaster interference such as vehicle traffic and construction vibration, and is suitable for continuous monitoring and grading early warning of disasters with time evolution characteristics, such as highway slope slip, roadbed settlement, abnormality in bridge abutment connection area and the like.

Description

Technical Field

[0001] This invention belongs to the field of highway infrastructure safety monitoring technology, specifically relating to a method and system for monitoring and classifying early warning of highway DAS disaster evolution. Background Technology

[0002] Distributed Acoustic Sensing (DAS) is a novel distributed sensing technology that utilizes sensing optical fibers laid along highways or existing communication optical cables as continuous sensing media. By demodulating the phase, intensity, or frequency changes of backscattered light in the fiber, it obtains the vibration response or strain rate response that varies with time at different spatial locations along the fiber. Compared with traditional discrete accelerometers, vibration pickups, or strain gauges, DAS technology has significant advantages such as long monitoring distance (up to tens of kilometers), continuous spatial sampling (sampling intervals down to the meter level), resistance to electromagnetic interference, and suitability for utilizing existing communication optical cables. Therefore, it is particularly suitable for continuous safety monitoring of long-distance highway infrastructure.

[0003] Existing methods for anomaly monitoring using DAS typically divide continuously acquired DAS data into fixed-length time windows, calculating energy, spectrum, time-frequency maps, statistical features, or deep learning features within each window, and then using these features to determine whether an anomaly has occurred at a specific spatial location. These methods can detect vibration changes with high amplitude or prominent features within a single time window, but the processing results between different time windows are usually independent, lacking cross-window information correlation and comprehensive analysis capabilities. For catastrophic processes such as slope slippage, progressive roadbed deformation, and instability in bridge abutment connection zones, the early response may only manifest as a slight increase in localized low-frequency vibrations, which gradually maintain, intensify, expand, or migrate across multiple consecutive time windows, exhibiting rapid changes in vibration intensity and impact range as instability approaches. If the monitoring system only issues alarms based on the classification results of a single time window, it is difficult to confirm whether anomalies in multiple consecutive windows belong to the same catastrophic event, and it is also difficult to determine whether the event is currently in its nascent, ongoing, developing, accelerating, stable, or receding stage.

[0004] Meanwhile, the sources of vibration interference along highways are numerous, vary greatly in intensity, and have complex propagation patterns. Vehicles traveling along the road can create continuously moving, strong vibration trajectories; construction machinery can generate long-term periodic vibrations at fixed locations; heavy rainfall can cause a large-scale increase in background energy; and falling rocks or accidental collisions can produce impacts with high amplitude but extremely short duration. If the system triggers an alarm based solely on abnormal amplitudes at a single moment or the classification results of a single window, it is easy to misreport short-term strong interference as a high-level disaster warning. Conversely, some slowly developing slope slippage or roadbed settlement may have low early vibration amplitudes, but the duration and affected area gradually increase; using a single threshold judgment can easily lead to missed detections.

[0005] Furthermore, the spatial coordinates in DAS data are essentially distance coordinates along the length of the optical fiber, and are not necessarily equivalent to the engineering connections on the road structure. Adjacent optical fiber sampling points may be located in locations that are not connected in the engineering structure, such as: optical cables turning back near bridge abutments, having excess cable coils, entering trenches from slope sections, crossing different bridge spans, or passing through different structural types. If the spatial spread of anomalies is judged solely based on the proximity of optical fiber coordinates, sections that are not connected in the engineering structure may be mistakenly considered to be propagating continuously, resulting in false associations and false alarms. Although vehicle vibrations can form a strong temporal continuity in the road direction, their center migration velocity, frequency band composition, and regional morphology are usually different from catastrophic processes such as slope slippage, structural cracking, or roadbed deformation. Existing methods for event tracking based on the spatial continuity of optical fibers cannot fully utilize these differences in engineering structures to distinguish between catastrophic and non-catastrophic events.

[0006] Regarding early warning output, existing common methods typically set multiple anomaly thresholds, outputting different levels of alarms when the anomaly score exceeds different thresholds. However, because actual DAS signals are affected by various factors such as traffic flow, weather conditions, fiber optic coupling status, equipment gain, and on-site noise, the anomaly score for the same disaster event may fluctuate around the threshold, leading to frequent increases and decreases in warning levels. Some systems, to avoid missed detections, only allow one-way escalation of warning levels and do not allow automatic rollback. As a result, the warning level remains unchanged for a long time after a short-term interference event ends, causing false alarms. Other systems directly update the warning level in real time based on the detection results of the current window, which results in highly unstable warning states, making it difficult for engineers to accurately determine whether a disaster is continuing to develop.

[0007] In summary, the existing DAS highway monitoring technology has at least the following technical defects: (1) Single-window independent detection is difficult to reflect the time evolution process of disasters; (2) Judging the propagation of anomalies based solely on the proximity of fiber optic coordinates is unreliable and ignores the structural relationship of highway engineering; (3) The warning level is prone to frequent jumps or long-term false alarms due to signal fluctuations.

[0008] Therefore, this invention proposes a disaster evolution analysis method and system for continuous highway DAS monitoring. Summary of the Invention

[0009] The present invention aims to overcome the above-mentioned defects in existing DAS highway monitoring technology and provide a method and system for highway DAS disaster evolution monitoring and graded early warning.

[0010] Specifically, the present invention aims to solve the following technical problems: First, it addresses the problem that single-window anomaly detection results in continuous highway DAS monitoring are insufficient to reflect the development process of disasters. The system needs to determine whether different anomaly areas belong to the same anomaly source or the same disaster process across multiple consecutive time windows, and continuously update the event's time range, spatial range, vibration intensity, main frequency band, and state changes, thereby forming an event chain that can be continuously updated as new data arrives.

[0011] Second, the system addresses the unreliability of judging anomaly propagation solely based on optical fiber spatial proximity. The system needs to map optical fiber sampling locations to specific highway structural sections and utilize factors such as whether structural sections are connected, whether influence propagation is permissible, whether the response sequence is reasonable, whether the apparent propagation speed is within a reasonable range, and whether it crosses blocking boundaries to verify the engineering rationality of the spatial propagation and migration process of the event chain.

[0012] Third, the system addresses the issues of short-term, strong interference easily triggering high-level warnings, slow-moving disasters being difficult to escalate in a timely manner, and warning levels repeatedly fluctuating around thresholds. The system needs to determine the evolutionary stage of an event based on its persistence, intensity changes, spatial expansion, apparent migration, frequency band changes, propagation consistency, and decay trend. It should also employ a warning state machine to set different escalation, maintenance, and rollback conditions, ensuring that the warning level continuously, stably, and interpretably reflects the development of the disaster.

[0013] In a first aspect, this invention proposes a method for monitoring and hierarchical early warning of highway DAS disaster evolution, which includes the following steps: Continuously collect DAS vibration data along the highway and establish a mapping relationship between the optical fiber spatial sampling location and the highway structural section; The DAS vibration data is processed using a sliding time window, and one or more event regions are generated within each window based on the abnormal response map. Cross-window association between the event region of the current window and the historical event chain: the event region is converted to structural coordinates according to the mapping relationship, the reachable path is determined based on the structural segment diagram, the association cost is calculated for the combination with reachable path, and the current event region is added to the corresponding event chain when the association cost meets the preset association conditions, and a new candidate event chain is established when the association conditions are not met. Calculate the evolutionary characteristics of the established event chain; The propagation consistency of the event chain is verified using the aforementioned structural segment diagram, and a propagation consistency score is calculated. The evolutionary stages of the event chain are determined based on the aforementioned evolutionary characteristics and propagation consistency scores; Based on the evolution stage and its duration, the early warning level is controlled by an early warning state machine. The upgrade and rollback conditions in the early warning state machine adopt different duration window number thresholds, and the rollback confirmation window number is not less than the upgrade confirmation window number. Output early warning results and save the evolution process record of the event chain.

[0014] Preferably, the event region includes at least a window number, the start and end points of time within the window, the start and end points of structural coordinates, the corresponding set of structural segments, peak intensity, average intensity, main frequency bands, frequency band energy ratio, regional credibility, and data quality markers; the event chain at least stores the creation time, the most recent update time, the cumulative number of missing windows, the current structural segment range, evolution characteristics, propagation consistency score, the current evolution stage, and the current warning level.

[0015] Preferably, in the cross-window association, the current event region Chain of historical events Related costs Represented as: ,in, Due to differences in spatial location, Due to regional boundary differences, Due to frequency band differences, Differences in intensity and trend, Differences in the structural segment relationships; , , , , They are respectively , , , , The corresponding weight coefficients, and satisfying All weight coefficients are greater than or equal to 0 and less than or equal to 1; when If the current event region is less than or equal to the preset association threshold and there is no better matching conflict, the current event region will be added to the corresponding event chain.

[0016] More preferably, the spatial location difference The distance between the current region center coordinates and the event chain prediction center coordinates is obtained by normalization; the region boundary difference is... The frequency band difference is obtained by the ratio of the length of the overlapping interval between the current region and the end region of the event chain; The difference in intensity and trend is obtained by calculating the Euclidean distance or cosine similarity of the energy distribution vectors of the two main frequency bands. The difference is obtained from the relative rate of change of the average or peak intensity of the two; the structural segment relationship is different. The weight is obtained based on the structural distance penalty value between corresponding structural segments of the two regions; , , , , The values ​​are 0.25, 0.20, 0.15, 0.20, and 0.20 respectively, and the preset association threshold is 0.6.

[0017] Preferably, the propagation consistency verification is based on at least several of the following: structural reachability consistency, response order consistency, apparent expansion speed consistency, structural boundary consistency, and frequency band variation consistency. Represented as: ,in, To ensure structural reachability consistency, a score is awarded based on whether a reachable path exists in the structural segment graph. To ensure consistency in response order, a score is awarded based on whether the spatial order in which event regions appear aligns with the allowed expansion direction of the structure diagram. To ensure consistent apparent expansion speed, a score is awarded based on whether the migration speed of the event chain falls within a preset speed range. To ensure structural boundary consistency, a score is awarded based on whether preset transition conditions are met when an event crosses the structural boundary. To ensure consistency in frequency band variations, a score is awarded based on whether the frequency band energy variation patterns before and after an abnormal event crosses different structural media conform to engineering experience; weighting. , , , , The values ​​are 0.30, 0.20, 0.20, 0.20, and 0.10, respectively.

[0018] Further preferred, when When it is determined to be a propagation conflict, When it is determined that the credibility of the dissemination is low, When it is determined to be effective transmission, If the propagation is deemed to have high consistency, the system will not advance to the disaster phase and will output the structural segment where the conflict occurred and the cause of the conflict.

[0019] Further preferably, the evolutionary characteristics include at least one of event persistence, vibration intensity growth rate, spatial expansion rate, center migration speed, frequency band change, and event chain stability; the evolutionary stages include pending confirmation, transient anomaly, continuous monitoring, development, acceleration, stabilization, decay, and propagation conflict; the early warning state machine controls the generation, escalation, maintenance, rollback, or cancellation of the early warning level.

[0020] Preferably, the construction of the structural segment map includes the following steps: based on highway design data and geographic information system, the highway is divided into independent physical monitoring units as nodes; based on field surveys and geographic information topology, engineering connection relationships between nodes are established as edges, and edges are marked with blocking attributes at locations crossing empty areas or isolation zones; combined with optical cable laying records, a mapping table between optical fiber coordinates and structural nodes is established using manual tapping calibration, known equipment excitation, or vehicle positioning calibration; based on the cross-segment attenuation characteristics of vibration signals obtained from field tests, the edge attributes of the structural segment map are verified and corrected.

