Tunnel structure integrated monitoring system based on distributed and point sensor fusion

Abstract

The application relates to the technical field of tunnel structure variable monitoring, and discloses a tunnel structure integrated monitoring system based on distributed and point type sensing fusion, which comprises an abnormal space fingerprint generation module, an expected response generation module, a response relationship judgment module and an abnormal section correction module. According to residual strain after temperature correction, an initial abnormal section with left and right boundaries and an abnormal peak position is determined, and an abnormal space fingerprint is generated. An expected response state is generated in combination with sensor monitoring attributes. A boundary contraction correction interval is formed according to response state conversion inside the boundary, a boundary expansion correction interval is formed according to time correlation of response outside the boundary and distributed abnormal space continuity, and a correction boundary is determined by using continuously distributed residual strain and gradient attenuation state thereof in the correction interval. The application can improve tunnel abnormal section positioning accuracy and reduce boundary misjudgment caused by occasional non-response and local independent abnormality.

Description

Technical Field

[0001] This invention relates to the field of tunnel structure variable monitoring technology, and in particular to an integrated tunnel structure monitoring system based on the fusion of distributed and point-based sensing. Background Technology

[0002] As a crucial link between highways, the structural safety of tunnels plays a vital role in ensuring highway traffic safety and smooth flow. However, tunnel structures have a service life of decades or even centuries. The combined effects of environmental erosion, material aging, and the long-term effects of loads inevitably lead to the accumulation of damage and weakening of the structure and system, potentially even triggering catastrophic accidents in extreme cases. The main defects in the lining of civil engineering structures include lining cracks, water leakage, exposed reinforcement, voids, misalignment, damage, and concrete spalling. Therefore, to ensure the safety, integrity, and serviceability of tunnel structures during operation, it is urgent to employ effective technical means to accurately diagnose their long-term performance changes and effectively assess their operational status, thereby guaranteeing safe structural operation.

[0003] In existing tunnel structure monitoring, although distributed fiber optic sensing can continuously acquire strain anomaly information distributed along the tunnel mileage, its anomaly boundaries and centers are easily affected by factors such as temperature changes, local measurement fluctuations, and insignificant spatial attenuation characteristics. Relying solely on distributed strain results can easily lead to the expansion or contraction of the anomaly range or location deviation. At the same time, existing point sensors are limited by their deployment locations, different monitoring objects, and different measurement directions, and are usually only used for local auxiliary verification. It is difficult to establish a clear spatial and temporal correspondence with the distributed monitoring results, and they cannot effectively participate in the correction of anomaly sections. Summary of the Invention

[0004] The technical problem to be solved by this invention is that the anomaly range identified by distributed optical fiber in existing tunnel monitoring is easily interfered with and the positioning is inaccurate, and point-based sensing data is difficult to effectively participate in the correction of anomaly boundaries and anomaly centers. To this end, we propose an integrated tunnel structure monitoring system based on the fusion of distributed and point sensing.

[0005] To achieve the above objectives, this application adopts the following technical solution: an integrated monitoring system for tunnel structures based on the fusion of distributed and point-based sensing, comprising: an anomaly space fingerprint generation module. The initial anomalous section with left and right boundaries and anomalous peak positions is determined based on the residual strain after temperature correction, and an anomalous spatial fingerprint is generated. Expected Response Generation Module Using the left and right boundaries as the boundaries to be corrected respectively, the expected response state is generated based on the abnormal space fingerprint and sensor monitoring attributes, and the high expected response sensor is identified. Response Relationship Determination Module Based on the comparison results of the measured response of the high expected response sensor with the corresponding expected response state in multiple consecutive diagnostic time windows, the matched response sensor and the continuously missing response sensor are determined. Abnormal section correction module Along the direction from the abnormal peak position to the boundary to be corrected, a boundary contraction correction interval is formed based on the spatial transition relationship between the matching response sensor inside the boundary and the persistently missing response sensor; based on the temporal correlation between the high expected response sensor outside the boundary that continuously meets the expected response state and the matching response sensor inside the boundary to be corrected, the matching response sensor outside the boundary is determined; when the distributed anomaly between the boundary to be corrected and the matching response sensor outside the boundary maintains spatial continuity, a boundary expansion correction interval is formed by combining it with the persistently missing response sensor outside the boundary. When the same boundary to be corrected forms only one of the two correction intervals mentioned above, the correction boundary is determined within that correction interval based on the continuously distributed residual strain after temperature correction and its gradient decay state; otherwise, the original position of the boundary to be corrected is retained.

[0006] Preferably, it also includes a sensor monitoring and management module, which establishes sensor monitoring attribute records for each point sensor. The sensor monitoring attributes include structural location, measurement direction and monitoring object, and further record the installation mileage and the cross section where it is located.

[0007] Preferably, the expected response generation module determines point sensors in the inner and outer neighborhoods of each boundary to be corrected; the expected response state includes the expected response direction, the expected response time window, and the minimum effective response level. When the measured response direction matches the expected response direction, the measured response is within the expected response time window, and the measured response intensity reaches the minimum effective response level, it is determined to meet the expected response state; the high expected response sensor is a point sensor that is in an effective measurement state and the expected response state indicates that a response should be given.

[0008] Preferably, the response relationship determination module identifies high-expected-response sensors that meet the expected response state in multiple consecutive diagnostic time windows as matched response sensors, and identifies high-expected-response sensors that are continuously in an effective measurement state in multiple consecutive diagnostic time windows and whose measured response intensity does not reach the minimum effective response level as continuously missing response sensors; when the measured response intensity of a high-expected-response sensor reaches the minimum effective response level but the measured response direction is opposite to the expected response direction, the corresponding measured response is determined as an opposite-direction response.

[0009] Preferably, the abnormal segment correction module, along the direction from the abnormal peak position to the boundary to be corrected, determines the mileage range between the sensors on both sides of the transition as the boundary contraction correction interval when the first transition from the matched response sensor to the continuously missing response sensor occurs; it also determines the high expected response sensor outside the boundary to be corrected that continuously meets the expected response state and whose response satisfies the time correlation with the matched response sensor inside the boundary to be corrected as the boundary-outside matched response sensor; and when the distributed anomalies between the boundary to be corrected and the boundary-outside matched response sensor remain spatially continuous, the mileage range between the outermost boundary-outside matched response sensor and its nearest continuously missing response sensor is determined as the boundary expansion correction interval.

