Underwater tunnel structure health monitoring multi-source fusion and quality enhancement system and method

Abstract

The invention relates to the technical field of tunnel monitoring, and discloses an underwater tunnel structure health monitoring multi-source fusion and quality enhancement system and method, and the system comprises a sensing layer which is used for collecting the multi-source physical state information of a tunnel structure, and comprises the following subsystems: a structure stress state monitoring subsystem which employs a fiber bragg grating concrete strain gauge, the device is arranged at the maximum bending moment position of a tunnel roof and a longitudinal construction joint of a side wall to monitor concrete strain in real time. According to the system, a full-dimensional monitoring network covering tunnel structure stress, deformation joint multi-dimensional deformation, water leakage, differential settlement, internal water accumulation and overlying riverbed scouring and silting is constructed through cooperative arrangement of point type sensors and distributed sensing optical cables, accurate monitoring of key point positions is achieved, large-range continuous monitoring is achieved, and the monitoring precision is improved. The problem that traditional monitoring only focuses on a single dimension and has a monitoring blind area is thoroughly solved.

Description

Technical Field

[0001] This invention relates to the field of tunnel monitoring technology, specifically to a multi-source fusion and quality enhancement system and method for underwater tunnel structural health monitoring. Background Technology

[0002] As a core hub for cross-river and cross-sea transportation, the structural safety of underwater tunnels directly impacts the continuity of transportation and the safety of public life and property. Underwater tunnels operate in complex and harsh environments, facing multiple risks and challenges: the tunnel structure is prone to stress concentration and deformation due to water pressure and soil lateral pressure; the Ω-shaped waterstops at expansion joints are susceptible to aging and damage over time, leading to water leakage, which in turn corrodes the reinforcing steel and reduces the structural bearing capacity; sections traversing soft soil layers are prone to uneven settlement, resulting in abnormal pipe joint connections; low-lying areas inside the tunnel may accumulate water due to poor drainage, affecting traffic safety; and changes in the scouring and sedimentation of the overlying riverbed may alter the stress state at the tunnel roof, even causing structural damage. Failure to monitor and address these risks in a timely manner can lead to continuous degradation of structural performance, ultimately resulting in safety accidents. Therefore, constructing a precise, comprehensive, and efficient structural health monitoring system has become a core requirement for the operation and maintenance management of underwater tunnels.

[0003] While underwater tunnel structural health monitoring technology has made some progress, significant shortcomings remain in practical applications. Monitoring dimensions are limited, and blind spots exist. Traditional monitoring often focuses on single physical quantities, lacking coordinated monitoring of multiple risks such as structural stress, deformation, leakage, settlement, water accumulation, and erosion / deposition by the overlying riverbed. For example, some systems only monitor the deformation of the tunnel structure itself, neglecting the impact of overlying riverbed erosion / deposition on the tunnel's stress; or they only monitor whether leakage occurs, failing to simultaneously acquire the leakage rate and extent. This single-dimensional monitoring model struggles to reflect the coupling mechanism between the tunnel structure and its surrounding environment, easily leading to missed risk assessments and an inability to comprehensively evaluate the overall health status of the tunnel. Therefore, this invention provides a multi-source fusion and quality enhancement system and method for underwater tunnel structural health monitoring. Summary of the Invention

[0004] To address the aforementioned shortcomings of existing technologies, this invention provides a multi-source fusion and quality enhancement system and method for underwater tunnel structure health monitoring, comprehensively improving monitoring coverage, data quality reliability, diagnostic and early warning accuracy, and operation and maintenance management efficiency.