[0021] Further preferably, the edges in the structural segment diagram store connection type, spatial distance, allowed expansion direction, apparent expansion speed range, structural transformation attribute, fiber coupling reliability, and blocking attribute; the apparent expansion speed range is determined comprehensively based on historical disaster database, highway design speed, and physical motion laws, including: 0.01 m / s to 0.5 m / s for gradual slope slippage or roadbed settlement process, vehicle traffic disturbance determined according to highway design speed, and rockfall impact or local collapse event greater than 50 m / s.

[0022] Preferably, the warning levels are four levels: blue, yellow, orange, and red, with upgrades and rollbacks performed step-by-step. Upgrading from blue to yellow requires maintaining a development phase for at least two windows and a propagation consistency of at least 0.6; upgrading from yellow to orange requires entering an acceleration phase or continuing to develop within a high-risk area with a propagation consistency of at least 0.75; upgrading from orange to red requires rapid expansion of the high-risk area or simultaneous rapid growth in intensity and spatial range with a propagation consistency of at least 0.75. The number of confirmation windows for yellow rollback is three, for orange it is four, and for red it is five.

[0023] Preferably, the event chain allows for a maximum of 2 consecutive missing windows. The event chain remains active while the number of consecutive missing windows does not exceed 2. Once the number exceeds 2, the event chain enters a decay or termination judgment. The minimum duration of the event region is 0.5s, and the minimum spatial length is 3m.

[0024] Secondly, embodiments of the present invention provide a highway DAS disaster evolution monitoring and graded early warning system, comprising: The DAS data acquisition and spatial calibration unit is used to continuously acquire DAS vibration data along the highway and establish a mapping relationship between the optical fiber spatial sampling position and the highway structural section. The window event region generation unit is used to process the DAS vibration data using a sliding time window, and generate one or more event regions in each window based on the abnormal response map. The event chain construction and management unit is used to associate the event area of ​​the current window with the historical event chain across windows, convert the event area to structural coordinates according to the mapping relationship, determine the reachable path and calculate the association cost based on the structural segment diagram, and add the event area to the existing event chain, create a new event chain, or keep it missing for a short time according to the association result. The event chain evolution feature calculation unit is used to calculate the evolution features of the event chain; The structural propagation relationship and consistency verification unit is used to establish a structural segment diagram with highway structural segments as nodes and engineering connections or permissible influence relationships between segments as edges, and to score the propagation consistency of the event chain. The disaster evolution stage determination unit is used to determine the evolution stage of the event chain based on the evolution characteristics and propagation consistency score. The early warning state control unit is used to control the early warning level through an early warning state machine according to the evolution stage and its duration. The upgrade condition and rollback condition in the early warning state machine adopt different duration window number thresholds, and the rollback confirmation window number is not less than the upgrade confirmation window number. The warning results and historical record unit is used to output warning results and save the evolution process record of the event chain.

[0025] Thirdly, embodiments of the present invention provide an electronic device, including: one or more processors; and a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any implementation of the first aspect.

[0026] Fourthly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method as described in any of the implementations of the first aspect.

[0027] Fifthly, embodiments of this application provide a computer program product, wherein the computer program product includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps described in the first aspect of embodiments of this application. The computer program product may be a software installation package.

[0028] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) This invention organizes abnormal regions in multiple independent time windows into a continuous development process of the same event through event chains, enabling the system to observe the persistence, enhancement, expansion, migration, stabilization, and decline of anomalies, rather than processing the detection results of each window in isolation. Therefore, for early-stage disasters with low amplitude but continuous development and gradual expansion, the system can gradually increase its level of attention based on the evolution trend; for a single high-amplitude impact, the system will not directly raise the alert level to a high level based solely on the results of a single window.

[0029] (2) This invention utilizes the reachability relationships, response sequence, apparent expansion rate, and structural boundaries between highway structural sections to verify the spatial changes of events, avoiding the misinterpretation of simple proximity on fiber optic coordinates as continuous propagation on the engineering structure. For rapid vehicle movement, simultaneous anomalies in multiple unconnected sections, fiber optic return position response jumps, and complex responses near structural transition boundaries, the system can accurately identify the differences between these responses and the disaster propagation process.

[0030] (3) This invention controls the escalation, maintenance, and rollback of the warning level according to the continuous evolution stage, and sets different conditions and different duration window requirements for escalation and rollback. This can effectively reduce false escalation caused by single high-amplitude interference, and also avoid frequent changes in the warning level caused by abnormal scores fluctuating near the threshold. After the event subsides, the system can gradually remove the warning when the continuous decline or contraction conditions are met, so that the warning output is consistent with the actual evolution process.

[0031] (4) This invention outputs the event chain, affected section, evolution stage, propagation consistency, warning level, and reasons for level changes, enabling monitoring personnel to understand why the system upgrades, maintains, rolls back, or suspends automatic upgrades, and to determine key review locations accordingly. Compared to simply providing anomaly scores or categories, this output method is more suitable for continuous safety monitoring and on-site handling of highway infrastructure.

[0032] (5) This invention can also identify inconsistencies between the fiber coordinates and the engineering structure in scenarios such as fiber optic return, cable coiling, cross-structure laying, or changes in fiber optic cable coupling status, based on the structural segment diagram, thereby reducing the risk of false association and false escalation caused by the fiber optic cable layout. At the same time, since the early warning results are output to the highway structural segments, monitoring personnel can directly locate the roadbed, slope, abutment, bridge span, or tunnel segments that need to be verified, significantly improving on-site handling efficiency. Attached Figure Description

[0033] The accompanying drawings are included to provide a further understanding of the embodiments and are incorporated in and constitute a part of this specification. The drawings illustrate embodiments and, together with the description, serve to explain the principles of the invention. Other embodiments and many anticipated advantages of the embodiments will be readily recognized as they become better understood through reference to the following detailed description. Elements in the drawings are not necessarily to scale. The same reference numerals refer to corresponding similar parts.

[0034] Figure 1 This is a flowchart illustrating the highway DAS disaster evolution monitoring and graded early warning method according to an embodiment of the present invention. Figure 2 This is the overall flowchart of the highway DAS disaster evolution monitoring and graded early warning method of the present invention; Figure 3 This is a schematic diagram illustrating the mapping relationship between the spatial location of the optical fiber and the road structure section in this invention; Figure 4 This is a schematic diagram illustrating the cross-window event region association and event chain formation process of the present invention; Figure 5 This is a schematic diagram of the structural segment diagram and the relationship between the propagation consistency verification of the present invention; Figure 6 This is a flowchart of the disaster evolution determination and early warning status control of the present invention; Figure 7 This is a schematic diagram of the overall structure of the highway DAS disaster evolution monitoring and hierarchical early warning system, which is an embodiment of the present invention. Detailed Implementation

[0035] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0036] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0037] To overcome the shortcomings of existing DAS highway monitoring technologies, this invention discloses a disaster evolution analysis method for continuous highway DAS monitoring. This method should not directly equate the detection results of a single window with disaster conclusions. Instead, it should organize responses from different windows belonging to the same anomaly source or the same disaster process into an event chain, reflecting the complete development process of the abnormal event. Simultaneously, it requires mapping the spatial location of optical fibers to highway structural sections and using engineering connections between structural sections, permissible expansion directions, apparent expansion speeds, and blocking boundaries to determine whether the event expansion has engineering rationality. Finally, it requires controlling the warning level through a warning state machine with different escalation and rollback conditions, so that the warning output can stably reflect the continuous evolution state of the event.

[0038] Furthermore, this method needs to avoid being misinterpreted as a simple software classification rule. Its processing object is continuous vibration or strain rate data along the highway collected by DAS demodulation equipment, and its output objects are the event chain, propagation consistency, evolution stage, and warning level corresponding to the highway structural section. Therefore, this invention solves the engineering and technical problems of inconsistency between DAS spatial coordinates and highway engineering structures, the difficulty of single-window anomaly results reflecting disaster evolution, and the tendency for warning levels to be erroneously escalated or frequently change.

[0039] To facilitate understanding of the technical solution of this invention, several key terms used in this invention will first be defined and explained.

[0040] Event region: refers to the abnormal vibration region with continuous time and spatial range, defined by the abnormal response map within a sliding time window. The event region includes at least the window number, the start and end points of the time within the window, the start and end points of the structural coordinates, the set of corresponding structural segments, peak intensity, average intensity, main frequency band, frequency band energy ratio, regional confidence level, and data quality markers.

[0041] Event chain: A sequence object formed by associating multiple event regions belonging to the same anomaly source or the same catastrophic process within consecutive or near-consecutive time windows. Event chains record the complete process of anomaly events from their emergence to their maintenance, enhancement, expansion, migration, stabilization, and decay. In addition to storing the event regions of each window, the event chain also stores the creation time, last update time, cumulative number of missing windows, current structural segment range, evolution characteristics, propagation consistency score, current evolution stage, and current warning level.

[0042] Structural segment diagram: This refers to a graph structure with highway structural segments as nodes and engineering connections or permissible influence relationships between segments as edges. Nodes can correspond to slope units, roadbed segments, abutments, bridge spans, tunnel segments, culverts, retaining structure units, or other monitoring units. Edges can store connection type, spatial distance, permissible expansion direction, apparent expansion speed range, structural transformation attributes, fiber coupling confidence level, and blocking attributes.

[0043] Propagation consistency refers to whether the structural reachability, response sequence, apparent expansion rate, structural boundary crossing method, and frequency band changes conform to preset engineering relationships when an event region expands, migrates, or influences another structural segment within a continuous window. Propagation consistency is used to determine whether the spatial changes in an event chain are reasonable for a highway structure, rather than referring to the consistency of the physical wave velocity of vibration waves in the material.

[0044] Apparent expansion velocity: refers to the rate at which the center, boundary, or affected area of ​​an event changes along structural or road mileage coordinates. This velocity is used to describe the change process of the anomalous affected area on the engineering structure and is not equivalent to the propagation speed of sound waves in optical fibers or the elastic wave velocity of structural materials.

[0045] In a first aspect, embodiments of the present invention disclose a method for monitoring and classifying early warning of highway DAS disaster evolution, such as... Figure 1 As shown, the method includes the following steps: S1. Continuously collect DAS vibration data along the highway and establish a mapping relationship between the optical fiber spatial sampling location and the highway structural section; S2. DAS vibration data is processed using a sliding time window, and one or more event regions are generated within each window based on the abnormal response map. S3. Cross-window association between the event area of ​​the current window and the historical event chain: Convert the event area to structural coordinates according to the mapping relationship, determine the reachable path based on the structural segment diagram, calculate the association cost for combinations with reachable paths, add the current event area to the corresponding event chain when the association cost meets the preset association conditions, and establish a new candidate event chain when the association conditions are not met. S4. Calculate the evolution characteristics of the established event chain; S5. Use the structural segment diagram to verify the propagation consistency of the event chain and calculate the propagation consistency score. S6. Determine the evolutionary stage of the event chain based on evolutionary characteristics and propagation consistency scores; S7. Based on the evolution stage and its duration, the early warning level is controlled by the early warning state machine. The upgrade conditions and rollback conditions in the early warning state machine adopt different duration window thresholds. The rollback confirmation window number is not less than the upgrade confirmation window number. S8. Output the early warning results and save the evolution process record of the event chain.

[0046] Specifically, this invention first continuously collects DAS vibration data along the highway and establishes a mapping relationship between the optical fiber spatial sampling location and the highway structural sections. The system uses a sliding time window to process continuous data, generating one or more abnormal event regions within each window. The event regions are not the final early warning conclusions, but rather the basic objects for subsequent cross-window correlation and evolution analysis.

