Intelligent tunnel construction monitoring and self-adaptive supporting system based on multi-modal sensing and self-adaptive method

Abstract

The invention discloses an intelligent tunnel construction monitoring and self-adaptive supporting system based on multi-mode perception and a self-adaptive method.The system comprises a plurality of distributed optical fibers located in a tunnel, and the distributed optical fibers are connected with a data acquisition system through anti-bending optical fiber jumpers; a plurality of adjustable anchor rod assemblies are arranged at preset positions on the wall of a tunnel, anchor rods of the adjustable anchor rod assemblies are connected with an assembled pressure-bearing central pivot cylinder, the inner wall of the assembled pressure-bearing central pivot cylinder is connected with a pressure-bearing core column through a micro jack, the assembled pressure-bearing central pivot cylinder is sleeved with the pressure-bearing core column, and the pressure-bearing core column is connected with the pressure-bearing core column. And the data acquisition system is positioned on the assembled pressure-bearing central cylinder. According to the intelligent tunnel construction monitoring and self-adaptive supporting system based on multi-modal sensing and the self-adaptive method, efficient and accurate supporting of tunnel surrounding rock is achieved, the safety of tunnel construction and operation is improved, and meanwhile the supporting cost is reduced.

Description

Technical Field

[0001] The present invention relates to an intelligent tunnel construction monitoring and adaptive support system and an adaptive method based on multimodal perception, and relates to the field of tunnels. Background Art

[0002] Throughout the life cycle of a tunnel project, surrounding rock stability is the core issue that determines construction safety and operational reliability. Due to the hidden nature and geological uncertainty of underground projects, the mechanical behavior of the surrounding rock presents significant spatiotemporal evolution characteristics: excavation unloading triggers stress redistribution, leading to crack expansion, groundwater infiltration weakens the rock mass strength, and construction blasting disturbances aggravate structural surface loosening. These dynamic changes cause the surrounding rock to enter a non-steady-state process of continuous adjustment from the initial equilibrium state. Traditional tunnel support methods are mostly fixed support structures with preset parameters, such as common anchor support, whose parameters such as length, spacing, and anchoring force are determined before construction. Support parameters are preset based on engineering analogy or empirical formulas, such as the use of Φ22 threaded steel anchors arranged at equal intervals. Its design logic is based on the assumption of a "static geological model." However, in actual projects, key parameters such as the degree of rock joint development, ground stress direction, and groundwater occurrence status often deviate significantly from the survey report, resulting in the support structure facing polarization risks:

[0003] When the actual surrounding rock strength falls short of design expectations, the rigid support structure cannot compensate for the bearing capacity shortfall through parameter adjustment. A typical example is a heavily weathered section of a deep tunnel where fixed-length anchors were unable to penetrate the loosening zone, causing the entire support structure to sink, ultimately leading to primary lining collapse and surface subsidence. In Class II / III surrounding rock sections with strong self-stabilizing properties, excessive support not only results in ineffective investment of materials like steel and concrete, but also disrupts the original geostress field due to excessive grouting pressure, artificially creating new weak zones.

[0004] Existing support systems generally lack closed-loop feedback capabilities: ① The sensing dimension is single, relying on manual inspections and point displacement meters, making it difficult to capture the three-dimensional stress field reconstruction process; ② The data fusion level is low, and multi-source data such as vibrating string sensors and fiber optic monitoring have not formed an effective joint interpretation mechanism; ③ The control lag is strong, and the average response period from monitoring anomalies to support reinforcement is as long as 72 hours.

[0005] Therefore, there is an urgent need to develop an intelligent tunnel construction monitoring and adaptive support system with multimodal perception, which can more conveniently maintain the tunnel and provide a theoretical basis for the research and development of a new generation of adaptive support technology. Summary of the Invention

[0006] Purpose of the invention: In order to overcome the deficiencies in the prior art, the present invention provides an intelligent tunnel construction monitoring and adaptive support system and adaptive method based on multimodal perception, so as to achieve efficient and precise support of tunnel surrounding rock, improve the safety of tunnel construction and operation, and reduce support costs.