[0010] Preferably, the abnormal section correction module starts from the side of the formed boundary contraction correction interval or boundary expansion correction interval closest to the abnormal peak position, and sequentially checks the distributed optical fiber measuring points in the direction pointing outward of the abnormal section; along the direction pointing outward of the abnormal section, when the residual strain anomaly degree of a certain distributed optical fiber measuring point and multiple consecutive distributed optical fiber measuring points outside it continuously decays to below the effective level of the anomaly, and the absolute value of the residual strain gradient of each distributed optical fiber measuring point is successively less than the absolute value of the residual strain gradient of the adjacent distributed optical fiber measuring point on its peak side, the mileage position corresponding to the first distributed optical fiber measuring point that meets the above conditions along the inspection direction is determined as the correction boundary.

[0011] Preferably, the system further includes a distributed optical fiber sensing module, a point sensor module, and a data acquisition and processing module; the distributed optical fiber sensing module continuously acquires raw strain field data and temperature field data corresponding to the tunnel mileage location along the tunnel axis; the point sensor module acquires local structural response data of key tunnel sections, key structural parts, or existing defects; the data acquisition and processing module adds sampling time and corresponding spatial location information to the raw strain field data, temperature field data, and local structural response data, so that the distributed optical fiber measurement results and the point sensor measurement results are compared according to the corresponding spatial location and diagnostic time window.

[0012] Preferably, the anomaly spatial fingerprint generation module includes a temperature correction unit, an initial anomaly segment identification unit, and a fingerprint generation unit; the temperature correction unit performs temperature correction on the original strain field data according to the temperature change corresponding to the distributed optical fiber measurement location to obtain the temperature-corrected residual strain; the initial anomaly segment identification unit determines the mileage range covered by the corresponding distributed optical fiber measurement point as the initial anomaly segment when the residual strain anomaly degree of multiple consecutive distributed optical fiber measurement points reaches the effective level of anomaly and maintains spatial continuity; the anomaly spatial fingerprint includes at least the anomaly peak position, left boundary, right boundary, left strain gradient attenuation rate, right strain gradient attenuation rate, anomaly duration, anomaly change direction, and the structural part to which the anomaly belongs.

[0013] Preferably, the system further includes an anomaly center correction module. This module determines the location of the distributed anomaly peak based on the degree of residual strain anomaly at the distributed fiber optic measuring points within the corrected anomaly section. It also determines the normalized measured response intensity based on the degree to which the measured response intensity of each matched response sensor exceeds the corresponding lowest effective response level. The cross-section where the matched response sensor with the highest normalized measured response intensity is located is determined as the point-type response peak cross-section. When a stable spatial offset is formed between the point-type response peak cross-section and the location of the distributed anomaly peak, and the corresponding monitored object has a direct correspondence with the structural response reflected by the current anomaly spatial fingerprint, the center correction search range is determined by the point-type response peak cross-section, and the mileage location corresponding to the local maximum value of the residual strain anomaly within this range is determined as the corrected anomaly center.

[0014] Preferably, it also includes an anomaly type discrimination module, which determines the anomaly type based on the corrected anomaly segment, the corrected anomaly center, the anomaly spatial fingerprint, and the measured response state of the point sensor. The anomaly type includes structural deformation anomaly, crack propagation anomaly, and transient vibration disturbance. The anomaly type discrimination module is also used to verify the opposite direction response. When the opposite direction response still cannot be explained after verification of the sensor installation direction and measurement direction, the output of the deterministic anomaly type is paused and a measurement direction conflict mark is generated.

[0015] The technical effects and advantages of this invention are as follows: This invention corrects the strain data obtained from distributed optical fibers by applying temperature to form an initial anomalous section with left and right boundaries and anomalous peak positions. It then combines sensor monitoring attributes to generate a predicted response state. Only point sensors with clear response qualifications and in an effective measurement state are selected for anomalous section verification. A boundary contraction correction interval is formed using the spatial transformation relationship between the inner boundary matching response and the continuously missing response. A boundary expansion correction interval is formed using the temporal correlation of responses inside and outside the boundary and the spatial continuity of distributed anomalous structures. Subsequently, within the formed correction interval, the specific correction boundary is determined based on the continuously distributed temperature-corrected residual strain and its gradient decay state. Thus, the point sensing results are used to determine the correction direction of the anomalous boundary and define the possible range of the boundary, while the distributed optical fiber results are used to determine the actual boundary position in continuous space. This improves the accuracy of anomalous section identification even with limited point sensor deployment, reduces erroneous contraction caused by occasional non-response of a single sensor and erroneous expansion caused by independent local anomalies outside the boundary, and enhances adaptability to asymmetrical distribution of tunnel anomalies through independent correction of the left and right boundaries. Attached Figure Description

[0016] The disclosure of this invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. In the drawings, the same reference numerals are used to refer to the same parts: Figure 1 This is a schematic diagram of the overall workflow of the system of the present invention; Figure 2 This is a schematic diagram of the spatial relationship between the initial abnormal section and the boundary neighborhood of the present invention; Figure 3 This is a schematic diagram illustrating the formation principle of the boundary contraction correction interval of the present invention; Figure 4 This is a schematic diagram illustrating the formation principle of the boundary extension correction interval of the present invention; Figure 5 This is a schematic diagram of the planar arrangement of the tunnel structure monitoring sensors of the present invention. Detailed Implementation

[0017] It is readily understood that, based on the technical solution of this invention, those skilled in the art can propose various interchangeable structural methods and implementations without altering the essential spirit of the invention. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this invention and should not be considered as the entirety of the invention or as limitations or restrictions on the technical solution of this invention.

[0018] like Figures 1 to 5 As shown, this embodiment provides an integrated monitoring system for tunnel structures based on the fusion of distributed and point-based sensing, including a distributed optical fiber sensing module, a point-based sensing module, a data acquisition and processing module, a sensor monitoring and management module, an anomaly spatial fingerprint generation module, an expected response generation module, a measured response extraction module, a response relationship determination module, an anomaly section correction module, an anomaly center correction module, an anomaly type discrimination module, and a status level output module.