[0047] When a new time window arrives, the system does not immediately treat the newly detected event region as an independent catastrophic event. Instead, it compares it with historical event chains that are still active. The system first eliminates obviously unreachable regions and event chain combinations based on the structural segment diagram, and then calculates the association cost for reachable combinations. The association cost comprehensively considers the spatial location of the event region, its boundary overlap or expansion direction, frequency band composition, intensity variation trend, and structural segment relationships. If the current event region meets the association conditions with an event chain, the event region is added to that event chain; if the current event region does not meet the association conditions with any active event chain, a new candidate event chain is established; if an active event chain temporarily has no matching region within a few consecutive windows, it remains in a short-term missing state instead of immediately terminating.

[0048] After the event chain is established, the system calculates evolutionary characteristics such as persistence, vibration intensity growth rate, spatial range expansion rate, event center migration speed, frequency band variation, and stability based on the changes in continuous event regions within the chain. Subsequently, the system uses a highway structural segment map to verify the propagation consistency of the event chain, determining whether the process of an event region developing from one structural segment to another conforms to structural reachability, permissible expansion direction, apparent expansion speed range, and structural boundary conditions.

[0049] The system determines whether an event is in a stage of pending confirmation, transient anomaly, continuous monitoring, development, acceleration, stabilization, decline, or propagation conflict based on the event chain evolution characteristics and propagation consistency. The early warning state machine then determines whether the current early warning level should be generated, upgraded, maintained, rolled back, lifted, or automatically upgraded based on the evolution stage, stage duration, propagation consistency, risk attributes of the area, and historical early warning levels.

[0050] This invention does not rely solely on abnormal amplitude, single-window classification results, or fiber optic coordinate continuity for early warning. Instead, it combines continuous event chain evolution, structural propagation consistency verification, and an early warning state machine. The system only raises the warning level when the abnormal event is continuous in time, conforms to structural expansion relationships spatially, and meets escalation conditions in terms of state. Conversely, the system gradually lowers the warning level when the event intensity decreases, the impact range shrinks, or propagation stops and the fallback conditions are consistently met. For events with high amplitudes but unreasonable propagation relationships, the system does not directly escalate to a high-level disaster warning but instead outputs propagation conflict or verification prompts.

[0051] Figure 2 This is a flowchart of the overall process of the highway DAS disaster evolution monitoring and graded early warning method according to an embodiment of the present invention. It sequentially represents continuous DAS data acquisition, fiber optic location and structural segment calibration, sliding time window processing, event area generation, structural reachability screening, event chain construction and management, evolution feature calculation, propagation consistency verification, disaster evolution stage determination, early warning state machine control, early warning result output, and historical record storage.

[0052] In one specific embodiment, the present invention first performs DAS data acquisition and spatial calibration steps, such as... Figure 2 As shown.

[0053] Step S1: Continuous DAS vibration data acquisition and spatial calibration The system deploys sensing optical fibers along the highway to be monitored. These fibers can be specifically laid near road shoulders, ditches, slope toes, retaining structures, bridge approach roads, or tunnel sidewalls, or existing communication optical cables that meet the monitoring requirements can be used. The deployment path and coupling method of the sensing optical fibers should be determined based on the monitoring target and site conditions to ensure that the vibration transmission efficiency between the optical fiber and the structure to be monitored meets the monitoring requirements.

[0054] The DAS demodulation device transmits probe light pulses into the optical fiber at a set pulse repetition frequency and receives backscattered light signals from various spatial locations within the fiber. By demodulating the phase, intensity, or frequency changes of the backscattered light, the vibration or strain rate response varying with time at different spatial locations along the optical fiber can be obtained.

[0055] Continuous vibration data is generated based on the changes in backscattered light at different fiber positions. The raw data can be represented as follows: ,in Indicates the spatial sampling position along the fiber direction. The sampling time is indicated, and the data values ​​represent phase changes, strain rates, vibration amplitudes, or other vibration-related physical quantities converted from the original optical signals.

[0056] Before the system is put into operation, a mapping table between the spatial locations of optical fibers and the locations of highway structures needs to be established. This mapping table can be established using one or more combinations of the following methods: Manual tapping calibration: Vibration excitation is generated at known locations on the highway structure (such as milestones, bridge abutments, tunnel entrances, etc.) by manually tapping the structures. The response position of this excitation on the fiber optic distance coordinate system is then recorded, establishing a correspondence between the fiber optic distance coordinates and the highway structure location. This method is simple to operate, low in cost, and suitable for fine calibration of local sections.

[0057] Vehicle positioning and calibration: Vehicles equipped with GPS positioning devices travel along the highway, and the DAS system records the vehicle's vibration trajectory on fiber optic distance coordinates. Combined with GPS timestamps and location information, a correspondence between the fiber optic distance coordinates and the highway mileage is established. This method is highly efficient and suitable for rapid calibration of long-distance highways.

[0058] Known equipment excitation: This method utilizes the characteristic vibrations generated by fixed equipment (such as toll booths, information boards, and lighting poles) at known locations along the highway to identify their corresponding positions on fiber optic distance coordinates. This approach requires no additional excitation source and can fully utilize existing infrastructure.

[0059] Construction Records: Based on the optical cable laying construction records, including the optical cable route design diagram, laying mileage markers, junction box locations, slack cable coil locations, and optical cable laying methods such as conduit / direct burial / cable tray, a correspondence between optical fiber distance coordinates and highway mileage is established. The construction records provide the original location information of the optical cable during laying and are one of the basic data sources for establishing the mapping table.

[0060] Fiber Optic Cable Geographic Information System (GIS) Data: Based on fiber optic cable laying records and GIS data, the spatial route of the fiber optic cable and its spatial relationship with the highway structure are determined. GIS data can be cross-checked with on-site survey results.

[0061] On-site surveying: Using surveying equipment such as total stations and RTK-GPS, the coordinates of ground markers along the sensing fiber optic route (such as cable manholes, junction boxes, cable remnant wells, bridge abutment ends, tunnel entrances, etc.) are measured in real time. Combined with highway design coordinates and mileage markers, a precise mapping between fiber optic distance coordinates and highway structure locations is established. On-site surveying can correct potential errors in construction records and improve mapping accuracy.

[0062] A structural segment can be a continuous roadbed segment, an independent slope unit, a bridge span, a bridge abutment, a tunnel section, a culvert, or a retaining wall unit. The mapping table must store at least the following information: fiber optic distance range, corresponding road mileage range, corresponding structural segment number, structural type (such as roadbed, slope, abutment, bridge span, etc.), and fiber optic coupling status (normal, loose, broken fiber, etc.).

[0063] In scenarios involving cable reversals, cable coiling, or cross-structure laying, the same fiber optic distance neighborhood may correspond to different engineering locations, or adjacent locations in engineering may not be adjacent in fiber optic coordinates. For example... Figure 3 As shown, the present invention converts fiber coordinates into structural coordinates through the above mapping table. All subsequent propagation consistency checks are based on the structural segment relationships, rather than simply using the sequence number or proximity of fiber sampling points.

[0064] Figure 3 This diagram illustrates the mapping between the spatial location of optical fibers and highway structural sections according to an embodiment of the present invention. It represents the mapping relationship between the DAS fiber distance coordinates and roadbed sections, abutment sections, slope sections, excess cable, or turnaround points, as well as the formation methods of structural reachability relationships, permissible expansion directions, and structural boundary conditions. In the diagram, solid lines represent normal calibration mappings or connections that can participate in the establishment of structural relationships, while dashed lines represent mapping relationships on the fiber coordinates that may be adjacent but require interruption or verification of the engineering structural relationships.

[0065] After continuously receiving DAS data, the system marks abnormal missing values, saturation values, and data indicating obvious equipment failures. Background correction can be performed using historical data from the same time period, stable data at the beginning of the current window, or a recursive background model. To reduce the impact of traffic flow and equipment gain changes, the background center value and scale value can be estimated separately according to spatial location, and then the current vibration response can be converted into a relative background deviation. The system can also decompose the signal into several frequency bands, such as low-frequency, mid-frequency, and high-frequency, to retain the difference information between low-frequency gradual changes, mid-frequency structural responses, and high-frequency impacts.

[0066] Step S2: Sliding time window processing and event area generation The system is based on length as The sliding window process handles continuous DAS data streams. Adjacent windows overlap, and the window movement step size is... ,in Less than or equal to The purpose of setting overlapping windows is to prevent the disaster response from being truncated when it happens to be at the window boundary, and to provide a common observation interval for subsequent cross-window associations, so as to ensure information continuity between adjacent windows.

[0067] Window length The selection can be based on the monitoring target. For example, a short window (e.g., 10 seconds) can be used to capture transient responses to rockfalls, impacts, and cracks, while a long trend window (e.g., 60 seconds) can be used to observe long-term trends in slope slippage, roadbed deformation, and structural vibration. In a preferred embodiment, the system can run multiple windows of different lengths in parallel to accommodate the monitoring needs of events at different time scales.

[0068] In the Within each window, the system calculates the window anomaly response map based on indicators such as background deviation, energy in each frequency band, rate of change of time, and spatial continuity. Where d is the fiber optic distance coordinate or structural coordinate. This is the time within the window.

[0069] In a preferred embodiment, the calculation steps for the anomaly response diagram are as follows: First, the system performs background normalization on the raw DAS data, converting the current vibration response into a standardized deviation relative to the background. Secondly, the frequency band energy of each spatial location at each time is calculated according to the three frequency bands of low frequency, medium frequency and high frequency; Then, the time-space points that exceed the anomaly threshold are marked to form a binary map of candidate anomalies; Finally, connected component analysis is performed on the candidate anomalies, and spatially adjacent and temporally continuous anomalies are clustered into candidate event regions.

[0070] The system filters candidate event regions obtained from connected component analysis. Regions with durations less than a preset value (preferably...) are deleted. () or the space length is less than the preset value (preferred) Isolated regions, which are usually caused by random noise or occasional perturbations, do not have continuous tracking value.

[0071] For the reserved region, calculate the peak intensity, average intensity, main frequency band (i.e., the energy concentration band), temporal slope (i.e., the slope of intensity change over time), spatial length, and regional confidence.

[0072] The event area is denoted as ,in Indicates the window number. This indicates the region number within the window (multiple unconnected event regions may exist within the same window). Each event region stores all the above descriptive values, serving as the base object for subsequent cross-window associations.

[0073] Each event region not only stores its boundaries but also the descriptive data required for subsequent associations, including the region's center location, spatial length, duration, peak intensity, average intensity, intensity slope over time, main frequency bands, frequency band energy ratios (the proportion of each frequency band's energy to the total energy), spatial response shape (e.g., clusters, stripes, movement trajectories), the set of structural segments corresponding to the region, region confidence level, and data quality markers. The generation of event regions is the input for subsequent processing in this invention. The focus of this invention is on how to form event chains from event regions within a continuous window and combine this with structural propagation consistency for evolutionary early warning.

[0074] It should be noted that the generation of event regions is only the starting point of the process in this invention. The core innovation of this invention lies in how to associate event regions in a continuous window into an event chain and combine it with structural propagation consistency for evolution warning.

[0075] Step S3: Construction and Management of Cross-Window Event Chains An event chain is used to represent a group of event regions that are considered to belong to the same anomaly source or the same catastrophic process within multiple consecutive or nearly consecutive windows. The event chain is in the first The form after processing each window can be represented as follows: ,in The window number for creating the event chain.

