Distributed fiber based tunnel lining thermal-mechanical coupling damage identification and positioning method and system

Abstract

The application discloses a tunnel lining thermal-mechanical coupling damage identification and positioning method and system based on a distributed optical fiber, the application is characterized in that a distributed optical fiber network is arranged along the circumferential direction and the longitudinal direction at the inner surface of the lining and the key positions between layers, and the full-distributed temperature and strain data are acquired in combination with a demodulator; the core of the application is a built-in decoupling algorithm of a deep residual network (ResNet) based on a sparrow search algorithm (SSA) optimization; the algorithm introduces residual connection, a LeakyReLU activation function and a SSA global optimization strategy to construct a high-precision nonlinear mapping model in view of the cross-sensitivity problem; the application can accurately remove the non-damage deformation caused by temperature, and restore the real mechanical strain with high fidelity, so that the early identification of the stress concentration area and the micro crack is realized; meanwhile, the thermal expansion and cold contraction and the structural actual damage are effectively distinguished, the defects of the limited monitoring range and the easy misjudgment of the traditional monitoring method are overcome, and the accurate damage positioning and reliable monitoring of the full scale of the tunnel are realized.

Description

Technical Field

[0001] This invention belongs to the field of tunnel engineering structural health monitoring and distributed optical fiber sensing technology. In particular, it relates to a method and system for performing multi-field decoupling analysis of tunnel lining structures using distributed optical fibers under complex thermo-mechanical coupling environments, thereby achieving accurate identification and spatial positioning of damage characteristics. Background Technology

[0002] Deep-buried tunnels are complex underground engineering structures situated in environments of high ground stress and high geothermal temperature. Unlike conventional tunnels, their lining structures not only bear mechanical loads from the surrounding rock compression but are also subjected to the combined effects of geothermal and concrete hydration heat fields over extended periods. During construction and operation, the lining structure experiences passive stress from the surrounding rock pressure while simultaneously experiencing internal constraint stresses due to thermal expansion and contraction caused by temperature changes. This unique thermo-mechanical coupling environment is considered a key contributing factor to tunnel structural cracking, spalling, and even instability. However, the engineering community currently lacks sufficient understanding of the lining damage evolution mechanism under such complex multi-field coupling effects and remains concerned that the combined effects of thermal and mechanical stresses may subtly weaken the structural bearing capacity, thereby causing irreversible impacts on the tunnel's life-cycle safety.

[0003] To more accurately assess the health status of tunnel linings under thermo-mechanical coupling conditions, it is essential to monitor their stress-strain distribution patterns in real time using effective monitoring methods. While extensive research has been conducted on tunnel monitoring, existing methods primarily rely on traditional point sensors such as vibrating wire strain gauges and multi-point displacement gauges. These sensors not only have monitoring blind spots and struggle to capture randomly occurring localized cracks, but also suffer from low survival rates in high-temperature and high-humidity environments. Although distributed fiber optic sensing technology has been gradually introduced in recent years due to its advantages of long distance and full distribution, most existing applications have not yet solved the "cross-sensitivity" problem of optical fibers in thermo-mechanical coupling fields—that is, the inability to effectively distinguish whether monitoring data originates from structural stress damage or simple temperature drift. Existing monitoring methods struggle to decouple and analyze thermo-mechanical coupling effects, leading to high false alarm rates and an inability to accurately locate damage sources. Therefore, it is necessary to develop a distributed fiber optic-based method for identifying and locating thermo-mechanical coupling damage in tunnel linings. This method should employ precise decoupling algorithms to eliminate temperature interference, enabling comprehensive perception and accurate location of the true damage state of the lining structure. Summary of the Invention

[0004] The purpose of this invention is to address the problems of existing methods for monitoring the health of tunnel lining structures, which mainly rely on point sensors, resulting in blind spots and difficulty in capturing localized micro-cracks. Furthermore, to address the issue that existing distributed fiber optic monitoring technologies struggle to eliminate temperature interference with mechanical strain measurements (cross-sensitivity effect) in complex geothermal environments, leading to high false alarm rates and inaccurate positioning, this invention proposes a distributed fiber optic-based method for identifying and locating thermal-mechanical coupling damage in tunnel lining structures. This method aims to achieve full-scale, high-precision damage perception and location for tunnel lining structures under high geothermal and high-stress environments.