[0005] This invention provides the following technical solution: a multi-source fusion and quality enhancement system for underwater tunnel structural health monitoring, including... The perception layer, used to collect multi-source physical state information of the tunnel structure, consists of the following subsystems: Structural stress state monitoring subsystem: Fiber optic grating concrete strain gauges are used and deployed at the location of maximum bending moment in the tunnel roof and at the longitudinal construction joints of the sidewalls to monitor concrete strain in real time; Multidimensional deformation monitoring subsystem for expansion joints: Low-modal fixed-point densely distributed strain sensing optical cables are deployed along the expansion joints of the buried section of the tunnel to monitor joint misalignment, cracking and three-dimensional deformation. Expansion joint leakage monitoring subsystem: It adopts strain-sensitive leakage sensing optical cable and wet-bulb gauze-encased temperature-sensitive optical cable, which are laid along the Ω waterstop. The leakage location and leakage rate are located by the sudden change in strain or temperature drop of the optical cable when it absorbs water and expands. Structural uneven settlement monitoring subsystem: A fiber optic grating hydrostatic level is deployed in the soft soil section to monitor the relative settlement of each pipe section; Tunnel water accumulation monitoring subsystem: Fiber optic sensors based on active heating technology are deployed along the road surface in low-lying sections of the tunnel to dynamically monitor the water accumulation range inside the tunnel by identifying the temperature difference between the water and air interfaces. The overlying riverbed scouring and deposition monitoring subsystem uses an internally heated fiber optic temperature measuring tube, which is deployed in the riverbed at the top of the tunnel. It achieves real-time sensing of the thickness of the riverbed scouring and deposition and changes in water level by monitoring the temperature differences at the interfaces of sediment, water and air. The data transmission layer, used to realize the real-time acquisition and reliable transmission of monitoring data, consists of the following subsystems: Fiber optic demodulation equipment includes a dense distributed fiber optic demodulator and a fiber optic grating demodulator, which are used to demodulate the signals of distributed optical cables and point fiber optic grating sensors, respectively. Network transmission equipment: including industrial switches and 4G / 5G wireless communication modules, forming a redundant communication network that combines wired and wireless communication to upload data to cloud servers or local data centers. The data fusion and quality enhancement layer is used for fusion processing and quality optimization of multi-source heterogeneous monitoring data, and consists of the following subsystems: Data cleaning module: used to filter out noise, compensate for temperature drift, and identify and remove abnormal data; Multi-source data fusion module: Based on the spatiotemporal alignment algorithm, stress, deformation, leakage, settlement, water accumulation and scouring and silting data are fused under a unified time and space reference to construct a multi-dimensional feature dataset that reflects the coupling state of tunnel structure and environment; Quality Enhancement and Credibility Assessment Module: Employing a confidence rule base and evidence reasoning methods, this module quantitatively assesses the reliability of the fused data, thereby improving the confidence level of the monitoring results. The intelligent analysis and early warning layer is used to realize intelligent diagnosis and risk warning of tunnel structural status, and consists of the following subsystems: Structural health assessment model: Based on machine learning algorithms, a multi-parameter joint safety assessment model is established to classify and evaluate the overall safety status of tunnel structures; Dynamic early warning mechanism: It adopts a multi-level threshold early warning strategy based on interval evidence reasoning, combined with historical data trend analysis and real-time monitoring data, to achieve risk classification early warning from "normal" to "dangerous", and automatically associates and triggers corresponding response plans; The 3D visualization and interaction layer provides an immersive monitoring and management interface and consists of the following subsystems: Lightweight BIM engine: Integrates tunnel 3D information model to achieve dynamic association and bidirectional driving with monitoring data; Integrated cockpit interface: Supports real-time display of monitoring data, historical data review, early warning information push, structural health report generation and export functions, and has multiple interactive modes such as automatic inspection, fixed-point inspection and data monitoring.

[0006] Preferably, in the expansion joint leakage monitoring subsystem, the strain-sensitive leakage sensing optical cable uses a highly absorbent polymer as the encapsulation material. The axial compressive strain of the optical cable changes due to water absorption and expansion. The densely distributed fiber optic demodulator locates the strain anomaly zone by demodulating the strain distribution of the optical cable, thus determining the leakage event and approximate leakage rate. The wet-bulb gauze-encapsulated temperature-sensitive optical cable creates a temperature reduction zone at the leakage point through the evaporation and heat absorption effect of the outer wet-bulb gauze. The densely distributed fiber optic demodulator locates the temperature anomaly zone by demodulating the temperature distribution of the optical cable, thus determining the leakage location. The system is configured to comprehensively utilize the strain anomaly and temperature anomaly signals for cross-verification and precise location of leakage events.

[0007] Preferably, the machine learning algorithm used in the structural health assessment model includes at least one of support vector machine, decision tree and random forest, used to analyze the multidimensional feature dataset and realize millimeter-level and sub-millimeter-level progressive classification calculation of tunnel joint deformation and safety status classification; The dynamic early warning mechanism's multi-level threshold early warning strategy specifically divides the risk level into five levels: normal, attention, early warning, alarm, and danger, based on the assessment results and preset interval thresholds. It also automatically associates and invokes the corresponding emergency response plan preset in the three-dimensional visualization interaction layer for different levels.

[0008] Preferably, the fiber optic sensor based on active heating technology in the tunnel water accumulation range monitoring subsystem includes a thermally conductive metal baseband, a temperature-sensing optical cable spirally wound on the baseband, and a heating strip deployed at the edge of the baseband. During operation, the heating band is energized to uniformly heat the entire sensor. Due to the difference in thermal conductivity, the part submerged in water heats up at a lower rate than the part exposed to air, resulting in a distinct temperature inflection point at the water boundary. The densely distributed fiber optic demodulator identifies this temperature inflection point by analyzing the temperature distribution curve of the temperature sensing cable, achieving sub-meter level accuracy positioning of the water boundary.

[0009] Preferably, the internally heated fiber optic temperature measuring tube in the overlying riverbed scour and sedimentation monitoring subsystem is made by spirally winding a temperature sensing optical cable along the outer wall of a PVC measuring tube, and a heating conductor is encapsulated inside the measuring tube. The subsystem includes two working modes: natural temperature method and active heating method. When the natural temperature difference between the media is significant, the temperature distribution is directly measured to determine the interface. When the temperature difference is not significant, electricity is applied to the heating conductor to amplify the temperature gradient at the interface by utilizing the difference in thermal conductivity between water, sediment, and air, thereby achieving synchronous and accurate positioning and thickness change monitoring of the sediment-water interface and the water-air interface.

