A method for real-time monitoring and evaluation of health of underground structure segment

Abstract

This invention relates to a method for real-time monitoring and evaluation of the health of underground structural tunnel segments. The method includes the following steps: first, designing and deploying a segment health monitoring system and sensors; second, monitoring and calibrating sensor status; third, implementing adaptive monitoring and intelligent early warning mechanisms; fourth, establishing a health scoring system and maintenance decisions; fifth, conducting joint prediction and crack monitoring; sixth, data analysis and system optimization; seventh, long-term trend prediction and decision optimization; and finally, decision feedback and system optimization. The advantages of this invention are that by deploying a multi-functional fiber optic sensor array at key stress-bearing locations of underground structural tunnel segments, multiple monitoring parameters such as stress, cracks, temperature, and deformation can be acquired simultaneously, overcoming the problem of existing technologies that monitor only a single physical quantity and cannot reflect the overall health status of the tunnel segments. Through a unified sensor link, multi-parameter fusion monitoring is achieved, reducing the number of sensors and improving the deployment efficiency and data consistency of the monitoring system.

Description

Technical Field

[0001] This invention belongs to the field of civil engineering structural health monitoring technology, specifically relating to a method for real-time monitoring and evaluation of the health of underground structural segments. Background Technology

[0002] Fiber optic sensing technology, due to its resistance to electromagnetic interference, corrosion resistance, long-distance deployment, and applicability to underground structures, has been widely used in structural health monitoring of tunnels, underground utility tunnels, and underground structural segments. Currently, fiber optic gratings (FBGs) and distributed fiber optic sensing (DFOS) technologies can achieve high-precision acquisition of multiple physical quantities such as stress, temperature, and displacement, giving them irreplaceable advantages in long-term service monitoring of underground structures. However, with extended monitoring cycles, complex factors in the underground environment, such as temperature and humidity cycles, soil pressure changes, microcrack evolution, and material creep, can lead to performance degradation of the fiber optic sensing links themselves, affecting the reliability of monitoring data. For example, patent CN209264013U proposes a comprehensive utility tunnel health monitoring system based on fiber optic gratings, which uses various fiber optic sensors for cracks, settlement, and displacement to achieve real-time monitoring of the utility tunnel's structural condition. While the system offers advantages in sensor integration and remote monitoring, its primary monitoring targets are municipal utility tunnels, which differ significantly from underground tunnel segments in terms of structural type, stress characteristics, and environmental conditions. Furthermore, its monitoring methods rely heavily on single-parameter threshold judgments, lacking analysis of the coupling relationships between multiple monitoring parameters and failing to establish a unified health measurement system. Additionally, the scheme does not address the degradation mechanisms of fiber optic sensors or their bonding interfaces during long-term service, nor does it propose methods for correcting the time-varying accuracy of the sensor link.

[0003] Patent CN101713691A proposes using distributed optical fiber sensing technology to monitor the health of tunnel structures by deploying optical fibers in the lining structure to monitor tunnel stress and deformation. While this approach has promoted the application of optical fibers in tunnel engineering, the solution focuses primarily on fiber deployment technology and data acquisition systems, neglecting issues such as fiber signal drift, decreased strain transmission efficiency, and environmental factors during long-term service. Furthermore, this technology does not propose a multi-parameter fusion analysis method, relying instead on single-parameter determination of structural status, making it difficult to accurately reflect the overall health status of shield tunnel segments under complex loads and environmental conditions. US Patent US7536911B2 (and related publications) discloses a type of optical fiber coil-type structural monitoring device, primarily used for local strain and vibration testing, suitable for local monitoring scenarios in aerospace, bridge, and other components. Although this type of optical fiber sensor has high sensitivity, its monitoring range is often limited to a single physical quantity, making it difficult to meet the comprehensive monitoring needs of shield tunnel segments for multiple parameters such as stress, cracks, temperature, and deformation. Meanwhile, this technology lacks a framework for multi-source data fusion and health assessment, and it also fails to address compensation mechanisms for sensor performance changes under long-term underground environmental conditions. A review of existing technologies reveals that whether it's integrated utility tunnel monitoring systems, distributed fiber optic monitoring technology for tunnels, or local fiber optic sensing devices, they primarily address the "data acquisition" problem, while exhibiting significant shortcomings in multi-parameter fusion, quantitative assessment of health status, long-term stability identification of sensor links, and dynamic threshold adaptive early warning. Especially in the long-term service environment of shield tunnel underground structures, factors such as temperature and humidity cycles, soil pressure fluctuations, microcrack propagation, environmental corrosion, and material creep all affect fiber optic sensing signals. Existing technologies generally lack effective sensor status identification, signal accuracy correction, and adaptive health scoring systems based on multiple parameters.