[0005] To achieve the objectives of this invention, a method for identifying and locating thermo-mechanical coupling damage in tunnel lining based on distributed optical fibers is disclosed. Sensing optical fibers are deployed in key stress-bearing parts and deformation-sensitive areas of the tunnel lining (such as the arch crown, arch waist, and sidewalls) to form a continuous sensing loop. The sensing optical fibers include strain-sensing fibers for sensing mechanical deformation and temperature-compensating fibers for sensing ambient temperature. A distributed optical fiber demodulator is connected to the sensing optical fibers to acquire mixed frequency shift (affected by both temperature and strain) and temperature frequency shift data. The distributed optical fiber demodulator is connected to a data processing terminal, which incorporates a thermo-mechanical coupling decoupling algorithm module and a damage feature extraction module. This ultimately enables decoupling analysis of the thermo-mechanical coupling effect, eliminating non-damaging interference caused by temperature changes (such as material free expansion), and accurately identifying and locating the actual damage to the lining structure caused by the coupling effect of surrounding rock mechanical loads and temperature-induced thermal stress.

[0006] Specifically, the following steps are included:

[0007] Step 1: Under the condition that the tunnel lining structure is undamaged, read the Brillouin center frequency and temperature baseline value of each sensing fiber as the initial zero state of the method.

[0008] Step 2: Turn on the distributed fiber optic demodulator for real-time monitoring, and simultaneously acquire the hybrid frequency shift data of the strain sensing fiber and the temperature frequency shift data of the temperature compensation fiber; the data acquisition covers the entire length of the tunnel lining.

[0009] Step 3: Transmit the collected fiber optic data to the data processing terminal. Use the thermo-mechanical coupling decoupling algorithm to eliminate the free expansion deformation of the lining material caused by temperature changes and the thermal drift of the fiber refractive index, and calculate the true mechanical strain distribution of the lining structure.

[0010] Step 4: As the tunnel service environment changes, continuously monitor and calculate the strain distribution curve along the optical fiber path, and perform noise reduction and smoothing on the data.

[0011] Step 5: When the actual mechanical strain value or strain gradient at a certain location exceeds the preset safety threshold, it is determined that damage has occurred in that area.

[0012] Step 6: Based on the mapping relationship between fiber length and tunnel mileage, and combined with the geometric topology of fiber optic deployment, calculate the specific spatial coordinates (mileage and circumferential angle) of the damage point to complete the damage location.

[0013] Furthermore, the deployment path of the sensing optical fibers is optimized based on the tunnel cross-sectional geometry and the direction of the principal axis of geological stress. The sensing optical fibers construct an orthogonal grid-like topology structure with "multiple closed loops and longitudinal connections" on the lining surface. The circumferential sensing optical fibers are deployed in a closed loop along the tunnel outline, and their deployment spacing is set differently according to the stability of the surrounding rock: the spacing is increased to 0.5m-0.8m in deformation-sensitive areas such as fault fracture zones or biased sections, and is set to 1.0m-2.0m in general surrounding rock sections. The longitudinal connecting optical fibers are deployed parallel to the tunnel axis at the arch crown (0°), left and right arch waists (±45°), left and right arch feet (±135°), and invert arch (180°).

[0014] Furthermore, in step 2, the strain sensing fiber adopts a tight-buffered type and is fixed in the lining structure by bonding or pre-embedding with high-strength epoxy resin along its entire length to ensure that the fiber and the structure are coordinated in deformation; the temperature compensation fiber adopts a loose-buffered type and is arranged parallel to the strain sensing fiber, and does not bear mechanical stress.

[0015] Furthermore, in step 2, the sampling frequency and spatial resolution of the distributed fiber optic demodulator are determined according to the monitoring accuracy requirements.

[0016] Furthermore, in step 2, the distributed fiber demodulator adopts Brillouin optical time-domain reflectometry (BOTDR) or Brillouin optical time-domain analysis (BOTDA) and has dual-channel or multi-channel input interfaces, enabling it to simultaneously acquire Brillouin frequency shift signals and Raman scattering temperature signals.

[0017] Furthermore, in step 3, the thermal-mechanical coupling decoupling algorithm module adopts a deep residual network ResNet optimized based on the Sparrow Search Algorithm (SSA). By constructing a deep network topology structure including an input layer, multiple stacked residual blocks, and an output layer, a residual connection mechanism and the LeakyReLU activation function are introduced to establish a deep nonlinear mapping model between the Brillouin scattering spectral characteristics (Brillouin frequency shift and signal intensity change) of the sensing fiber and the physical field (temperature and strain). The Sparrow Search Algorithm is used to globally adaptively optimize the initial weights and biases of the neural network to overcome the gradient vanishing and local extremum problems, thereby separating the pure mechanical strain value with high accuracy. The damage feature extraction module is used to calculate the strain gradient along the fiber path.