[0010] Preferably, the multi-source data fusion module specifically performs the following operations: Establish a unified tunnel spatial coordinate system and timestamp to align the geographical locations of all sensors with the spatiotemporal data of the monitoring data; normalize the monitoring parameters from different subsystems to eliminate the influence of dimensions; apply data association algorithms to explore the spatiotemporal coupling relationship and evolution law among multiple parameters such as stress, deformation, leakage, and settlement, and form a fusion index to describe the overall performance degradation of the structure.

[0011] Preferably, the lightweight BIM engine in the three-dimensional visualization interaction layer loads a three-dimensional information model including the tunnel civil structure, electromechanical equipment pipelines and the sensing network of the perception layer. The integrated cockpit interface dynamically associates and binds the real-time monitoring data of the perception layer, the evaluation results and early warning information of the intelligent analysis and early warning layer with the corresponding components, equipment or sensor elements in the BIM model, so as to realize one-click positioning and highlighting of the location of defects and equipment status in the three-dimensional scene.

[0012] Preferably, it also includes an operation and maintenance management and decision support layer, which is connected to the intelligent analysis and early warning layer and the three-dimensional visualization interaction layer, and specifically includes: Equipment asset database: used to manage the identity information, installation location, maintenance records, and lifespan prediction data of all sensors; Contingency Plan Management Module: Stores standardized emergency response procedures, resource allocation plans, and expert recommendations for different warning levels, allowing managers to quickly access and execute them when a warning is triggered.

[0013] A multi-source fusion and quality enhancement method for underwater tunnel structural health monitoring is described below: S1. Coordinate the deployment of fiber optic grating point sensors and densely distributed sensing optical cables to construct a multi-source sensing network covering stress, deformation, leakage, settlement, water accumulation and siltation. S2. Collect multi-source monitoring data, clean and spatiotemporally register it, and then use an evidence-based reasoning fusion method to output enhanced structural state data with a comprehensive confidence score. S3. Input the enhanced data into the pre-trained deep learning evaluation model, combine multi-level thresholds and multi-source data to make a comprehensive judgment, diagnose the structural health status and trigger graded early warnings; S4. Real-time display of monitoring data, early warning information and assessment results on the lightweight BIM 3D platform, automatic generation of inspection tasks and health reports, providing visual decision support for tunnel operation and maintenance.

[0014] Compared with the prior art, the present invention has the following beneficial effects: (1) The system constructs a comprehensive monitoring network covering tunnel structural stress, multidimensional deformation of expansion joints, water leakage, uneven settlement, internal water accumulation, and erosion and siltation of the overlying riverbed through the coordinated deployment of point sensors and distributed sensing optical cables. This achieves both precise monitoring of key points and large-scale continuous monitoring, completely solving the problem of traditional monitoring focusing only on a single dimension and having blind spots. In particular, for the core risk points unique to underwater tunnels such as leakage, riverbed erosion and siltation, and internal water accumulation, a customized sensing solution is adopted to ensure that the monitoring covers the entire scenario of the coupling between the tunnel structure and the surrounding environment, providing comprehensive data support for structural safety assessment.

[0015] (2) Through a three-level processing mechanism of data fusion and quality enhancement layers, the data is first cleaned to filter out noise, compensate for drift, and remove anomalies. Then, deep fusion of multi-source heterogeneous data is achieved through spatiotemporal alignment and correlation analysis. Finally, a confidence rule base and evidence reasoning method are used to conduct quantitative evaluation of credibility, outputting enhanced data with a comprehensive confidence score. This design effectively solves the pain points of traditional monitoring data, such as large noise interference, multi-source data conflict, and lack of quantitative basis for reliability, ensuring the accuracy and consistency of monitoring data and providing a high-quality data foundation for subsequent health diagnosis and early warning.

[0016] (3) The structural health assessment model is based on machine learning algorithms and combined with multi-source fusion enhanced data, which can realize the fine classification calculation of joint deformation and the graded evaluation of structural health status; the dynamic early warning mechanism adopts a five-level threshold strategy, and realizes graded early warning from normal to dangerous through comprehensive analysis of real-time data and historical trends. Compared with the traditional single threshold early warning mode, this design can not only accurately identify the current structural status, but also predict the risk evolution trend, and automatically associate with emergency response plans, solving the problems of low accuracy, delayed response and lack of targeted disposal guidance in traditional early warning, and greatly improving the initiative and effectiveness of risk prevention and control. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the system of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. In order to keep the following description of the embodiments of this disclosure clear and concise, detailed descriptions of known functions and known components are omitted to avoid unnecessarily obscuring the concept of the present invention.

