A soft rock support system for shallow-buried tunnels in high-altitude permafrost regions and its construction method

Abstract

This invention relates to the field of tunnel surrounding rock support technology, and particularly to a soft rock support system and construction method for shallow-buried tunnels in high-altitude permafrost areas. The support system includes a temperature-adaptive and stable composite arch frame structure, comprising a rigid outer protective temperature-regulating layer, an elastic component, and an arch frame. The rigid outer protective temperature-regulating layer includes a steel protective layer, an insulation layer, and a temperature-regulating layer arranged radially from the outside to the inside. The elastic component includes an elastic element and a buffer layer, with both ends of the elastic element abutting against the inner side of the rigid outer protective temperature-regulating layer and the outer side of the arch frame, respectively. The buffer layer fills the space between the rigid outer protective temperature-regulating layer and the arch frame. This invention, through the temperature regulation and adaptive deformation capability of the temperature-adaptive and stable composite arch frame structure, achieves dynamic adaptation of the support system to freeze-thaw deformation, temperature fluctuations, and complex geological conditions, thereby improving the safety and long-term stability of tunnel construction.

Description

Technical Field

[0001] This invention relates to the field of tunnel surrounding rock support technology, and in particular to a soft rock support system for shallow buried tunnels in high-altitude permafrost areas and its construction method. Background Technology

[0002] As infrastructure construction continues to extend to higher altitudes, the demand for shallow-buried tunnels in permafrost regions is increasing. High-altitude permafrost regions possess unique and complex geological and climatic conditions, posing numerous severe challenges to tunnel construction.

[0003] On the one hand, high-altitude areas experience extremely low temperatures and significant diurnal temperature variations, with annual temperature ranges often reaching tens of degrees Celsius. Under these harsh climatic conditions, the soft rock surrounding shallow-buried tunnels frequently undergoes freeze-thaw cycles. Soft rock expands in volume when frozen and contracts when thawing, repeating this cycle repeatedly, causing severe damage to the rock mass structure, a significant reduction in strength, and consequently, substantial deformation, threatening the structural stability of the tunnel. Traditional support structures are mostly designed for normal temperatures or general geological conditions and are ill-suited to the unique freeze-thaw deformation characteristics of soft rock in high-altitude permafrost regions, making them highly susceptible to support failure.

[0004] On the other hand, the geomechanical properties of permafrost regions at high altitudes are quite unique. Parameters such as elastic modulus, Poisson's ratio, and cohesion differ significantly from those of soft rock in conventional areas. Furthermore, the distribution range of permafrost, the thickness of the freeze-thaw layer, and the complex groundwater flow conditions all make it difficult to predict changes in surrounding rock stress during tunnel excavation. Conventional surveying methods and design approaches cannot accurately obtain relevant parameters, resulting in poor matching between the support system design and actual working conditions.

[0005] Furthermore, existing tunnel support systems have limited thermal insulation measures. Most do not consider how to effectively prevent external low temperatures from penetrating the tunnel interior and maintain a relatively stable temperature environment for the surrounding rock, thus failing to effectively reduce frequent freeze-thaw cycles caused by temperature fluctuations in soft rock. At the same time, the support structure lacks sufficient stress adjustment capabilities; when local soft rock deformation is uneven, there is a lack of effective means to redistribute stress, easily leading to stress concentration and accelerating damage to the support structure.

[0006] Furthermore, the monitoring system is inadequate, making it difficult to capture the complex and variable deformation, temperature, and pressure data of soft rock in a comprehensive and high-precision manner. The grouting system also fails to adequately consider the problem of grout freezing in low-temperature environments, as well as the need for flexible grouting based on different pore structures and fracture development levels in soft rock, resulting in a significant reduction in the effectiveness of grouting reinforcement.

[0007] Therefore, the existing technology has the following problems:

[0008] First, the rigid connection of traditional support structures is difficult to adapt to the repeated expansion and contraction of soft rock caused by freeze-thaw cycles, resulting in stress concentration between the support body and the surrounding rock, which can easily lead to structural cracking or failure.

[0009] Secondly, conventional insulation materials only slow down the intrusion of low temperatures from the outside by insulating, and cannot actively regulate the temperature of the surrounding rock. The temperature difference between day and night will still cause frequent freeze-thaw cycles in soft rock, exacerbating rock mass damage.

[0010] In addition, grouting reinforcement relies on manual experience to set parameters, without taking into account the dynamic changes in the pore structure and fissure development of soft rock in permafrost areas. The grout is prone to freezing in low-temperature environments, resulting in unstable grouting effects.

[0011] Finally, the monitoring systems mostly use a single sensor and are sparsely distributed, resulting in insufficient data acquisition accuracy. They are unable to capture millimeter-level deformation and temperature field changes in the surrounding rock in a timely manner, and therefore cannot provide a reliable basis for support decisions. Summary of the Invention

[0012] To address the shortcomings of existing technologies, the purpose of this invention is to provide a soft rock support system for shallow-buried tunnels in high-altitude permafrost regions. By utilizing the temperature regulation and adaptive deformation capabilities of the temperature-adaptive composite arch structure, the support system can dynamically adapt to freeze-thaw deformation, temperature fluctuations, and complex geological conditions, thereby improving the safety and long-term stability of tunnel construction.

