Tunnel toughness implementation method and system based on dynamic learning and optimization

Abstract

The invention belongs to the technical field of tunnels, and discloses a tunnel toughness realization method and system based on dynamic learning and optimization, and the tunnel structure realizes the fundamental transformation of tunnel operation and maintenance from passive response to active prevention. Through multidisciplinary cross innovation, bionics, intelligent materials, deep learning and other technologies are applied to tunnel engineering, a systematic solution is provided for the tunnel engineering, and development and transformation of the tunnel engineering technology are promoted. And remarkable economic benefits are created: through active prevention and self-repairing functions, the maintenance cost of the whole life cycle of the tunnel can be remarkably reduced. And the frequent maintenance and large-scale improvement cost caused by diseases is reduced. Meanwhile, the intelligent and automatic operation and maintenance mode greatly improves the operation and maintenance efficiency, reduces the manual intervention and traffic interruption time, reduces the influence on the society and economy, and creates remarkable economic benefits.

Description

Technical Field

[0001] This invention belongs to the field of tunnel technology, and particularly relates to a method and system for achieving tunnel resilience based on dynamic learning and optimization. Background Technology

[0002] Existing technology 1: CN104852980A ("Tunnel Monitoring System")

[0003] The patent discloses a tunnel monitoring system, which has a multi-level preprocessing unit and an information processing unit, and emphasizes that the system has the characteristics of "strong self-healing ability" and "easy expansion and maintenance".

[0004] Existing technical problems:

[0005] 1. Although "self-healing ability" is mentioned, it mainly focuses on the robustness and self-organizing network of the monitoring system's software / hardware, and does not cover the active repair function after cracks or deformations occur in the structural body (tunnel lining).

[0006] 2. Although this system has abundant monitoring data, it does not explicitly propose a scheme for controlling the inflow of grouting or automatic repair materials based on a deep learning decision module, and lacks closed-loop control for structural repair.

[0007] 3. Its monitoring system functions are biased towards "detection and early warning" and lack an execution module for "structural adaptive repair or reinforcement", which is still different from the "grouting module" and "self-healing material" emphasized in this invention.

[0008] 4. Although there is room for expansion, the actual disclosure does not reflect a comprehensive consideration of crack propagation, material damage evolution, environmental-structural coupling (the effects of temperature and humidity), and the response mechanism of self-healing materials.

[0009] Existing technology 2: CN110847972A ("Tunnel Safety Detection and Early Warning System for Subway Construction")

[0010] This patent discloses a safety detection and early warning system for tunnels during subway construction, which includes video image monitoring, sensor data acquisition, communication module, control module, data storage module, and alarm module.

[0011] Existing technical problems:

[0012] 1. This patent mainly targets safety monitoring during the construction phase (such as personnel safety, equipment status, dust exceeding standards, etc.), focusing more on construction safety management rather than the structural health and repair of the tunnel structure during its long-term operation.

[0013] 2. The system focuses on "monitoring + early warning" rather than "structural reinforcement / self-repair", and does not involve grouting reinforcement, self-healing material injection, or the decision module's instruction and feedback mechanism for repair execution.

[0014] 3. Although it has sensing and video monitoring capabilities, it discloses relatively little information on structural stress, strain, environmental coupling (temperature and humidity), and long-term deformation trends, which is insufficient to support the need for structural toughness enhancement.

[0015] 4. The lack of a grouting module or similar automatic maintenance execution module makes it impossible to form a closed-loop control from "monitoring-decision-execution-feedback", which does not overlap with the "adaptive intelligent grouting system" sought in this invention.

[0016] The above analysis shows that although the two existing technologies have made some progress in the field of tunnel monitoring and early warning systems, neither of them covers the complete closed loop of "structural toughness enhancement", "intelligent decision-driven grouting repair" and "self-healing material injection" emphasized by this invention. Therefore, there are still technological gaps in real-time repair, intelligent decision-making and comprehensive operation and maintenance of structural health. Summary of the Invention

[0017] To address the problems existing in the prior art, this invention provides a method and system for achieving tunnel toughness based on dynamic learning and optimization.

[0018] This invention is implemented as follows: a tunnel resilience realization system based on dynamic learning and optimization includes:

[0019] The stress monitoring module, strain monitoring module, and temperature and humidity monitoring module are connected to the central control module to collect real-time data on structural stress, deformation, environmental erosion, and damage evolution, generating a multi-dimensional monitoring dataset containing "spatial damage characteristics + temporal evolution trend."

[0020] The central control module is connected to the decision-making module and the grouting module to coordinate the monitoring and control process;

[0021] The decision-making module analyzes monitoring data based on a deep learning model. The input consists of 200 sets of continuous time-series data output from the stress monitoring, strain monitoring, and temperature and humidity monitoring modules. A CNN layer extracts spatial features such as crack distribution and stress concentration, a BiLSTM layer captures the evolution of damage over time, and an attention layer dynamically assigns feature weights. The output is a tunnel toughness index and repair strategy. When the toughness index is <0.6, grouting parameters are automatically generated, and the model supports dynamic adjustment of hyperparameters based on historical repair results. Upon receiving the instructions, the grouting module automatically executes the grouting operation, injecting cement-based intelligent self-healing materials into the abnormal areas to achieve automatic reinforcement and self-repair of the structure.