[0004] Therefore, there is an urgent need for a monitoring method that can be used in the complex service environment of underground structures such as subway shield tunnel segments and jacking tunnel segments. This method should integrate multiple physical quantities for monitoring, have sensor status recognition capabilities, realize multi-parameter fusion analysis, and establish a unified health measurement index system. This would solve the problems of insufficient long-term monitoring accuracy, incomplete health assessment, and single early warning mechanism in existing technologies, thereby better meeting the safety requirements of subway tunnel structures throughout their entire life cycle. Summary of the Invention

[0005] This invention aims to address the problem that existing underground structural segments lack a health assessment method for multi-parameter fusion analysis oriented towards complex service environments.

[0006] The present invention solves the above-mentioned technical problems through the following technical means:

[0007] A method for real-time monitoring and evaluation of the health of underground tunnel segments includes the following steps: design of a segment health monitoring system and sensor deployment, sensor status monitoring and calibration, adaptive monitoring and intelligent early warning mechanism, health scoring system and maintenance decision, joint prediction and crack monitoring, data analysis and system optimization, long-term trend prediction and decision optimization, and decision feedback and system optimization. In the adaptive monitoring and intelligent early warning mechanism step, the system monitors parameters; if the parameters exceed a threshold, an alarm is issued; otherwise, the process proceeds to the next step. In the health scoring system and maintenance decision step, the health score determines whether to issue an emergency maintenance recommendation, strengthen monitoring and inspection, and proceed to the next step, or proceed directly to the next step. After completing the decision feedback and system optimization step, the process returns to the sensor status monitoring and calibration step, thus creating a loop.

[0008] This invention deploys a multi-functional fiber optic sensor array at key stress-bearing locations of underground tunnel segments. This array can simultaneously acquire multiple monitoring parameters, including stress, cracks, temperature, and deformation, overcoming the limitations of existing technologies that monitor only a single physical quantity and cannot reflect the overall health status of the tunnel segments. By achieving multi-parameter fusion monitoring through a unified sensor link, the number of sensors is reduced, improving the deployment efficiency and data consistency of the monitoring system.

[0009] Preferably, the specific process of the segment health monitoring system design and sensor deployment steps includes: multi-functional fiber optic sensor array integrated design: deploying multi-functional fiber optic sensor arrays in the key stress paths and potential weak areas of each underground structural segment, simultaneously performing stress monitoring, temperature monitoring, crack monitoring, and deformation monitoring; sensor deployment locations: deploying sensors at key axial stress locations, key circumferential stress locations, crack-sensitive areas, and deformation-sensitive areas, with multiple multi-functional sensor array nodes deployed in each ring of segments, and each node containing multiple monitoring sub-modules.

[0010] Preferably, in the step of designing and deploying the health monitoring system for the pipe segment, the sensor array node includes an optical fiber demodulation module, a microprocessor, a wireless communication module, and a local data cache and breakpoint retransmission module; all monitoring data are automatically uploaded to the cloud through a star or chain-type Internet of Things architecture.

[0011] Preferably, the sensor status monitoring and calibration steps specifically include: sensor health status monitoring: fiber optic sensor sensitivity monitoring, temperature sensor accuracy calibration, signal attenuation monitoring, multi-sensor data verification, and sensor accuracy correction; sensor automatic calibration and feedback mechanism: the system executes an automatic calibration program under preset standard environment or known load conditions. When an error is detected in the sensor, the system automatically corrects the output data based on a preset algorithm to ensure data accuracy. Through the Internet of Things platform, the sensor status information is fed back in real time to ensure that faulty sensors can be replaced or calibrated in a timely manner, avoiding data quality degradation.