[0013] To achieve the above objectives, the embodiments of the present invention provide the following technical solutions:

[0014] A soft rock support system for shallow buried tunnels in high-altitude permafrost areas includes a temperature-adaptive and stable composite arch frame structure, wherein the temperature-adaptive and stable composite arch frame structure includes a rigid outer temperature-regulating layer, an elastic component, and an arch frame.

[0015] The rigid outer protective temperature regulating layer is a modular structure located on the radial outer side of the tunnel. The rigid outer protective temperature regulating layer includes a steel protective layer on the radial outer side and an insulation layer and a temperature regulating layer on the radial inner side.

[0016] The elastomeric component is located between the rigid outer protective temperature-regulating layer and the arch frame. The elastomeric component includes an elastic element and a buffer layer. The two ends of the elastic element abut against the inner side of the rigid outer protective temperature-regulating layer and the outer side of the arch frame, respectively. The buffer layer fills the space between the rigid outer protective temperature-regulating layer and the arch frame.

[0017] The arch frame is a modular steel arch frame located on the radial inner side of the tunnel.

[0018] Optionally, the insulation layer is made of polyurethane foam, the temperature regulating layer is made of phase change insulation material, the elastic element is made of nickel-titanium shape memory alloy wire, and the buffer layer is made of natural rubber. This combination of materials not only improves the thermal insulation and stress regulation performance of the support system, but also enhances its ability to adapt to freeze-thaw deformation of soft rock.

[0019] Optionally, a monitoring and data transmission system is also included. This system comprises a sensor network and a transmission module. The sensor network includes sensors located at the tunnel entrances and exits, geological transition zones, and fault fracture zones along the tunnel's axial direction, as well as sensors located at the arch crown, side arch waists, and arch feet on the same tunnel cross-section. The sensor network includes displacement sensors, temperature sensors, and pressure sensors. A cloud computing platform equipped with professional data analysis software is used to perform real-time analysis and processing of massive amounts of monitoring data, displaying it in a visual manner, such as plotting deformation-time curves and temperature distribution cloud maps. This allows engineers to intuitively grasp the tunnel's condition, adjust construction strategies promptly, and achieve scientific and intelligent construction decision-making.

[0020] Optionally, it also includes a grouting system, which includes a storage tank, a stirring motor, and a liquid level sensor. The storage tank is used to store grouting materials, the stirring motor is installed in the storage tank to stir the grouting materials, and the liquid level sensor is installed in the storage tank to detect the liquid level of the grouting materials.

[0021] This invention also provides a construction method for the soft rock support system of shallow buried tunnels in high-altitude permafrost areas as described above, including the following steps:

[0022] The process involves acquiring the geological and geometric parameters of the tunnel to be constructed; inputting these parameters into a pre-trained machine learning model to output the parameters of the temperature-adaptive and stabilizing composite arch structure and the timing of grouting; assembling the temperature-adaptive and stabilizing composite arch structure based on the output parameters; and sequentially installing a modular rigid outer protective temperature-regulating layer, elastomer components, and modular steel arches after each section of excavation during tunnel excavation; deploying a sensor network along the tunnel's axial and circumferential directions and transmitting sensor data to a data processing center in real time via a wireless transmission module; and initiating grouting when monitoring data shows that the surrounding rock deformation reaches the predicted grouting timing. This entire construction method closely integrates numerical simulation, machine learning, and on-site construction, achieving optimized design of the support system and precise control of construction timing, improving construction efficiency and quality, and reducing construction risks.

[0023] Optionally, the geological parameters include rock strength, rock unit weight, rock porosity, rock elastic modulus, rock deformation modulus, rock mass RQD value, rock water content, and rock mass freeze-thaw volume expansion rate. The tunnel geometric parameters include tunnel surface area, tunnel maximum width, tunnel maximum height, and tunnel depth. By obtaining these detailed geological and geometric parameters, the machine learning model can more accurately simulate the tunnel excavation and support process, analyze the mechanical behavior and surrounding rock deformation patterns under different working conditions, and thus derive a more reasonable support system design scheme to ensure the safety and stability of tunnel construction.