[0076] In addition to storing the event region sequence, the event chain also stores the following management information: created window number, most recently updated window number, current activity status (active, missing, fading, ended), cumulative number of missing windows, current covered structural segment range, current evolution stage (pending confirmation, transient anomaly, continuous monitoring, development, acceleration, stable, fading, propagation conflict) and current warning level (blue, yellow, orange, red or no warning).

[0077] When the After the event area of ​​each window is generated, the system first determines whether the current event area and the end areas of each historical event chain are in the same segment, adjacent segments, or reachable through an allowed path, i.e., whether there is a structural reachability relationship, based on the structural segment diagram.

[0078] Specifically, the system transforms the current event region to its corresponding set of structural segments using a mapping table, and also transforms the end regions of historical event chains to their corresponding structural segments. Then, it queries the structural segment graph to determine if there is a reachable path (i.e., a connection through edges in the graph) between two structural segments. If the structural segment corresponding to the current region is unreachable from the end segment of a historical event chain (i.e., there is no connecting path, or all paths are blocked by blocking attribute markers), then this combination is not included in the subsequent candidate association list.

[0079] The purpose of structural accessibility pre-screening is to quickly eliminate candidate combinations that are unlikely to be related in engineering before calculating the association cost, thereby reducing unnecessary computation.

[0080] The associated cost is not simply based on spatial distance, but rather on a comprehensive consideration of regional location, boundary overlap or expansion direction, frequency band response, vibration intensity trend, and structural segment relationships.

[0081] For the current region and event chain The associated cost can be expressed as: The components and their calculation methods are as follows: (1) Spatial location differences The predicted center coordinates of the event chain are obtained by normalizing the absolute distance between the current region center coordinates and the center coordinates predicted by the event chain based on historical motion trends. The predicted center coordinates of the event chain can be extrapolated based on the motion trajectories of the centers of the most recent window regions (such as uniform motion assumptions or quadratic motion assumptions).

[0082] (2) Differences in regional boundaries This is calculated by the ratio of the overlap length between the current region and the region at the end of the event chain in structural coordinates. A higher overlap ratio indicates that the two regions are more spatially continuous and have smaller boundary differences.

[0083] (3) Frequency band differences The similarity is obtained from the Euclidean distance or cosine similarity between the energy distribution vectors of the two main frequency bands. If the energy distributions of the two are similar, it indicates that the vibration sources may be the same; if the frequency band distributions are significantly different, they may originate from different vibration sources.

[0084] (4) Differences in intensity and trend The value is obtained from the relative rate of change of the average or peak intensity of the two regions. If the intensity of the current region changes in the same direction as the intensity of the event chain (e.g., continuously increasing or continuously decreasing), the difference is small; if the intensity changes in opposite directions (e.g., the event chain continuously increases while the current region suddenly decreases), the difference is large.

[0085] (5) Differences in the relationship between structural segments The penalty value is calculated based on the structural distance between corresponding structural segments of two regions. Specifically, the penalty value is minimal if the two regions are in the same structural segment; moderate if they are in adjacent segments; and substantial if they are in separate segments (with other segments in between). The structural distance penalty value can be determined based on the shortest path length in the structural segment diagram.

[0086] Before calculation, all differences are normalized to the same numerical range (e.g., between 0 and 1) to ensure that the components are comparable.

[0087] in, , , , , They are respectively , , , , The corresponding weight coefficients, and satisfying Each weight coefficient is greater than or equal to 0 and less than or equal to 1.

[0088] In the preferred embodiment, each weight can be taken as follows: , , , , The weight values ​​mentioned above reflect the importance of each dimension in the association determination—spatial location and structural segment relationships are the most important, while frequency bands have relatively lower weights.

[0089] When the association cost D is less than or equal to the preset association threshold (preferably 0.6), that is... If there is no better match conflict, the system will add the current event region to the corresponding event chain.

[0090] If multiple current event regions may match the same event chain, the system performs the matching based on the principle of minimizing total cost (i.e., the Hungarian algorithm or a similar optimal matching algorithm).

[0091] When a historical event is indeed likely to split into multiple spatial branches (e.g., the impact of a disaster expands and differentiates into multiple independent regions), the system allows the event chain to generate branch sub-chains and records the source relationship between the main chain and the branch chains. If multiple historical event chains converge into one region in the current window, the system can temporarily maintain multiple candidate associations, and subsequent windows will determine whether to merge them based on the propagation direction, regional morphology, and structural path.

[0092] If the current event region does not meet the association conditions with any of the active event chains (i.e., the association cost is greater than the threshold or none of them pass the structural accessibility pre-screening), the system establishes a new candidate event chain for the event region. The newly created event chain is in a pending confirmation state for the first few windows. It is only converted into an active event chain when the event continues to appear in subsequent windows, or when its intensity, spatial range, and structural risk meet the rapid confirmation conditions.

[0093] If an event chain does not have a matching event region in the current window, the system does not immediately terminate the event chain, but instead increments its missing window count. This is because a catastrophic event may temporarily weaken to below the detection threshold within a window, or be split at the window boundary, but this does not mean the event has ended.

[0094] In a preferred embodiment, the system allows for a number of consecutive missing windows in the event chain. During the waiting period, the system predicts the possible spatial location and intensity range of the event region based on the previous center migration, expansion direction, and intensity trend of the event chain. If an event region that meets the predicted location and structural propagation conditions reappears in subsequent windows, the event chain can be restored, but its continuity score decreases according to the number of missing windows. After exceeding the allowed number of missing windows, the system will either initiate a fading or terminate the event chain assessment.

[0095] Figure 4 This is a schematic diagram of cross-window event region association and event chain formation in an embodiment of the present invention. It is used to represent the association relationship between the (n-1)th window event region, the historical active event chain, and the nth window event region, as well as the processing procedures of structural reachability filtering, association cost calculation, adding existing event chains, creating new candidate event chains, event chain updating, short-term missing event retention, splitting and merging or ending, and outputting active event chains.

[0096] like Figure 4 As shown, the process of cross-window event region association and event chain formation embodies the core idea of ​​this invention: organizing discrete window detection results into a continuous event evolution sequence.

[0097] Step S4: Calculation of Event Chain Evolution Features Once the event chain is formed, the system no longer focuses on the abnormal score of a single window, but instead calculates the changes of the event over a continuous period of time.

[0098] Event persistence This is used to indicate the extent to which an event chain maintains a valid response within the most recent observation interval. Persistence can be calculated by considering the following factors: the number of valid windows (the number of windows in the event chain that match the event region), the total number of observation windows, the length of consecutive valid windows, the number of missing windows, and the confidence level of each association.

[0099] Events that appear consecutively in multiple windows and are stably associated have high persistence, while events that occur only once or are frequently interrupted have low persistence. Persistence can be calculated using a weighted summation method, where the length of the consecutive effective window is given a higher weight.

[0100] Vibration intensity growth rate This system is used to represent the trend of average intensity, peak intensity, or characteristic frequency band energy of an event region over time. It can perform linear fitting, robust trend estimation, or exponential smoothing on the intensity sequences of several recent regions within an event chain, and then calculate the current growth rate.

[0101] To avoid misjudgment caused by a single peak, the growth rate is not obtained by subtracting two adjacent windows directly, but rather by using the trend of multiple consecutive windows. For example, the least squares method can be used to linearly fit the intensity values ​​of the most recent 5 windows, and the slope of the fitted line can be used as the growth rate.

[0102] Space expansion rate This system is used to describe the extent to which the influence of an event changes over time. It can calculate the changes in the spatial starting point, ending point, center, and region length separately.

[0103] Specifically, if one side of the event area remains stable while the other side continues to move towards adjacent structural sections, it indicates that the event has directional expansion characteristics, which is often a typical manifestation of disaster development; if the center of the area moves rapidly while the length of the area remains basically unchanged, it may correspond to moving vibration sources such as vehicles; if the length of the area increases while the center remains basically stable, it may correspond to local anomalies spreading to both sides.

[0104] Center migration speed The migration velocity is obtained by dividing the displacement of the event region's center position within a continuous window by the corresponding time interval. Structural coordinates or road mileage coordinates are used to calculate the migration velocity, with corrections made based on fiber optic return relationships.

[0105] Different types of anomalous events have different typical migration speed ranges. For example, the migration speed of passing vehicles is typically tens of meters per second (corresponding to the design speed of highways); the migration speed of slope slippage is typically on the order of millimeters per second to centimeters per second; while the migration speed of rockfall impacts can be extremely high. The center migration speed is one of the important characteristics that distinguishes catastrophic events from non-catastrophic disturbances.

[0106] Frequency band change Used to represent changes in the main frequency band or the energy center of the frequency band. Some disasters may gradually develop mid-to-high frequency fractures or impact components from a slow low-frequency response, while vehicle responses typically exhibit speed-related frequency band characteristics. Therefore, frequency band changes can serve as auxiliary information for distinguishing between evolutionary states and interference.

[0107] Event chain stability This is used to evaluate the association quality, smoothness of boundary changes, and consistency of feature changes among the most recent windows. If the event region exhibits large, irregular jumps between adjacent windows, or if the frequency band and intensity patterns change repeatedly, the stability is reduced; if the event regions of each window maintain a smooth and gradual change in position, shape, and features, the stability is higher.

[0108] The evolutionary characteristics of an event chain can be uniformly represented as: This serves as the input for determining subsequent evolutionary stages and controlling early warning states.

[0109] Step S5: Establishing structural propagation relationships and verifying propagation consistency To determine whether the spatial changes in the event chain conform to the highway structure conditions, the system establishes a structural segment diagram. Nodes in the structural segment diagram correspond to roadbed segments, slope units, bridge components, tunnel segments, culverts, or other monitoring units, while edges indicate direct connections or permissible influence extension relationships between two structural segments.

[0110] The specific steps for constructing the structural segment diagram are as follows: Step A, Node Division: Based on highway design data and a Geographic Information System (GIS), the highway is divided into independent physical monitoring units, which serve as nodes on the structural section map. For example, slopes are divided into independent landslide units based on geological exploration maps; bridge abutments and individual bridge span units are divided based on bridge design drawings; and different roadbed segment units are divided based on roadbed design data.

[0111] Step B, Initial Edge Construction: Based on the site survey and geographic information topology, establish the engineering connection relationships between each node (i.e., the edges in the figure). If two slope units are geographically connected and the hydrogeological conditions allow for displacement expansion, then establish a bidirectional or unidirectional edge that allows for the expansion of influence; if the optical cable crosses an empty area or isolation zone, then mark the edge as a blocking attribute; if two structural sections are physically connected but belong to different types of structures (such as roadbed and bridge abutment), then mark them as structural transformation edges.

[0112] Step C, On-site Mapping and Parameter Calibration: Based on the fiber optic cable laying records, vibration excitation at known locations is generated in different structural sections using methods such as manual tapping calibration, known equipment excitation, or vehicle GPS positioning tests. The system records the distance coordinates of the fiber optic response, thereby establishing an accurate mapping table between fiber optic coordinates and structural nodes.

[0113] Step D, Parameter Dynamic Correction: Based on the cross-segment attenuation characteristics of the vibration signal obtained from the field test, the edge attributes of the structural segment diagram (such as the frequency band change rules across the boundary, structural conversion type, apparent expansion velocity range, etc.) are verified and corrected, and finally a structural segment diagram that can be used for propagation consistency verification is generated.

[0114] Each edge in the structural segment diagram stores the following attributes: connection type (direct connection, indirect connection, transformation connection), spatial distance, allowed expansion direction (bidirectional, unidirectional, or specific direction), apparent expansion speed range, structural transformation attribute, fiber coupling reliability, and blocking attribute (normal, blocking, requires verification).