[0018] Furthermore, in step 3, the decoupling calculation is based on the trained SSA-ResNet deep learning model; the multidimensional spectral feature vectors of the collected Brillouin frequency shift and intensity change are input into the algorithm module of the data processing terminal. Instead of relying on the principle of linear superposition, the decoupled temperature field and the real strain field data are directly output through the nonlinear deduction of the deep network, eliminating the temperature cross-sensitivity effect.

[0019] Furthermore, in step 5, the safety threshold is determined based on the ultimate tensile strain of the tunnel lining concrete. If the monitored local tensile strain exceeds the cracking limit of the concrete, it is determined that a crack has started.

[0020] To achieve the objectives of this invention, this invention also discloses a system for identifying and locating thermal-mechanical coupling damage in tunnel lining based on distributed optical fiber, comprising a distributed sensing optical fiber network deployed along the surface or interior of the tunnel lining, a distributed optical fiber demodulator connected to the sensing optical fiber, and a data processing terminal communicatively connected to the demodulator.

[0021] Compared with existing technologies, the significant advancements of this invention are: 1) It solves the problem of "cross-sensitivity" in complex thermo-mechanical coupling environments, achieving high-precision decoupling. Existing technologies mostly rely on the principle of linear superposition or traditional BP neural networks, which are difficult to distinguish between temperature and strain or are prone to getting trapped in local extrema. This invention innovatively proposes a deep residual network (ResNet) decoupling model based on the Sparrow Search Algorithm (SSA) optimization. By establishing a deep nonlinear mapping between Brillouin scattering spectral features and the physical field, it overcomes the gradient vanishing problem, eliminates temperature drift interference, and accurately separates the true mechanical strain of the lining structure; 2) It breaks through the blind zone limitation of traditional point monitoring, achieving full-scale, gridded precise damage localization; Unlike the single-point monitoring of traditional vibrating wire strain gauges, this invention adopts an orthogonal grid-like fiber optic topology structure, combined with the mapping relationship between fiber length and tunnel mileage. This layout not only covers the entire length of the tunnel and key stress-bearing parts, eliminating monitoring blind zones, but also accurately obtains the specific spatial coordinates of damage points through inversion calculation; 3) It constructs an integrated intelligent identification system of "monitoring-decoupling-localization," significantly reducing the false alarm rate. This invention integrates dual-channel synchronous acquisition (BOTDR / BOTDA), data cleaning, deep learning decoupling, and damage feature extraction modules. By embedding a pre-trained SSA-ResNet model into the data processing terminal, it can output the real stress-strain distribution in real time under changing tunnel service environments (such as high ground temperature and hydration heat), effectively avoiding misjudgments of structural damage caused by changes in environmental temperature. It is suitable for the full life cycle safety assessment of deeply buried tunnels with high ground stress.

[0022] To more clearly illustrate the functional characteristics and structural parameters of the present invention, further explanation is provided below in conjunction with the accompanying drawings and specific embodiments. Attached Figure Description

[0023] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0024] Figure 1 This is a diagram of the thermal-mechanical coupling decoupling network topology based on SSA-ResNet;

[0025] Figure 2 This is a schematic diagram of the cross-sectional structure of the sensing fiber;

[0026] Figure 3 This is a schematic diagram of the overall structure of the experimental model;

[0027] Figure 4 yes Figure 3 Cross-sectional sensor arrangement diagram of section AA;

[0028] Figure 5 yes Figure 3 Longitudinal sensor layout diagram of the tunnel in the BB section. Detailed Implementation

[0029] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and 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.

[0030] This invention discloses a method and system for identifying and locating thermo-mechanical coupling damage in tunnel lining based on distributed optical fibers. The system mainly includes a distributed sensing optical fiber network embedded or surface-mounted in the tunnel lining concrete structure, a distributed optical fiber demodulator connected to the optical fibers, and a data processing terminal. In terms of deployment, sensing optical fibers are deployed along the circumferential and longitudinal directions on the inner surface and key stress locations between layers of the lining structure, forming a fully covered sensing network. Utilizing the photosensitivity of the distributed optical fibers, the system simultaneously acquires the fully distributed temperature and strain field data of the lining structure under complex environments. Regarding core data processing, the data processing terminal incorporates a thermo-mechanical coupling effect decoupling algorithm based on a deep residual network (ResNet) optimized by the Sparrow Search Algorithm (SSA). This algorithm addresses the cross-sensitivity and nonlinear error issues of fiber optic sensors under multi-field coupling by constructing a deep nonlinear mapping model based on the relationship between spectral features (Brillouin frequency shift and intensity variation) and physical fields (temperature, strain). By introducing a residual connection mechanism and the LeakyReLU activation function into the network topology, it effectively solves the gradient vanishing and network degradation problems in the deep feature extraction process. Simultaneously, leveraging the robustness and global optimization characteristics of the sparrow search algorithm, it adaptively optimizes the initial weights and biases of the neural network, overcoming the tendency of traditional backpropagation algorithms to get trapped in local extrema. Based on this algorithm, this invention can accurately eliminate non-damaging free expansion deformation caused by temperature changes, faithfully reconstruct the true mechanical strain from mixed signals, and accurately extract local strain anomalies induced by structural damage. It enables early identification of stress concentration areas and micro-cracks, effectively distinguishing between thermal expansion and contraction effects and substantial structural damage. This overcomes the limitations of traditional point sensors, such as limited monitoring range and susceptibility to misjudgment under complex thermo-mechanical fields, achieving precise damage localization across the entire scale of tunnel lining, resulting in more comprehensive and reliable monitoring results.