[0019] The distributed fiber optic sensing module includes a strain measurement unit and a temperature measurement unit. The strain measurement unit continuously acquires raw strain field data corresponding to the tunnel mileage location along the tunnel axis, while the temperature measurement unit continuously acquires temperature field data corresponding to the raw strain field data. Specifically, the strain measurement unit can use a distributed strain sensing fiber optic cable continuously deployed along the sidewalls, arch waist, or other parts of the structure to be monitored, and the temperature measurement unit can use a temperature sensing fiber optic cable deployed corresponding to the distributed strain sensing fiber optic cable. By assigning corresponding mileage locations to the raw strain field data and temperature field data, the temperature influence at each distributed fiber optic measurement location can be determined during subsequent temperature correction.

[0020] Point-type sensing modules are deployed at key sections, critical structural parts, or existing defects in the tunnel, and include one or more of the following: section convergence monitoring units, displacement monitoring units, crack monitoring units, vibration monitoring units, and water leakage monitoring units. The section convergence monitoring unit includes a section convergence sensor to acquire changes in tunnel section convergence; the displacement monitoring unit includes a laser displacement meter to acquire changes in crown settlement or local displacement; the crack monitoring unit includes a crack sensor to acquire changes in crack width; the vibration monitoring unit includes a vibration acceleration sensor to acquire transient dynamic response; and the water leakage monitoring unit includes a water leakage monitoring sensor to acquire water leakage triggering status and leakage location.

[0021] The data acquisition and processing module includes a data receiving unit and a spatiotemporal reference unification unit. The data receiving unit receives measurement data output from the distributed fiber optic sensing module and the point sensor module, respectively. The spatiotemporal reference unification unit adds sampling time, sensor identification, installation mileage, cross-section and structural location information to each measurement data, enabling the distributed fiber optic measurement results and the point sensor measurement results to be compared according to the corresponding spatial location and diagnostic time window.

[0022] A diagnostic time window refers to the time range used to determine the responses of distributed fiber optic devices and point sensors within the same stage of an anomaly development. Diagnostic time windows can be continuously divided according to a fixed time length, or they can be established using the occurrence time of the distributed anomaly as a time reference. Multiple consecutive diagnostic time windows can be consecutive or partially overlapping, but the same time window division method is used in the same anomaly determination process to ensure comparability of response states across different diagnostic time windows.

[0023] The sensor monitoring and management module includes an attribute recording unit and a sensor status determination unit, and manages the sensor monitoring attribute library. The attribute recording unit establishes a corresponding sensor monitoring attribute record for each point sensor and stores this record in the sensor monitoring attribute library. Sensor monitoring attributes include at least installation mileage, location section, structural component, measurement direction, and monitoring object. The structural component can include at least one of the following: arch crown, arch waist, sidewall, invert arch, construction joint, and existing crack location. The measurement direction indicates the structural response direction that the point sensor can directly perceive, and can include at least one of the following: lateral convergence direction, vertical displacement direction, crack opening direction, and vibration measurement direction. The monitoring object can include at least one of the following: cross-sectional convergence, local displacement, crack width, vibration acceleration, and water leakage status.

[0024] The sensor status determination unit is used to determine whether the point sensor is in an effective measurement state. An effective measurement state refers to a state where the point sensor is not experiencing continuous communication interruptions, data freezes, prolonged saturation of measurement values, or significantly exceeding the device's measurement range. The sensor monitoring attribute record may also include the effective influence range of the point sensor, which defines the adjacent area that the point sensor can serve as a basis for spatial verification. Through this method, whether a point sensor can participate in the boundary correction of the current anomaly is no longer determined solely by its distance from the anomaly section. It also requires consideration of its structural location, measurement direction, monitored object, and effective measurement state to determine whether the point sensor can reflect the current anomaly, thereby avoiding the use of point sensors with no clear response relationship to the current anomaly as the basis for boundary correction.

[0025] The anomaly spatial fingerprint generation module includes a temperature correction unit, an initial anomaly segment identification unit, and a fingerprint generation unit. The temperature correction unit corrects the original strain field data based on the temperature changes corresponding to each distributed fiber measurement location, obtaining the temperature-corrected residual strain. The temperature-corrected residual strain refers to the remaining strain component after subtracting the temperature strain component determined according to the corresponding temperature change from the original strain change. It is used to characterize structural strain changes that cannot be directly explained by the current temperature change. In this embodiment, the degree of residual strain anomaly refers to the degree of deviation of the temperature-corrected residual strain from the reference residual strain at the corresponding mileage location. In one implementation, the degree of residual strain anomaly can be determined based on the absolute value of the difference between the temperature-corrected residual strain and the corresponding reference residual strain.

[0026] The temperature-corrected residual strain is determined by subtracting the temperature strain component generated by the corresponding temperature change from the original strain change. The temperature strain component can be obtained based on the temperature change and the temperature correction coefficient determined by the sensor calibration results or the historical stable operation data of the tunnel.

[0027] The initial anomaly section identification unit determines the initial anomaly section based on the temperature-corrected residual strain continuously distributed along the tunnel mileage. Specifically, when the residual strain anomaly level of multiple consecutive distributed optical fiber measuring points reaches the anomaly effective level and remains spatially continuous within at least one diagnostic time window, the mileage range covered by the corresponding consecutive distributed optical fiber measuring points is determined as the initial anomaly section. The anomaly effective level referred to in this embodiment is the criterion used to distinguish the anomaly level of residual strain from normal background fluctuations. The anomaly effective level can be determined based on the residual strain fluctuation range during historical stable operation phases or a preset anomaly threshold.

[0028] The initial anomalous section has a left boundary, a right boundary, and an anomalous peak position. To ensure clear directional relationships during subsequent boundary correction, this embodiment specifies that, along the direction of increasing tunnel mileage, the starting point of the initial anomalous section is the left boundary, the ending point is the right boundary, and the anomalous peak position is located between the left and right boundaries. For the left boundary, the direction towards the anomalous peak position is the inward direction of the left boundary, and the direction away from the anomalous peak position is the outward direction of the left boundary; for the right boundary, the direction towards the anomalous peak position is the inward direction of the right boundary, and the direction away from the anomalous peak position is the outward direction of the right boundary. Therefore, the outward direction of the left boundary corresponds to the direction of decreasing tunnel mileage, and the outward direction of the right boundary corresponds to the direction of increasing tunnel mileage. Subsequent contractions towards the anomalous peak position or expansions towards the outward direction of the anomalous section are determined according to the above directional relationships.