[0012] Preferably, the specific process of the adaptive monitoring and intelligent early warning mechanism includes: multi-parameter adaptive threshold adjustment and over-limit early warning: the system adaptively adjusts the multi-parameter thresholds based on real-time monitoring data and historical data, and dynamically corrects the alarm thresholds of stress, temperature, cracks, and deformation parameters by combining the health score of the tunnel segment, crack development, and earth pressure changes. It also monitors temperature, deformation, crack development, and earth pressure parameters and sets thresholds for each. The system collects data in real time and compares it with the preset thresholds. Once a parameter exceeds the limit, an alarm is automatically triggered, and the alarm information is transmitted to relevant personnel through the Internet of Things platform.

[0013] Preferably, the actual parameters monitored by the adaptive monitoring and intelligent early warning mechanism are compared with the thresholds to determine the response and decision: for minor alarms, a reminder is given to follow up and monitor; if there are further changes, the system will push again; for emergency alarms, emergency repairs are organized immediately.

[0014] Preferably, the specific process of the health scoring system and maintenance decision-making includes: multi-parameter influence assessment and adaptive weight determination: using an Internet of Things monitoring system to retrieve historical monitoring data within a certain time period, including axial and circumferential stress time series. and Crack width and crack growth rate w(t) , dw / dt Segment deformation or misalignment Segment surface temperature T(t) Simultaneously organize the engineering event data for the corresponding time period, constructing each time period and each monitoring loop as a sample, denoted as... and using the formula Normalization and dimensionless processing of the indicators yield a feature matrix with uniform dimensions: In the formula, k=1,2,3,4 correspond to stress, crack, temperature, and deformation, respectively. j For sample number, and These represent the time series of axial and circumferential stresses, respectively. w(t) and dw / dtThis represents the relationship between crack width and crack growth rate. T(t) Indicates the surface temperature of the tube segment. This indicates segment deformation or misalignment displacement. Indicates label j The corresponding dimensionless values ​​of stress, crack, temperature, or deformation.

[0015] Preferably, the specific process of the health scoring system and maintenance decision-making further includes: impact calculation, which includes statistical correlation method, or machine learning feature importance method, or statistical correlation combined with machine learning feature importance method; wherein statistical correlation: defining risk labels. ( Using binary, : 0 - Safe, 1 - Defective or already repaired; or use multi-level. : Safe, sub-healthy, dangerous; or use expert scoring. (Indicates the level of risk), calculates each indicator and risk label. Pearson correlation coefficient or mutual information between them: or Four influence indicators were obtained. The larger the value, the stronger the correlation between the parameter and the problem or maintenance event. Sort from largest to smallest; Machine learning feature importance method: Training a risk prediction model ,enter Output predicted risk Random forests and XGBoost can be used. After the model is trained, the "importance" index of each feature is extracted. For tree models, the "feature importance score" can be used directly; the machine learning method used utilizes feature masks to obtain the contribution of each parameter, thus yielding the same result. Four importance values, ranked in order; where These represent the influence indices corresponding to stress, cracks, temperature, and deformation, respectively.

[0016] Preferably, the specific process of the health scoring system and maintenance decision-making further includes: adaptive weight calculation and normalization: using formulas Impact index Convert to weights Therefore, Segment health score calculation: For the current time t, each monitoring unit obtains Four types of dimensionless indicators, using formulas Calculate the damage risk index; where This indicates the proportion of the current stress to the design allowable value. This indicates the proportion of the crack width to the allowable limit. This indicates the degree to which the temperature deviates from the design temperature. This indicates the proportion of deformation or misalignment to the allowable value. These represent the weights corresponding to stress, temperature, crack, and deformation, respectively. Indicators representing damage risk.

[0017] Preferably, the specific process of the health scoring system and maintenance decision-making further includes: using formulas Calculate health status when ≥0.8 indicates good health and requires no maintenance; when 0.5≤ <0.8 indicates sub-health, and monitoring is recommended; when <0.5 indicates poor health, and immediate repair is recommended; where Indicates the current Stress at any moment Indicates the current Temperature at any moment Indicates the current The cracks of time Indicates the current Transformation in a moment Indicates the maximum allowable stress value. Indicates the maximum allowable temperature value. This represents the maximum allowable crack value. Indicates the maximum allowable deformation value. Rate your health.