[0007] Technical Solution: To address the aforementioned technical issues, the present invention proposes an intelligent tunnel construction monitoring and adaptive support system based on multimodal sensing. The system comprises several optical fibers located within the tunnel, connected to a data acquisition system via anti-bend fiber jumpers. Several adjustable anchor rod assemblies are positioned at preset locations on the tunnel wall. The anchor rods of the adjustable anchor rod assemblies are connected to an assembled pressure-bearing central cylinder. The inner wall of the assembled pressure-bearing central cylinder is connected to a support sleeve via micro-jacks. The support sleeve is mounted on a pressure-bearing core column. The data acquisition system is located on the assembled pressure-bearing central cylinder. The adaptive support system comprises a pressure-bearing core column, an assembled pressure-bearing central cylinder, an adjustable anchor rod assembly, and micro-jacks. An optical fiber demodulator is positioned on the exterior of the pressure-bearing core column to monitor in real time the signals received from changes in various physical fields in the surrounding rock. Serving as the mechanical coupling hub of the support system, 24 micro-jacks are arranged in a circular array within the interior of the assembled pressure-bearing central cylinder, connected to a data processing center. When pressure on a specific part of the lining structure exceeds a preset threshold, the data processing center activates a micro-jack at the corresponding location, fine-tuning the lining structure locally and increasing the support strength in that area to resist the pressure of the surrounding rock. The intelligent control system presets optimization conditions and optimizes the tension ratio of its adjustable anchor bolts, avoiding the overall load-bearing efficiency degradation caused by uneven individual tension in traditional prestressed anchor bolts. Through the cylinder-anchor linkage mechanism, the support range can be dynamically expanded to a small extent as the excavation face advances, avoiding the "one-step" redundancy of traditional support.

[0008] Preferably, the optical fiber is a distributed optical fiber, which is arranged along the axial direction of the tunnel, with a group of optical fibers arranged every 30°.

[0009] Preferably, a fine-tuning screw mechanism is provided on the jack installation module, which can fine-tune the installation angle and depth in the radial and axial directions.

[0010] Preferably, each of the jacks is provided with a PLC submodule.

[0011] An adaptive method for intelligent tunnel construction monitoring and adaptive support system based on multimodal perception includes the following steps:

[0012] S1. Based on the specific tunnel environment and monitoring requirements, select a distributed optical fiber with high sensitivity, wear resistance, and adaptability to complex geological conditions. Determine the hardware configuration of the data acquisition system, including processor performance, memory capacity, and storage media, to ensure it can efficiently operate the edge computing module. Build the hardware platform for the intelligent control system, including servers and communication modules, and install the basic software of the digital twin engine to prepare for the subsequent establishment of a mapping relationship library between surrounding rock degradation patterns and support parameters.

[0013] S2. Use high-precision measuring instruments to measure along the tunnel axis, accurately marking the fiber optic laying positions every 30° to ensure that the marking points are evenly distributed;

[0014] S3. Accurately install the fiber optic interrogator on the outside of the pressure-bearing core column, and ensure that the connection between the fiber optic interrogator and the distributed optical fiber is accurate during the installation process;

[0015] S4. Inside the modular, assembled pressure-bearing central cylinder, precisely arrange 24 sets of micro-jacks in a circular array. Connect these jacks to the data processing center via dedicated control lines, which are insulated and waterproofed.

[0016] S5. Install the adjustable anchor assembly at the predetermined location on the tunnel wall according to the design requirements. During the installation process, ensure that the anchor depth meets the standard and the angle between the anchor and the tunnel wall is accurate. Accurately set the initial tension of the adjustable anchor using the optimization solution preset by the intelligent control system.

[0017] S6. To achieve intelligent response control of surrounding rock conditions during tunnel excavation, a mapping database was first established between surrounding rock monitoring characteristics (strain mutation, seepage excess, pressure change) and optimal support parameters (anchor arrangement, shotcrete thickness, arch form, etc.) through numerical simulation and historical case studies. During construction, a distributed fiber optic system collected surrounding rock state data in real time. The signal processing and feature extraction module analyzed the current deformation and seepage patterns and input them into the intelligent control system. The intelligent control system used the current surrounding rock state as input and invoked the multi-objective optimization module. The multi-objective optimization module first constructed a "surrounding rock state-support parameter" mapping database based on historical data. Input-output feature pairs were constructed through numerical simulation, measured data, and empirical design samples. A nonlinear prediction model was established using machine learning methods. State variables such as surrounding rock deformation, seepage, and disturbance rate were mapped to corresponding support response targets. A multi-objective optimization function was then constructed based on this mapping relationship. The solution was solved using an improved particle swarm optimization algorithm to output the optimal control solution that satisfied both support stability and economic efficiency, thus establishing a multi-objective optimization module with feedback regulation capabilities. This module is linked with the distributed fiber optic sensing system to form a perception-decision-control closed loop, realizing dynamic response to tunnel surrounding rock disturbances and adaptive adjustment of the support force field.