[0024] Optionally, the structural parameters of the temperature-adaptive and stable composite arch frame include the thickness of the steel protective layer, the thickness of the polyurethane foam, the thickness of the phase change insulation material, the deformation of the elastic element, the elastic coefficient of the elastic element, the elastic modulus of the buffer layer, the arch frame spacing, and the cross-sectional dimensions of the steel arch frame. The thickness of the steel protective layer determines the ability of the rigid outer protective layer to resist external impacts and maintain the temperature environment; the thickness of the polyurethane foam and the thickness of the phase change insulation material directly affect the thermal insulation effect of the rigid outer protective layer. The deformation and elastic coefficient of the elastic element reflect the elasticity and preload of the elastic component when adapting to soft rock deformation, while the elastic modulus of the buffer layer reflects the buffering performance of the buffer layer. The arch frame spacing and the cross-sectional dimensions of the steel arch frame are related to the support strength and overall stability of the arch frame. Through precise calculation of these parameters using machine learning models, the temperature-adaptive and stable composite arch frame structure can achieve the best support effect under different geological conditions, ensuring the structural safety of the tunnel, improving the economy and efficiency of construction, and avoiding problems such as support failure or resource waste caused by unreasonable parameters.

[0025] Optionally, the training of the machine learning model includes the following steps: constructing a database containing different geological conditions and tunnel geometric parameters, the database being generated through numerical simulation of the tunnel excavation and support process, the numerical simulation establishing a three-dimensional tunnel model based on parameters such as rock strength, rock unit weight, rock porosity, rock elastic modulus, rock deformation modulus, rock mass RQD value, rock water content, rock mass freeze-thaw volume expansion rate, tunnel surface area, tunnel maximum width, tunnel maximum height, and tunnel burial depth; normalizing the database, mapping the input and output data to a preset numerical range, and dividing it into training, validation, and test sets; using a long short-term memory network to process the time series data of the surrounding rock temperature field, extracting dynamic features, and then combining it with a random forest algorithm for model training, outputting the parameters of the temperature-adaptive and stable composite arch structure and the grouting timing.

[0026] Optionally, the training of the machine learning model further includes: extracting a subset of data from the training set using a bootstrap sampling method to construct a random forest decision tree, and selecting features and thresholds based on the principle of minimizing mean squared error when splitting nodes; and adjusting the number of hidden layers in the long short-term memory network and the number of trees in the random forest using cross-validation until the prediction accuracy of the model on the test set meets a preset threshold.

[0027] Optionally, based on real-time monitoring data of the soft rock pore structure and fracture development, the grouting flow rate and pressure parameters can be dynamically adjusted to adapt to the grouting needs of different areas.

[0028] One or more technical solutions provided in the embodiments of the present invention have at least the following technical effects or advantages:

[0029] 1. The soft rock support system for shallow buried tunnels in high-altitude permafrost areas of the present invention mainly includes a temperature-adaptive and stable composite arch frame structure, comprising a rigid outer protective temperature-regulating layer, an elastic component, and an arch frame. The rigid outer protective temperature-regulating layer, as a rigid support layer, is located on the radial outer side of the tunnel. A steel protective layer, an insulation layer, and a temperature-regulating layer are sequentially arranged. The elastic component is located between the rigid outer protective temperature-regulating layer and the arch frame, comprising an elastic element and a buffer layer. The two ends of the elastic element abut against the inner side of the rigid outer protective temperature-regulating layer and the outer side of the arch frame, respectively. The buffer layer fills the space between the support layer and the arch frame. The arch frame is located on the radial inner side of the tunnel. The temperature-adaptive and stable composite arch structure adopts a multi-layer design. The outer steel protective layer, the insulation layer (polyurethane foam), and the temperature-regulating layer (phase change insulation material) form a sandwich composite panel, which integrates low-temperature resistant steel, high-efficiency insulation materials, and intelligent temperature-regulating phase change materials. It can not only resist external impacts but also regulate the temperature of the surrounding rock and reduce the frequency of freeze-thaw cycles in soft rock. The middle elastic component utilizes the composite properties of elastic elements (nickel-titanium shape memory alloy wire) and buffer layers (natural rubber) to flexibly respond to the deformation of soft rock caused by temperature changes. The inner steel arch further strengthens the overall support force. All parts work together to fully adapt to the complex and variable geological and climatic conditions of high-altitude permafrost areas and enhance the stability of the support system.

[0030] 2. By numerically simulating different working conditions, the mechanical behavior of the support structure and the deformation law of the surrounding rock are analyzed to obtain the initial values ​​of structural parameters and grouting timing. Then, combined with machine learning training models, a large amount of data is used to learn the complex relationship between geological and tunnel parameters and support requirements. These two methods complement each other, enabling new projects to quickly and accurately predict various parameters and grouting timing based on actual conditions, achieving optimized design of the support system, improving construction accuracy, and saving manpower, material resources, and time costs.

[0031] 3. The monitoring and data transmission system integrates fiber optic displacement sensors, thermistor temperature sensors, and piezoelectric pressure sensors, strategically positioned along the tunnel's axial and circumferential directions. This allows for comprehensive and high-precision capture of millimeter-level deformation of the surrounding rock, dynamic changes in the temperature field, and contact pressure information, providing detailed first-hand data for understanding the tunnel's real-time status. Utilizing a low-power wireless transmission module, the data is aggregated in real-time to the data processing center. A cloud computing platform, coupled with specialized software, performs in-depth analysis of the massive amounts of monitoring data and provides visual visualization. Engineers can then accurately assess the stability of the surrounding rock and the working status of the support structure. Furthermore, by combining machine learning predictions of grouting timing, scientific and intelligent construction decisions can be made, avoiding problems such as blind grouting or untimely support.