[0018] The advantages of this invention are: (1) The present invention deploys a multi-functional fiber optic sensor array at key stress-bearing parts of underground structural segments, which can simultaneously acquire multiple monitoring parameters such as stress, cracks, temperature and deformation, overcoming the problem of the existing technology that monitors only a single physical quantity and cannot reflect the overall health status of the segments. By realizing multi-parameter fusion monitoring through a unified sensing link, the number of sensors is reduced, and the deployment efficiency and data consistency of the monitoring system are improved; (2) This invention comprehensively identifies the baseline response (initial zero-point wavelength and its offset over time), temperature drift characteristics and long-term drift of the fiber optic sensor link, constructs sensor status evaluation index, and monitors and calibrates the sensor in a timely manner during the data processing process. This can effectively solve the problem of signal error accumulation caused by temperature and humidity cycles, bonding interface aging and other factors during long-term service of fiber optic sensors, and improve the long-term stability of monitoring data. (3) This invention utilizes historical monitoring data, fault samples, and correlation analysis or machine learning methods to rank the influence of multiple monitoring parameters and generate adaptive weight coefficients accordingly, thereby achieving dynamic updates to the health rating model. This method overcomes the shortcomings of traditional monitoring systems that rely on fixed thresholds or human experience for status judgment, enabling health assessment results to adaptively adjust with changes in the environment and structural status, thus improving the scientificity and accuracy of the assessment; (4) This invention constructs a real-time anomaly identification and early warning mechanism based on multi-parameter monitoring data. It can automatically identify various risk events such as stress over-limit, crack propagation, and abnormal temperature rise, and provide operation and maintenance personnel with risk contribution information of specific parameters. Through the linkage analysis of health score and risk information, the system can prompt relevant maintenance information, providing a basis for daily inspection of underground structural segments, structural reinforcement, and operational risk identification, thereby improving the safety assurance capability of subway tunnels throughout their entire life cycle. Attached Figure Description

[0019] Figure 1 This is a flowchart of the real-time health monitoring and evaluation method for underground structural tunnel segments according to the first embodiment of the present invention; Figure 2 This is a schematic diagram of the arrangement of fiber optic monitoring points for underground structural tunnel segments according to the first embodiment of the present invention; Figure 3 The arrangement of monitoring points for circular underground structural segments is shown in the first embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Example 1: See Figure 1 This application provides a method for real-time monitoring and evaluation of the health of underground tunnel segments, the specific steps of which include: S1. Segment Health Monitoring System Design and Sensor Deployment: S1a. Multifunctional Fiber Optic Sensor Array Integration Design; S1b. Sensor Deployment Location; S1c. Sensor Data Acquisition and Transmission Mechanism; The specific process of S1 is as follows: A multifunctional fiber optic sensor array (FBG array) is deployed in the critical stress path and potentially weak areas of each underground structural segment. Each sensor array can simultaneously output multiple monitoring parameters, including: monitoring axial stress. and circumferential stress (Monitoring longitudinal and circumferential stress changes in the pipe segment using axially and circumferentially deployed FBG sensors; employing a multi-wavelength FBG series structure allows for continuous monitoring of multiple measurement points from a single optical fiber), temperature monitoring. (Using a temperature-compensated FBG, ambient and structural temperatures are collected via an independent temperature grid), crack width and crack evolution. Monitoring (using FBG crack gauges and surface-mounted optical fibers to monitor crack development trends), displacement and deformation of tunnel segments. Monitoring (using displacement-type FBG or fiber optic distributed sensing (BOTDA / BOTDR) to monitor local opening deformation and misalignment displacement); to improve monitoring efficiency and structural condition identification capabilities, this system employs multi-point multiplexing and intelligent collaboration of sensor arrays, including dense multi-node FBG deployment, and 8–12 FBG measuring points connected in series on the same fiber (the lower limit is used in this embodiment under normal working conditions, and the upper limit is used under complex working conditions; other numbers are also possible in engineering practice) capable of independently acquiring multiple parameters; see reference Figure 2 Based on the stress characteristics of underground tunnel segments, finite element analysis results, and engineering experience, sensors are deployed at key locations of axial stress, key locations of circumferential stress, crack-sensitive areas, and deformation-sensitive areas. The method in this application involves deploying 3-5 multi-functional sensor array nodes per ring segment (the lower limit is used in this embodiment under normal working conditions, and the upper limit under complex working conditions; other numbers are also possible in actual engineering). Each node contains several monitoring sub-modules. Each sensor is connected via an IoT acquisition node. The sensor node includes: an optical fiber demodulation module (completing the conversion from grating reflection wavelength to digital signal, capable of identifying different wavelengths of FBG to distinguish multiple parameters), a microprocessor (performing signal processing, filtering, and compression), a wireless communication module (transmitting real-time data to the cloud platform), and a local data caching and breakpoint retransmission module (avoiding data loss due to the tunnel environment). All monitoring data is automatically uploaded to the cloud via a star or chain-like IoT architecture. The cloud performs multi-parameter fusion, health scoring, stress-crack correlation analysis, long-term trend prediction, and risk level identification.