[0022] Furthermore, the central control module:

[0023] The control involves deploying a distributed sensor network at potential weak points in the tunnel, including various types of sensors such as stress sensors, strain sensors, humidity sensors, and temperature sensors; the sensors are connected via wired or wireless means to form a sensing network covering the entire tunnel structure;

[0024] The system can sense the stress, strain, humidity, temperature and other state parameters of the tunnel structure in real time, and transmit the collected data to a deep learning-based decision-making system. Through real-time analysis of this data, abnormal changes in the tunnel structure can be detected in a timely manner, providing a basis for subsequent decision-making and repair.

[0025] Furthermore, the grouting module:

[0026] (1) Anomaly detection and instruction reception

[0027] The biomimetic intelligent sensing system monitors the tunnel structure in real time. When it detects any abnormalities in the tunnel structure (including cracks or excessive deformation), it transmits the abnormal information to the central control module.

[0028] The deep learning-based decision system analyzes and judges abnormal information. If it determines that grouting repair is needed, it issues a command to start the adaptive intelligent grouting system and transmits the command to the central control module. The central control module then forwards the command to the grouting module.

[0029] (2) Grouting system start-up

[0030] After receiving the start command, the grouting module automatically starts the adaptive intelligent grouting system; this includes starting the grouting pump and mixing equipment to ensure that the grouting system is in normal working condition;

[0031] (3) Material preparation

[0032] According to the preset formula and proportion, various raw materials (including cement, additives, and water) of cement-based intelligent self-healing material are put into a mixing device for thorough mixing to prepare grouting material that meets the requirements; during the mixing process, the mixing time and speed parameters must be strictly controlled to ensure the uniformity and performance of the material.

[0033] (4) Grouting pipeline connection and inspection

[0034] Connect the grouting pipeline to the grouting pump and the abnormal parts; during the connection process, ensure that the pipeline connection is tight and there is no leakage.

[0035] Conduct a comprehensive inspection of the grouting pipeline, including its unobstructed flow and sealing. A pressure test can be performed by injecting a small amount of water into the pipeline to check for leaks or blockages. If any problems are found, repair them promptly.

[0036] (5) Precision grouting

[0037] Start the grouting pump and inject the prepared cement-based intelligent self-healing material into the abnormal parts of the tunnel structure through the grouting pipeline; during the grouting process, it is necessary to make precise control according to the actual situation of the abnormal parts (including crack size and depth) and the preset grouting parameters (including grouting pressure, grouting volume and grouting speed);

[0038] Real-time monitoring of various parameters during the grouting process, including grouting pressure and grouting volume; if abnormal parameters are detected, timely adjustment of the grouting pump's operating status to ensure a stable and reliable grouting process;

[0039] (6) Grouting completion and pipeline cleaning

[0040] When the preset grouting volume or grouting pressure is reached, stop the grouting pump; observe the grouting situation at abnormal locations and confirm whether the grouting effect meets the requirements;

[0041] After grouting is completed, the grouting pipeline should be cleaned in a timely manner; any remaining grouting material in the pipeline should be drained, and the pipeline should be rinsed with clean water to prevent the grouting material from solidifying in the pipeline and affecting the next use.

[0042] Furthermore, the decision-making module:

[0043] 1) Data reception and preprocessing

[0044] The decision-making module is connected to the central control module and receives real-time tunnel structure status data transmitted by the bionic intelligent sensing system, including stress, strain, displacement, and temperature.

[0045] The received data is preprocessed, including data cleaning (removing noise and outliers) and data normalization (converting data of different units to a unified unit) to improve data quality and usability.

[0046] Data cleaning - removing noise and outliers

[0047] Moving average method: used to smooth data and remove high-frequency noise; for a set of data arranged in chronological order. ;

[0048] Its m-th order moving average The calculation formula is:

[0049] in Indicates rounding down;

[0050] 2) Tunnel Model Construction and Update

[0051] Based on the tunnel's design parameters and geological conditions, a three-dimensional model of the tunnel is constructed; this model should accurately reflect the tunnel's structural characteristics and mechanical properties.

[0052] As tunnel construction and operation progress, actual data is continuously collected to update and optimize the tunnel model in order to improve its accuracy and reliability.

[0053] 3) Intelligent simulation and analysis

[0054] The preprocessed data is input into the tunnel model for intelligent simulation and analysis; the structural response of the tunnel under different working conditions is simulated, including stress and deformation.

[0055] Advanced numerical simulation methods (including finite element method and boundary element method) and artificial intelligence algorithms (including neural network and genetic algorithm) are used to conduct in-depth analysis of simulation results and uncover the patterns and potential problems behind the data.

[0056] 4) Operation and maintenance strategy formulation

[0057] Based on the results of intelligent simulation and analysis, determine whether there are any abnormalities in the tunnel structure and the severity of the abnormalities;

[0058] If it is determined that the adaptive intelligent grouting system needs to be activated, the grouting parameter settings should be further determined, including grouting pressure, grouting volume, and grouting material type. The parameter settings should take into account the actual situation of abnormal parts, the analysis results of the tunnel model, and previous experience data.