[0029] The fingerprint generation unit generates an anomaly spatial fingerprint based on the initial anomaly segment. The anomaly spatial fingerprint includes at least the anomaly peak location, left boundary, right boundary, left strain gradient decay rate, right strain gradient decay rate, anomaly duration, anomaly change direction, and the structural location to which the anomaly belongs, and may further include temperature-corrected residual strain.

[0030] In one implementation, distributed fiber optic measuring points can be arranged sequentially from the location of the abnormal peak towards the outer edge of the corresponding boundary. The residual strain gradient is determined based on the residual strain difference between two adjacent distributed fiber optic measuring points and the corresponding mileage distance. Then, the strain gradient attenuation rate on the corresponding side is determined based on the degree of change in the absolute value of the adjacent residual strain gradients. Thus, the same calculation method can be applied to both the left and right boundaries without changing the subsequent judgment logic due to the different mileage directions of the left and right boundaries.

[0031] The expected response generation module includes a boundary neighborhood establishment unit, a candidate point sensor determination unit, an expected response state generation unit, and a high expected response sensor determination unit. The boundary neighborhood establishment unit uses the left and right boundaries of the initial anomalous segment as boundaries to be corrected, and establishes an inner neighborhood and an outer neighborhood for each boundary to be corrected. The inner neighborhood is located within the initial anomalous segment and adjacent to the corresponding boundary to be corrected, while the outer neighborhood is located outside the initial anomalous segment and adjacent to the corresponding boundary to be corrected.

[0032] In one implementation, the inner and outer neighborhoods can be determined according to a preset mileage range based on the boundary to be corrected, or they can be determined based on the spacing between adjacent point sensors. Since the specific width of the boundary neighborhood can be adjusted according to the tunnel length, the density of point sensors, and the monitoring scale, the above method is an alternative determination method.

[0033] The candidate point sensor determination unit selects point sensors whose installation mileage is located in the inner or outer neighborhood from the sensor monitoring attribute library, and determines the selected point sensors as candidate point sensors. The expected response state generation unit generates expected response states for each candidate point sensor based on the mileage deviation of the candidate point sensor relative to the abnormal peak position, the mileage deviation relative to the boundary to be corrected, the cross section, the structural location, the measurement direction, and the monitored object.

[0034] The expected response status includes at least the expected response validity, expected response direction, expected response time window, and expected response intensity level. Expected response validity indicates whether a candidate point sensor should exhibit a identifiable measurement change under the current anomalous conditions; expected response direction indicates whether the corresponding measured value should increase, decrease, remain stable, or exhibit a short-term peak; expected response time window indicates the time range within which the corresponding point response should occur; and expected response intensity level indicates the expected degree of change in the point measured value relative to the normal reference state, and includes the minimum effective response level used to determine whether an effective response has been formed.

[0035] For example, when an anomaly spatial fingerprint corresponds to a sidewall or arch area and has continuous residual strain, the cross-sectional convergence sensor located at the relevant cross-section can have the expected response of continuous change in convergence displacement; when the anomaly peak position is close to the existing crack position and has a sudden change in local strain gradient, the corresponding crack sensor can have the expected response of increased crack width; when the anomaly duration is short and the residual strain is low, the corresponding vibration acceleration sensor can have the expected response of short-term vibration peak.

[0036] The high-expected-response sensor determination unit identifies high-expected-response sensors from candidate point sensors. Specifically, a candidate point sensor is identified as a high-expected-response sensor when it simultaneously meets the following conditions: the expected response validity of the candidate point sensor indicates that it should produce an effective response under the current abnormal conditions; the structural part of the candidate point sensor is the same as the structural part to which the current abnormality belongs, or the two belong to related parts that can transmit the current structural deformation; the measurement direction of the candidate point sensor can reflect the structural response corresponding to the current abnormality change direction; the monitoring object of the candidate point sensor can reflect the structural state that the current abnormality may correspond to; and the candidate point sensor is in an effective measurement state.

[0037] Therefore, only point sensors that demonstrate a clear response capability under the current anomalous conditions are identified as high-expected-response sensors. Point sensors with no sustained expected response are not considered high-expected-response sensors for boundary contraction or expansion determination, but their measurement results can still be used for subsequent anomaly type identification. If the measurement direction of a point sensor is unrelated to the structural response direction corresponding to the current anomaly, even if the point sensor is located near the anomaly boundary, its lack of measurement change will not be used as a basis for boundary contraction.

[0038] The measured response extraction module includes a measured state extraction unit. This unit acquires measured data from each high-expected-response sensor according to a unified diagnostic time window and extracts the corresponding measured response state based on the monitoring object of different point sensors. For cross-sectional convergence sensors, the measured response state can include the direction, amplitude, and duration of convergence displacement change; for laser displacement gauges, the measured response state can include the direction, amplitude, and duration of local displacement change; for crack sensors, the measured response state can include the direction, amplitude, and duration of crack width change; for vibration acceleration sensors, the measured response state can include the vibration peak value, peak time, and vibration duration; and for water leakage monitoring sensors, the measured response state can include the leakage trigger state, trigger location, and duration.

[0039] The response relationship determination module includes a matching response determination unit, a persistent missing response determination unit, and a direction-opposite response determination unit. The matching response determination unit compares the measured response state of the high-expected-response sensor with the corresponding expected response state. When the measured response direction of the high-expected-response sensor meets the expected response direction within multiple consecutive diagnostic time windows, the measured response is within the corresponding expected response time window, and the measured response intensity reaches the minimum effective response level, the high-expected-response sensor is determined to be a matching response sensor.

[0040] The persistent missing response determination unit is used to determine whether a high-expected-response sensor has formed a persistent missing response. Specifically, when the measured response intensity of a high-expected-response sensor fails to reach the minimum effective response level within multiple consecutive diagnostic time windows, and the high-expected-response sensor remains in an effective measurement state within the aforementioned multiple consecutive diagnostic time windows, it is determined to be a persistent missing response sensor.