[0022] See Figure 2 The diagram shows the layout of fiber optic monitoring points for rectangular underground tunnel segments. Details are as follows:

[0023] Figure 3 This is the arrangement of monitoring points for circular underground tunnel segments in the first embodiment of the present invention. Figure 2 Another structural form of observation point arrangement.

[0024] S2. Sensor Status Monitoring and Calibration: S2a. Sensor Health Status Monitoring; S2b. Automatic Sensor Calibration and Feedback Mechanism; The specific process of S2 is as follows: First, calibration is performed using a standard load to ensure the accuracy of the fiber optic sensor; the temperature compensation performance of the temperature sensor is verified to ensure its measurement accuracy; the transmitted signal is monitored to ensure that the signal attenuation is within an acceptable range and does not affect the data accuracy; after completing the above steps, errors are monitored by cross-validation of data from multiple sensors. When a data deviation is detected, it is automatically marked as abnormal and an alarm is triggered. When a sensor's accuracy is found to have decreased, the system will automatically start the calibration mechanism or remind engineers to replace the sensor (the system executes an automatic calibration program through preset standard environments or known load conditions. When an error is detected in the sensor, the system automatically corrects the output data based on a preset algorithm to ensure the accuracy of the data). The system also provides real-time feedback on the sensor's status information through an IoT platform to ensure that faulty sensors can be replaced or calibrated in a timely manner to avoid data quality degradation.

[0025] S3. Adaptive Monitoring and Intelligent Early Warning Mechanism: S3a. Multi-parameter Adaptive Threshold Adjustment and Over-limit Early Warning; S3b. Alarm Push and Feedback Mechanism; The specific process of S3 is as follows: The system monitors data in real time (stress, temperature, cracks, deformation, earth pressure, etc.) and adaptively adjusts the thresholds of multiple parameters based on historical data; the system also dynamically corrects the alarm thresholds of parameters such as stress, temperature, cracks, and deformation by combining factors such as the health score of the tunnel segment, crack development, and changes in earth pressure; if ,like ,like ,like , like The system collects data in real time and compares it with preset thresholds. Once a parameter exceeds the limit, an alarm is automatically triggered, and alarm information, including alarm type, location, and degree of exceedance, is pushed to the equipment of engineers via various means such as APP, SMS, and email. and These represent axial stress and circumferential stress, respectively. and These represent the maximum design-defined stresses for axial stress and circumferential stress, respectively. Indicates the temperature at the time of monitoring. This indicates the threshold value for the set temperature. This indicates the deformation value at the monitoring time. The threshold representing the deformation value, This indicates the crack value at the monitoring time. This indicates the threshold value set for the crack. Indicates the earth pressure at the monitoring time. This indicates the set threshold for earth pressure.

[0026] S4. Parameter Exceedance Judgment; The specific process of S4 is as follows: Determine the alarm level. If the alarm level is a minor alarm, remind the system to monitor it. If there is any further change, the system will push again. If the alarm level is an emergency alarm, organize emergency repairs immediately. After confirming that the system is fault-free and no alarm occurs, proceed to step S5. If no alarm occurs, proceed directly to step S5.