[0019] Please see Figure 1 The system adopts a modular, hierarchical architecture, with each layer precisely linked through standardized interfaces to form a complete closed loop of "data acquisition - transmission - processing - analysis - display - decision-making". The perception layer, acting as the data acquisition terminal, acquires multi-source physical state information of the tunnel structure and surrounding environment through various fiber optic sensors and sensing cables. The data transmission layer receives data from the perception layer and achieves real-time transmission through demodulation equipment and redundant communication networks. The data fusion and quality enhancement layer cleans, fuses, and optimizes the credibility of multi-source heterogeneous data, outputting high-quality structural state data. The intelligent analysis and early warning layer performs health diagnosis and tiered early warning based on the enhanced data. The 3D visualization and interaction layer uses BIM models to dynamically link data and scenes, providing a visual management interface. The operation and maintenance management and decision support layer integrates equipment assets and contingency plan resources to support operation and maintenance decisions. The collaborative operation of each layer ensures the comprehensiveness and accuracy of monitoring data and the efficiency of decision-making.

[0020] (I) Perception Layer: Construction of Multi-Source Perception Network and Implementation of Data Acquisition The sensing layer, through the coordinated deployment of point sensors and distributed sensing optical cables, enables comprehensive monitoring of tunnel structure stress, deformation, leakage, settlement, water accumulation, and erosion and deposition of the overlying riverbed. The specific implementation of each subsystem is as follows: Structural Stress State Monitoring Subsystem: This subsystem employs fiber optic grating concrete strain gauges, precisely deployed according to the mechanical characteristics of the tunnel structure. Sensors are fixed in the area of ​​maximum bending moment in the tunnel roof and at longitudinal construction joints in the sidewalls, ensuring close contact between the sensors and the concrete structure to capture real-time strain changes caused by structural stress. The sensors convert strain signals into optical signals, which are transmitted to the data transmission layer via optical fiber, enabling real-time sensing of the structural stress state.

[0021] The multi-dimensional deformation monitoring subsystem for expansion joints utilizes low-modal, fixed-point, densely distributed strain-sensing optical cables, continuously deployed along all expansion joints in the buried section of the tunnel. The optical cables maintain the same orientation as the expansion joints and are securely fixed. When a joint experiences misalignment, cracking, or three-dimensional deformation, it induces strain changes in the sensing optical cables. The distributed optical cables capture the strain distribution along the joint's length, thereby reversing the multi-dimensional deformation state of the expansion joint and achieving precise deformation monitoring.

[0022] The expansion joint leakage monitoring subsystem employs a strain-sensitive leakage sensing optical cable and a wet-bulb gauze-encased temperature-sensitive optical cable, both deployed along the Ω-shaped waterstop to form a complementary monitoring network. The encapsulation material of the strain-sensitive leakage sensing optical cable expands upon contact with water, causing a sudden change in axial compressive strain. By demodulating the strain distribution, abnormal strain zones can be located, allowing for the determination of leakage events and leakage rates. Similarly, the outer gauze of the wet-bulb gauze-encased temperature-sensitive optical cable absorbs water, creating a temperature drop zone at the leakage point due to evaporation and heat absorption. By demodulating the temperature distribution, abnormal temperature zones can be located. The system comprehensively analyzes both strain and temperature anomaly signals, cross-validating leakage events to improve the accuracy of leakage location.

[0023] The structural uneven settlement monitoring subsystem employs fiber optic grating static levels. In sections where the tunnel traverses soft soil layers, these levels are installed at regular intervals on top of the tunnel segments to ensure they are horizontally level and have a uniform reference. By monitoring the relative elevation changes between the levels, the relative settlement of the tunnel segments is calculated, allowing for timely detection of uneven settlement trends.

[0024] The tunnel water accumulation monitoring subsystem utilizes fiber optic sensors based on active heating technology, deployed along the road surface in low-lying sections of the tunnel. Each sensor consists of a thermally conductive metal base strip, a temperature-sensing optical cable spirally wound around the base strip, and a heating strip at the edge of the base strip. During operation, the heating strip is energized, causing the sensor to heat up uniformly. Due to the difference in thermal conductivity between water and air, the sensor portion submerged in water heats up at a slower rate than its portion in air, creating a distinct temperature inflection point at the water accumulation boundary. By analyzing the temperature distribution curve of the temperature-sensing optical cable and identifying this inflection point, precise location of the water accumulation boundary can be achieved.

[0025] The overlying riverbed scour and sedimentation monitoring subsystem consists of internally heated fiber optic temperature measuring tubes deployed in the riverbed at the top of the tunnel. The tubes are constructed by spirally winding temperature-sensing optical cables along the outer wall of a PVC tube, with a heating conductor encapsulated inside. This subsystem supports two operating modes: natural temperature method and active heating method. When the natural temperature difference between sediment, water, and air is significant, the interface between each medium can be determined by directly measuring the temperature distribution. When the natural temperature difference is not significant, the heating conductor is energized to amplify the temperature gradient at the interface by utilizing the difference in thermal conductivity between different media, thereby accurately locating the sediment-water interface and the water-air interface, and achieving real-time monitoring of riverbed scour and sedimentation thickness and water level changes.