[0115] The allowed expansion direction can be bidirectional, or it can be set to a direction with different weights based on the slope aspect, water flow direction, structural force path, or historical event patterns.

[0116] The apparent expansion velocity range is not equivalent to the elastic wave velocity in the material, but rather the reasonable velocity range of the anomalous influence region observed in the system as it expands or migrates along the structure.

[0117] The apparent expansion rate range is determined based on the following criteria: the apparent expansion rate range for different types of structural sections and different types of disasters is mainly determined based on highway engineering experience, physical motion laws, and historical monitoring data. Specifically: (1) Progressive slope slippage or roadbed settlement process: In the early initiation and development stages, this type of disaster manifests as slow stress release and microcrack propagation. Based on historical disaster databases and long-term monitoring experience, its apparent propagation rate is usually extremely low, which can be set between 0.01 m / s and 0.5 m / s. The system identifies slow disasters by tracking this low-speed propagation process.

[0118] (2) Vehicle movement interference: The migration speed of the vibration center caused by vehicles is limited by the statutory design speed of the highway section and historical traffic flow statistics. For example, if the speed limit of a certain highway is 60~120 km / h, then its apparent spread speed range can be determined to be 16 m / s to 33.3 m / s. When the system detects the migration of the event area within this speed range, it can determine it as vehicle movement interference through propagation consistency verification, thereby suppressing the false triggering of high-level warnings.

[0119] (3) Rockfall impact or local collapse events: These sudden events are instantaneous, and their apparent expansion speed is extremely high or affects adjacent sections in a very short time (such as within a short window of 10 seconds). Based on the free fall physical model and the impact diffusion range of historical rockfall events, the upper limit of its speed range can be set to greater than 50 m / s.

[0120] By combining the aforementioned historical data with physical laws, the system can assign a reasonable apparent expansion velocity range to each edge in the structural segment diagram, serving as the basis for calculating the propagation consistency score.

[0121] Blocking boundaries are used to indicate that although two fiber optic locations are geographically adjacent, their corresponding structures should not be considered as a continuous catastrophic extension. Typical blocking boundaries include: fiber optic slack sections, fiber optic cable turnaround points, crossings of voids (such as suspended sections between bridges and the ground), isolation locations between different bridge spans, and unconnected slope units.

[0122] A structural transition boundary is used to indicate that an event can cross the boundary, but the speed, frequency band, or intensity should change accordingly after crossing. Typical structural transition boundaries include: roadbed entering bridge abutment, slope entering retaining wall, and open-cut tunnel entering tunnel.

[0123] For sections where propagation parameters cannot be determined, a wider allowable range can be set and the weight of structural verification can be reduced, rather than simply judging them as unreachable.

[0124] Figure 5 This diagram illustrates the structural segment diagram and propagation consistency verification in an embodiment of the present invention. It represents the relationships between the highway structural segment diagram, spatial changes in candidate event chains, structural accessibility, response order, apparent expansion speed, structural boundary conditions, frequency band changes, and propagation consistency scores. It also indicates that when propagation conditions are met, the event chain is allowed to continue or escalate; when propagation conditions are not met, propagation conflict occurs or automatic escalation is suspended. Solid lines in the diagram represent structural relationships that normally participate in propagation consistency verification, while dashed lines represent blocking boundaries or relationships requiring verification. These types of relationships are not used as direct evidence for automatic event chain continuation or automatic early warning escalation.

[0125] like Figure 5 As shown, the relationship between the structural segment diagram and the propagation consistency verification embodies the core idea of ​​this invention: to convert the spatial relationship of optical fibers into an engineering structural relationship for verification.

[0126] The system checks whether the structural segment changes in event regions within an event chain, whether adjacent windows or areas separated by several windows, conform to the preset structural propagation relationship. Propagation consistency is comprised of the following five dimensions: (1) Structural consistency can be achieved Based on the structural segment diagram, determine whether there is a reachable path from one structural segment to another in the event region. If the reachable path is unobstructed, assign a high score (e.g., 1.0); if it is unreachable (no path exists or all paths are blocked), assign a value of 0.

[0127] (2) Consistency of response order The scoring is based on whether the spatial order in which the event areas appear over time is consistent with the allowed expansion direction in the structural segment diagram. If the events appear sequentially in the allowed direction (e.g., from the upstream slope to the downstream slope), the score is higher; if the order of appearance is opposite to the allowed direction or disordered, the score is lower.

[0128] (3) Consistency of apparent expansion rate The scoring is based on whether the central migration speed of the event chain falls within a reasonable speed range defined in the structural segment diagram. Higher scores (e.g., 1.0) are assigned to speeds within the range, and lower scores are assigned to speeds further away from the range.

[0129] (4) Consistency of structural boundaries The scoring is based on whether the event meets the preset physical transformation conditions when crossing the structural boundary. For example, when entering the bridge abutment from the roadbed, if the frequency band and intensity changes are as expected, the score is higher; if it jumps directly to the middle of an unconnected bridge span, the score is lower.

[0130] (5) Consistency of frequency band variation The scoring is based on whether the frequency band energy change pattern before and after the abnormal event crosses different structural media (such as from soil roadbed to concrete bridge abutment) conforms to engineering experience.

[0131] Propagation Consistency Score The calculation formula is: In the preferred embodiment, each weight can be taken as follows: , , , , The weights and thresholds mentioned above are one feasible example. In actual engineering projects, they can be adjusted according to the road segment type, optical cable laying method, structural risk level, historical event samples, or manual calibration results. They do not constitute a limitation on the scope of protection of this invention.

[0132] Based on the numerical range of the propagation consistency score, the system makes the following classification judgment: when When this occurs, the system determines that there is a propagation conflict in the event chain; the spatial changes in the event chain are seriously inconsistent with the structural propagation relationship, which may indicate data errors, fiber mapping errors, or aliasing of different events.

[0133] when At that time, it was determined that the credibility of the propagation was low; the spatial changes in the event chain conformed to the structural propagation relationship, but there were several inconsistencies that required further observation.

[0134] when When the event chain changes spatially, it is considered to be a valid propagation event; the spatial changes of the event chain basically conform to the structural propagation relationship and can be treated as a valid evolutionary event.

[0135] when When the propagation is highly consistent, the spatial changes in the event chain fully conform to the structural propagation relationship, indicating a high degree of credibility.

[0136] The propagation consistency score is not only used for post-event evaluation but also participates in the correlation determination of event chains. If the current event region is similar to a historical event chain in signal characteristics, but joining the event chain would cause obvious structural propagation conflicts, the system will reject the association or retain it as a low-confidence candidate. If an event exhibits high-intensity vibrations in multiple consecutive windows, but its spatial location jumps between multiple unreachable structural segments, the system will identify it as a propagation conflict or data anomaly, and will not directly escalate it to the disaster stage.

[0137] Step S6: Determining the Stage of Cataclysmic Evolution The system divides the event chain state into eight stages: pending confirmation, transient anomaly, continuous monitoring, development, acceleration, stabilization, decay, and propagation conflict.

[0138] Specifically, pending confirmation: the initial state of the newly created event chain. At this point, only one or a few windows have appeared for the event, which is insufficient to confirm it as a valid event; Transient anomalies: These events occur only within a single window, or have insufficient persistence or low correlation stability. Such events are typically sporadic disturbances and not worth continuous tracking. Continued monitoring: The event appears consecutively in at least two windows, and the consistency of propagation meets the minimum effective conditions (e.g., Even if the intensity has not yet increased significantly, it is worth continuing to observe; Development: The event occurs consecutively for at least three windows, and at least one of the vibration intensity growth rate, spatial spread rate, or frequency band change reaches the development condition, and the propagation consistency meets the effective propagation condition (e.g., ); Acceleration: The intensity increase and spatial expansion accelerate simultaneously, or the event enters a high-risk structural segment and maintains reasonable propagation, with a high level of consistency in propagation (e.g., ); Stable: The intensity, spatial extent, and migration speed of the event remain stable over a certain period of time, without further amplification or significant decline; Regression: The intensity of the event continues to decrease, the spatial range shrinks, and the propagation stops, and the above conditions are maintained for the prescribed number of windows; Propagation conflict: The event has high vibration intensity or long duration, but there is a significant conflict in structural accessibility, response sequence, or apparent propagation rate.

[0139] A newly created event chain first enters the confirmation phase. If an event occurs within only one window, or has insufficient persistence or low correlation stability, the system maintains it as a transient anomaly. If an event occurs consecutively for at least two windows, and propagation consistency meets the minimum valid condition, for example... Even if the intensity has not yet increased significantly, it can still enter a phase of continuous monitoring.

[0140] When an event occurs consecutively for at least three windows, and at least one of the vibration intensity growth rate, spatial spread rate, or frequency band change reaches the development condition, and the propagation consistency meets the effective propagation condition, for example... The system classifies the event as being in the development stage. If the intensity increase and spatial expansion accelerate simultaneously, or if the event enters a high-risk structural segment and maintains reasonable propagation with a high level of consistency, for example... The system classifies the event as an acceleration phase. The determination of development and acceleration uses multiple consecutive window conditions; it does not allow for cross-level transitions based on a single peak value.

[0141] When the intensity, spatial extent, and migration speed of an event remain stable within a certain period, the system marks it as a stable phase. Stability does not mean the risk has been eliminated, but rather that there is no trend of further escalation in the current phase. If the event intensity continues to decrease, the spatial extent shrinks, propagation ceases, and these conditions persist for a specified window period, the system enters a decay phase. After the decay state persists for a period of time, the event chain ends and the event becomes a historical event.

[0142] If an event has high vibration intensity or a long duration, but there are significant conflicts in structural accessibility, response sequence, or apparent propagation speed (i.e., a propagation consistency score below 0.4), the system classifies it as a propagation conflict state. A propagation conflict state is not directly equivalent to normal interference, as it may also originate from fiber mapping errors, unknown structural anomalies, or the aliasing of multiple events. The system outputs the structural segment where the conflict occurred and the cause of the conflict for manual verification or verification by other sensors.

[0143] The number of windows, thresholds, growth conditions, and risk section correction conditions for each of the above evolution stages can be configured according to the road segment type, optical cable laying method, structural risk level, and historical monitoring data. The stage determination result is not directly equivalent to the final warning level, but serves as the input to the warning state machine. The warning state machine, in conjunction with the stage holding time, propagation consistency, and historical warning levels, determines whether to generate, upgrade, maintain, revert, or cancel the warning.

[0144] Step S7: Control of warning level escalation, maintenance, and rollback The early warning status control unit converts the event evolution stage into an engineering early warning level. In a preferred embodiment, the early warning level uses four levels: blue, yellow, orange, and red, corresponding to the four levels of attention, warning, alert, and emergency, respectively.

[0145] The relationship between warning levels and evolution stages is not a simple one-to-one correspondence. It is also necessary to consider the risk attributes, structural type, consistency of propagation, and duration of the state of the road segment where the event occurs.

[0146] The system sets escalation and rollback conditions for each level of alert. When an event progresses from no alert to blue alert (a level of concern), it needs to maintain a continuous monitoring phase for at least two windows. When an event escalates from blue to yellow, it needs to maintain a development phase for at least two windows, and the consistency of propagation must not be lower than the effective propagation threshold (e.g., ...). When an event escalates from yellow to orange, it needs to enter an accelerated phase, or be in a developing phase within a high-risk area and continue to expand, with a high level of consistency in its propagation (e.g., ...). When an event escalates from orange to red, it requires a rapid expansion of the high-risk area, or a simultaneous rapid increase in both intensity and spatial extent while maintaining highly consistent propagation.