[0032] Advantages of additional aspects of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In addition, the spacing or dimensions between components are exaggerated to show the position of each component, and the schematic diagrams are for illustrative purposes only.

[0034] Figure 1 This is a schematic diagram of the temperature-adaptive and stable composite arch frame structure provided in an embodiment of the present invention;

[0035] Figure 2 This is a flowchart illustrating the implementation of the support system provided in this embodiment of the invention;

[0036] In the diagram: 1. Modular rigid outer protective temperature-regulating layer; 2. Steel protective layer; 3. Insulation layer and temperature-regulating layer; 4. Buffer layer; 5. Elastic component; 6. Modular steel arch frame; Detailed Implementation

[0037] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0038] I. The soft rock support system for shallow-buried tunnels in high-altitude permafrost areas includes a temperature-adaptive and stable composite arch frame structure. This structure comprises a rigid outer temperature-regulating layer, an elastomer component, and an arch frame. The rigid outer temperature-regulating layer is a modular rigid support structure located radially outside the entire composite arch frame structure. It includes a radially outer steel protective layer 2 and a radially inner insulation layer and temperature-regulating layer 3. The elastomer component is located between the rigid outer temperature-regulating layer and the arch frame. It includes an elastic element 5 and a buffer layer 4. The two ends of the elastic element 5 abut against the inner side of the rigid outer temperature-regulating layer and the outer side of the arch frame, respectively. The buffer layer 4 fills the space between the rigid outer temperature-regulating layer and the arch frame. The arch frame is a modular steel arch frame 6 located radially inside the tunnel.

[0039] The outermost steel protective layer 2 uses high-strength, low-temperature resistant steel (such as Q355D steel) to directly resist the pressure of the surrounding rock and the impact of frozen soil blocks. The middle insulation layer (such as polyurethane foam) slows down the intrusion of external low temperatures through its low thermal conductivity. The inner temperature-regulating layer (such as paraffin-based phase change material) absorbs or releases heat through a phase change process, actively regulating the temperature of the surrounding rock and reducing the frequency of freeze-thaw cycles. The elastic component is located between the rigid outer protective temperature-regulating layer and the modular steel arch 6. It is composed of elastic elements 5 (such as nickel-titanium shape memory alloy wire) and a buffer layer 4 (such as natural rubber). The pre-tension of the elastic elements 5 offsets the shrinkage deformation of the soft rock, and the buffer layer 4 absorbs the expansion stress. The modular steel arch 6 is located on the innermost side and provides stable support through standardized splicing to maintain the tunnel outline. The radial distribution and synergistic effect of each layer significantly improve the adaptability of the support system to freeze-thaw deformation and temperature fluctuations.

[0040] Specifically:

[0041] (1) Temperature-adaptive and stable composite arch frame structure:

[0042] Modular Rigid Outer Sheath Temperature-Regulating Layer 1 (Outermost): This layer uses a high-strength, low-temperature resistant steel-polyurethane foam-phase change insulation material sandwich composite panel as its base material. The outer steel layer uses low-temperature resistant, high-strength Q355D steel, possessing excellent impact and compressive strength, capable of withstanding external forces such as soft rock deformation and frozen soil impacts in high-altitude areas. The middle layer of polyurethane foam is a highly efficient insulation material, slowing down the intrusion of external low temperatures into the tunnel interior. The inner layer of phase change insulation material plays a crucial role in temperature regulation, undergoing solid-liquid or solid-solid phase changes within a specific temperature range. It absorbs and stores heat during the day and releases it when temperatures plummet at night, thus maintaining a relatively stable surrounding rock temperature environment and reducing frequent freeze-thaw cycles caused by large diurnal temperature variations in soft rock. The phase change material can be paraffin-based; the phase change temperature of paraffin can be flexibly adjusted by changing the carbon chain length, effectively matching the temperature environment of high-altitude permafrost regions. To enhance the thermal conductivity of paraffin wax, graphite filler is added to it, forming thermally conductive pathways within the paraffin matrix. This improves the overall thermal conductivity of the material, allowing the paraffin wax to absorb and release heat more quickly during phase change, thus enhancing its ability to regulate the temperature of the tunnel surrounding rock. Understandably, other materials, such as fatty acids, can also be used in phase change materials.

[0043] This material combination not only improves the thermal insulation and stress regulation performance of the support system, but also enhances its ability to adapt to freeze-thaw deformation in soft rock. The overall structure is designed as an arched modular unit, with each unit spliced ​​together by high-strength bolts, facilitating rapid on-site assembly and disassembly to meet the needs of different stages of tunnel construction.