[0031] Specifically, sensing optical fibers are deployed in key stress-bearing parts and deformation-sensitive areas of the tunnel lining (such as the arch crown, arch waist, and sidewalls) to form a continuous sensing loop. The sensing optical fibers include strain sensing fibers for sensing mechanical deformation and temperature compensation fibers for sensing ambient temperature. The sensing optical fibers are connected to a distributed optical fiber demodulator to acquire mixed frequency shift (affected by both temperature and strain) and temperature frequency shift data. The distributed optical fiber demodulator is connected to a data processing terminal, which has a thermo-mechanical coupling decoupling algorithm module and a damage feature extraction module embedded inside. Ultimately, it achieves decoupling analysis of the thermo-mechanical coupling effect, eliminates non-damaging interference caused by temperature changes (such as free material expansion), and accurately identifies and locates the actual damage to the lining structure caused by the coupling effect of mechanical loads and thermal stress of the surrounding rock.

[0032] Specifically, the following steps are included:

[0033] Step 1: Under the condition that the tunnel lining structure is undamaged, read the Brillouin center frequency and temperature baseline value of each sensing fiber as the initial zero state of the method.

[0034] Step 2: Turn on the distributed fiber optic demodulator for real-time monitoring, and simultaneously acquire the hybrid frequency shift data of the strain sensing fiber and the temperature frequency shift data of the temperature compensation fiber; the data acquisition covers the entire length of the tunnel lining.

[0035] Step 3: Transmit the collected fiber optic data to the data processing terminal. Use the thermo-mechanical coupling decoupling algorithm to eliminate the free expansion deformation of the lining material caused by temperature changes and the thermal drift of the fiber refractive index, and calculate the true mechanical strain distribution of the lining structure.

[0036] Step 4: As the tunnel service environment changes, continuously monitor and calculate the strain distribution curve along the optical fiber path, and perform noise reduction and smoothing on the data.

[0037] Step 5: When the actual mechanical strain value or strain gradient at a certain location exceeds the preset safety threshold, it is determined that damage has occurred in that area.

[0038] Step 6: Based on the mapping relationship between fiber length and tunnel mileage, and combined with the geometric topology of fiber optic deployment, calculate the specific spatial coordinates (mileage and circumferential angle) of the damage point to complete the damage location.

[0039] Specifically, the deployment path of the sensing optical fiber is optimized based on the tunnel cross-sectional geometry and the direction of the principal axis of geological stress. The sensing optical fiber constructs an orthogonal grid-like topology structure with "multiple closed loops and longitudinal connections" on the lining surface. The circumferential sensing optical fiber is deployed in a closed loop along the tunnel outline, and its deployment spacing is set differently according to the stability of the surrounding rock: it is densified to 0.5m-0.8m in deformation-sensitive areas such as fault fracture zones or biased sections, and set to 1.0m-2.0m in general surrounding rock sections. The longitudinal connecting optical fiber is deployed parallel to the tunnel axis at the arch crown (0°), left and right arch waists (±45°), left and right arch feet (±135°), and invert arch (180°).

[0040] Specifically, in step 2, the strain sensing fiber adopts a tight-buffered type and is fixed in the lining structure by bonding or pre-embedding with high-strength epoxy resin along its entire length to ensure that the fiber and the structure are coordinated in deformation; the temperature compensation fiber adopts a loose-buffered type and is arranged parallel to the strain sensing fiber, and does not bear mechanical stress.

[0041] Specifically, in step 2, the sampling frequency and spatial resolution of the distributed fiber optic demodulator are determined according to the monitoring accuracy requirements.

[0042] Specifically, in step 2, the distributed fiber demodulator uses Brillouin optical time-domain reflectometry (BOTDR) or Brillouin optical time-domain analysis (BOTDA) and has dual-channel or multi-channel input interfaces, enabling it to simultaneously acquire Brillouin frequency shift signals and Raman scattering temperature signals.