[0026] (II) Data Transmission Layer: Implementation of Demodulation and Reliable Transmission of Monitoring Data The data transmission layer is responsible for converting the optical signals collected by the sensing layer into digital signals and transmitting them to the back-end processing system through a redundant communication network. The specific implementation is as follows: Fiber Optic Demodulation Equipment Implementation: The system is configured with a densely distributed fiber optic demodulator and a fiber Bragg grating demodulator, each adapted to different types of sensor signals. The densely distributed fiber optic demodulator connects to various distributed sensing optical cables, demodulating distributed signals such as strain and temperature transmitted through the cables and outputting continuous monitoring data along the cable's length. The fiber Bragg grating demodulator connects to point sensors such as fiber Bragg grating concrete strain gauges and fiber Bragg grating hydrostatic levels, demodulating the optical signals from each sensor and outputting discrete monitoring data for each monitoring point. The two demodulators operate synchronously to ensure parallel acquisition and synchronous output of multi-source data.

[0027] Network transmission equipment implementation: The network transmission equipment consists of an industrial switch and a 4G / 5G wireless communication module, constructing a redundant communication network combining wired and wireless connections. Demodulated digital monitoring data is first transmitted to the industrial switch, and then prioritized for uploading to a cloud server or local data center via the wired network. When the wired network fails, the system automatically switches to the 4G / 5G wireless communication module to continue data transmission via the wireless network, preventing data loss. This redundant communication design ensures the reliability and real-time performance of monitoring data transmission, meeting the continuous requirements of tunnel health monitoring.

[0028] (III) Data Fusion and Quality Enhancement Layer: Data Processing and Optimization Implementation This layer cleans, fuses, and enhances the quality of multi-source heterogeneous monitoring data, outputting structural state data with confidence scores. The specific implementation is as follows: Data cleaning module implementation: The data cleaning module employs targeted processing strategies to address potential noise, temperature drift, and abnormal data in multi-source monitoring data. It filters out noise caused by environmental interference using filtering algorithms; it uses temperature compensation algorithms to eliminate the impact of temperature changes on sensor measurement accuracy and correct drift data; and based on preset reasonable thresholds and data trend analysis, it identifies and removes abnormal data caused by sensor malfunctions or transmission anomalies, ensuring the reliability of the original monitoring data.

[0029] Implementation of the multi-source data fusion module: First, a unified tunnel spatial coordinate system and timestamp are established, and the installation location coordinates of all sensors are spatiotemporally aligned with the monitoring data to ensure that data from different sources are comparable in time and space. Then, the monitoring parameters of different dimensions (such as strain, deformation, temperature, etc.) are normalized to eliminate the influence of dimensional differences. Finally, data association algorithms are applied to explore the spatiotemporal coupling relationship and evolution law between multiple parameters such as stress, deformation, leakage, and settlement, and to integrate the scattered single-source data into a multi-dimensional feature dataset that can describe the overall performance degradation of the structure, thereby achieving deep fusion of multi-source data.

[0030] Implementation of the Quality Enhancement and Credibility Assessment Module: This module employs a confidence rule base and evidence-based reasoning methods to quantitatively assess the reliability of the fused data. The confidence rule base contains various pre-defined rules based on engineering experience and historical data, used to determine the credibility of data under different monitoring scenarios. The evidence-based reasoning method comprehensively analyzes the support of multi-source data, eliminating data conflicts and uncertainties. Through this module, enhanced structured state data with a comprehensive confidence score is output, clarifying the reliability of the data and providing high-quality data support for subsequent health assessments.

[0031] (iv) Intelligent Analysis and Early Warning Layer: Implementation of Health Diagnosis and Tiered Early Warning Based on the enhanced structural state data, this layer enables intelligent diagnosis and graded early warning of the tunnel structure's health status. The specific implementation is as follows: Structural health assessment model implementation: The model employs machine learning algorithms such as support vector machines, decision trees, and random forests, and is pre-trained using historical monitoring data and structural defect cases. Enhanced data, including data fusion and the output of the quality enhancement layer, is input into the model. By analyzing multi-dimensional feature data, the model achieves precise classification and calculation of tunnel joint deformation and a graded evaluation of the overall structural health status, clearly defining the current health level of the structure.

[0032] Dynamic early warning mechanism implementation: A multi-level threshold early warning strategy based on interval evidence reasoning is adopted, with five preset risk threshold intervals: "normal, attention, warning, alarm, and danger". The system combines enhanced results of real-time monitoring data with historical data trend analysis to comprehensively determine whether the structural status exceeds the corresponding threshold interval through multi-source data: when the structural status is in the normal interval, only data is recorded; when it enters the attention or higher interval, the system automatically triggers the corresponding level of early warning, and simultaneously calls the emergency response plan in the 3D visualization interaction layer and the operation and maintenance management and decision support layer to promptly remind managers to take countermeasures.