[0041] Multiple consecutive diagnostic time windows can be three or more. Using multiple consecutive diagnostic time windows to determine persistent missing responses can reduce the error boundary correction caused by single data fluctuations, short-term data omissions, and temporary lags in local structural responses.

[0042] The opposite-direction response determination unit is used to determine whether the measured response direction of a high-expected-response sensor is opposite to the expected response direction. When the measured response intensity of a high-expected-response sensor reaches the minimum effective response level, but its measured response direction is opposite to the expected response direction, it is determined to be an opposite-direction response. Opposite-direction responses are not used as evidence of boundary contraction but are used for subsequent anomaly type verification.

[0043] The abnormal section correction module is the core processing module of this embodiment, including a boundary establishment unit to be corrected, a boundary contraction evidence generation unit, a boundary expansion evidence generation unit, a boundary correction interval determination unit, a specific boundary determination unit, and a bilateral independent correction unit.

[0044] The boundary establishment unit determines the left and right boundaries of the initial anomalous segment as two independent boundaries to be corrected, and for each boundary, determines whether evidence of boundary contraction or boundary expansion is formed. Therefore, the left and right boundaries have independent correction results, and the correction direction of one boundary does not determine the correction direction of the other boundary.

[0045] When a persistently missing response sensor exists in the inner neighborhood of the boundary to be corrected, the boundary contraction evidence generation unit does not directly contract the boundary to be corrected based on this. Instead, it further determines whether there is a matching response sensor between the persistently missing response sensor and the abnormal peak position. When there is at least one matching response sensor between the persistently missing response sensor and the abnormal peak position, boundary contraction evidence corresponding to the boundary to be corrected is generated.

[0046] Specifically, boundary contraction evidence indicates a spatial transition from a matched response to a persistently missing response in the spatial direction from the location of the anomaly peak to the boundary to be corrected. Matched response sensors indicate that the local structural response corresponding to the current anomaly extends at least to its location, while persistently missing response sensors indicate that, although a valid response should have appeared near its location based on the structural location, measurement direction, and monitored object, no corresponding measurement change has occurred within multiple consecutive diagnostic time windows. By simultaneously utilizing the matched response on the anomaly peak side and the persistently missing response on the boundary to be corrected side to form boundary contraction evidence, misjudgments caused by shrinking the initial anomaly segment due to the lack of response from a single point sensor near the boundary can be reduced.

[0047] The boundary correction interval determination unit sorts the high-expected-response sensors located between the abnormal peak position and the boundary to be corrected in sequence according to the direction from the abnormal peak position to the boundary to be corrected. Specifically, when the first transition from a matched response state to a persistently missing response state occurs among the sorted high-expected-response sensors, the matched response sensor before the first transition is determined as the peak-side boundary reference sensor, and the persistently missing response sensor after the first transition is determined as the boundary-side boundary reference sensor.

[0048] When multiple matching response sensors exist between a persistently missing response sensor and an abnormal peak location, the matching response sensor closest to the persistently missing response sensor is selected as the peak-side boundary reference sensor. The boundary correction interval determination unit determines the mileage range between the peak-side boundary reference sensor and the boundary-side boundary reference sensor as the boundary contraction correction interval.

[0049] Therefore, point sensors are used to determine that the boundary needs to shrink towards the abnormal peak position and to limit the continuous spatial range where the real boundary may exist, rather than directly determining the installation position of a certain point sensor as the corrected boundary position.

[0050] The boundary expansion evidence generation unit determines whether a high-expected-response sensor exists in the outer neighborhood of the boundary to be corrected. When a high-expected-response sensor in the outer neighborhood reaches the minimum effective response level within multiple consecutive diagnostic time windows, its measured response state matches the corresponding expected response state generated based on the current anomaly spatial fingerprint, and its response time at least partially overlaps with the response time of the matching response sensor inside the boundary to be corrected, and both are within the anomaly duration range of the initial anomaly segment, the outer high-expected-response sensor is identified as an outside-boundary matching response sensor.

[0051] When there is at least one matching response sensor inside the boundary to be corrected and at least one matching response sensor outside the boundary to be corrected, the boundary expansion evidence generation unit further determines whether the distributed anomaly between the boundary to be corrected and the matching response sensor outside the boundary maintains spatial continuity.

[0052] When the distributed anomalies between the boundary to be corrected and the matching response sensor outside the boundary maintain spatial continuity, boundary expansion evidence corresponding to the boundary to be corrected is generated. If the residual strain anomalies at multiple consecutive distributed fiber optic measurement points decrease to below the effective anomaly level, and then subsequently reach the effective anomaly level again near the matching response sensor outside the boundary, a spatial interruption is determined between the two responses, and the matching response sensor outside the boundary is temporarily not used to form boundary expansion evidence for the current boundary to be corrected. This process further reduces the possibility of independent local anomalies outside the boundary being incorrectly incorporated into the current initial anomaly segment.

[0053] As a supplementary processing method, if there is no persistently missing response sensor outside the outermost boundary matching response sensor, then after boundary expansion evidence is formed, the boundary correction interval determination unit continues to search for high-expectation response sensors along the outer direction of the boundary to be corrected. When there is a persistently missing response sensor outside the outermost boundary matching response sensor, the mileage range between the outermost boundary matching response sensor and its nearest persistently missing response sensor is determined as the boundary expansion correction interval. If there is no persistently missing response sensor outside the outermost boundary matching response sensor, the boundary expansion confirmation interval is determined based on the location of the outermost boundary matching response sensor and the preset effective influence range. The specific boundary determination unit analyzes the residual strain anomaly degree and strain gradient decay state within the boundary expansion confirmation interval in the same way as the boundary expansion correction interval, and determines candidate correction boundaries. When the candidate correction boundary is continuously verified in subsequent diagnostic time windows, it is determined as the corrected boundary; before the candidate correction boundary is continuously verified, the original boundary to be corrected is retained.

[0054] The boundary-matched response sensor is used to indicate that the response of the anomaly-related structure has at least extended to its location, while the outer continuous missing response sensor is used to indicate that the response of the anomaly-related structure should not continue to extend completely outward, so the true anomaly boundary is more likely to be located between the two.