[0059] 5) Strategy evaluation and optimization, strategy transmission and execution.

[0060] Furthermore, the strategy is evaluated and optimized:

[0061] The established operation and maintenance strategies are evaluated to analyze their feasibility and effectiveness; the advantages and disadvantages of various strategies can be compared by simulating the tunnel structure response under different strategies.

[0062] Based on the assessment results, the operation and maintenance strategy will be optimized and adjusted; the strategy will be continuously improved to enhance the efficiency and safety of tunnel operation and maintenance.

[0063] Furthermore, the policy transmission and execution are as follows:

[0064] The established operation and maintenance strategy is transmitted to the central control module, which then forwards the relevant instructions to the corresponding execution modules to guide the modules to work together and achieve scientific operation and maintenance of the tunnel.

[0065] Another objective of this invention is to provide a method for achieving tunnel resilience based on dynamic learning and optimization, comprising:

[0066] Step 1: Monitor tunnel stress data using stress sensors via a stress monitoring module;

[0067] Step 2: Monitor tunnel deformation data using strain sensors via strain monitoring module;

[0068] Step 3: Monitor tunnel temperature and humidity data using temperature and humidity sensors via a temperature and humidity monitoring module;

[0069] Step 4: When the bionic intelligent sensing system detects an anomaly in the tunnel structure, the central control module, through the grouting module, issues a command to the deep learning-based decision-making system. The adaptive intelligent grouting system is then automatically activated, injecting cement-based intelligent self-healing material into the abnormal area through the grouting pipeline to complete precise grouting reinforcement and achieve self-repair of the tunnel.

[0070] Step 5: Receive data transmitted from the biomimetic intelligent sensing system through the decision module, and perform intelligent simulation and analysis in conjunction with the tunnel model; based on the analysis results, formulate corresponding operation and maintenance strategies, such as whether to start the adaptive intelligent grouting system and the parameter settings for grouting;

[0071] Step 6: Use the communication module to connect to the Internet via wireless communication equipment for network communication.

[0072] Another object of the present invention is to provide a computer device including a memory and a processor, the memory storing a computer program, which, when executed by the processor, causes the processor to perform the steps of the tunnel resilience implementation method based on dynamic learning and optimization.

[0073] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the tunnel resilience implementation method based on dynamic learning and optimization.

[0074] Another objective of this invention is to provide an information data processing terminal for implementing the tunnel resilience realization system based on dynamic learning and optimization.

[0075] Based on the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solution to be protected by this invention are as follows:

[0076] (1) Promote technological change in the industry: realize the fundamental transformation of tunnel operation and maintenance from "passive response" to "proactive prevention"; through multidisciplinary cross-innovation, provide systematic solutions for tunnel engineering.

[0077] (2) Create significant economic benefits: Through proactive prevention and self-repair, the maintenance cost of the tunnel throughout its entire life cycle can be significantly reduced; intelligent and automated operation and maintenance can greatly improve efficiency and reduce manual intervention and traffic interruption time.

[0078] (3) Enhance engineering and social value: enhance the service resilience and durability of tunnels in complex environments, and ensure public safety; meet the huge demand for treatment of defects such as water leakage in domestic infrastructure construction, and have broad market prospects.

[0079] (4) Bionic intelligent sensing system ("sensory nerve"): Deploy a distributed sensor network in the potential weak links of the tunnel to sense the structural status in real time.

[0080] (5) Adaptive intelligent grouting system ("self-repair mechanism"): integrates cement-based intelligent self-repairing materials and automatic grouting equipment; when the sensor detects an abnormality, the system can automatically start and complete precise grouting reinforcement.

[0081] (6) Deep learning-based decision-making system ("intelligent brain"): Utilizes technologies such as digital twins to establish tunnel models for intelligent simulation and decision-making. The system can learn from historical data and disaster experience, continuously optimize the decision-making model, and achieve enhanced resilience.

[0082] (7) Modularization and Precision ("Precision Diagnosis and Treatment"): Adopting a modular and scalable architecture, it can flexibly adapt to different engineering needs. Through technologies such as three-dimensional monitoring of grout diffusion, the visualization and precise control of grout diffusion can be achieved.

[0083] The expected benefits and commercial value of the technical solution of this invention after transformation are as follows:

[0084] Creating significant economic benefits: Through proactive prevention and self-repair, the maintenance cost of the tunnel throughout its entire life cycle can be significantly reduced; intelligent and automated operation and maintenance can greatly improve efficiency and reduce manual intervention and traffic interruption time.

[0085] Enhancing engineering and social value: improving the tunnel's resilience and durability in complex environments, ensuring public safety; meeting the huge demand for the treatment of defects such as water leakage in domestic infrastructure construction, with broad market prospects. Attached Figure Description

[0086] Figure 1 This is a block diagram of a method and system for achieving tunnel resilience based on dynamic learning and optimization, provided in an embodiment of the present invention.