[0147] To prevent the warning level from fluctuating around the threshold, different thresholds and durations are used for escalation and rollback conditions. Once an event reaches a higher warning level, a single window's intensity decrease is insufficient to immediately cause it to roll back. The system only lowers the warning level step by step when the intensity decreases, the spatial extent shrinks, or propagation ceases for a sustained period reaching the rollback window count.

[0148] In a preferred embodiment, the number of rollback confirmation windows is three for yellow alerts, four for orange alerts, and five for red alerts. The number of rollback confirmation windows is not less than the number of escalation confirmation windows, and the number of rollback confirmation windows is longer for higher-level alerts. This asymmetric hysteresis control mechanism ensures the stability of the alert output.

[0149] When propagation consistency suddenly decreases, the system takes different actions depending on the cause: if the event reaches the structural transition boundary and causes parameter changes (such as moving from the roadbed to the bridge abutment), the system can maintain the warning and adjust the propagation parameters; if the event area jumps to an unreachable structural section, the system stops automatic upgrades and enters a propagation conflict state, while outputting conflict information for manual review; if the fiber optic coupling status decreases or the data quality is insufficient, the system maintains the most recent trust level and outputs a data quality prompt.

[0150] The risk attributes of road sections can be adjusted to modify the conditions for state transition. For example, in high-risk slopes, bridge abutments, or tunnel entrances, the number of duration windows required for escalation can be reduced, but basic propagation consistency requirements must still be met. Risk correction cannot replace event evolution judgment; it only adjusts the speed of early warning response.

[0151] Figure 6 This is a flowchart of the disaster evolution determination and early warning status control in an embodiment of the present invention. It is used to show that after the event chain is input, the system determines whether it is a valid evolution event based on the evolution characteristics and propagation consistency, and further generates or upgrades the early warning, maintains the current early warning, rolls back or cancels the early warning, suspends the upgrade or reviews the warning through early warning status control, and finally outputs the early warning result and records the reason.

[0152] like Figure 6 As shown, the control rules for an early warning state machine may include the following: When an event is in a state of no warning, if the event enters the continuous monitoring phase and maintains the set number of windows, the system will generate a blue warning; if the event chain ends or the credibility is insufficient, the system will remain in a state of no warning.

[0153] When an event is under a blue alert, if the event enters the development stage and maintains the set number of windows, and the consistency of propagation reaches an effective threshold, the system will upgrade the alert level to yellow; if the event is still trackable but does not meet the yellow conditions, the system will remain in blue; if several consecutive windows show a decrease in intensity, a shrinkage in spatial range, or no matching event area, the system can roll back or lift the blue alert.

[0154] When an event is under a yellow alert, if the event enters an acceleration phase, or continues to develop and expand reasonably within a high-risk area, the system will upgrade the alert level to orange; if the event is still in the development phase but has not accelerated further, the system will remain yellow; if the conditions for fading continuously reach the number of yellow rollback windows, the system will roll back level by level.

[0155] When an event is under an orange alert, if the event rapidly expands in a high-risk area, or if the intensity and spatial range increase rapidly at the same time and the propagation is highly consistent, the system will upgrade the alert level to red; if the high intensity or high-risk state continues, the system will remain orange; if the conditions for fading continuously reach the number of orange rollback windows, the system will roll back level by level.

[0156] When an event is under a red alert, if the high-risk status continues or the rollback conditions are not met during the handling period, the system remains red; if the rollback conditions reach the red rollback window number consecutively, the system does not immediately lift the alert, but rolls back level by level in the order of red, orange, yellow, and blue.

[0157] In the aforementioned state machine, the number of rollback confirmation windows is no less than the corresponding number of upgrade confirmation windows. The system preferably adopts a step-by-step upgrade and rollback approach to avoid alert level jumps caused by a single window's anomaly or recovery.

[0158] Step S8: Output of early warning results, online updates and historical records The system outputs the following information for each event chain: unique identifier, first occurrence time, last update time, current structural segment, current spatial impact range, main vibration frequency band, event duration, current evolution stage, propagation consistency score, current warning level, and reason for level change.

[0159] The reasons for the change in level can be specifically described as follows: the number of continuous windows reaches the upgrade condition, the spatial influence range continues to expand, the apparent expansion rate is consistent with the slope section range, the event intensity continues to decrease and meets the retreat condition, or it crosses the boundary of an inaccessible structure and stops automatic upgrading.

[0160] The monitoring platform can link on-site video, inspection tasks, traffic control, and information notifications based on the warning level. For events with transmission conflicts or insufficient data quality, the system provides suggested fiber optic locations and structural sections for review, facilitating on-site personnel to quickly pinpoint the problem area.

[0161] For completed event chains, the system saves their complete evolution process, including the event areas of each window, evolutionary feature sequences, changes in propagation consistency scores, changes in warning levels and their causes, etc., so as to facilitate subsequent event review, system parameter optimization, or use as training and update data for machine learning models.

[0162] After system startup, the fiber space-structure segment mapping table, structure segment diagram, window parameters, event chain association parameters, and early warning state transition parameters are loaded. The DAS demodulation device continuously writes new data into the circular buffer. When the data in the buffer reaches the length of a complete window, the window data processing unit reads the current window and begins calculation. At the same time, the acquisition unit continues to receive subsequent data to avoid blocking data acquisition during the processing.

[0163] When a window finishes processing, the system saves the current active event chain, a summary of the completed event chains, alert status, and necessary intermediate features. When the next window arrives, the system does not need to reread all historical DAS data; it only needs to read the most recent state of the active event chain and a limited historical region reserved for trend calculation.

[0164] Therefore, this method can be continuously run in a streaming manner, and its storage capacity mainly depends on the number of active events and the historical length of each event chain, rather than monitoring the total running time, which has good engineering practicality and scalability.

[0165] To facilitate implementation of this invention by those skilled in the art, a set of preferred parameter examples are provided below. It should be noted that these parameters are only used to illustrate one feasible approach. In actual engineering, adjustments can be made based on the fiber optic laying method, DAS equipment sampling rate, road segment structure type, traffic interference intensity, historical monitoring data, and management system. Parameter adjustments do not alter the technical essence of this invention.

[0166] Assume the DAS data has a preprocessed temporal sampling rate of 1000 Hz and a spatial sampling interval of 1 m. The system uses a short window length. = 10 s, window step size = 2 s, with adjacent windows having an 8 s overlap. For slower-developing slope events, parallel processing can be used. = 60 s long trend window to observe changes in intensity and spatial range.

[0167] The minimum duration of the event region is 0.5 s, and the minimum spatial length is 3 m. The allowed number of consecutive missing windows in the event chain is [not specified]. = 2 windows. The association threshold D of the event chain ≤ 0.6. The effective threshold C for propagation consistency ≥ 0.6, and the high consistency propagation threshold C ≥ 0.75.

[0168] In the warning state machine, the escalation confirmation conditions for blue, yellow, and orange are two consecutive windows meeting the corresponding escalation conditions, respectively; the rollback confirmation conditions for yellow, orange, and red are three, four, and five consecutive windows meeting the corresponding rollback conditions, respectively.

[0169] The present invention will now be described in detail through several specific embodiments.

[0170] Example 1: Monitoring of Progressive Slope Slippage Sensor fibers are laid near the toe of the highway slope, and the slope is divided into multiple structural sections (such as A1, A2, A3, A4, A5, etc.) along the road direction and slope units. The system continuously collects DAS data and uses overlapping time windows to generate event regions.

[0171] When a short-duration, low-frequency local response first appears in a certain slope section (e.g., A3), the system establishes an event chain to be confirmed. In subsequent windows, the response continues to appear in the same section, its persistence increases, and the propagation consistency meets the minimum valid condition, at which point the event chain enters the continuous attention phase.

[0172] As the slip process continued, the upper boundary of the anomalous area and the boundary along the road gradually expanded, and the low-frequency energy continuously intensified. The event chain management unit connected the event areas in these continuous windows to the same event chain, and the evolution characteristic calculation unit obtained positive intensity growth rate and spatial expansion rate. The structural propagation consistency verification unit found that the newly added affected sections were connected to the original sections, the expansion direction conformed to the slope structure relationship, and the apparent expansion rate was within the allowable range (e.g., 0.01~0.5 m / s), thus the propagation consistency remained at a high level. The evolution stage changed from continuous monitoring to development, and the warning level was raised from blue to yellow.

[0173] If the intensity growth rate and spatial expansion rate further increase in subsequent windows, and the event extends towards the pre-defined high-risk structural sections, the system will classify the event as an acceleration phase and raise the warning level to orange or red. If the vibration intensity continues to decrease and the affected area gradually shrinks after on-site measures are taken, and the conditions for reversal are met, the system will gradually lower the warning level, rather than immediately lifting the warning after a single window returns to normal.

[0174] The following example, using a specific set of window data, illustrates the change in warning level as the state evolves continuously. This example is not intended to limit the scope of protection: In the first window In the middle, structural section The average intensity is approximately 1.2, and the spatial length is approximately... For local anomalies with a propagation consistency of approximately 0.60, the system treats them as unconfirmed event chains and does not directly output warnings. In the second window The anomaly is still located in In this section, the average intensity increased to approximately 1.4, and the spatial length increased to approximately [missing value]. The propagation consistency was approximately 0.65, which the system determined to be a phase of continued attention and generated a blue alert. In the third window In the middle, the range of anomalies is from Expand to to The average strength is approximately 1.8, and the spatial length is approximately... The propagation consistency was approximately 0.72, and the system determined that it had entered the development stage and upgraded it to a yellow alert. In the fourth window The anomaly is still located in to The average strength is approximately 2.3, and the spatial length is approximately... The propagation consistency is approximately 0.76, and the system maintains a yellow alert level. In the fifth window In the middle, the anomaly further expanded to to The average strength is approximately 3.1, and the spatial length is approximately... The propagation consistency was approximately 0.81, which the system determined to be an acceleration phase and upgraded to an orange alert. In the sixth window In the middle, the anomaly intensity and spatial range continued to increase, with an average intensity of approximately 3.8 and a spatial length of approximately The propagation consistency was approximately 0.85, and the system was upgraded to a red alert. Then, in the seventh window In the middle, the average intensity decreased to about 2.6, and the spatial length shrank to about However, the propagation consistency remains at approximately 0.83, and the system determines that the event has entered a stable phase and maintains a red alert. Eighth window and the ninth window In the middle, the abnormal range shrank to to and The average intensity decreased to approximately 1.7 and 1.1 respectively, and the system gradually reduced the warning level after meeting the continuous retreat conditions; The tenth window When there is no longer a continuous event area, the event chain ends, and the system removes the warning.

[0175] As can be seen from this embodiment, the system does not directly output a high-level warning based on the peak intensity of a single window. Instead, it gradually increases the warning level only after the event area appears continuously, the spatial range expands reasonably along the structural segment, the propagation consistency meets the conditions, and the evolution stage gradually escalates. After the event intensity decreases and the spatial range shrinks, the system does not immediately lift the warning, but gradually lowers it after continuous backtracking confirmation.

[0176] Example 2: Suppression of vehicle movement vibration When a vehicle passes through a monitored section of road, a high-intensity event area moving along the road direction will be formed in the DAS data. The event chain management unit can associate vehicle responses in a continuous window into an event chain, but this event chain usually exhibits the following characteristics: rapid movement of the center position, relatively stable area length, high main frequency band, and movement speed close to the vehicle's speed (e.g., 16~33.3 m / s).

[0177] The system combines the aforementioned evolutionary pattern with structural propagation consistency verification. Although vehicle events have good accessibility on the road structure segment map, their center migration speed is significantly higher than the apparent expansion speed range (0.01~0.5 m / s) of progressive disasters on slopes or roadbeds, and the events do not exhibit a continuous expansion of their local area. Therefore, the system can maintain them in a state of transient anomaly or mobile disturbance, preventing them from entering the disaster development stage. After the vehicle leaves, the event chain ends quickly, without triggering a high-level warning.