[0055] After forming the boundary contraction correction interval or the boundary expansion correction interval, the specific boundary determination unit re-extracts the distributed optical fiber residual strain data within the corresponding boundary correction interval, and analyzes the residual strain continuity and strain gradient attenuation state in the direction from the abnormal peak position to the outside of the abnormal section.

[0056] In this embodiment, point sensors are used to determine the boundary correction direction and boundary correction interval. The specific correction boundary is determined by the continuous distributed optical fiber measurement results within the boundary correction interval. Furthermore, the specific boundary determination unit starts from the side of the boundary correction interval closest to the abnormal peak position and sequentially checks the distributed optical fiber measurement points in a direction pointing outwards from the abnormal segment.

[0057] Along the direction pointing outward from the abnormal section, when the residual strain anomaly of a certain distributed optical fiber measuring point and multiple consecutive distributed optical fiber measuring points outside it continuously decreases to below the effective level of the anomaly, and the absolute value of the residual strain gradient of each distributed optical fiber measuring point is less than the absolute value of the residual strain gradient of the adjacent distributed optical fiber measuring point on its peak side, the mileage position corresponding to the first distributed optical fiber measuring point that meets the above conditions along the inspection direction is determined as the correction boundary.

[0058] If a sequence of distributed fiber optic measurement points with a residual strain anomaly continuously below the effective level of the anomaly has not yet formed within the boundary correction interval, the first location where the strain gradient attenuation state changes from a significant change to a stable low level can be identified as a candidate correction boundary, and further verification can be performed in subsequent diagnostic time windows. Using multiple consecutive distributed fiber optic measurement points to determine the correction boundary can reduce the impact of noise from individual measurement points, changes in local bonding state, or occasional measurement fluctuations on the specific boundary location.

[0059] The two independent correction units determine the correction method for the corresponding boundary based on the formation results of the correction intervals for the left and right boundaries, as well as the specific boundary determination results. When the same boundary to be corrected forms only one of the boundary contraction correction intervals or the boundary expansion correction intervals, the specific boundary determination unit determines the correction boundary within the formed correction interval and updates the corresponding boundary to be corrected based on the determined correction boundary. When the same boundary to be corrected forms both correction intervals simultaneously or neither is formed, the original position of the boundary to be corrected is retained. If only candidate correction boundaries awaiting verification are obtained within the formed correction intervals, the original boundary to be corrected is retained until the candidate correction boundaries are continuously verified. The left and right boundaries are judged independently.

[0060] When the same boundary to be corrected simultaneously generates conflicting evidence of boundary contraction and boundary expansion, the bilateral independent correction unit does not change the boundary to be corrected and marks it as a boundary to be confirmed. The boundary will be re-evaluated after new point-based sensor data is obtained in the subsequent diagnostic time window.

[0061] By processing the left and right boundaries independently, the corrected abnormal section can exhibit different results, such as the left boundary shrinking while the right boundary remains, the left boundary remaining while the right boundary expands, both boundaries shrinking simultaneously, or both boundaries expanding simultaneously. This adapts to the situation where actual structural anomalies are asymmetrically distributed along the tunnel axis.

[0062] The anomaly center correction module includes a peak stability offset determination unit and a center position determination unit. After correcting the left and right boundaries of the initial anomaly section, the peak stability offset determination unit redetermines the distributed anomaly peak position within the corrected anomaly section based on the degree of residual strain anomaly of the distributed optical fiber measuring points, and determines whether there is a stable spatial offset between the mileage position corresponding to the point response peak section and the distributed anomaly peak position.

[0063] For multiple matched response sensors that can reflect the current abnormal structural response, the normalized measured response intensity is determined based on the degree to which the measured response intensity of each matched response sensor exceeds its corresponding minimum effective response level. The section containing the matched response sensor with the highest normalized measured response intensity is determined as the point response peak section. When no matched response sensor can be used to determine the point response peak section, no anomaly center correction is performed, and the newly determined distributed anomaly peak position within the corrected anomaly segment is taken as the current anomaly center. The normalized measured response intensity referred to in this embodiment is used to characterize the degree to which the measured response of a point sensor exceeds the corresponding minimum effective response level of that point sensor, so that the responses of point sensors with different monitoring objects or different dimensions can be compared on a uniform scale. When the spatial offset between the mileage position corresponding to the point response peak section and the distributed anomaly peak position within the corresponding diagnostic time window remains in the same direction within multiple consecutive diagnostic time windows, and the monitored object of the high-expected response sensor corresponding to the point response peak section has a direct correspondence with the structural response reflected by the current anomaly spatial fingerprint, the point response peak section is determined as the anomaly center correction reference position.

[0064] Specifically, the center location determination unit determines the center correction search range with the anomaly center correction reference position as the center, and re-finds the local maximum value of the residual strain anomaly degree among the distributed optical fiber measurement points within the center correction search range, and determines the mileage position corresponding to the local maximum value as the corrected anomaly center.

[0065] Therefore, point sensors are used to provide the offset direction and spatial reference of the anomaly center, and the continuous measurement results of the distributed optical fiber are used to determine the specific location of the anomaly center, so that the corrected anomaly center still corresponds to the continuous mileage location of the distributed optical fiber.

[0066] When the spatial offset occurs only in a single diagnostic time window, or when the offset direction changes in consecutive diagnostic time windows, the anomaly center correction module will take the distributed anomaly peak position re-determined based on the corrected anomaly segment as the current anomaly center and mark the anomaly center as pending confirmation.

[0067] The anomaly type discrimination module includes a type relationship matching unit and a measurement conflict verification unit. The type relationship matching unit determines the anomaly type based on the corrected anomaly segment, the corrected anomaly center, the anomaly spatial fingerprint, and the measured response status of the point sensor.

[0068] In this embodiment, the anomaly effectiveness level is used to determine whether the distributed optical fiber measurement results form an initial anomaly segment that requires further analysis, and the structural anomaly determination level is used to determine whether the degree of residual strain anomaly after temperature correction reaches a level that can support continuous structural anomaly. The structural anomaly determination level is higher than the anomaly effectiveness level. When the degree of residual strain anomaly within the corrected anomaly segment continuously reaches the structural anomaly determination level, and the cross-sectional convergence sensor or laser displacement gauge continuously shows changes that conform to the corresponding expected response state, the anomaly type is determined to be structural deformation anomaly.