[0027] S5. Health rating system and maintenance decision-making; S5a. Multi-parameter influence assessment and adaptive weight determination; S5b. Segment health rating calculation; The specific steps of S5 are as follows: Using the Internet of Things monitoring system, retrieve historical monitoring data within a certain time period (in this embodiment, the past 3-5 years), including axial and circumferential stress time series. and Crack width and crack growth rate w(t) , dw / dt Segment deformation or misalignment Segment surface temperature T(t) Simultaneously, it organizes engineering event data for that period (including whether crack repair, reinforcement, or leakage control events have occurred, the "health level" assessed by experts, and the inspection records of the operating unit), and defines input features. and output risk labels ( Using binary, : 0 - Safe, 1 - Defective or already repaired; or use multi-level. : Safe, sub-healthy, dangerous; or use expert scoring. (Indicating the level of risk), using the formula For the k The indexes (stress, cracks, temperature, deformation) are normalized using range to obtain a characteristic matrix with uniform dimensions: Then, the formula is derived using statistical correlation. or Calculate each indicator and risk label The influence index is obtained by using the Pearson correlation coefficient or mutual information between them. ,according to Sort from largest to smallest; alternatively, a risk prediction model can be trained using machine learning feature importance methods. ,enter Output predicted risk Random forests or XGBoost can be used. After the model is trained, the "importance" index of each feature can be extracted using a tree model or a feature mask. ; Obtain the influence index Then use the formula Calculate weights And obtain the weight vector. (The weights can be updated periodically, recalculated annually or after a certain number of event samples are added; the system has self-learning and adaptive capabilities); after obtaining the calculated weights, the current time... t Each monitoring unit (such as a single-ring segment) receives Four types of dimensionless indicators, and using formulas Calculate the damage risk index; finally, use the formula Calculate the health score; where These are the weighting coefficients for stress, temperature, cracks, and deformation on health status. Indicates the current Stress at any moment Indicates the current Temperature at any moment Indicates the current The cracks of time Indicates the current Transformation in a moment Indicates the maximum allowable stress value. Indicates the maximum allowable temperature value. This represents the maximum allowable crack value. Indicates the maximum allowable deformation value. Rate your health. Indicators representing damage risk This indicates the proportion of the current stress to the design allowable value. This indicates the proportion of the crack width to the allowable limit. This indicates the degree to which the temperature deviates from the design temperature. This indicates the proportion of deformation or misalignment to the allowable value.

[0028] S6. Judgment; The specific process of S6 is as follows: Determine the health score range in step S5. If the health score is not less than 0.8, it means the system is healthy and no maintenance is required. Proceed directly to step S7. If the health score is between 0.5 and 0.8, it means the system is in a sub-healthy state. It is recommended to monitor and conduct regular checks. Proceed to step S7 based on experience. If the health score is less than 0.5, it means the system is in poor health and it is recommended to perform maintenance immediately.

[0029] S7. Joint Prediction and Crack Monitoring; S7a. Joint Prediction Model Building; S7b. Intelligent Crack Monitoring and Early Warning; The specific steps of S7 are as follows: Combining the health score in step S5, the system will display the degree of influence of each parameter on the health score and sort them according to the degree of influence. The system will prioritize parameters with greater influence on the health score based on the degree of influence of the parameters; Combining stress and crack monitoring data, a joint prediction model will be established to predict the occurrence time and location of cracks; The system will assess the crack development trend based on the prediction data and issue early warnings in advance; At the same time, crack development will be monitored in real time, and early warning information will be generated based on the prediction model.

[0030] S8, Data Analysis and System Optimization; S8a, Data Storage and Cloud Analysis; S8b, System Optimization and Self-Learning; The specific steps of S8 are: store all data in the cloud platform, and use big data analysis models to predict trends and assess the health changes of the pipe segments; the system performs self-optimization based on real-time data feedback, and adjusts the decision-making model to improve accuracy and adaptability.

[0031] S9. Long-term trend prediction and decision optimization; The specific steps of S9 are: The system uses historical monitoring data to analyze the health change trend of the tunnel segments and provide a basis for future maintenance.

[0032] S10, Decision Feedback and System Optimization; The specific steps of S10 are: the system automatically optimizes the decision support model based on maintenance feedback and data changes to ensure the accuracy of subsequent decisions.