[0018] To comprehensively consider factors such as support stability, cost, surrounding rock deformation, and construction efficiency, an optimal support parameter combination is generated and control instructions are sent to the adaptive support system. The following multi-objective optimization function is used: support cost f1(x), surrounding rock deformation f2(x), construction disturbance time f3(x), and support stability f4(x). The decision variables x represent anchor length, spacing, shotcrete thickness, and steel arch arrangement. The optimized support parameter set is compared with a mapping library or construction specifications to check for safety redundancy. Control instructions are automatically generated and transmitted to the adaptive support system control module to drive the actuators for automatic adjustment. After the support system is installed, it continues to monitor surrounding rock changes to determine whether they meet the preset stability threshold. If not, it automatically enters the next round of optimization. The system automatically adjusts the support configuration based on the instructions and, combined with subsequent monitoring data, forms a closed-loop control loop to ensure surrounding rock stability and construction safety.

[0019] In the present invention, the data acquisition system includes a signal processing module, which performs feature extraction on the input signal. A mapping database is provided in the signal processing module. The signal processing module is connected to the multi-objective optimization module, which is connected to the support parameter generation module. The support parameter generation module is connected to the signals of each micro-jack, and the movement of the micro-jack is controlled by the support parameter generation module.

[0020] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0021] (1) This invention uses a distributed fiber optic sensing network to accurately capture coupled information on multiple physical quantities, such as surrounding rock deformation, temperature, and seepage, achieving real-time identification at the millimeter level. Based on a large amount of monitoring data, a deep learning algorithm is used to rapidly analyze surrounding rock degradation patterns. Unlike traditional methods, which often take several hours to generate support strategies, this invention accelerates this process to seconds, greatly improving decision-making efficiency.

[0022] (2) During the support process, the anchor rod-micro jack fusion unit adopted by the present invention enables it to flexibly and accurately adjust the support stiffness according to the actual deformation of the surrounding rock. When the surrounding rock deforms, the adaptive support system can respond quickly, provide initial support force, and further optimize the support stiffness to achieve dynamic matching with the surrounding rock deformation. This avoids material waste caused by excessive support, ensures that the support structure can always function stably, and protects the structural safety of the tunnel.

[0023] (3) The built-in monitoring system can regularly evaluate the support effectiveness attenuation rate and extend the service life of the structure through the prestress compensation mechanism. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 Flowchart of the intelligent tunnel construction monitoring method based on multimodal perception.

[0025] Figure 2 This is a structural diagram of the intelligent tunnel adaptive support system.

[0026] Figure 3 Schematic diagram of the fine-tuning screw mechanism structure.

[0027] In the figure: tunnel 1; distributed optical fiber 2; data acquisition system 3; support sleeve 4; pressure-bearing core column 5; assembled pressure-bearing central cylinder 6; adjustable anchor rod assembly 7; micro jack 8; anti-bending optical fiber jumper 9; ball joint 81; screw wheel 82; screw adjustment shaft 83; nut 84; universal joint 85; fixing sleeve 86. DETAILED DESCRIPTION

[0028] The present invention will be further described below with reference to the accompanying drawings.

[0029] like Figures 1 to 3 As shown, the present invention is an intelligent tunnel construction monitoring and adaptive support system based on multimodal perception, including a plurality of optical fibers located in the tunnel 1, the optical fibers are connected to the data acquisition system through anti-bending optical fiber jumpers, and a plurality of adjustable anchor rod assemblies are provided at preset positions on the tunnel wall, the anchor rods of the adjustable anchor rod assemblies are connected to the assembled pressure-bearing central cylinder, the inner wall of the assembled pressure-bearing central cylinder is connected to the support sleeve through a micro jack, the support sleeve is mounted on the pressure-bearing core column, and the data acquisition system is located on the assembled pressure-bearing central cylinder.