[0043] Specifically, in step 3, the thermal-mechanical coupling decoupling algorithm module adopts a deep residual network ResNet optimized based on the Sparrow Search Algorithm (SSA). By constructing a deep network topology structure including an input layer, multiple stacked residual blocks, and an output layer, a residual connection mechanism and the LeakyReLU activation function are introduced to establish a deep nonlinear mapping model between the Brillouin scattering spectral characteristics (Brillouin frequency shift and signal intensity change) of the sensing fiber and the physical field (temperature and strain). The initial weights and biases of the neural network are globally adaptively optimized using the Sparrow Search Algorithm to overcome the gradient vanishing and local extremum problems, thereby separating the pure mechanical strain value with high precision. The damage feature extraction module is used to calculate the strain gradient along the fiber path.

[0044] Specifically, in step 3, the decoupling calculation is based on the trained SSA-ResNet deep learning model; the multidimensional spectral feature vectors of the collected Brillouin frequency shift and intensity change are input into the algorithm module of the data processing terminal. Instead of relying on the principle of linear superposition, the decoupled temperature field and the real strain field data are directly output through the nonlinear deduction of the deep network, eliminating the temperature cross-sensitivity effect.

[0045] This paper presents an overview of the decoupling algorithm. To overcome the cross-sensitivity effect and nonlinear error of traditional fiber optic sensors in the measurement of temperature and strain as dual parameters, this study proposes a decoupling model based on a deep residual neural network (ResNet) optimized by the Sparrow Search Algorithm (SSA). This algorithm aims to solve the problems of traditional BP neural networks easily getting trapped in local minima and the degradation of deep network training by fusing deep feature extraction with a global optimization strategy, thereby achieving high-precision reconstruction of physical fields (temperature and strain).

[0046] The theoretical basis and mathematical model of the decoupling algorithm: The decoupling model constructed in this study is not a traditional fully connected feedforward network, but introduces a residual learning mechanism and combines it with a nonlinear activation function to construct a nonlinear mapping between high-dimensional spectral features and physical quantities.

[0047] The deep residual mapping mechanism introduces residual connections between hidden layers to address the vanishing and degradation problems in deep networks. Let the input of a layer be x, and the desired latent mapping be H(x). Then, the network layer learns by fitting the residual mapping F(x) = H(x) − x. Its mathematical expression is:

[0048]

[0049] in, and The first Layer and First The output feature vector of the layer, It is a non-linear activation function. This structure allows information to propagate directly across layers, significantly improving the ability to extract deep features.

[0050] To enhance the model's nonlinear expressive power and avoid the "dead ReLU" phenomenon, this study employs the LeakyReLU function. Compared to traditional sigmoid or ReLU functions, LeakyReLU preserves a small gradient in the negative half-region. (Usually taken as 0.01), ensuring the neuron's ability to continuously update under negative input:

[0051]

[0052] The algorithm design and architecture of the decoupling algorithm: In view of the strong nonlinear response characteristics of fiber optic sensors, this study designed a specific SSA-ResNet topology and training strategy.

[0053] The network topology, Input Layer (Nin=4): Input feature vector X=[ΔνB,ΔI,Tref,∇ε]T, where ΔνB is the Brillouin frequency shift change of the distributed optical fiber, ΔI is the change in Brillouin scattering signal intensity, Tref is the reference temperature field data acquired by the auxiliary temperature-sensing optical fiber, and ∇ε is the strain gradient feature calculated at the previous time step. This input vector aims to comprehensively characterize the multidimensional response state of the optical fiber under a thermo-mechanical coupling field.

[0054] Residual Hidden Layers: Four stacked residual blocks are set up with a node distribution of [512, 512, 512, 64]. This "wide-depth" structure helps to capture high-dimensional nonlinear coupling features.

[0055] Output Layer ): Output the predicted physical quantity vector Y= (Strain and temperature). For example... Figure 1 As shown.

[0056] Fusion optimization strategy of decoupling algorithms: SSA-Adam hybrid optimization. In order to balance global optimization capability and local convergence speed, this study adopts a two-stage hybrid optimization strategy:

[0057] Phase 1: SSA Global Optimization Initialization. Leveraging the robustness and fast convergence of the Sparrow Search algorithm, the initial weights W and bias b of the neural network are globally optimized using the network prediction error as the fitness function to obtain the optimal initial parameter space and avoid the model falling into local optima.

[0058] Phase Two: Adam Adaptive Gradient Descent. Based on the initial parameters optimized by SSA, the Adam optimizer is used for refined backpropagation training.

[0059] Loss Function: The mean squared error (MSE) is used as the objective function.

[0060]

[0061] Introduced Regularization term To suppress overfitting.