[0033] This invention deploys a multi-functional fiber optic sensor array at key stress-bearing locations of underground tunnel segments. This array can simultaneously acquire multiple monitoring parameters, including stress, cracks, temperature, and deformation, overcoming the limitations of existing technologies that monitor only a single physical quantity and cannot reflect the overall health status of the tunnel segments. A unified sensor link enables multi-parameter fusion monitoring, reducing the number of sensors and improving the deployment efficiency and data consistency of the monitoring system. By comprehensively identifying the baseline response (initial zero-point wavelength and its time-varying shift), temperature drift characteristics, and long-term drift of the fiber optic sensor link, a sensor status evaluation index is constructed. Timely monitoring and calibration of the sensors during data processing effectively solves the problem of signal error accumulation caused by temperature and humidity cycles and bonding interface aging during long-term service of fiber optic sensors, improving the long-term stability of monitoring data. Historical monitoring data, fault samples, and correlation analysis or machine learning methods are used to rank the influence of multiple monitoring parameters and generate adaptive weight coefficients, enabling dynamic updates to the health rating model. This method overcomes the shortcomings of traditional monitoring systems that rely on fixed thresholds or human experience for status judgment, allowing health evaluation results to adaptively adjust with environmental and structural changes, improving the scientific rigor and accuracy of the assessment. This invention also constructs a real-time anomaly identification and early warning mechanism based on multi-parameter monitoring data. It can automatically identify various risk events such as stress exceeding limits, crack propagation, and abnormal temperature rise, and provide maintenance personnel with risk contribution information for specific parameters. Through the linkage analysis of health scores and risk information, the system can provide relevant maintenance information, offering a basis for daily inspections of underground tunnel segments, structural reinforcement, and operational risk identification, thereby enhancing the safety assurance capabilities of subway tunnels throughout their entire life cycle.

[0034] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Terms such as "upper," "lower," "left," "right," "front," and "rear" used in the invention are merely for clarity of description and are not intended to limit the scope of the invention. Changes or adjustments to their relative relationships, without substantially altering the technical content, should also be considered within the scope of the invention.

[0035] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for real-time monitoring and evaluation of the health of underground tunnel segments, characterized in that, The real-time monitoring and evaluation method includes the following steps: design and sensor deployment of the segment health monitoring system, sensor status monitoring and calibration, adaptive monitoring and intelligent early warning mechanism, health scoring system and maintenance decision, joint prediction and crack monitoring, data analysis and system optimization, long-term trend prediction and decision optimization, and decision feedback and system optimization. In the adaptive monitoring and intelligent early warning mechanism step, the system monitors parameters; if a parameter exceeds a threshold, an alarm is issued; otherwise, the process proceeds to the next step. In the health scoring system and maintenance decision step, the health score determines whether to issue an emergency maintenance recommendation, strengthen monitoring and inspection, and proceed to the next step, or proceed directly to the next step. After completing the decision feedback and system optimization step, the process returns to the sensor status monitoring and calibration step, thus creating a loop.

2. The method for real-time monitoring and evaluation of the health of underground tunnel segments according to claim 1, characterized in that, The specific process of designing and deploying sensors for the segment health monitoring system includes: integrated design of multi-functional fiber optic sensor arrays: deploying multi-functional fiber optic sensor arrays in the key stress paths and potential weak areas of each underground structural segment to simultaneously monitor stress, temperature, cracks, and deformation; sensor deployment locations: deploying sensors at key axial stress locations, key circumferential stress locations, crack-sensitive areas, and deformation-sensitive areas, with multiple multi-functional sensor array nodes deployed in each ring of segments, and each node containing multiple monitoring sub-modules.

3. The method for real-time monitoring and evaluation of the health of underground tunnel segments according to claim 2, characterized in that, In the design and sensor deployment of the segment health monitoring system, the sensor array nodes include an optical fiber demodulation module, a microprocessor, a wireless communication module, a local data cache and breakpoint retransmission module; all monitoring data are automatically uploaded to the cloud through a star or chain-type Internet of Things architecture.

4. The method for real-time monitoring and evaluation of the health of underground tunnel segments according to claim 1, characterized in that, The specific process of the sensor status monitoring and calibration steps includes: sensor health status monitoring: fiber optic sensor sensitivity monitoring, temperature sensor accuracy calibration, signal attenuation monitoring, multi-sensor data verification, and sensor accuracy correction; sensor automatic calibration and feedback mechanism: the system executes an automatic calibration program under preset standard environment or known load conditions. When an error is detected in the sensor, the system automatically corrects the output data based on a preset algorithm to ensure data accuracy. Through the Internet of Things platform, the sensor status information is fed back in real time to ensure that faulty sensors can be replaced or calibrated in a timely manner to avoid data quality degradation.