[0044] The elastomeric component (middle section) is located between the modular rigid outer protective temperature-regulating layer 1 and the modular steel arch frame 6. It is entirely composed of nickel-titanium shape memory alloy wires and natural rubber. In low-temperature environments, the shape memory alloy wires apply a moderate preload to the rubber according to a pre-set shape, counteracting the shrinkage deformation of the soft rock caused by freeze-thaw cycles. When the temperature rises and the soft rock expands, the alloy wires adapt to the deformation, working in conjunction with the rubber to buffer the expansion force and maintain a stable support state. It is understood that the elastomeric component 5 can also be replaced with other shape memory alloys (such as copper-aluminum-nickel alloys).

[0045] Modular steel arch frame 6 (innermost): Provides further support strength for the entire support structure, while also supporting the modular rigid outer protective temperature regulating layer 1 and the elastomer components, ensuring the stability of the tunnel's internal profile, and working in conjunction with the outer rigid outer protective temperature regulating layer and the middle elastomer components to jointly bear the surrounding rock pressure.

[0046] (2) Monitoring and data transmission system:

[0047] Sensor Network: Integrating various types of sensors, including fiber Bragg grating displacement sensors, thermistor temperature sensors, and piezoelectric pressure sensors, the network is rationally laid out along the tunnel's axial and circumferential directions. The fiber Bragg grating displacement sensors are densely deployed to accurately capture millimeter-level deformations in soft rock; thermistors monitor changes in the surrounding rock temperature field in real time to determine freeze-thaw conditions; and piezoelectric pressure sensors are installed at the interface between the support structure and the soft rock to obtain real-time contact pressure.

[0048] Data transmission and processing: Data collected by sensors is transmitted in real time to the data processing center at the tunnel entrance via a low-power, interference-resistant wireless transmission module. A cloud computing platform equipped with professional data analysis software is used to analyze and process the massive amounts of monitoring data in real time. Visual displays, such as deformation-time curves and temperature distribution cloud maps, allow engineers to intuitively understand the tunnel's condition.

[0049] (3) Grouting system:

[0050] Grouting Material Storage and Transportation: Equipped with multiple storage tanks to store different types of grouting materials, such as low-temperature early-strength cement grout and quick-setting chemical grout. Through an intelligent pipeline system, different proportions of mixed grout are precisely prepared and transported to various grouting points within the tunnel according to preset programs or on-site instructions. The pipelines are wrapped with insulation material and heated cables to prevent the grout from freezing during transportation.

[0051] Grouting equipment: High-pressure grouting pump set is adopted, which has adjustable flow and pressure functions. According to the pore structure and fracture development of soft rock, the grouting parameters are flexibly controlled to ensure that the grout is evenly diffused and fills the voids in soft rock, thereby enhancing the strength of the surrounding rock.

[0052] During the grouting process, the high-pressure grouting pump unit precisely mixes different proportions of grout according to a preset program or on-site instructions and delivers it to various grouting points within the tunnel. This ensures that the grout is evenly diffused and fills the voids in the soft rock, enhancing the strength of the surrounding rock. Grouting parameters such as flow rate and pressure can be flexibly adjusted according to the pore structure and fracture development of the soft rock to achieve the best reinforcement effect. This grouting system design fully considers the fluidity of the grout in low-temperature environments and the geological characteristics of soft rock, improving the reliability and effectiveness of grouting reinforcement.

[0053] II. Methods for Determining Structural Parameters and Grouting Timing

[0054] (1) Numerical simulation calculation:

[0055] Numerical simulation calculations were performed to simulate the tunnel excavation and support process under different geological conditions and tunnel cross-section requirements. The mechanical behavior and surrounding rock deformation law of the temperature-adaptive and stable composite arch frame structure under different working conditions were analyzed. The reasonable values ​​of various structural parameters (such as the thickness of the steel protective layer 2, the mechanical performance parameters of the elastic component, the spacing and cross-sectional dimensions of the steel arch frame, etc.) under different conditions were calculated and inverted. The optimal grouting effect was determined when the deformation of the surrounding rock of the tunnel reached a certain level, that is, the timing of grouting was determined.

[0056] (2) Database establishment and machine learning training:

[0057] The results obtained from numerical simulations were compiled into a database, covering various geological conditions, tunnel cross-sectional parameters, structural parameters, and corresponding grouting timing. A machine learning algorithm combining LSTM (Long Short-Term Memory) and Random Forest was used to train a model on this database. The trained model can then quickly predict and calculate various parameters of the temperature-adaptive and stable composite arch structure and the most suitable grouting timing based on new geological conditions and tunnel design requirements. This provides a scientific basis for actual engineering projects, enabling optimized design of the support system and precise control of construction timing.

[0058] The training data for the machine learning algorithm combining LSTM (Long Short-Term Memory) and Random Forest is as follows:

[0059] Input data: rock strength, rock unit weight, rock porosity, rock elastic modulus, rock deformation modulus, rock mass RQD value (rock quality index, which reflects the degree of joint and fracture development), rock water (ice) content, volume expansion rate of rock mass corresponding to different freeze-thaw cycles, tunnel surface area, maximum tunnel width, maximum tunnel height, tunnel burial depth, and surrounding rock temperature field variation data (temperature changes over time).