[0069] When a sudden change in local strain gradient exists within the corrected anomalous section, and the crack width of the corresponding crack sensor continues to increase, the anomalous type is determined to be a crack propagation anomaly.

[0070] When the duration of distributed strain anomaly is short, the vibration acceleration sensor synchronously shows a short-term peak, the degree of residual strain anomaly after temperature correction reaches the effective level of anomaly but is lower than the structural anomaly judgment level, and the cross-sectional convergence, local displacement and crack width do not show continuous changes, the anomaly type is determined to be transient vibration disturbance.

[0071] When the original strain change has a significant time correlation with the corresponding temperature change, the degree of residual strain anomaly after temperature correction reaches the effective level of anomaly but is lower than the level for determining structural anomalies, and there is no continuous change in cross-sectional convergence, local displacement, and crack width, the anomaly type is determined to be a temperature-related disturbance.

[0072] When the degree of residual strain anomaly increases slowly and continuously, and the trigger position of the leakage monitoring sensor is located within the corrected anomaly section or near the corrected anomaly boundary, the anomaly type is determined to be leakage-induced deterioration anomaly.

[0073] The measurement conflict verification unit handles situations where a high-expected-response sensor reaches the minimum effective response level, but the measured response direction is opposite to the expected response direction. When the opposite response cannot be explained by verification of the sensor installation direction and measurement direction, the anomaly type discrimination module suspends the output of the deterministic structural damage type and generates a measurement direction conflict flag to avoid outputting incorrect anomaly causes based on contradictory measurement results.

[0074] The state level output module includes a state level determination unit and a result output unit. The state level determination unit determines the tunnel structure state level based on the corrected abnormal section length, corrected abnormal center, abnormal type, residual strain abnormality degree, abnormal duration, and the number of matched response sensors. The result output unit outputs the corresponding state level and abnormality information. When generating a measurement direction conflict marker, the state level output module does not use the abnormal type as a necessary input for state level determination. Instead, it determines the current state level based on the corrected abnormal section length, residual strain abnormality degree, abnormal duration, and the number of matched response sensors, and simultaneously outputs the measurement direction conflict marker in the output result.

[0075] In this embodiment, the structural state level can include normal, attention, warning, and severe warning. When only a short-term distributed anomaly exists and no persistent residual strain or persistent point response is formed, the current anomaly is recorded as a short-term disturbance event. When the corrected anomaly segment persists but the point response support is low, the attention level is output. When the corrected anomaly segment persists and there are one or more structural deformation class matching response sensors consistent with the expected response state, the warning level is output. When the temperature-corrected residual strain, crack width, cross-sectional convergence, or local displacement continues to increase, the severe warning level is output.

[0076] Through the above implementation method, this embodiment first uses a distributed optical fiber sensing module to form an initial anomalous segment, and then establishes boundary neighborhoods for the left and right boundaries of the initial anomalous segment. Within the boundary neighborhood, only high-expected-response sensors whose expected response validity, structural location, measurement direction, monitored object, and effective measurement state all meet the current anomalous response conditions are allowed to participate in boundary correction. Furthermore, the spatial transition relationship between the matching response inside the boundary and the persistent missing response is used to form evidence of boundary contraction, and the spatial continuity relationship between the matching response inside the boundary and the persistent response outside the boundary that meets the corresponding expected response state is used to form evidence of boundary expansion.

[0077] After evidence of boundary contraction or expansion is generated, point sensors are used to define the boundary correction direction and interval, while distributed fiber optic continuous measurement results are used to determine the specific boundary location within the boundary correction interval. Thus, even with limited point sensor deployment, point measurement results can be used to reverse-correct distributed anomaly sections, while avoiding directly substituting the installation location of discrete point sensors for continuous anomaly boundaries.

[0078] By using the matched response on the abnormal peak side and the persistent missing response on the boundary side to form evidence of boundary contraction, the possibility of erroneous contraction caused by occasional non-response of a single point sensor can be reduced. By using the persistent responses, time correlations, and spatial continuity of distributed residual strain that conform to the expected response states inside and outside the boundary to form evidence of boundary expansion, the possibility of erroneous expansion caused by independent events outside the boundary can be reduced. By using the left and right boundaries to make independent corrections, the adaptability to asymmetric anomaly spatial distribution can be improved, so that the corrected anomaly section is more consistent with the spatial range of the actual response of the tunnel structure.

[0079] The technical scope of this invention is not limited to the content described above. Those skilled in the art can make various modifications and variations to the above embodiments without departing from the technical concept of this invention, and all such modifications and variations should fall within the protection scope of this invention.

Claims

1. An integrated monitoring system for tunnel structures based on the fusion of distributed and point-based sensing, characterized in that, include: Anomaly space fingerprint generation module: The initial anomalous section with left and right boundaries and anomalous peak positions is determined based on the residual strain after temperature correction, and an anomalous spatial fingerprint is generated. Expected response generation module: Using the left and right boundaries as the boundaries to be corrected respectively, the expected response state is generated based on the abnormal space fingerprint and sensor monitoring attributes, and the high expected response sensor is identified. Response Relationship Determination Module: Based on the comparison results of the measured response of the high expected response sensor with the corresponding expected response state in multiple consecutive diagnostic time windows, the matched response sensor and the continuously missing response sensor are determined. Abnormal section correction module: Along the direction from the abnormal peak position to the boundary to be corrected, a boundary contraction correction interval is formed based on the spatial transition relationship between the matching response sensor inside the boundary and the persistently missing response sensor; based on the temporal correlation between the high expected response sensor outside the boundary that continuously meets the expected response state and the matching response sensor inside the boundary to be corrected, the matching response sensor outside the boundary is determined; when the distributed anomaly between the boundary to be corrected and the matching response sensor outside the boundary maintains spatial continuity, a boundary expansion correction interval is formed by combining it with the persistently missing response sensor outside the boundary. When the same boundary to be corrected forms only one of the two correction intervals mentioned above, the correction boundary is determined within that correction interval based on the continuously distributed residual strain after temperature correction and its gradient decay state; otherwise, the original position of the boundary to be corrected is retained.