[0062] Training techniques: Introduce an Early Stopping mechanism (patience=20) and combine it with 5-FoldCross Validation to ensure the generalization performance of the model.

[0063] To comprehensively quantify the accuracy and reliability of decoupling algorithms, this study employs multi-dimensional evaluation metrics, including mean squared error (MSE), mean absolute error (MAE), and coefficient of determination (COP). )

[0064]

[0065] Comparative experiments show that the fusion algorithm outperforms the traditional matrix inversion method and the single BP neural network model in terms of both decoupling accuracy and convergence stability.

[0066] Specifically, in step 5, the safety threshold is determined based on the ultimate tensile strain of the tunnel lining concrete. If the monitored local tensile strain exceeds the concrete cracking limit, it is determined that a crack has started.

[0067] Example

[0068] Figure 2 This is a schematic diagram of various types of optical fibers. Figure 3 This is a schematic diagram of the overall structure of the experimental method. Figure 4 yes Figure 3 AA section in the middle, Figure 5 yes Figure 3The diagram shows the BB profile. In the attached figures, reference numeral 1 represents the tunnel's inner contour (clearance area); reference numeral 2 represents the tunnel lining structure (the monitored object, concrete material); reference numeral 3 represents the longitudinal sensing fiber optic cable (responsible for connecting the various circumferential monitoring rings and monitoring longitudinal settlement and bending); reference numeral 4 represents the distributed fiber optic demodulator (a BOTDA / DTS integrated device responsible for emitting lasers and receiving scattered signals); reference numeral 5 represents the data processing terminal (a computer with a built-in thermo-mechanical decoupling algorithm and damage identification module); reference numeral 6 represents the circumferential sensing fiber optic cable (used to tightly couple the fiber optic cable to the lining surface to transmit strain); and reference numeral 7 represents the fiber optic junction box / protection box (used for fiber optic splice protection and signal conversion).

[0069] In one embodiment, Figure 2 This is a schematic diagram of the cross-sectional structure of a preferred sensing optical fiber according to an embodiment of the present invention. The diagram details the microscopic composition and multi-layered encapsulation structure within the optical cable, specifically demonstrating a composite structure including the fiber core, cladding, buffer layer, tensile reinforcement, and outer sheath. This structure aims to illustrate that the selected optical fiber, while ensuring good optical transmission performance, possesses sufficient mechanical strength and corrosion resistance to adapt to the complex temperature and pressure environment inside the tunnel lining.

[0070] In one embodiment, Figure 3 This is a schematic diagram of the overall structure of the test model and the testing method. The main body of the test method of this invention is a model device for simulating the stress state of a deeply buried tunnel. The core structure of this device includes the inner contour of the tunnel 1 and the tunnel lining structure 2 surrounding it. In order to simulate the real engineering environment, the tunnel lining structure 2 is usually also provided with surrounding rock simulation material or loading boundary.

[0071] To achieve real-time acquisition and analysis of the health status of the tunnel lining structure 2 under thermo-mechanical coupling environment, this method is equipped with a distributed fiber optic demodulator 4 and a data processing terminal 5. The distributed fiber optic demodulator 4 is connected to the sensing fiber optic cable pre-embedded in the lining structure via fiber optic patch cords, forming a complete monitoring loop. It can acquire strain and temperature distribution data along the fiber optic cable in real time and transmit the data to the data processing terminal 5 for processing and imaging.

[0072] In one embodiment, Figure 4 yes Figure 3The cross-sectional sensor layout diagram of section AA mainly shows the circumferential monitoring network. Circumferential sensing optical fibers 6 are deployed circumferentially along the tunnel's inner contour 1 on the inner surface or inside the tunnel lining structure 2. The circumferential sensing optical fibers 6 are tightly attached to the tunnel lining structure 2 using a specific fixing method, ensuring accurate sensing of lining deformation and achieving strain transfer under thermo-mechanical coupling. This cross-sectional layout is mainly used to monitor the radial contraction, expansion deformation, and circumferential stress concentration of the tunnel structure under thermo-mechanical coupling, and to detect potential longitudinal crack opening displacement.

[0073] In one embodiment, Figure 5 yes Figure 3 The diagram of the longitudinal sensor layout in the BB section of the tunnel mainly shows the longitudinal monitoring network. In the tunnel lining structure 2 area, longitudinal sensing fibers 3 are deployed to accurately locate axial damage to the tunnel. These fibers extend along the tunnel axis and are arranged in a grid or S-shape reciprocating pattern on the bottom plate of the tunnel lining structure 2. This arrangement effectively identifies longitudinal bending and tensile cracking damage caused by uneven settlement, thermal expansion, or floor heave. All sensing fibers deployed on the lining (including the leads of longitudinal sensing fibers 3 and some circumferential sensing fibers 6) converge at the fiber optic junction box / protection box 7. This junction box / protection box 7 protects the fiber optic splices and organizes the wiring, and connects to an external distributed fiber optic demodulator 4 via an outgoing optical cable.