[0060] Output data: thickness of the outer steel protective layer 2 of the temperature-comfortable composite arch frame, thickness of polyurethane foam, thickness of phase change insulation material, maximum deformation of nickel-titanium shape memory alloy wire in the elastomer component, elastic coefficient, elastic modulus of rubber, spacing of steel arch frames, cross-sectional length of steel arch frames, cross-sectional width of steel arch frames, grouting timing (i.e., when the maximum surrounding rock deformation reaches before grouting begins).

[0061] III. Working principle of the support system:

[0062] (1) During the tunnel excavation process, the temperature-adaptive and stable composite arch frame structure immediately plays its role. The rigid outer protective temperature-regulating layer resists external impact and surrounding rock pressure, the elastic component adapts to the deformation of the surrounding rock and adjusts the stress distribution, and the steel arch frame provides internal support.

[0063] (2) The monitoring and data transmission system monitors the surrounding rock deformation, temperature and contact pressure in real time and transmits the data to the data processing center for analysis. Based on the analysis results, the engineers judge the working status of the support structure and the stability of the surrounding rock.

[0064] (3) When the grouting time is predicted based on monitoring data and machine learning, the grouting system is started and the appropriate grout is injected into the surrounding rock according to the set parameters to reinforce the surrounding rock and further improve the stability of the tunnel. All parts of the support system work together to ensure the safe construction and operation of the soft rock section of the shallow buried tunnel in the high-altitude permafrost area.

[0065] Specific construction methods are as follows: Figure 2 As shown:

[0066] Step 1: Based on a large amount of existing geological survey data of high-altitude permafrost areas, set working conditions, construct multi-scenario tunnel models, carry out numerical simulation calculations, obtain the optimal values ​​of various parameters, and accumulate massive amounts of computational data.

[0067] 1. Model Establishment and Working Condition Setting: Utilizing professional numerical simulation software (such as ANSYS, FLAC3D, etc.), and based on a large amount of existing geological survey data from high-altitude permafrost regions, the model covers the variation range of parameters such as rock strength, unit weight, porosity, elastic modulus, deformation modulus, rock mass RQD value, and water (ice) content in different strata. Combined with different tunnel cross-sectional dimensions (including area, maximum width, maximum height, and burial depth), a rich variety of 3D tunnel models are constructed. For each model, multiple working conditions are set, fully considering factors such as seasonal temperature changes, different surrounding rock types, and complex groundwater runoff, ensuring that the simulation scenarios are comprehensive and representative.

[0068] 2. Simulation Process and Data Recording: The simulation covers the entire process of tunnel excavation and support, accurately simulating the actual tunnel advance distance and dynamically updating the stress and strain state of the surrounding rock in real time. At each stage of excavation, the various components of the temperature-stability and support composite arch structure are incorporated into the model according to the design concept, and are given realistic material properties and mechanical boundary conditions. Detailed mechanical response data and surrounding rock deformation data for each part of the temperature-stability and support composite arch structure under different working conditions are recorded. The simulation duration covers the entire tunnel construction cycle and a certain operational period, accumulating a massive amount of simulation results data.

[0069] 3. Parameter Analysis and Initial Database Construction: Through data analysis, the optimal range of values ​​for structural parameters such as the thickness of the steel protective layer 2, the mechanical performance parameters of the elastomeric components (e.g., maximum deformation of nickel-titanium wire, elastic coefficient, and elastic modulus of rubber), the spacing of the steel arch frames, and the cross-sectional dimensions were determined under different geological conditions and tunnel cross-section combinations. A curve showing the relationship between surrounding rock deformation and grouting effect (precisely quantified by the reduction in the deformation rate of the surrounding rock after grouting) was plotted to obtain data on the optimal grouting timing. Based on this, an initial parameter database containing rich information was compiled, providing a solid data foundation for subsequent machine learning training.

[0070] Step 2: Improve the database, construct Long Short-Term Memory (LSTM) and Random Forest machine learning models, and conduct model training.

[0071] 1. Database Setup and Data Preprocessing: A dedicated database for tunnel support was created using efficient database management systems such as MySQL and SQL Server. Data obtained from numerical simulations were categorized and entered, with comprehensive and accurate labels added to each data set, covering key information such as geological conditions, tunnel cross-sectional parameters, structural parameters, and grouting timing. A complete index and relational structure were built for easy retrieval and retrieval, resulting in a high-quality dataset with a clear structure and complete data. The dataset specifically includes input data: rock strength... Rock density Rock porosity Rock elastic modulus Rock deformation modulus RQD value of rock mass, water (ice) content of rock. Volume expansion rate of rock mass corresponding to different freeze-thaw cycles Tunnel surface area Maximum tunnel width Tunnel maximum height Tunnel depth Surrounding rock temperature field variation data Output data: Thickness of the outer steel protective layer of the temperature-stability composite arch frame (2mm). Polyurethane foam thickness Thickness of phase change insulation material The maximum deformation of nickel-titanium shape memory alloy wire in elastomeric components Elasticity coefficient Elastic modulus of rubber Spacing of steel arches Length of steel arch frame cross section , width of steel arch frame section Grouting timing The data above is normalized to map data of different dimensions to the interval [0,1] or [-1,1], effectively avoiding training bias caused by differences in data dimensions. The dataset is then divided into training, validation, and test sets according to a reasonable ratio (e.g., 70%, 20%, 10%).