2. The integrated tunnel structure monitoring system based on the fusion of distributed and point-based sensing as described in claim 1, characterized in that: It also includes a sensor monitoring and management module, which establishes sensor monitoring attribute records for each point sensor. The sensor monitoring attributes include structural location, measurement direction and monitoring object, and further record the installation mileage and the cross section where it is located.

3. The integrated monitoring system for tunnel structures based on the fusion of distributed and point-based sensing according to claim 2, characterized in that: The expected response generation module determines point sensors in the inner and outer neighborhoods of each boundary to be corrected; the expected response state includes the expected response direction, the expected response time window, and the minimum effective response level. When the measured response direction matches the expected response direction, the measured response is within the expected response time window, and the measured response intensity reaches the minimum effective response level, it is determined to meet the expected response state; the high expected response sensor is a point sensor that is in an effective measurement state and the expected response state indicates that it should respond.

4. The integrated monitoring system for tunnel structures based on the fusion of distributed and point-based sensing according to claim 3, characterized in that: The response relationship determination module will identify high-expected-response sensors that meet the expected response state in multiple consecutive diagnostic time windows as matched response sensors, and high-expected-response sensors that are continuously in an effective measurement state in multiple consecutive diagnostic time windows and whose measured response intensity does not reach the minimum effective response level as continuously missing response sensors; when the measured response intensity of a high-expected-response sensor reaches the minimum effective response level but the measured response direction is opposite to the expected response direction, the corresponding measured response will be identified as an opposite-direction response.

5. The integrated tunnel structure monitoring system based on the fusion of distributed and point-based sensing according to claim 4, characterized in that: The abnormal segment correction module, along the direction from the abnormal peak position to the boundary to be corrected, determines the mileage range between the sensors on both sides of the transition as the boundary contraction correction interval when the first transition from a matched response sensor to a continuously missing response sensor occurs. It also determines the high-expected response sensors outside the boundary to be corrected that continuously meet the expected response state and whose responses satisfy the time correlation with the matched response sensors inside the boundary to be corrected as boundary-outside matched response sensors. When the distributed anomalies between the boundary to be corrected and the boundary-outside matched response sensors remain spatially continuous, the mileage range between the outermost boundary-outside matched response sensor and its nearest continuously missing response sensor is determined as the boundary expansion correction interval.

6. The integrated tunnel structure monitoring system based on the fusion of distributed and point-based sensing according to claim 5, characterized in that: The abnormal section correction module starts from the side of the boundary contraction correction interval or boundary expansion correction interval near the abnormal peak position and sequentially checks the distributed optical fiber measurement points in the direction pointing outward of the abnormal section. In the direction pointing outward of the abnormal section, when the residual strain anomaly of a certain distributed optical fiber measurement point and multiple consecutive distributed optical fiber measurement points outside it continuously decreases to below the effective level of the anomaly, and the absolute value of the residual strain gradient of each distributed optical fiber measurement point is less than the absolute value of the residual strain gradient of the distributed optical fiber measurement point adjacent to its peak side, it is determined that the distributed optical fiber measurement point meets the boundary judgment condition, and the mileage position corresponding to the first distributed optical fiber measurement point that meets the boundary judgment condition along the inspection direction is determined as the correction boundary.

7. The integrated tunnel structure monitoring system based on the fusion of distributed and point-based sensing according to claim 1, characterized in that: It also includes a distributed fiber optic sensing module, a point-type sensing module, and a data acquisition and processing module; The distributed optical fiber sensing module continuously acquires raw strain field data and temperature field data corresponding to the tunnel mileage location along the tunnel axis. The point-type sensing module acquires local structural response data of key sections, key structural parts, or existing defects in the tunnel; the data acquisition and processing module adds sampling time and corresponding spatial location information to the original strain field data, temperature field data, and local structural response data, so that the distributed fiber optic measurement results and the point-type sensor measurement results can be compared according to the corresponding spatial location and diagnostic time window.

8. The integrated tunnel structure monitoring system based on the fusion of distributed and point-based sensing according to claim 7, characterized in that: The anomaly spatial fingerprint generation module includes a temperature correction unit, an initial anomaly segment identification unit, and a fingerprint generation unit. The temperature correction unit performs temperature correction on the original strain field data based on the temperature change corresponding to the distributed optical fiber measurement location to obtain the temperature-corrected residual strain. The initial anomaly segment identification unit determines the mileage range covered by the corresponding distributed optical fiber measurement point as the initial anomaly segment when the residual strain anomaly degree of multiple consecutive distributed optical fiber measurement points reaches the effective level of anomaly and maintains spatial continuity. The anomaly spatial fingerprint includes at least the anomaly peak position, left boundary, right boundary, left strain gradient attenuation rate, right strain gradient attenuation rate, anomaly duration, anomaly change direction, and the structural part to which the anomaly belongs.

9. The integrated monitoring system for tunnel structures based on the fusion of distributed and point-based sensing according to claim 5, characterized in that: It also includes an anomaly center correction module, which determines the location of the distributed anomaly peak based on the degree of residual strain anomaly of the distributed optical fiber measuring points in the corrected anomaly section, and determines the normalized measured response intensity based on the degree of excess of the measured response intensity of each matched response sensor relative to the corresponding lowest effective response level, and determines the section where the matched response sensor with the largest normalized measured response intensity is located as the point response peak section. When a stable spatial offset is formed between the peak section of the point response and the peak position of the distributed anomaly, and the corresponding monitored object has a direct correspondence with the structural response reflected by the current anomaly spatial fingerprint, the center correction search range is determined by the peak section of the point response, and the mileage position corresponding to the local maximum value of the residual strain anomaly degree within this range is determined as the corrected anomaly center.

10. The integrated monitoring system for tunnel structures based on the fusion of distributed and point-based sensing according to claim 9, characterized in that: It also includes an anomaly type discrimination module, which determines the anomaly type based on the corrected anomaly segment, the corrected anomaly center, the anomaly spatial fingerprint, and the measured response state of the point sensor. The anomaly type includes structural deformation anomaly, crack propagation anomaly, and transient vibration disturbance. The anomaly type discrimination module is also used to verify the opposite direction response. When the opposite direction response still cannot be explained after verification of the sensor installation direction and measurement direction, the output of the deterministic anomaly type is paused and a measurement direction conflict mark is generated.

Images