[0074] The specific implementation steps of this invention for identifying and locating thermo-mechanical coupling damage in tunnel linings are as follows:

[0075] Step A, Method Setup and Initial Calibration; according to Figures 3 to 5 As shown, the tunnel model is completed. Longitudinal sensing fibers 3 and circumferential sensing fibers 6 are properly laid out on the tunnel lining structure 2 and connected to the fiber optic junction box / protection box 7. The distributed fiber optic demodulator 4 is connected to the data processing terminal 5, and the optical path loss is checked to ensure there are no breaks. Before applying loads and temperature fields, the initial center frequency (or wavelength) data of all distributed fibers is read as the baseline for subsequent analysis.

[0076] Step B: Applying mechanical loads. Simulated in-situ stress is applied to the tunnel model using an external loading device. After the structural deformation stabilizes, fiber optic data is collected using a distributed fiber optic demodulator 4 to obtain the lining strain distribution under pure mechanical action.

[0077] Step C: Apply thermal loads. While maintaining or changing the mechanical loads, apply a temperature field inside the tunnel outline 1 or outside the lining to simulate high ground temperature or high temperature environment during a fire, and set different temperature gradient conditions.

[0078] Step D: Real-time monitoring of thermo-mechanical coupling data. Due to the superposition of temperature and stress fields, the fiber optic data is scanned once every predetermined time interval by the distributed fiber optic demodulator 4. The data processing terminal 5 records the frequency shift changes of the fiber optic cable at different locations in real time.

[0079] Step E, Damage Identification and Localization: The collected data is processed as follows:

[0080] Location: Utilizing the time-domain reflection principle of optical fibers, based on the distance of the anomalous frequency shift peak along the fiber, combined with... Figure 3 and Figure 4 The topological structure of the central ring sensing fiber 6 and the longitudinal sensing fiber 3 accurately locates the specific geometric position of the tunnel lining structure 2.

[0081] Identification: Analyze the abrupt change characteristics of the frequency shift curve. If a local strain surge occurs in a certain section of the longitudinal sensing fiber 3 or the circumferential sensing fiber 6 and exceeds the ultimate tensile strain of the concrete, it is determined that a crack or damage has occurred at that location.

[0082] Step F, Cyclic Testing and Evaluation: By changing the temperature conditions and mechanical load combinations, repeat steps B to E to establish a cumulative damage evolution model of tunnel lining structure 2 under thermo-mechanical coupled cyclic action.

[0083] It should be noted that, in this document, relational terms such as "first" and "second" are used only 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 process, method, article, or apparatus.

[0084] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for identifying and locating thermo-mechanical coupling damage in tunnel lining based on distributed optical fibers, characterized in that, Sensing optical fibers are deployed in key stress-bearing areas and deformation-sensitive zones of the tunnel lining to form a continuous sensing loop. The sensing optical fibers include strain sensing fibers for sensing mechanical deformation and temperature compensation fibers for sensing ambient temperature. The sensing optical fibers are connected to a distributed optical fiber demodulator to acquire mixed frequency shift and temperature frequency shift data. The distributed optical fiber demodulator is connected to a data processing terminal, which has a built-in thermo-mechanical coupling decoupling algorithm module and a damage feature extraction module. Ultimately, it achieves decoupling analysis of the thermo-mechanical coupling effect, eliminates non-damaging interference caused by temperature changes, and accurately identifies and locates the actual damage to the lining structure caused by the coupling effect of mechanical loads and temperature thermal stress in the surrounding rock.

2. The method for identifying and locating thermal-mechanical coupling damage in tunnel lining based on distributed optical fiber according to claim 1, characterized in that, Specifically, the following steps are included: Step 1: Under the condition that the tunnel lining structure is undamaged, read the Brillouin center frequency and temperature baseline value of each sensing fiber as the initial zero state of the method. Step 2: Turn on the distributed fiber demodulator for real-time monitoring and simultaneously acquire the hybrid frequency shift data of the strain sensing fiber and the temperature frequency shift data of the temperature compensation fiber. The data collection covers the entire length of the tunnel lining; Step 3: Transmit the collected fiber optic data to the data processing terminal. Use the thermo-mechanical coupling decoupling algorithm to eliminate the free expansion deformation of the lining material caused by temperature changes and the thermal drift of the fiber refractive index, and calculate the true mechanical strain distribution of the lining structure. Step 4: As the tunnel service environment changes, continuously monitor and calculate the strain distribution curve along the optical fiber path, and perform noise reduction and smoothing on the data. Step 5: When the actual mechanical strain value or strain gradient at a certain location exceeds the preset safety threshold, it is determined that damage has occurred in that area. Step 6: Based on the mapping relationship between fiber length and tunnel mileage, and combined with the geometric topology of fiber optic deployment, calculate the specific spatial coordinates of the damage point to complete the damage location.