[0072] 2. Model Architecture Construction: The machine learning model is built based on an advanced architecture that combines Long Short-Term Memory Network (LSTM) and Random Forest, giving full play to the advantages of LSTM in processing time series data (surrounding rock temperature field change data) and the ability of Random Forest to process complex multi-feature data.

[0073] 3. Model Training and Optimization Iteration: During model training, the unique gating mechanism of LSTM (input gate, forget gate, output gate) is used to perform in-depth processing on the surrounding rock temperature field change data. Through complex formula calculations and neuron operations, key information of past temperature changes is selectively remembered, highly representative dynamic features are extracted, and these features are precisely concatenated with other non-time series input data along the feature dimension to form a comprehensive feature vector.

[0074] The core principle formula of Long Short-Term Memory (LSTM) networks is:

[0075] Input Gate:

[0076] Forgotten Gate:

[0077] Output gate:

[0078] Memory unit:

[0079] Hidden state:

[0080] in, It is the input at the current moment (i.e., the input data in the dataset at the current moment). and These are the hidden state and the memory unit from the previous time step, respectively. They combine information from the input data of previous time steps and are affected by the current input. The influence of this, in turn, affects the hidden state at the current moment. and memory unit . It is a weight matrix. These are bias vectors, parameters that need to be learned during model training. They determine how the input data and hidden states interact to extract key features from the input data. It is the sigmoid function, and its mathematical expression is: Through these gating mechanisms, LSTM selectively memorizes key information about past temperature changes and extracts dynamic features.

[0081] Subsequently, the random forest begins constructing decision trees. Each decision tree uses a bootstrap sampling method to sample with replacement from the complete training set, generating a subset of the dataset with the same number of samples as the original training set. Assume the original training set has a sample size of... (The sample here includes input and output data from the database), for the first... Training subset of trees Each time from the original training set When a sample is drawn from the pool, the probability of that sample being drawn is always 1 / 3. Repeat this process. That's how I got it. .

[0082] Decision tree construction and split point finding: When splitting at each node of the decision tree, a feature set is provided. Randomly select a feature subset from it (Usually take size) Taking a regression tree as an example, if the sample set of the current node is... (Including input and output parameters), for a selected feature Try different splitting thresholds Calculate the mean square error (MSE) after splitting. Let... Split into and The actual value is The predicted value is Before the split After splitting , choose to smallest and As the basis for splitting, the splitting continues until a stopping condition is met (such as the tree depth reaching a preset value, or the number of node samples being too small), thus completing the construction of the decision tree. During this process, the true value... It could represent any parameter in the output data, such as , Etc., predicted value It is the corresponding value predicted by the model based on the input data (various input parameters).

[0083] Comprehensive prediction and hyperparameter tuning: Each tree for each sample Give the predicted value Final predicted value ,here This represents the total number of decision trees. (Using...) Cross-validation divides the training set into equal parts. Shares, used in turn One training pass, one validation pass. Adjust the number of hidden layers in the LSTM during each training iteration. Number of trees in a random forest Equal hyperparameters, calculate the loss function (such as mean squared error) on the validation set. Find the hyperparameter combination that minimizes the loss, repeat the training and validation process until the model achieves the required accuracy and stability on the test set, and the training is complete.

[0084] Step 3: Conduct geological and tunnel parameter surveys to obtain parameters of the tunnel to be built, and use machine learning to calculate the parameters of the composite arch frame structure.

[0085] 1. Geological and Tunnel Parameter Survey: For the tunnel to be constructed, comprehensive geological information along the tunnel route is obtained using exploration techniques and equipment, including rock strength. Rock density Rock porosity Rock elastic modulus Rock deformation modulus RQD value of rock mass, water (ice) content of rock. Volume expansion rate of rock mass corresponding to different freeze-thaw cycles Key parameters, and accurately measure the tunnel surface area. Maximum tunnel width Tunnel maximum height Tunnel depth By using geometric parameters and setting up dense monitoring points, the temperature field changes of the surrounding rock can be continuously monitored over a long period of time. This ensures that the data obtained is accurate, comprehensive, and timely.