[0087] Figure 2 This is a flowchart of the grouting module method provided in an embodiment of the present invention.

[0088] Figure 3This is a flowchart of the decision module method provided in an embodiment of the present invention.

[0089] Figure 4 This is a flowchart of a method and system for achieving tunnel resilience based on dynamic learning and optimization, provided in an embodiment of the present invention.

[0090] Figure 1 The module consists of: 1. Stress monitoring module; 2. Strain monitoring module; 3. Temperature and humidity monitoring module; 4. Central control module; 5. Grouting module; 6. Decision module; 7. Communication module. Detailed Implementation

[0091] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0092] The tunnel resilience implementation method and system based on dynamic learning and optimization in this embodiment consists of an embedded hardware system and an intelligent decision-making software system. The hardware includes a distributed sensor network, a central controller, a grouting pump system, and a communication gateway; the software includes a deep learning analysis platform and an adaptive control program. All modules form a closed-loop system via industrial Ethernet and wireless communication, achieving full automation from tunnel condition perception to self-repair control. The system is installed in the inner lining and arch structure of the subway tunnel section and can operate online for extended periods.

[0093] Fiber optic strain sensors, piezoelectric stress sensors, and temperature and humidity sensors are deployed at key locations in the tunnel to collect structural stress, displacement, and environmental parameters in real time. Data from all sensing nodes is initially processed by edge computing nodes and then transmitted to the central control module at a sampling frequency of 10 times per second. The central control module synchronously uploads the data to a cloud database, providing training and inference input for subsequent deep learning algorithms.

[0094] The decision-making algorithm in this embodiment is based on a hybrid model of convolutional neural networks and long short-term memory networks. The convolutional layers are responsible for extracting spatial features, while the long short-term memory layers are used to identify the changing trends of the time series and output a structural safety index. When the safety index is below 0.3, the algorithm determines the coordinates of the abnormal location and the risk level through an attention weight mechanism, and automatically generates grouting parameters, including pressure, flow rate, and duration, to achieve a fully automatic closed loop of "perception-judgment-control".

[0095] Upon receiving a repair command, the central control module activates the grouting module. The mixing device blends the cement-based intelligent self-healing material in a specific ratio, maintaining uniform flow. The system controls a servo grouting pump based on grouting parameters output by the algorithm, precisely injecting the material into the affected area via a dedicated grouting pipeline. The entire grouting process monitors pressure and flow curves in real time and adaptively adjusts the grouting speed based on feedback, ensuring the repair material fully fills the cracks and achieves precise reinforcement.

[0096] After grouting is completed, the central control module feeds back the repaired sensor data to the cloud. The deep learning model automatically updates the weight parameters based on the differences between the data before and after repair, achieving continuous learning and self-optimization. The system outputs a tunnel structural health assessment report through a visualization platform, providing a scientific basis for subsequent operation and maintenance decisions. Actual testing has shown that the system demonstrates a significant improvement in toughness during tunnel micro-crack repair and long-term structural health maintenance.

[0097] like Figure 1 As shown in the figure, an embodiment of the present invention provides a method and system for achieving tunnel resilience based on dynamic learning and optimization, comprising:

[0098] Stress monitoring module 1, strain monitoring module 2, temperature and humidity monitoring module 3, central control module 4, grouting module 5, decision-making module 6, communication module 7;

[0099] Stress monitoring module 1, connected to central control module 4, is used to monitor tunnel stress data through stress sensors;

[0100] Strain monitoring module 2, connected to central control module 4, is used to monitor tunnel deformation data through strain sensors;

[0101] Temperature and humidity monitoring module 3 is connected to central control module 4 and is used to monitor tunnel temperature and humidity data through temperature and humidity sensors.

[0102] The central control module 4 is connected to the stress monitoring module 1, strain monitoring module 2, temperature and humidity monitoring module 3, grouting module 5, decision-making module 6, and communication module 7, and is used to control the normal operation of each module.

[0103] Grouting module 5, connected to central control module 4, is used to issue instructions based on deep learning when the bionic intelligent sensing system detects an abnormality in the tunnel structure. The adaptive intelligent grouting system will automatically start, injecting cement-based intelligent self-healing material into the abnormal part through the grouting pipeline to complete precise grouting reinforcement and realize the tunnel's self-repair.

[0104] The decision module 6, connected to the central control module 4, is used to receive data transmitted by the bionic intelligent sensing system, perform intelligent simulation and analysis in conjunction with the tunnel model, and formulate corresponding operation and maintenance strategies based on the analysis results, such as whether to start the adaptive intelligent grouting system and the parameter settings for grouting.

[0105] Communication module 7, connected to central control module 4, is used to connect to the Internet for network communication via wireless communication equipment.

[0106] The tunnel resilience realization system based on dynamic learning and optimization in this embodiment consists of multiple collaborative modules, including a stress monitoring module 1, a strain monitoring module 2, a temperature and humidity monitoring module 3, a central control module 4, a grouting module 5, a decision-making module 6, and a communication module 7. These modules form a closed-loop control system via industrial Ethernet and wireless communication networks. The system deploys various types of sensors at key stress-bearing locations such as the tunnel arch, sidewalls, and floor slab to achieve full-space, multi-dimensional data perception. The central control module 4, as the core node, is responsible for data acquisition, command distribution, and logic control, ensuring the coordinated consistency of monitoring, decision-making, and repair processes.