3. The method for identifying and locating thermal-mechanical coupling damage in tunnel lining based on distributed optical fiber according to claim 1, characterized in that, The deployment path of the sensing fiber is optimized based on the tunnel cross-sectional geometry and the direction of the principal axis of geological stress. The sensing optical fibers construct an orthogonal grid-like topology structure with "multiple closed loops and longitudinal connections" on the lining surface. The circumferential sensing optical fibers are laid out in a closed loop along the tunnel outline, and their spacing is set differently according to the stability of the surrounding rock: the spacing is increased to 0.5m-0.8m in the fault fracture zone or deformation-sensitive area of ​​the biased section, and set to 1.0m-2.0m in the general surrounding rock section. The longitudinal connecting optical fibers are laid out parallel to the tunnel axis at the arch crown, left and right arch waists, left and right arch feet, and invert arch positions.

4. The method for identifying and locating thermal-mechanical coupling damage in tunnel lining based on distributed optical fiber according to claim 1, characterized in that, In step 2, the strain sensing fiber adopts a tight-buffered type and is fixed in the lining structure by bonding or pre-embedding with high-strength epoxy resin along its entire length to ensure that the fiber and the structure are coordinated in deformation; the temperature compensation fiber adopts a loose-buffered type and is arranged parallel to the strain sensing fiber, and does not bear mechanical stress.

5. The method for identifying and locating thermal-mechanical coupling damage in tunnel lining based on distributed optical fiber according to claim 1, characterized in that, In step 2, the sampling frequency and spatial resolution of the distributed fiber optic demodulator are determined according to the monitoring accuracy requirements.

6. The method for identifying and locating thermal-mechanical coupling damage in tunnel lining based on distributed optical fiber according to claim 1, characterized in that, In step 2, the distributed fiber demodulator uses Brillouin optical time-domain reflectometry (BOTDR) or Brillouin optical time-domain analysis (BOTDA) and has dual-channel or multi-channel input interfaces, enabling it to simultaneously acquire Brillouin frequency shift signals and Raman scattering temperature signals.

7. The method for identifying and locating thermal-mechanical coupling damage in tunnel lining based on distributed optical fiber according to claim 1, characterized in that, In step 3, the thermal-mechanical coupling decoupling algorithm module adopts a deep residual network ResNet optimized based on the Sparrow Search Algorithm (SSA). By constructing a deep network topology containing an input layer, multiple stacked residual blocks, and an output layer, a residual connection mechanism and the LeakyReLU activation function are introduced to establish a deep nonlinear mapping model between the Brillouin scattering spectral features of the sensing fiber and the physical field. The Sparrow Search Algorithm is used to globally adaptively optimize the initial weights and biases of the neural network, overcoming the gradient vanishing and local extremum problems, thereby separating the pure mechanical strain values ​​with high precision. The damage feature extraction module is used to calculate the strain gradient along the fiber path.

8. The method for identifying and locating thermal-mechanical coupling damage in tunnel lining based on distributed optical fiber according to claim 1, characterized in that, In step 3, the decoupling calculation is based on the trained SSA-ResNet deep learning model; the multidimensional spectral feature vectors of Brillouin frequency shift and intensity change are input into the algorithm module of the data processing terminal, and the decoupled temperature field and real strain field data are directly output through the nonlinear deduction of the deep network, eliminating the temperature cross-sensitivity effect.

9. The method for identifying and locating thermal-mechanical coupling damage in tunnel lining based on distributed optical fiber according to claim 1, characterized in that, In step 5, the safety threshold is determined based on the ultimate tensile strain of the tunnel lining concrete. If the monitored local tensile strain exceeds the concrete cracking limit, it is determined that a crack has started.

10. A system for identifying and locating thermo-mechanical coupling damage in tunnel lining based on distributed optical fiber, the system being based on the method for identifying and locating thermo-mechanical coupling damage in tunnel lining based on distributed optical fiber according to any one of claims 1-9, characterized in that, It includes a distributed sensing fiber optic network laid along the surface or inside of the tunnel lining, a distributed fiber optic demodulator connected to the sensing fiber optics, and a data processing terminal that is communicatively connected to the demodulator.

Images