[0033] (V) 3D Visualization Interaction Layer: Visualization and Interaction Implementation This layer provides managers with an immersive and visual monitoring and management platform through a lightweight BIM engine and integrated cockpit interface, as implemented below: Lightweight BIM Engine Implementation: The engine loads a 3D information model containing the tunnel's civil structure, electromechanical equipment pipelines, and the sensing network of the perception layer. The BIM model is then lightweighted to ensure smooth 3D display and interaction even on ordinary hardware. The engine supports dynamic association and bidirectional driving between monitoring data and the BIM model; changes in monitoring data can be fed back to the corresponding location on the model in real time, and monitoring data and equipment information at the corresponding location can be quickly queried through model operations.

[0034] Integrated cockpit interface implementation: The interface integrates multiple functional modules, including a real-time monitoring data display module, a historical data backtracking module, an early warning information push module, and a structural health report generation and export module. The interface supports multiple interactive modes, including automatic inspection, fixed-point inspection, and data monitoring: The automatic inspection mode allows users to navigate and display the status of key monitoring points in the BIM 3D scene according to a preset route; the fixed-point inspection mode allows managers to locate and highlight the corresponding defects, equipment status, and monitoring data by clicking on components, equipment, or sensor elements in the BIM model; the data monitoring mode can refresh various monitoring curves and statistical charts in real time, intuitively presenting the trend of structural status changes.

[0035] (vi) Operation and Maintenance Management and Decision Support Layer: Implementation of Operation and Maintenance and Decision Support This layer works in conjunction with the intelligent analysis and early warning layer and the 3D visualization and interactive layer to provide full-process support for tunnel operation and maintenance. The specific implementation is as follows: Equipment asset database implementation: The database stores the identity information, installation location, calibration records, maintenance history, and lifespan prediction data of all sensors. By monitoring the sensor's operating status and historical maintenance data in real time, the system automatically predicts the remaining lifespan of the sensors. When a sensor approaches its lifespan threshold or malfunctions, it promptly sends maintenance reminders to ensure the continuous and stable operation of the sensing network.

[0036] The contingency plan management module stores standardized emergency response procedures, resource allocation plans, and expert recommendations for different warning levels. When the intelligent analysis warning layer triggers a tiered warning, the module automatically calls up the corresponding level of contingency plan, which is displayed in the dashboard interface of the 3D visualization interactive layer for managers to quickly view and execute. It also supports managers in adjusting and optimizing the contingency plan according to actual conditions, improving emergency response efficiency.

[0037] Complete system workflow The underwater tunnel structure health monitoring multi-source fusion and quality enhancement system of the present invention is used to monitor tunnel health, and achieves full-process monitoring, diagnosis and decision support through the following steps; Multi-source sensing network construction: Based on the characteristics of the tunnel structure and monitoring needs, fiber optic point sensors (such as fiber optic concrete strain gauges and fiber optic hydrostatic levels) and densely distributed sensing optical cables (such as strain-sensitive leakage sensing optical cables and temperature-sensitive optical cables) are deployed in a coordinated manner to ensure that the sensors and optical cables cover key monitoring dimensions such as tunnel stress, deformation, leakage, settlement, water accumulation and erosion and siltation of the overlying riverbed, and to build a comprehensive and blind-spot-free multi-source sensing network.

[0038] Data Acquisition, Cleaning, and Fusion Enhancement: Various sensors and optical cables in the perception layer acquire multi-source monitoring data in real time. This data is then demodulated into digital signals by the fiber optic demodulation equipment in the data transmission layer and transmitted to the data fusion and quality enhancement layer via a redundant communication network. First, the data cleaning module filters out noise, compensates for temperature drift, and removes outlier data. Then, the multi-source data fusion module performs spatiotemporal alignment, normalization, and correlation analysis on the cleaned data to achieve multi-source data fusion. Finally, the quality enhancement and credibility assessment module uses evidence-based reasoning methods to output enhanced structural state data with a comprehensive confidence score.

[0039] Health Diagnosis and Tiered Early Warning: Enhanced structural status data is input into a pre-trained machine learning health assessment model. The model combines multi-source data and multi-level thresholds to make a comprehensive judgment and diagnose the health status level of the tunnel structure. If the structural status exceeds the preset threshold, the system automatically triggers the corresponding level of early warning, and simultaneously calls the emergency response plan of the operation and maintenance management and decision support layer, pushing the early warning information to management personnel through the 3D visualization interaction layer.

[0040] Visualization and Decision Support: The lightweight BIM engine in the 3D visualization and interactive layer dynamically links real-time monitoring data, early warning information, and health assessment results with the tunnel's 3D model, displaying them in real-time on the integrated dashboard interface. Managers can view relevant information through automatic and fixed-point inspection modes, and the system automatically generates inspection tasks and structural health reports. Maintenance personnel, based on the displayed results, early warning information, and invoked emergency plans, formulate and execute maintenance decisions, achieving scientific and visualized management of tunnel operation and maintenance.