[0107] The stress monitoring module 1 employs a distributed fiber optic stress sensor array to measure the axial and circumferential stress changes in the tunnel lining in real time; the strain monitoring module 2 uses grating strain gauges to acquire micro-deformation information; and the temperature and humidity monitoring module 3 is deployed in the tunnel lining and the surface transition zone to sense changes in environmental temperature and humidity and predict the material performance degradation trend. All sensor data are synchronously acquired, noise filtered, and feature extracted by the central control module 4 to form a multi-dimensional state parameter set, providing high-precision input for subsequent intelligent decision-making.

[0108] The decision module 6 embeds a deep learning-based multimodal fusion algorithm model, integrating convolutional neural networks and temporal memory network structures, to identify potential structural anomalies from stress, strain, and environmental parameters. When the risk index calculated by the algorithm exceeds a threshold, it automatically generates grouting control commands and transmits them to the central control module 4. The central control module 4 then triggers the grouting module 5, simultaneously setting the grouting pressure, flow rate, and duration, achieving data-driven active control.

[0109] Upon receiving the command, grouting module 5 automatically starts the grouting pump and mixing unit, prepares cement-based intelligent self-healing material according to the proportions issued by central control module 4, and precisely delivers the material to the abnormal area through grouting pipelines. During the grouting process, the system monitors the pressure and flow curves in real time and automatically adjusts the grouting parameters based on feedback to ensure the material fully penetrates the cracks and forms a dense structure. After repair, communication module 7 uploads the pre- and post-repair monitoring data to the cloud platform, and decision module 6 updates the model weights based on the feedback results, achieving a self-learning and self-optimizing closed-loop maintenance system for tunnel toughness.

[0110] The central control module provided in this embodiment of the invention:

[0111] The control involves deploying a distributed sensor network at potential weak points in the tunnel, including various types of sensors such as stress sensors, strain sensors, humidity sensors, and temperature sensors; the sensors are connected via wired or wireless means to form a sensing network covering the entire tunnel structure;

[0112] The system can sense the stress, strain, humidity, temperature and other state parameters of the tunnel structure in real time, and transmit the collected data to a deep learning-based decision-making system. Through real-time analysis of this data, abnormal changes in the tunnel structure can be detected in a timely manner, providing a basis for subsequent decision-making and repair.

[0113] like Figure 2 As shown, the grouting module provided in this embodiment of the invention:

[0114] S101, Anomaly Detection and Command Reception

[0115] The biomimetic intelligent sensing system monitors the tunnel structure in real time. When it detects any abnormalities in the tunnel structure (including cracks or excessive deformation), it transmits the abnormal information to the central control module.

[0116] The deep learning-based decision system analyzes and judges abnormal information. If it determines that grouting repair is needed, it issues a command to start the adaptive intelligent grouting system and transmits the command to the central control module. The central control module then forwards the command to the grouting module.

[0117] S102, Grouting system started.

[0118] After receiving the start command, the grouting module automatically starts the adaptive intelligent grouting system; this includes starting the grouting pump and mixing equipment to ensure that the grouting system is in normal working condition;

[0119] S103, Material Preparation

[0120] According to the preset formula and proportion, various raw materials (including cement, additives, and water) of cement-based intelligent self-healing material are put into a mixing device for thorough mixing to prepare grouting material that meets the requirements; during the mixing process, the mixing time and speed parameters must be strictly controlled to ensure the uniformity and performance of the material.

[0121] S104 Grouting Pipeline Connection and Inspection

[0122] Connect the grouting pipeline to the grouting pump and the abnormal parts; during the connection process, ensure that the pipeline connection is tight and there is no leakage.

[0123] Conduct a comprehensive inspection of the grouting pipeline, including its unobstructed flow and sealing. A pressure test can be performed by injecting a small amount of water into the pipeline to check for leaks or blockages. If any problems are found, repair them promptly.

[0124] S105, Precision Grouting

[0125] Start the grouting pump and inject the prepared cement-based intelligent self-healing material into the abnormal parts of the tunnel structure through the grouting pipeline; during the grouting process, it is necessary to make precise control according to the actual situation of the abnormal parts (including crack size and depth) and the preset grouting parameters (including grouting pressure, grouting volume and grouting speed);

[0126] Real-time monitoring of various parameters during the grouting process, including grouting pressure and grouting volume; if abnormal parameters are detected, timely adjustment of the grouting pump's operating status to ensure a stable and reliable grouting process;

[0127] S106, Grouting completion and pipeline cleaning

[0128] When the preset grouting volume or grouting pressure is reached, stop the grouting pump; observe the grouting situation at abnormal locations and confirm whether the grouting effect meets the requirements;

[0129] After grouting is completed, the grouting pipeline should be cleaned in a timely manner; any remaining grouting material in the pipeline should be drained, and the pipeline should be rinsed with clean water to prevent the grouting material from solidifying in the pipeline and affecting the next use.