[0030] like Figure 2 As shown, an adaptive method for intelligent tunnel construction monitoring and adaptive support system based on multimodal perception includes the following steps:

[0031] Step (1): Based on the specific environment and monitoring requirements of the tunnel, select a distributed optical fiber 2 that is highly sensitive, wear-resistant, and adaptable to complex geological conditions. Determine the hardware configuration of the data acquisition system 3, including processor performance, memory capacity, storage media, etc., to ensure that it can efficiently run the edge computing module. Build the hardware platform of the intelligent control system, including servers, communication modules, etc. Install the engine basic software to prepare for the subsequent establishment of a mapping relationship library between surrounding rock degradation patterns and support parameters.

[0032] Step (2): Use a total station to set reference points every 5m along the tunnel axis, and use a laser projection locator to mark the fiber path every 30° in the circumferential direction. During the laying process, hot melt splicing technology is used to complete the connection between optical fibers. Before hot melt splicing, the optical fiber end face is cleaned and cut to ensure that the end face is flat and smooth. During the splicing process, parameters such as heating time, temperature and thrust amount are strictly controlled to ensure the splicing quality. For high-risk sections, such as areas with broken rocks, stress concentration or abundant groundwater, radial auxiliary fixators are added. The radial auxiliary fixators use special metal clips or adhesive materials to tightly fix the optical fiber to the rock surface, ensuring good adhesion between the optical fiber and the rock surface, and avoiding inaccurate monitoring data due to optical fiber shaking or displacement. Through the above measures, a high-precision three-dimensional sensing network is gradually formed to achieve all-round and real-time monitoring of the surrounding rock.

[0033] Step (3): A stainless steel threaded bottom base is preset on the outer surface of the pressure core column 5, and the fiber optic demodulator is installed by a mechanical arm. The anti-bending fiber optic jumper 9 is used to connect the distributed optical fiber 2, and silicone sealant is poured at the interface. The silicone sealant has good waterproof, dustproof and insulation properties, protecting the interface from the influence of the external environment and ensuring stable signal transmission. After the installation is completed, an optical power meter is used to detect the signal strength received by the fiber optic demodulator. By detecting the signal strength, it is determined whether the optical fiber connection is normal and whether there is any loss in the signal transmission. Based on the platform simulation of surrounding rock pressure loading, different levels of pressure are applied to the pressure core column to simulate the surrounding rock under different stress conditions. The data output of the fiber optic demodulator is observed and compared with the simulated pressure data to verify the data synchronization rate. After multiple tests and adjustments, it is ensured that under various working conditions, the fiber optic demodulator can monitor the signals generated by the changes in the surrounding rock physical field in real time and stably, providing reliable data support for the adaptive support system.

[0034] Step (4): Use high-precision processing equipment to precisely process 24 sets of installation grooves on the inner side of the assembled pressure-bearing central cylinder 6. The size, position and accuracy of the installation grooves are strictly controlled in accordance with the design requirements to ensure that the micro-jack 8 can be accurately embedded. The control line uses a double-layer shielded armored cable. The double-layer shielding structure can effectively resist external electromagnetic interference and ensure the stability of signal transmission. At the joint, first perform heat shrink tubing packaging so that it can tightly wrap the joint after heating, providing good insulation and waterproof performance. Then perform epoxy resin potting. Epoxy resin has excellent electrical insulation performance, bonding performance and mechanical strength, which can further protect the joint and prevent moisture and impurities from intruding. After the packaging is completed, the control line is subjected to a withstand voltage test and the insulation resistance is measured to ensure that under extreme environments, such as high humidity, strong electromagnetic interference, etc., the signal transmission is still reliable and the micro-jack can accurately receive control instructions and work normally.

[0035] Step (5): Use an anchor drill to drill holes at the designed angle, install adjustable anchor bolts 7, and inject modified epoxy resin into the anchoring section. Initial prestressing is applied via a wireless hydraulic pump station. Vibrating wire sensors are used to monitor the tension distribution, and a particle swarm optimization algorithm is applied to adjust the tension difference between each anchor bolt to form a uniform prestressing field.