[0178] Example 3: Suppression of construction vibration at a fixed location Road construction equipment may generate long-term periodic vibrations in a fixed location. These events are characterized by high persistence, but the spatial range is usually stable, the frequency band is periodic, and the intensity varies with the operating cycle of the equipment.

[0179] After establishing the event chain, the system found that the event was persistent but its spatial expansion rate was close to zero, and its propagation range did not increase to adjacent structural sections. In this case, the system will not determine it as a catastrophic event solely based on its long duration. If the event intensity and spatial range remain stable over a long period, it will remain in a state of concern or stability. By combining construction plans, video surveillance, or an optional interference fingerprint database, it can be marked as a construction event. Even without the assistance of external information, the lack of reasonable spatial expansion is sufficient to prevent the warning from automatically escalating to a higher level.

[0180] Example 4: Anomaly Monitoring in the Connection Zone between Bridge Abutment and Roadbed The connection point between the bridge abutment and the roadbed belongs to the structural transition section. The system records the connection methods between the roadbed section, the abutment section, and the bridge span section in the structural section diagram, and sets different apparent expansion speed range, frequency band variation rules, and boundary crossing conditions for the structural transition edge compared to ordinary roadbeds.

[0181] When an event extends from the roadbed towards the abutment, the system allows it to cross the structural transition boundary, but requires that the response sequence, apparent extension rate, and frequency band change satisfy the corresponding transition conditions.

[0182] If an anomaly jumps directly from the roadbed to the middle of an unconnected bridge span, the system generates a propagation conflict and does not connect the two into the same event chain. If the anomaly appears gradually along the roadbed → abutment → bridge span end, the system considers the propagation path reasonable and updates the evolution stage based on the event intensity and expansion speed. This method effectively avoids erroneous associations caused by adjacency based solely on fiber optic distance.

[0183] Example 5: Treatment of Background Lifting Due to Rainfall When widespread rainfall causes background vibration uplift, low-frequency energy may increase simultaneously in multiple spatial locations. In this case, the event center usually does not exhibit a reasonable local expansion process, or the event region may occur simultaneously in multiple unconnected segments.

[0184] The system reduces the probability of catastrophic escalation by analyzing structural propagation conflicts, abrupt changes in regional scope, and background correlations. If rainfall itself may trigger slope catastrophic events, the system continues to track event chains with local enhancement and structural expansion characteristics, and outputs corresponding warnings when these chains meet the conditions of persistence, propagation consistency, and development.

[0185] Example 6: Handling Equipment Malfunctions and Insufficient Data Quality When a DAS device experiences a sudden gain change, time synchronization anomaly, or partial channel failure, the system suspends the automatic upgrade of the corresponding event chain based on the large-scale synchronization changes and data quality flags.

[0186] Once the equipment returns to normal, the system can re-associate based on the structural location and event characteristics before and after the recovery, but this will reduce the continuity score across the fault period. If insufficient data quality makes it impossible to reliably determine propagation consistency, the system will maintain the most recent reliable warning level and output data quality prompts to avoid erroneous upgrades or deletions due to equipment anomalies.

[0187] Example 7: Suppression of Misassociations Caused by Fiber Optic Turnback or Excess Cable Near certain highway bridge abutments, tunnel entrances, or communication wells, sensing optical cables may be folded back, have excess cable coiled, or be laid across structures. This results in adjacent sampling points in terms of fiber distance corresponding to non-adjacent roadbed sections, bridge abutment sections, bridge span mid-sections, or trench sections in terms of engineering structure. In such cases, vehicle impacts, falling object collisions, or equipment noise may manifest as continuous or near-adjacent anomalies in the fiber coordinate system.

[0188] If event association is based solely on the fiber sampling point sequence, fiber distance, or the adjacency relationship of abnormal areas on fiber coordinates, the system may misjudge two abnormal areas that are not connected in the engineering as a continuous extension of the same disaster event.

[0189] In this embodiment, the system first transforms the anomalous region from fiber optic coordinates to highway structural segments using a mapping table between fiber optic spatial locations and highway structural segments. Subsequently, the event chain construction and management unit queries the structural segment map to determine whether there is an allowed extension path between the structural segment corresponding to the current anomalous region and the end segment of a historical event chain. If two anomalous regions are adjacent in fiber optic coordinates, but their corresponding structural segments do not have a structural reachability relationship, or require crossing a boundary marked with a blocking attribute, the system refuses to directly associate them as the same event chain, or marks the candidate association as a propagation conflict.

[0190] For example, a fiber optic cable turns back near the bridge abutment, and the fiber optic distance... to The corresponding roadbed section in front of the bridge abutment, and the adjacent fiber optic cable distance. to This corresponds to the remaining cable in the middle of the bridge span. When high-amplitude anomalies appear in two locations in the current window, the system does not directly identify it as a catastrophic spread from the roadbed to the bridge span simply because the fiber optic cables are adjacent. Instead, it confirms, based on the structural segment diagram, that there is no reasonable propagation path between the two locations via the abutment and the end of the bridge span, thus stopping automatic escalation and outputting a verification prompt. If the anomaly appears sequentially in subsequent windows along the roadbed segment → abutment segment → end of the bridge span, and the response order, apparent spread rate, and frequency band change meet the conditions of the structural transition edge, the system allows it to be restored to the same event chain and outputs a warning level according to the evolution stage.

[0191] This implementation method avoids spatial misassociations caused by fiber optic cable reversals, cable coiling, or cross-structure laying. This embodiment demonstrates that structural segment diagrams and propagation consistency verification are not merely supplementary display information, but rather key technical means to correct misjudgments caused by inconsistencies between DAS fiber optic coordinates and the relationship between the coordinates and the highway engineering structure.

[0192] Example 8: Comparative Experiment with Existing Single-Window Thresholding Methods To verify the technical effectiveness of this invention, actual DAS monitoring data from a section of a mountainous highway slope and roadbed over 30 consecutive days were used for playback testing. During this period, on-site verification revealed the following: two early minor slope slippage events, approximately 150 instances of strong interference from heavy vehicles, and five instances of strong vibrations at the fixed positions of construction machinery.

[0193] The testing method involved inputting the same raw DAS data into both the existing single-window energy threshold method and the method of this invention, and comparing the monitoring effects of the two methods. The comparison results are shown in Table 1 below: Table 1 The comparative experiments above show that existing single-window methods, by directly equating the detection results of a single time window with catastrophic conclusions, are prone to misreporting non-catastrophic strong interferences such as vehicle traffic and construction as high-level hazards. This invention, through the evolutionary characteristics of the event chain, such as the center migration speed and spatial expansion rate, combined with structural propagation consistency verification, successfully filters out all high-level false alarms caused by vehicles and construction. Simultaneously, relying on the hysteresis control logic of the early warning state machine, it effectively eliminates frequent jumps in early warning levels near the threshold, significantly improving the early warning accuracy of the highway DAS catastrophic monitoring system.

[0194] The preferred embodiments described above employ a rule-based processing flow and can be implemented without relying on a trained model. In other embodiments, event region generation can utilize convolutional neural networks, spatiotemporal object detection networks, clustering algorithms, or other methods capable of outputting anomalous spatiotemporal regions and their characteristics. Matching event regions with event chains can also employ probabilistic data association, multiple hypothesis tracking, graph matching, or trained association models. Structural propagation relationships can be represented as explicit graph structures, segment adjacency lists, reachability matrices, rule bases, or transition probability matrices learned from historical data. Evolutionary stage determination can employ deterministic state machines, or hidden Markov models, conditional random fields, recurrent neural networks, temporal convolutional networks, or other sequential decision models.

[0195] Regardless of the alternative algorithm used, the three core processing steps—event chain, structural propagation consistency verification, and early warning state machine—must be retained. The alternative algorithm is only used to implement the specific methods of event region generation, association cost calculation, propagation consistency scoring, or stage determination; it does not change the technical essence of this invention, which uses continuous evolution of the event chain and structural propagation consistency to jointly drive hierarchical early warning.

[0196] The system can operate entirely on field edge devices, or field devices can perform data preprocessing and event region generation, and then transmit the event regions, frequency band characteristics, and necessary data fragments to the central platform for event chain management and early warning. To reduce communication load, field devices do not need to upload all raw DAS data; they can upload only the event regions, event chain status, and necessary event data fragments.

[0197] Further reference Figure 7 As an implementation of the methods shown in the above figures, this application provides an embodiment of a highway DAS disaster evolution monitoring and hierarchical early warning system. This system embodiment is similar to... Figure 1 and Figure 2 Corresponding to the method embodiments shown, the system can be specifically applied to various electronic devices.

[0198] Secondly, embodiments of the present invention also disclose a highway DAS disaster evolution monitoring and graded early warning system, such as... Figure 7As shown, the system of the present invention is divided into eight functional units according to the processing flow: DAS data acquisition and spatial calibration unit 71, window event area generation unit 72, event chain construction and management unit 73, event chain evolution feature calculation unit 74, structure propagation relationship and consistency verification unit 75, disaster evolution stage determination unit 76, early warning status control unit 77, and early warning result and historical record unit 78.

[0199] In one specific embodiment, the DAS data acquisition and spatial calibration unit 71 is used to continuously acquire DAS vibration data along the highway and establish a mapping relationship between the optical fiber spatial sampling position and the highway structural section; the DAS data acquisition and spatial calibration unit 71 is connected to the DAS demodulation equipment to receive continuous vibration data and save the correspondence between optical fiber distance, highway mileage, structural type, structural section number and optical fiber coupling state.

[0200] The window event region generation unit 72 is used to process DAS vibration data using a sliding time window. In each window, it generates one or more event regions based on the abnormal response map. The window event region generation unit 72 extracts the current sliding window from the continuous data stream and completes anomaly point processing, background suppression, frequency band decomposition, amplitude correction, and connected region identification.

[0201] The event chain construction and management unit 73 is used to associate the event region of the current window with the event chain of historical activities across windows, convert the event region to structural coordinates according to the mapping relationship, determine the reachable path and calculate the association cost based on the structural segment graph, and add the event region to the existing event chain, create a new event chain, or keep it short-term missing according to the association result. The event chain construction and management unit 73 receives the event region of the current window and the event chain of historical activities, and completes the matching, creation, short-term missing retention, restoration, splitting, merging and termination operations between the event region and the event chain according to the structural reachability and association cost.

[0202] The event chain evolution feature calculation unit 74 is used to calculate the evolution features of the event chain; the event chain evolution feature calculation unit 74 calculates the event persistence, enhancement, expansion, migration, frequency band change and decay features based on the attribute changes of continuous event regions in the event chain.

[0203] The structural propagation relationship and consistency verification unit 75 is used to establish a structural segment diagram with highway structural segments as nodes and engineering connections or permissible influence relationships between segments as edges, and to score the propagation consistency of the event chain; the structural propagation relationship and consistency verification unit 75 establishes a highway structural segment diagram and judges whether the spatial expansion and migration in the event chain are reasonable.

[0204] The catastrophe evolution stage determination unit 76 is used to determine the evolution stage of the event chain based on evolution characteristics and propagation consistency scores; the catastrophe evolution stage determination unit 76 updates the event status based on evolution characteristics and propagation consistency.