5. The method for real-time monitoring and evaluation of the health of underground tunnel segments according to claim 1, characterized in that, The specific process of the adaptive monitoring and intelligent early warning mechanism includes: multi-parameter adaptive threshold adjustment and over-limit early warning: the system adaptively adjusts the multi-parameter thresholds based on real-time monitoring data and historical data, and dynamically corrects the alarm thresholds for stress, temperature, cracks, and deformation parameters by combining the health score of the tunnel segment, crack development, and earth pressure changes. It also monitors temperature, deformation, crack development, and earth pressure parameters and sets thresholds for each. The system collects data in real time and compares it with the preset thresholds. Once a parameter exceeds the limit, an alarm is automatically triggered, and the alarm information is transmitted to relevant personnel through the Internet of Things platform.

6. The method for real-time monitoring and evaluation of the health of underground tunnel segments according to claim 5, characterized in that, The system compares the actual parameters monitored by the adaptive monitoring and intelligent early warning mechanism with the thresholds to determine the response and decision: for minor alarms, it reminds users to follow up and monitor; if there are further changes, the system will push the notification again; for emergency alarms, it immediately organizes emergency repairs.

7. The method for real-time monitoring and evaluation of the health of underground tunnel segments according to claim 1, characterized in that, The specific process of the health rating system and maintenance decision-making includes: multi-parameter influence assessment and adaptive weight determination; utilizing an IoT monitoring system to retrieve historical monitoring data within a certain time period, including axial and circumferential stress time series. and Crack width and crack growth rate w(t) , dw / dt Segment deformation or misalignment Segment surface temperature T(t) Simultaneously organize the engineering event data for the corresponding time period, constructing each time period and each monitoring loop as a sample, denoted as... and using the formula Normalization and dimensionless processing of the indicators yield a feature matrix with uniform dimensions: In the formula, k=1,2,3,4 correspond to stress, crack, temperature, and deformation, respectively. j For sample number, and These represent the time series of axial and circumferential stresses, respectively. w(t) and dw / dt This represents the relationship between crack width and crack growth rate. T(t) Indicates the surface temperature of the tube segment. This indicates segment deformation or misalignment displacement. Indicates label j The corresponding dimensionless values ​​of stress, crack, temperature, or deformation.

8. The method for real-time monitoring and evaluation of the health of underground tunnel segments according to claim 7, characterized in that, The specific process of the health scoring system and maintenance decision-making also includes: impact calculation, which includes statistical correlation method, or machine learning feature importance method, or statistical correlation combined with machine learning feature importance method; wherein statistical correlation: define risk label y, and calculate the Pearson correlation coefficient or mutual information between each indicator and risk label y. or Four influence indicators were obtained. The larger the value, the stronger the correlation between the parameter and the problem or maintenance event. Sort from largest to smallest; Machine learning feature importance method: Training a risk prediction model ,enter Output predicted risk Random forests and XGBoost can be used. After the model is trained, the "importance" index of each feature is extracted. For tree models, "feature importance scores" can be used directly; the machine learning method employed utilizes feature masks to obtain the contribution of each parameter, thus yielding... Four importance values, ranked in order; where These represent the influence indices corresponding to stress, cracks, temperature, and deformation, respectively.

9. A method for real-time monitoring and evaluation of the health of underground tunnel segments according to claim 8, characterized in that, The specific process of the health scoring system and maintenance decision-making also includes: adaptive weight calculation and normalization: using formulas... Impact index Convert to weights Therefore, there is Segment health score calculation: for the current moment t Each monitoring unit receives Four types of dimensionless indicators, using formulas Calculate the damage risk index; where This indicates the proportion of the current stress to the design allowable value. This indicates the proportion of the crack width to the allowable limit. This indicates the degree to which the temperature deviates from the design temperature. This indicates the proportion of deformation or misalignment to the allowable value. These represent the weights corresponding to stress, temperature, crack, and deformation, respectively. Indicators representing damage risk.

10. A method for real-time monitoring and evaluation of the health of underground tunnel segments according to claim 9, characterized in that, The specific process of the health scoring system and maintenance decision-making also includes: using formulas ,Right now Calculate health status when ≥0.8 indicates good health and requires no maintenance; when 0.5≤ <0.8 indicates sub-health, and monitoring is recommended; when <0.5 indicates poor health, and immediate repair is recommended; where Indicates the current Stress at any moment Indicates the current Temperature at any moment Indicates the current The cracks of time Indicates the current Transformation in a moment Indicates the maximum allowable stress value. Indicates the maximum allowable temperature value. This represents the maximum allowable crack value. Indicates the maximum allowable deformation value. Rate your health.

Images