[0036] Step (6): Simulate conditions under complex geological conditions such as a fault fracture zone, triggering the extension of anchor rod 7 and the pressurization of jack 8. By setting the corresponding geological parameters and mechanical boundary conditions in the model, the deformation and stress conditions of the surrounding rock in harsh environments such as fault fracture zones are simulated. The extension mechanism of anchor rod 7 and the pressurization system of jack 8 are manually triggered by the control device, and the response of the system is observed.

[0037] Among them, the specific assembly method for the modular assembled pressure cylinder structure is as follows:

[0038] (a) In the modular, assembled pressure-bearing cylinder structure, each module is equipped with an independent jack mounting unit. This structure consists of several adjustable mounting units, each of which includes a slide rail slot, a positioning bolt, and a fine-tuning bracket. These units can be individually adjusted or replaced based on construction requirements or changes in surrounding rock stress. The module connecting the jack to the pressure-bearing core column features a fine-tuning screw mechanism, including a spherical joint 81, which connects the anchor rod to the structural component and allows universal rotation within a certain angular range. A screw wheel 82 provides a non-slip outer surface for easy adjustment or tightening, allowing for fine-tuning the position of the screw rod, anchor end, or support structure. The screw adjustment shaft 83 rotates to drive axial movement, enabling linear displacement adjustment. A nut 84 locks the position after adjustment to prevent loosening or displacement during use. A universal joint 85 enables angular transmission or relative rotation between components, maintaining continuous force transmission. A fixed sleeve 86 connects the jack 8 to the fine-tuning screw mechanism and is fixed to the pressure-bearing component of the tunnel support system.

[0039] (b) Equally spaced guide locating bases are pre-set on the inner wall of the assembly tube. Each base connects to the corresponding jack mounting module via a dovetail groove structure, enabling adjustable jack installation positions and highly accurate repeatable positioning. A fine-tuning screw mechanism is incorporated into the jack mounting module to fine-tune the installation angle and depth in both radial and axial directions, ensuring precise contact between each jack and the surrounding rock.

[0040] (c) Each jack is equipped with a PLC submodule that receives instructions from the main control system and feeds back real-time pressure and displacement data.

[0041] An adaptive method for intelligent tunnel construction monitoring and adaptive support system based on multimodal perception includes the following steps:

[0042] S1. Through numerical simulation and historical cases, a mapping relationship database is established between surrounding rock monitoring characteristics including strain mutation, seepage limit exceedance, pressure change and optimal support parameters such as anchor arrangement, shotcrete thickness and arch form.

[0043] S2. Based on the specific tunnel environment and monitoring requirements, select a distributed optical fiber with high sensitivity, wear resistance, and adaptability to complex geological conditions. Determine the hardware configuration of the data acquisition system, including processor performance, memory capacity, and storage media, to ensure it can efficiently operate the edge computing module. Build the hardware platform for the intelligent control system, including servers and communication modules, and install the basic software of the digital twin engine to prepare for the subsequent establishment of a mapping relationship library between surrounding rock degradation patterns and support parameters.

[0044] S3. Use high-precision measuring instruments to measure along the tunnel axis, accurately marking the fiber optic laying positions every 30° to ensure that the marking points are evenly distributed;

[0045] S4. Accurately install the fiber optic interrogator on the outside of the pressure-bearing core column, and ensure that the connection between the fiber optic interrogator and the distributed optical fiber is accurate during the installation process;

[0046] S5. Inside the modular, assembled pressure-bearing central cylinder, 24 sets of micro-jacks are precisely arranged in a circular array. These jacks are connected to the data processing center via dedicated control lines, which are insulated and waterproofed.

[0047] S6. Install the adjustable anchor assembly at the predetermined location on the tunnel wall according to the design requirements. During the installation process, ensure that the anchor depth meets the standard and the angle between the anchor and the tunnel wall is accurate. Accurately set the initial tension of the adjustable anchor using the optimization solution preset by the intelligent control system.

[0048] S7. During the construction process, the distributed fiber optic system collects surrounding rock status data in real time, and the signal processing and feature extraction module analyzes the current deformation and seepage pattern, and inputs it into the intelligent control system. The intelligent control system uses the current state of the surrounding rock as input, calls the multi-objective optimization module, and comprehensively considers support stability, cost, surrounding rock deformation and construction efficiency factors to generate the optimal support parameter combination, and sends control instructions to the adaptive support system for adaptive adjustment.