[0130] like Figure 3 As shown, the decision module provided in this embodiment of the invention:

[0131] S201, Data Reception and Preprocessing

[0132] The decision-making module is connected to the central control module and receives real-time tunnel structure status data transmitted by the bionic intelligent sensing system, including stress, strain, displacement, and temperature.

[0133] The received data is preprocessed, including data cleaning (removing noise and outliers) and data normalization (converting data of different units to a unified unit) to improve data quality and usability.

[0134] Data cleaning - removing noise and outliers

[0135] Moving average method: used to smooth data and remove high-frequency noise; for a set of data arranged in chronological order. ;

[0136] Its m-th order moving average The calculation formula is:

[0137] in Indicates rounding down;

[0138] S202, Tunnel Model Construction and Update

[0139] Based on the tunnel's design parameters and geological conditions, a three-dimensional model of the tunnel is constructed; this model should accurately reflect the tunnel's structural characteristics and mechanical properties.

[0140] As tunnel construction and operation progress, actual data is continuously collected to update and optimize the tunnel model in order to improve its accuracy and reliability.

[0141] S203, Intelligent Simulation and Analysis

[0142] The preprocessed data is input into the tunnel model for intelligent simulation and analysis; the structural response of the tunnel under different working conditions is simulated, including stress and deformation.

[0143] Advanced numerical simulation methods (including finite element method and boundary element method) and artificial intelligence algorithms (including neural network and genetic algorithm) are used to conduct in-depth analysis of simulation results and uncover the patterns and potential problems behind the data.

[0144] S204, Operation and Maintenance Strategy Formulation

[0145] Based on the results of intelligent simulation and analysis, determine whether there are any abnormalities in the tunnel structure and the severity of the abnormalities;

[0146] If it is determined that the adaptive intelligent grouting system needs to be activated, the grouting parameter settings should be further determined, including grouting pressure, grouting volume, and grouting material type. The parameter settings should take into account the actual situation of abnormal parts, the analysis results of the tunnel model, and previous experience data.

[0147] S205, Strategy evaluation and optimization, strategy transmission and execution.

[0148] Strategy evaluation and optimization provided by embodiments of the present invention:

[0149] The established operation and maintenance strategies are evaluated to analyze their feasibility and effectiveness; the advantages and disadvantages of various strategies can be compared by simulating the tunnel structure response under different strategies.

[0150] Based on the assessment results, the operation and maintenance strategy will be optimized and adjusted; the strategy will be continuously improved to enhance the efficiency and safety of tunnel operation and maintenance.

[0151] The policy transmission and execution provided in this embodiment of the invention:

[0152] The established operation and maintenance strategy is transmitted to the central control module, which then forwards the relevant instructions to the corresponding execution modules to guide the modules to work together and achieve scientific operation and maintenance of the tunnel.

[0153] like Figure 4 As shown in the figure, an embodiment of the present invention provides a method for achieving tunnel resilience based on dynamic learning and optimization, which includes:

[0154] S301 uses stress sensors to monitor tunnel stress data through a stress monitoring module;

[0155] S302 uses a strain monitoring module to monitor tunnel deformation data using strain sensors.

[0156] S303 uses a temperature and humidity monitoring module to monitor tunnel temperature and humidity data using temperature and humidity sensors.

[0157] S304, when the bionic intelligent sensing system detects an anomaly in the tunnel structure through the grouting module, the deep learning-based decision system issues an instruction, and the adaptive intelligent grouting system is automatically activated. The cement-based intelligent self-healing material is injected into the abnormal area through the grouting pipeline to complete precise grouting reinforcement and realize the tunnel's self-repair.

[0158] S305 receives data transmitted from the biomimetic intelligent sensing system through the decision module, and performs intelligent simulation and analysis in conjunction with the tunnel model; based on the analysis results, it formulates corresponding operation and maintenance strategies, such as whether to start the adaptive intelligent grouting system and the parameter settings for grouting;

[0159] S306 uses a communication module to connect to the Internet for network communication via wireless communication equipment.

[0160] The working principle of this invention lies in constructing a dynamic resilience closed-loop system integrating sensing, analysis, and execution. First, through distributed stress monitoring, strain monitoring, and temperature and humidity monitoring modules, real-time data on the stress, deformation, and environment of the tunnel structure are collected, forming a high-dimensional time-series monitoring matrix. The system uses this data to describe the stress evolution and damage trends of the tunnel structure under long-term service conditions, providing data support for subsequent decision-making modules and realizing the transformation from single-point monitoring to multi-dimensional comprehensive sensing.

[0161] Secondly, the central control module dynamically analyzes multi-source monitoring data based on deep learning-based decision-making algorithms. Convolutional neural network layers extract spatial features such as crack distribution and stress concentration, while bidirectional long short-term memory network layers capture the evolution of damage over time. An attention mechanism assigns high weights to key features, thereby calculating the tunnel toughness index. When the toughness index falls below a set threshold, the system automatically generates a grouting repair strategy, including parameters such as grouting pressure, grouting volume, and grouting speed, enabling intelligent judgment and response to the structural health status.