[0205] The early warning status control unit 77 is used to control the early warning level through the early warning state machine according to the evolution stage and its holding time. The upgrade conditions and rollback conditions in the early warning state machine adopt different duration window number thresholds. The rollback confirmation window number is not less than the upgrade confirmation window number. The early warning status control unit 77 controls the generation, upgrade, holding, rollback, cancellation or suspension of automatic upgrade of the early warning level according to the event status and its duration.

[0206] The early warning result and historical record unit 78 is used to output early warning results and save the evolution process record of the event chain. The early warning result and historical record unit 78 sends the event location, scope of impact, evolution stage, propagation consistency, early warning level and reasons for level changes to the monitoring platform, and saves the complete evolution process of the event chain.

[0207] All the aforementioned units can be physically deployed within the same field edge computing device, or they can be deployed in a distributed manner: the field device completes DAS data acquisition and event region generation, and then transmits the event region data to the central platform via a communication network. The central platform then manages the event chain, verifies propagation consistency, and outputs early warnings. Distributed deployment reduces the computational load and communication bandwidth requirements of the field devices—the field devices do not need to upload all the original DAS data, only the event region, event chain status, and necessary event data fragments.

[0208] It is important to emphasize that the units in the system are not independent parallel modules, but rather organic components constituting the complete technical processing chain of this invention. The event region of the current window can only form a continuously evolving object after being managed by the event chain; the evolutionary characteristics of the event chain can only distinguish between true expansion and false expansion when combined with structural propagation consistency; the early warning state control does not directly use the current anomaly score, but rather uses an evolutionary stage that has undergone dual verification of continuity and structural rationality. The event chain, propagation consistency verification, and early warning state machine together constitute the core technical processing chain of this invention.

[0209] The functions and methods of the above modules correspond to each other, and will not be repeated here.

[0210] This invention relates to the fields of highway infrastructure safety monitoring, distributed fiber optic acoustic sensing, vibration signal processing, structural condition identification, and disaster early warning technology. Specifically, it relates to a method and system that utilizes distributed fiber optic acoustic sensing data along a highway to correlate abnormal vibration areas that occur during continuous monitoring across time windows, and combines this with the engineering connection relationships between highway structural sections, permissible expansion directions, apparent expansion rates, and structural boundary conditions to continuously determine the development process of abnormal events, thereby realizing the generation, maintenance, upgrading, and rollback of disaster early warning levels.

[0211] This invention is primarily applicable to the monitoring of disasters with time-evolutionary processes, such as highway slope slippage, roadbed settlement, progressive deformation of roadbed slopes, and anomalies in the connection zone between bridge abutments and roadbed. After establishing the corresponding structural segment relationships and propagation parameters, this invention can also be used for continuous monitoring of abnormal bridge vibrations, tunnel anomalies, culvert anomalies, retaining structure anomalies, rockfall impacts, and other abnormal events in highway infrastructure. This invention can also suppress or verify non-catastrophic vibrations such as vehicle traffic, road construction, machinery operation, background uplift from heavy rainfall, wind vibration, and accidental collisions, thereby reducing false alarms and frequent jumps in warning levels.

[0212] It should be noted that for slow-moving disasters such as roadbed settlement, progressive slope slippage, and loosening of bridge abutment connection zones, this invention does not require DAS to directly measure the static displacement of the structure. Instead, it utilizes dynamic and quasi-dynamic responses generated or altered during the disaster evolution process for identification. These dynamic or quasi-dynamic responses include microcrack propagation, particle friction slippage, changes in structural contact state, changes in dynamic response under vehicle traffic excitation, enhanced local low-frequency response caused by rainwater infiltration or groundwater action, and changes in vibration frequency band, duration, spatial influence range, or response center location resulting from the above factors. This invention is applicable to highway disaster evolution scenarios where detectable changes in vibration response, strain rate response, or vehicle excitation response can be generated in DAS data.

[0213] The highway DAS disaster evolution monitoring and graded early warning method and system provided by this invention can be widely applied to the continuous monitoring of highway slope slippage, roadbed settlement, progressive deformation of roadbed slopes, anomalies in the abutment-roadbed connection zone, abnormal bridge vibration, tunnel anomalies, culvert anomalies, retaining structure anomalies, rockfall impacts, and other abnormal events in highway infrastructure. This invention can be implemented using existing communication optical cables or specially laid sensing optical fibers, without requiring large-scale modifications to existing highway facilities. It has good economic efficiency and scalability, and is suitable for safety monitoring and disaster early warning scenarios for various types of highway infrastructure.

[0214] The above description is merely a preferred embodiment of the present invention and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention is not limited to the specific combination of the above-described technical features, but also includes other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in this invention.

Claims

1. A method for monitoring and hierarchical early warning of highway DAS disaster evolution, characterized in that, The method includes the following steps: Continuously collect DAS vibration data along the highway and establish a mapping relationship between the optical fiber spatial sampling location and the highway structural section; The DAS vibration data is processed using a sliding time window, and one or more event regions are generated within each window based on the abnormal response map. Cross-window association between the event region of the current window and the historical event chain: the event region is converted to structural coordinates according to the mapping relationship, the reachable path is determined based on the structural segment diagram, the association cost is calculated for the combination with reachable path, and the current event region is added to the corresponding event chain when the association cost meets the preset association conditions, and a new candidate event chain is established when the association conditions are not met. Calculate the evolutionary characteristics of the established event chain; The propagation consistency of the event chain is verified using the aforementioned structural segment diagram, and a propagation consistency score is calculated. The evolutionary stages of the event chain are determined based on the aforementioned evolutionary characteristics and propagation consistency scores; Based on the evolution stage and its duration, the early warning level is controlled by an early warning state machine. The upgrade and rollback conditions in the early warning state machine adopt different duration window number thresholds, and the rollback confirmation window number is not less than the upgrade confirmation window number. Output early warning results and save the evolution process record of the event chain.

2. The method according to claim 1, characterized in that, The event region includes at least the window number, the start and end times within the window, the start and end times of the structural coordinates, the corresponding set of structural segments, peak intensity, average intensity, main frequency bands, frequency band energy ratio, regional credibility, and data quality markers; the event chain stores at least the creation time, the most recent update time, the cumulative number of missing windows, the current structural segment range, evolution characteristics, propagation consistency score, the current evolution stage, and the current warning level.

3. The method according to claim 1, characterized in that, In the cross-window association, the current event area Chain of historical events Related costs Represented as: in, Due to differences in spatial location, Due to regional boundary differences, Due to frequency band differences, Differences in intensity and trend, Differences in the structural segment relationships; , , , , They are respectively , , , , The corresponding weighting coefficients; when If the current event region is less than or equal to the preset association threshold and there is no better matching conflict, the current event region will be added to the corresponding event chain.

4. The method according to claim 3, characterized in that, The spatial location difference The distance between the current region center coordinates and the event chain prediction center coordinates is obtained by normalization; the region boundary difference is... It is obtained from the ratio of the length of the overlapping interval between the current region and the end region of the event chain; The frequency band difference The difference in intensity and trend is obtained by calculating the Euclidean distance or cosine similarity of the energy distribution vectors of the two main frequency bands. The difference is obtained from the relative rate of change of the average or peak intensity of the two; the structural segment relationship is different. The weight is obtained based on the structural distance penalty value between corresponding structural segments of the two regions; , , , , The values ​​are 0.25, 0.20, 0.15, 0.20, and 0.20 respectively, and the preset association threshold is 0.

6.

5. The method according to claim 1, characterized in that, The propagation consistency verification is based on at least several of the following criteria: structural reachability consistency, response order consistency, apparent expansion speed consistency, structural boundary consistency, and frequency band variation consistency. Represented as: in, To ensure structural reachability consistency, a score is awarded based on whether a reachable path exists in the structural segment graph. To ensure consistency in response order, a score is awarded based on whether the spatial order in which event regions appear aligns with the allowed expansion direction of the structure diagram. To ensure consistent apparent expansion speed, a score is awarded based on whether the migration speed of the event chain falls within a preset speed range. To ensure structural boundary consistency, a score is awarded based on whether preset transition conditions are met when an event crosses the structural boundary. To ensure consistency in frequency band variations, a score is awarded based on whether the frequency band energy variation patterns before and after an abnormal event crosses different structural media conform to engineering experience; weighting. , , , , The values ​​are 0.30, 0.20, 0.20, 0.20, and 0.10, respectively.

6. The method according to claim 5, characterized in that, when When it is determined to be a propagation conflict, when When it is determined that the credibility of the dissemination is low, When it is determined to be effective transmission, The time frame indicates that the propagation exhibits a high degree of consistency. In the event of a propagation conflict, the system does not escalate to the disaster phase, but outputs the structural segment where the conflict occurred and the cause of the conflict.

7. The method according to claim 1, characterized in that, The evolutionary characteristics include at least one of event persistence, vibration intensity growth rate, spatial expansion rate, center migration speed, frequency band change, and event chain stability; the evolutionary stages include pending confirmation, transient anomaly, continuous monitoring, development, acceleration, stabilization, decay, and propagation conflict; the early warning state machine controls the generation, escalation, maintenance, rollback, or cancellation of the early warning level.

8. The method according to claim 1, characterized in that, The construction of the structural segment diagram includes the following steps: Based on highway design data and geographic information systems, the highway route is divided into independent physical monitoring units as nodes; Based on the on-site investigation and geographic information topology, the engineering connection relationships between each node are established as edges, and the edges are marked with blocking attributes at locations that cross empty areas or isolation zones. Based on the optical cable laying records, a mapping table between optical fiber coordinates and structural nodes was established by manual tapping calibration, known equipment excitation, or vehicle positioning calibration. Based on the cross-segment attenuation characteristics of vibration signals obtained from field tests, the edge attributes of the structural segment diagram are verified and corrected.

9. The method according to claim 1, characterized in that, The warning levels are divided into four levels: blue, yellow, orange, and red. Upgrades and downgrades are performed step by step. Upgrading from blue to yellow requires at least two windows of development and a propagation consistency of no less than 0.

6. Upgrading from yellow to orange requires entering an acceleration phase or continuing to develop within a high-risk area and a propagation consistency of no less than 0.

75. Upgrading from orange to red requires rapid expansion of the high-risk area or rapid growth in both intensity and spatial range and a propagation consistency of no less than 0.

75. The number of confirmation windows for downgrading from yellow is three, from orange it is four, and from red it is five.

10. A highway DAS disaster evolution monitoring and hierarchical early warning system, characterized in that, include: The DAS data acquisition and spatial calibration unit is used to continuously acquire DAS vibration data along the highway and establish a mapping relationship between the optical fiber spatial sampling position and the highway structural section. The window event region generation unit is used to process the DAS vibration data using a sliding time window, and generate one or more event regions in each window based on the abnormal response map. The event chain construction and management unit is used to associate the event area of ​​the current window with the historical event chain across windows, convert the event area to structural coordinates according to the mapping relationship, determine the reachable path and calculate the association cost based on the structural segment diagram, and add the event area to the existing event chain, create a new event chain, or keep it missing for a short time according to the association result. The event chain evolution feature calculation unit is used to calculate the evolution features of the event chain; The structural propagation relationship and consistency verification unit is used to establish a structural segment diagram with highway structural segments as nodes and engineering connections or permissible influence relationships between segments as edges, and to score the propagation consistency of the event chain. The disaster evolution stage determination unit is used to determine the evolution stage of the event chain based on the evolution characteristics and propagation consistency score. The early warning status control unit is used to control the early warning level through an early warning state machine according to the evolution stage and its duration. The upgrade condition and rollback condition in the early warning state machine adopt different duration window number thresholds, and the rollback confirmation window number is not less than the upgrade confirmation window number. The warning results and historical record unit is used to output warning results and save the evolution process record of the event chain.

Images