[0049] Adaptive adjustment includes the following methods:

[0050] S71 built a "surrounding rock state-support parameter" mapping database based on historical data. It constructed input-output feature pairs through numerical simulation, measured data, and empirical design samples. It also used machine learning methods to establish a nonlinear prediction model, mapping state variables such as surrounding rock deformation, seepage, and disturbance rate to corresponding support response targets.

[0051] S72 constructs a dual-objective optimization function that takes into account both support accuracy and support uniformity, and its form is as follows:

[0052]

[0053] Among them, P set is the initial target pressure set by particle swarm optimization according to the supporting force of each micro-jack, that is, P i ref is the target support force of the i-th micro-jack, obtained through step S71, and the parameters ω1 and ω2 are weight coefficients in the dual-objective optimization function, ω1 controls the support accuracy, and ω2 controls the support field uniformity;

[0054] S73 implements reading the value of each distributed optical fiber and converting it into P i se , calculate f(P) i , if |f(P) i -f(P) i-1 |>ε, start the corresponding micro jack, according to P i =P set -P i set , Δu i =κP i , adjust the micro jack;

[0055] S74 repeats step S73 to complete the adaptive adjustment of the micro jack.

[0056] This system adopts a hierarchical structure to achieve precise support control of micro jacks. The overall control strategy is divided into three parts: perception layer, optimization layer and control layer. It relies on distributed fiber optic sensors, particle swarm optimization algorithm and feedback controller to operate in coordination.

[0057] Perception layer (input layer): The system uses distributed optical fiber sensors to obtain continuously distributed perception data xi along the tunnel axis, including surrounding rock displacement, microstrain change rate, local disturbance rate, etc. The calculated interpretation model maps xi to the support reaction force P corresponding to the jack position. ref , which is also the support reaction force of the surrounding rock state monitored by the optical fiber monitoring system, serves as the perception layer (input value) of the optimization algorithm and the controller, and controls the actual output pressure value P of the micro jack in real time. act .

[0058] Optimization layer: In the optimization layer, the system constructs a dual-objective optimization function that takes into account both support accuracy and support uniformity. The form is as follows:

[0059]

[0060] Among them, P set is the supporting force P of each micro jack i ref , the initial target pressure set by particle swarm optimization is It can be used as the optimization layer of the optimization algorithm and controller. i ref According to the steps. The first term minimizes the deviation between the target pressure of each jack and the perceived target to ensure support accuracy; the second term minimizes the pressure difference rate between adjacent jacks to optimize the overall uniformity of the support field. Parameters ω1 and ω2 are weight coefficients in the dual-objective optimization function, which are used to control the relative importance of the two optimization objectives in the overall objective function. ω1 controls support accuracy, and ω2 controls support field uniformity. To avoid stress concentration, ω1 is set to 0.4 and ω2 is set to 0.6. They can be dynamically updated according to actual requirements.

[0061] Control layer: The bottom controller takes the target pressure output by the optimization layer as input P1 set , control the micro jack to perform precise support. The controller uses the real-time error P i =P set -P act Calculate and convert into displacement adjustment Δu i =κP i κ is a constant gain function representing the displacement adjustment required per unit pressure change. Its value can be obtained through calibration testing. The fine-tuning screw mechanism is driven to fine-tune the jack, ensuring the system's ability to quickly respond to local disturbances.

[0062] In each iteration described above, the system records the current optimal particle position and the corresponding function value f(P) i , when the difference between two consecutive iterative function values ​​satisfies:

[0063] |f(P) i -f(P) i-1 |<ε

[0064] If the change in the optimization function value is less than the threshold (initial setting 1e-4), or the maximum number of iterations is reached, the optimization is stopped and the current optimal target pressure P1 is output. set If the optimization function value does not converge or the target deviates too much, you can apply disturbances near it to perform fine-tuning. Where δ=Δu(-Δ max ,+Δ max ), Δ max Can be set to 0.01-0.05MPa, depending on the system's sensitivity to small adjustments.

[0065] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. An intelligent tunnel construction monitoring and adaptive support system based on multimodal perception, characterized by: It includes several optical fibers located in the tunnel, which are connected to the data acquisition system through anti-bending optical fiber jumpers. Several adjustable anchor rod assemblies are provided at preset positions on the tunnel wall. The anchor rods of the adjustable anchor rod assemblies are connected to the assembled pressure-bearing central cylinder. The inner wall of the assembled pressure-bearing central cylinder is connected to the support sleeve through a micro jack. The support sleeve is mounted on the pressure-bearing core column. The data acquisition system is located on the assembled pressure-bearing central cylinder.