[0162] Finally, the grouting module, acting as the execution unit, initiates the adaptive intelligent grouting system based on instructions from the decision-making module, injecting cement-based intelligent self-healing material into the abnormal areas. Microcapsules within the material rupture at the cracks, releasing a repair agent that reacts with cement hydration products to form a dense structure, thereby restoring the tunnel's overall load-bearing capacity. The effectiveness of the repair process is again captured by the monitoring module and fed back to the decision-making module, enabling model self-learning and parameter self-optimization, ultimately forming a dynamic closed-loop toughness enhancement system of "perception-decision-repair-feedback".

[0163] A specific embodiment of the present invention:

[0164] 1. Highway tunnel project in a mountainous area

[0165] Project Overview: Located in a mountainous area with complex geological conditions, including fault fracture zones and weak surrounding rock, the tunnel has experienced problems such as concrete structure cracking and water leakage during operation, seriously affecting the tunnel's safety and normal use.

[0166] Implementation process

[0167] Sensing system deployment: A distributed sensor network, including stress sensors, strain sensors, and humidity sensors, is deployed in potential weak points such as the tunnel's arch, sidewalls, and invert. The sensors wirelessly transmit the collected data to the decision-making system.

[0168] Decision-making system modeling and analysis: A tunnel model is established using digital twin technology, and real-time simulation and analysis are performed using data transmitted from the sensing system. When the system detects stress concentration and crack propagation in a certain section of the sidewall, the decision-making system determines, based on preset algorithms and historical experience, that an adaptive intelligent grouting system needs to be activated for reinforcement.

[0169] Grouting System Operation: After the decision-making system issues a command, the adaptive intelligent grouting system automatically starts. The grouting pump injects cement-based intelligent self-healing material into the crack through the grouting pipeline. At the same time, grout diffusion three-dimensional monitoring technology monitors the diffusion of the grout in real time to ensure that the grout fills the crack evenly and achieves the effect of precise grouting.

[0170] Implementation Results: After a period of operation, the expansion of cracks in the tunnel was effectively controlled, and the water leakage problem was significantly improved. Through proactive prevention and self-repair, the frequency and cost of manual maintenance were reduced, and the tunnel's traffic capacity and safety were improved.

[0171] 2. A subway tunnel project in a certain city

[0172] Project Overview: This subway tunnel is located in a bustling urban area with a complex surrounding environment, posing high requirements for construction and operation. During operation, due to factors such as changes in groundwater levels and train vibrations, the tunnel experienced problems such as excessive structural deformation and localized water leakage.

[0173] Implementation process

[0174] Modular Design and Installation: Based on the characteristics and requirements of subway tunnels, a modular and scalable architecture is adopted for the tunnel structure design. Sensing modules, grouting modules, and decision-making modules are designed and installed independently to ensure compatibility and collaborative operation between modules.

[0175] Real-time monitoring and decision-making: The biomimetic intelligent sensing system monitors the tunnel's structural deformation, stress, strain, and other parameters in real time. When the deformation of a section of the tunnel arch exceeds the warning value, the decision-making system quickly analyzes the cause and formulates a corresponding grouting reinforcement plan.

[0176] Intelligent grouting and effect evaluation: The adaptive intelligent grouting system performs grouting operations according to the instructions of the decision-making system. After grouting is completed, the grouting effect is evaluated through a combination of grout diffusion three-dimensional monitoring technology and manual inspection. The evaluation results show that the deformation of the tunnel structure is effectively controlled after grouting, meeting the design requirements.

[0177] Implementation Results: The implementation of this tunnel structure effectively solved the problems of subway tunnel defects and ensured the safe operation of the subway. At the same time, the modular design made the system installation and maintenance more convenient and efficient, reducing the impact of construction on the surrounding environment and traffic.

[0178] The specific application areas or related products of this invention.

[0179] (a) Application areas

[0180] Highway tunnels: Applicable to highway tunnels of various geological conditions and sizes, effectively solving common problems in highway tunnels during operation, such as water damage, concrete structure cracking, and excessive structural deformation, thereby improving the traffic safety and service level of highway tunnels.

[0181] Railway tunnels: This invention can be used for the construction, operation, and maintenance of railway tunnels, ensuring the safety and smooth operation of railway transportation. Especially in railway tunnels under complex geological conditions, the tunnel structure of this invention can leverage its advantages in proactive prevention and self-repair, reducing the impact of defects on railway operations.

[0182] Urban subway tunnels: Meeting the high requirements for safety and operational efficiency of urban subway tunnels. Capable of promptly addressing tunnel defects caused by factors such as changes in groundwater levels and train vibrations, ensuring the normal operation of the urban subway system.

[0183] Underground integrated utility tunnel: In the construction and operation of underground integrated utility tunnels, the tunnel structure of the present invention can realize real-time monitoring and self-repair of the tunnel structure, prevent pipeline damage and safety accidents caused by structural defects, and improve the reliability and durability of underground integrated utility tunnels.