[0041] The above embodiments are merely exemplary embodiments of the present invention and are not intended to limit the present invention. The scope of protection of the present invention is defined by the claims. Those skilled in the art can make various modifications or equivalent substitutions to the present invention within its spirit and scope of protection, and such modifications or equivalent substitutions should also be considered to fall within the scope of protection of the present invention.

Claims

1. A multi-source fusion and quality enhancement system for underwater tunnel structural health monitoring, characterized in that: include The perception layer, used to collect multi-source physical state information of the tunnel structure, consists of the following subsystems: Structural stress state monitoring subsystem: Fiber optic grating concrete strain gauges are used and deployed at the location of maximum bending moment in the tunnel roof and at the longitudinal construction joints of the sidewalls to monitor concrete strain in real time; Multidimensional deformation monitoring subsystem for expansion joints: Low-modal fixed-point densely distributed strain sensing optical cables are deployed along the expansion joints of the buried section of the tunnel to monitor joint misalignment, cracking and three-dimensional deformation. Expansion joint leakage monitoring subsystem: It adopts strain-sensitive leakage sensing optical cable and wet-bulb gauze-encased temperature-sensitive optical cable, which are laid along the Ω waterstop. The leakage location and leakage rate are located by the sudden change in strain or temperature drop of the optical cable when it absorbs water and expands. Structural uneven settlement monitoring subsystem: A fiber optic grating hydrostatic level is deployed in the soft soil section to monitor the relative settlement of each pipe section; Tunnel water accumulation monitoring subsystem: Fiber optic sensors based on active heating technology are deployed along the road surface in low-lying sections of the tunnel to dynamically monitor the water accumulation range inside the tunnel by identifying the temperature difference between the water and air interfaces. The overlying riverbed scouring and deposition monitoring subsystem uses an internally heated fiber optic temperature measuring tube, which is deployed in the riverbed at the top of the tunnel. It achieves real-time sensing of the thickness of the riverbed scouring and deposition and changes in water level by monitoring the temperature difference at the interface between sediment, water and air. The data transmission layer, used to realize the real-time acquisition and reliable transmission of monitoring data, consists of the following subsystems: Fiber optic demodulation equipment includes a dense distributed fiber optic demodulator and a fiber optic grating demodulator, which are used to demodulate the signals of distributed optical cables and point fiber optic grating sensors, respectively. Network transmission equipment: including industrial switches and 4G / 5G wireless communication modules, forming a redundant communication network that combines wired and wireless communication to upload data to cloud servers or local data centers. The data fusion and quality enhancement layer is used for fusion processing and quality optimization of multi-source heterogeneous monitoring data, and consists of the following subsystems: Data cleaning module: used to filter out noise, compensate for temperature drift, and identify and remove abnormal data; Multi-source data fusion module: Based on the spatiotemporal alignment algorithm, stress, deformation, leakage, settlement, water accumulation and scouring and silting data are fused under a unified time and space reference to construct a multi-dimensional feature dataset that reflects the coupling state of tunnel structure and environment; Quality Enhancement and Credibility Assessment Module: Employing a confidence rule base and evidence reasoning methods, this module quantitatively assesses the reliability of the fused data, thereby improving the confidence level of the monitoring results. The intelligent analysis and early warning layer is used to realize intelligent diagnosis and risk warning of tunnel structural status, and consists of the following subsystems: Structural health assessment model: Based on machine learning algorithms, a multi-parameter joint safety assessment model is established to classify and evaluate the overall safety status of tunnel structures; Dynamic early warning mechanism: It adopts a multi-level threshold early warning strategy based on interval evidence reasoning, combined with historical data trend analysis and real-time monitoring data, to achieve risk classification early warning from "normal" to "dangerous", and automatically associates and triggers corresponding response plans; The 3D visualization and interaction layer provides an immersive monitoring and management interface and consists of the following subsystems: Lightweight BIM engine: Integrates tunnel 3D information model to achieve dynamic association and bidirectional driving with monitoring data; Integrated cockpit interface: Supports real-time display of monitoring data, historical data review, early warning information push, structural health report generation and export functions, and has multiple interactive modes such as automatic inspection, fixed-point inspection and data monitoring.

2. The underwater tunnel structure health monitoring multi-source fusion and quality enhancement system according to claim 1, characterized in that: The strain-sensitive leakage sensing optical cable in the expansion joint leakage monitoring subsystem uses a highly absorbent polymer as the encapsulation material. The cable's axial compressive strain changes due to water absorption and expansion. The densely distributed fiber optic demodulator locates the strain anomaly zone by demodulating the cable's strain distribution, thus determining the leakage event and approximate leakage rate. The wet-bulb gauze-encapsulated temperature-sensitive optical cable utilizes the evaporative heat absorption effect of the outer wet-bulb gauze to create a temperature reduction zone at the leakage point. The densely distributed fiber optic demodulator locates the temperature anomaly zone by demodulating the cable's temperature distribution, thus determining the leakage location. The system is configured to comprehensively utilize the strain anomaly and temperature anomaly signals for cross-validation and precise location of leakage events.