2. The intelligent tunnel construction monitoring and adaptive support system based on multimodal perception according to claim 1 is characterized by: The optical fiber is a distributed optical fiber, which is arranged along the axial direction of the tunnel, with a group of optical fibers arranged every 30°.

3. The intelligent tunnel construction monitoring and adaptive support system based on multimodal perception according to claim 1 is characterized by: A fine-adjusting screw mechanism is provided on the micro-jack mounting module.

4. The intelligent tunnel construction monitoring and adaptive support system based on multimodal perception according to claim 1 is characterized by: The micro jacks are all provided with a PLC submodule.

5. An adaptive method for an intelligent tunnel construction monitoring and adaptive support system based on multimodal perception according to any one of claims 1 to 3, characterized in that: The following steps are involved: S1. Through numerical simulation and historical cases, a mapping relationship database is established between surrounding rock monitoring characteristics including strain mutation, seepage limit exceedance, pressure change and optimal support parameters such as anchor arrangement, shotcrete thickness and arch form. S2. Based on the specific tunnel environment and monitoring requirements, select a distributed optical fiber with high sensitivity, wear resistance, and adaptability to complex geological conditions. Determine the hardware configuration of the data acquisition system, including processor performance, memory capacity, and storage media, to ensure it can efficiently operate the edge computing module. Build the hardware platform for the intelligent control system, including servers and communication modules, and install the basic software of the digital twin engine to prepare for the subsequent establishment of a mapping relationship library between surrounding rock degradation patterns and support parameters. S3. Use high-precision measuring instruments to measure along the tunnel axis, accurately marking the fiber optic laying positions every 30° to ensure that the marking points are evenly distributed; S4. Accurately install the fiber optic interrogator on the outside of the pressure-bearing core column, and ensure that the connection between the fiber optic interrogator and the distributed optical fiber is accurate during the installation process; S5. Inside the modular, assembled pressure-bearing central cylinder, 24 sets of micro-jacks are precisely arranged in a circular array. These jacks are connected to the data processing center via dedicated control lines, which are insulated and waterproofed. S6. Install the adjustable anchor assembly at the predetermined location on the tunnel wall according to the design requirements. During the installation process, ensure that the anchor depth meets the standard and the angle between the anchor and the tunnel wall is accurate. Accurately set the initial tension of the adjustable anchor using the optimization solution preset by the intelligent control system. S7. During the construction process, the distributed fiber optic system collects surrounding rock status data in real time, and the signal processing and feature extraction module analyzes the current deformation and seepage pattern, and inputs it into the intelligent control system. The intelligent control system uses the current state of the surrounding rock as input, calls the multi-objective optimization module, and comprehensively considers support stability, cost, surrounding rock deformation and construction efficiency factors to generate the optimal support parameter combination, and sends control instructions to the adaptive support system for adaptive adjustment.

6. According to the adaptive method of the intelligent tunnel construction monitoring and adaptive support system based on multimodal perception according to claim 5, in step S7, the adaptive adjustment includes the following methods: S71 built a "surrounding rock state-support parameter" mapping database based on historical data. It constructed input-output feature pairs through numerical simulation, measured data, and empirical design samples. It also used machine learning methods to establish a nonlinear prediction model, mapping state variables such as surrounding rock deformation, seepage, and disturbance rate to corresponding support response targets. S72 constructs a dual-objective optimization function that takes into account both support accuracy and support uniformity, and its form is as follows: in, P set is the initial target pressure set by particle swarm optimization according to the supporting force of each micro-jack, that is, P i ref is the target support force of the i-th micro-jack, parameters ω1 and ω2 are weight coefficients in the dual-objective optimization function, ω1 controls the support accuracy, and ω2 controls the uniformity of the support field; S73 implements reading the value of each distributed optical fiber and converting it into P i se , calculate f(P) i , if |f(P) i -f(P) i-1 |>ε, start the corresponding micro jack, according to P i =P set -P i set , Δu i =κP i , adjust the micro jack; S74 repeats step S73 to complete the adaptive adjustment of the micro jack.

Images