[0184] (ii) Related Products

[0185] Bionic intelligent sensing sensors: including various types of stress sensors, strain sensors, humidity sensors, temperature sensors, etc., which are characterized by high precision, high reliability, and low power consumption, and can adapt to the harsh environmental conditions of tunnels.

[0186] Adaptive intelligent grouting equipment: an automatic grouting equipment that integrates cement-based intelligent self-healing materials, including grouting pumps, grout storage tanks, grouting pipelines, etc., which can realize the automation and precise control of grouting.

[0187] Deep learning-based decision-making system software: Decision-making system software developed using digital twin technology and deep learning algorithms can simulate and analyze the state of tunnel structures in real time and formulate scientific and reasonable operation and maintenance strategies.

[0188] Modular tunnel structure components: Tunnel structure components with modular design, such as sensing modules, grouting modules, and decision-making modules, are easy to install and maintain, and can be flexibly combined according to different engineering needs.

[0189] It should be noted that embodiments of the present invention can be implemented in hardware, software, or a combination of both. The hardware portion can be implemented using dedicated logic; the software portion can be stored in memory and executed by a suitable instruction execution system, such as a microprocessor or dedicated-design hardware. Those skilled in the art will understand that the above-described devices and methods can be implemented using computer-executable instructions and / or included in processor control code, for example, such code provided on a carrier medium such as a disk, CD, or DVD-ROM, a programmable memory such as read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The devices and modules of the present invention can be implemented by hardware circuitry such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field-programmable gate arrays, programmable logic devices, etc., or by software executed by various types of processors, or by a combination of the above-described hardware circuitry and software, such as firmware.

[0190] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions, and improvements made by those skilled in the art within the scope of the technology disclosed in the present invention, and within the spirit and principles of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A tunnel resilience realization system based on dynamic learning and optimization, characterized in that, include: The stress monitoring module, strain monitoring module, and temperature and humidity monitoring module are used to collect stress, deformation, and environmental data of the tunnel structure, forming a multi-dimensional monitoring dataset that includes spatial damage characteristics and temporal evolution trends. The central control module is connected to each monitoring module, decision-making module, and grouting module to coordinate the monitoring and control process; The decision module is used to analyze the multidimensional monitoring data based on a deep learning model, calculate the tunnel toughness index, and generate a repair instruction when the toughness index is below 0.

6. The grouting module is used to automatically perform grouting operations according to the repair instructions, injecting intelligent self-healing materials into abnormal parts to achieve structural self-healing and reinforcement. The deep learning model includes a convolutional neural network layer, a bidirectional long short-term memory network layer, and an attention layer, which are used to extract spatial features, temporal features, and assign feature weights, respectively. The model dynamically adjusts hyperparameters based on historical repair feedback.

2. The system as described in claim 1, characterized in that, The central control module constructs a distributed sensor network, with sensor nodes communicating via wired or wireless means to form a real-time sensing system covering the entire tunnel, used to dynamically map the structural stress distribution and deformation trends.

3. The system as described in claim 1, characterized in that, The communication module includes an edge computing terminal and a cloud database, which are connected via a wireless network to achieve dual-layer collaborative control of rapid local response and in-depth cloud analysis.

4. The system as described in claim 1, characterized in that, Grouting parameters include grouting pressure, grouting volume, and grouting speed. The central control module automatically adjusts these parameters based on real-time monitoring data to ensure grouting stability and repair effectiveness.

5. The system as described in claim 1, characterized in that, The intelligent self-healing material is a composite system of cement-based material and microencapsulated self-healing agent. When cracks appear, the microcapsules rupture to release the repair agent, which reacts with cement hydration products to generate a dense product, thus achieving self-healing of the cracks.

6. A deep learning decision algorithm for tunnel resilience control, characterized in that, Includes the following steps: Step 1: The input layer receives multi-dimensional time-series data from the stress, strain, and temperature and humidity sensing modules; Step 2: The convolutional neural network extracts the local spatial features of the data; Step 3: Extract time-dependent features using a bidirectional long short-term memory network; Step 4: The attention mechanism layer assigns feature weights and generates anomaly detection indicators. Step 5: The output layer outputs the grouting start signal and parameter instructions according to the judgment indicators; The abnormality judgment index is a real number between 0 and 1. When the value is greater than 0.7, the grouting action is triggered.

7. The algorithm as described in claim 6, characterized in that, The algorithm's loss function is the cross-entropy function, which is used to minimize the difference between the predicted structural state and the actual structural state.

8. The algorithm as described in claim 6, characterized in that, The algorithm optimization process uses an improved Adam optimizer, which achieves a dynamic balance between convergence speed and prediction accuracy through an adaptive learning rate adjustment mechanism.

9. The algorithm as described in claim 6, characterized in that, The input data consists of 200 continuously collected time-series samples, and the output results include the toughness index and the optimal grouting strategy parameters.

10. The algorithm as described in claim 6, characterized in that, The algorithm works in conjunction with the tunnel resilience management platform, which is used to visualize the tunnel's health status and automatically generate maintenance strategies, enabling closed-loop control of monitoring, decision-making, and repair.

Images