3. The underwater tunnel structure health monitoring multi-source fusion and quality enhancement system according to claim 1, characterized in that: The structural health assessment model employs machine learning algorithms including at least one of support vector machine, decision tree, and random forest to analyze the multidimensional feature dataset and achieve millimeter-level and sub-millimeter-level progressive classification calculation and safety status classification of tunnel joint deformation. The dynamic early warning mechanism's multi-level threshold early warning strategy specifically divides the risk level into five levels: normal, attention, early warning, alarm, and danger, based on the assessment results and preset interval thresholds. It also automatically associates and invokes the corresponding emergency response plan preset in the three-dimensional visualization interaction layer for different levels.

4. The underwater tunnel structure health monitoring multi-source fusion and quality enhancement system according to claim 1, characterized in that: The fiber optic sensor based on active heating technology in the tunnel water accumulation range monitoring subsystem includes a thermally conductive metal baseband, a temperature-sensing optical cable spirally wound on the baseband, and a heating strip deployed at the edge of the baseband. During operation, the heating band is energized to uniformly heat the entire sensor. Due to the difference in thermal conductivity, the part submerged in water heats up at a lower rate than the part exposed to air, resulting in a distinct temperature inflection point at the water boundary. The densely distributed fiber optic demodulator identifies this temperature inflection point by analyzing the temperature distribution curve of the temperature sensing cable, achieving sub-meter level accuracy positioning of the water boundary.

5. The underwater tunnel structure health monitoring multi-source fusion and quality enhancement system according to claim 1, characterized in that: The internally heated fiber optic temperature measuring tube in the overlying riverbed scour and sedimentation monitoring subsystem is made by spirally winding a temperature sensing optical cable along the outer wall of a PVC measuring tube, and a heating conductor is encapsulated inside the measuring tube. The subsystem includes two working modes: natural temperature method and active heating method. When the natural temperature difference between the media is significant, the temperature distribution is directly measured to determine the interface. When the temperature difference is not significant, electricity is applied to the heating conductor to amplify the temperature gradient at the interface by utilizing the difference in thermal conductivity between water, sediment, and air, thereby achieving synchronous and accurate positioning and thickness change monitoring of the sediment-water interface and the water-air interface.

6. The underwater tunnel structure health monitoring multi-source fusion and quality enhancement system according to claim 1, characterized in that, The multi-source data fusion module performs the following operations: Establish a unified tunnel spatial coordinate system and timestamp to align the geographical locations of all sensors with the spatiotemporal data of the monitoring data; normalize the monitoring parameters from different subsystems to eliminate the influence of dimensions; apply data association algorithms to explore the spatiotemporal coupling relationship and evolution law among multiple parameters such as stress, deformation, leakage, and settlement, and form a fusion index to describe the overall performance degradation of the structure.

7. The underwater tunnel structure health monitoring multi-source fusion and quality enhancement system according to claim 1, characterized in that: The lightweight BIM engine in the three-dimensional visualization interaction layer loads a three-dimensional information model containing the tunnel civil structure, electromechanical equipment pipelines, and the sensing network of the perception layer. The integrated cockpit interface dynamically associates and binds the real-time monitoring data of the perception layer, the evaluation results and early warning information of the intelligent analysis and early warning layer with the corresponding components, equipment or sensor elements in the BIM model, so as to realize one-click positioning and highlighting of the location of defects and equipment status in the three-dimensional scene.

8. The underwater tunnel structure health monitoring multi-source fusion and quality enhancement system according to claim 1, characterized in that, It also includes an operation and maintenance management and decision support layer, which is connected to the intelligent analysis and early warning layer and the 3D visualization and interaction layer, and specifically includes: Equipment asset database: used to manage the identity information, installation location, maintenance records, and lifespan prediction data of all sensors; Contingency Plan Management Module: Stores standardized emergency response procedures, resource allocation plans, and expert recommendations for different warning levels, allowing managers to quickly access and execute them when a warning is triggered.

9. A multi-source fusion and quality enhancement method for underwater tunnel structural health monitoring, characterized in that, The underwater tunnel structure health monitoring multi-source fusion and quality enhancement system according to any one of claims 1-8 is operated as follows: S1. Coordinate the deployment of fiber optic grating point sensors and densely distributed sensing optical cables to construct a multi-source sensing network covering stress, deformation, leakage, settlement, water accumulation and siltation. S2. Collect multi-source monitoring data, clean and spatiotemporally register it, and then use an evidence-based reasoning fusion method to output enhanced structural state data with a comprehensive confidence score. S3. Input the enhanced data into the pre-trained deep learning evaluation model, combine multi-level thresholds and multi-source data to make a comprehensive judgment, diagnose the structural health status and trigger graded early warnings; S4. Real-time display of monitoring data, early warning information and assessment results on the lightweight BIM 3D platform, automatic generation of inspection tasks and health reports, providing visual decision support for tunnel operation and maintenance.

Images