[0086] 2. Calculation of structural parameters for the temperature-adaptive and stable composite arch frame: The various parameters of the tunnel to be constructed are input into a pre-trained machine learning model according to a predetermined format. This model quickly and accurately calculates and outputs the thickness of the outer steel protective layer 2 of the temperature-adaptive and stable composite arch frame. Polyurethane foam thickness Thickness of phase change insulation material The maximum deformation of nickel-titanium shape memory alloy wire in elastomeric components Elasticity coefficient Elastic modulus of rubber Spacing of steel arches Length of steel arch frame cross section , width of steel arch frame section Grouting timing These key data provide a precise design basis for the construction of the support system.

[0087] Step 4: Assemble the temperature-adaptive and stable composite arch frame structure, deploy the monitoring and data transmission system, install the grouting system, and build a soft rock support system for shallow buried tunnels in high-altitude permafrost areas.

[0088] 1. Assembling a temperature-controlled and stable composite arch frame structure

[0089] Modular Rigid Outer Sheath Temperature-Regulating Layer 1 Fabrication and Installation: Based on model calculations, suitable low-temperature resistant, high-strength Q355D steel is selected as the outer layer steel. High-quality polyurethane foam with low thermal conductivity and high closed-cell rate is selected, along with suitable paraffin-based phase change insulation material (its phase change temperature is precisely adjusted according to the local temperature environment, and an appropriate amount of graphite filler is added to improve thermal conductivity). The polyurethane foam is cut and tightly bonded to the inside of the steel, then the prepared phase change material is evenly applied or poured in to create standard arched modules. After rigorous quality inspection, these modules are transported to the tunnel construction site. Inside the tunnel, high-precision positioning tools are used to accurately place the modules, and high-strength bolts are used to splice them according to the specified torque to ensure a tight and stable connection, effectively blocking external low temperatures and resisting impacts.

[0090] Elastomer component installation: Following the elastomer component parameters output from the model, nickel-titanium shape memory alloy wires are combined with natural rubber. Temperature, pressure, and time parameters are precisely controlled to ensure the alloy wires are evenly distributed within the rubber, guaranteeing that the alloy wires apply appropriate pre-tension to the rubber according to the preset shape even at low temperatures. The elastomer component is then accurately installed between the modular rigid outer protective temperature-regulating layer 1 and the modular steel arch frame 6. High-precision tools such as laser levels are used to calibrate the component's position, ensuring a tight fit and optimal stress distribution, effectively buffering soft rock deformation.

[0091] Modular steel arch frame 6 construction: Based on the calculated steel arch frame parameters, appropriate specifications of steel are selected to fabricate modular steel arch frames 6, strictly controlling dimensional parameters such as spacing, cross-sectional length, and height of the steel arch frames to enhance their durability. Inside the tunnel, cranes and other hoisting equipment are used to lift the steel arch frames one by one to the predetermined installation positions. They are then coordinated and calibrated with the outer rigid outer protective temperature-regulating layer and the intermediate elastic components. High-precision measuring instruments such as levels and total stations are used to monitor the verticality and horizontality of the steel arch frames in real time. Shims and bolt fine-tuning are used to ensure the steel arch frames are securely installed, providing solid internal support for the entire support structure and maintaining the stability of the tunnel outline.

[0092] 2. Deployment of Monitoring and Data Transmission System

[0093] Sensor network layout: Based on the tunnel's design axis, sensors are densely deployed in complex geological areas such as tunnel entrances and exits, geological transition zones, and fault fracture zones. Generally, one group is placed every 1-2 meters, while in normal geological sections, one group is placed every 3-5 meters. For axial layout, specialized drilling equipment is used to drill installation holes of appropriate depth in the tunnel wall. Fiber optic displacement sensors, thermistor temperature sensors, and piezoelectric pressure sensors are then accurately embedded sequentially and securely fixed with anchoring agent to ensure close contact between the sensors and the surrounding rock for accurate data acquisition. For circumferential layout, sensors are deployed at key locations on the same tunnel cross-section, such as the arch crown, side arch waists, and arch feet. The sensor spacing angle is set according to the tunnel radius and monitoring accuracy requirements (generally 30°-60°). Specially designed ring supports are used to firmly fix the sensors, ensuring comprehensive monitoring of surrounding rock deformation, temperature, and pressure.

[0094] Data transmission and processing system setup: Each sensor is equipped with a low-power, interference-resistant wireless transmission module, utilizing advanced and mature wireless communication technologies such as ZigBee and LoRa to ensure the stability and reliability of data transmission. Signal repeaters are installed every 30-50 meters at suitable locations on the tunnel wall, based on signal attenuation, to ensure stable and real-time transmission of sensor data to the data processing center at the tunnel entrance. A professional data processing center is built at the tunnel entrance, equipped with a high-performance server cluster and large-capacity storage devices to meet the storage and rapid processing needs of massive amounts of monitoring data. Professional data analysis software (such as a Python-based data analysis platform) is installed, and a large-screen visualization display is built. The transmitted data is connected to the processing software through an efficient data interface, enabling real-time reception, in-depth analysis, and intuitive display of monitoring data, providing engineers with accurate tunnel status information.

[0095] 3. Grouting system installation

[0096] Material storage and transportation system installation: Multiple stainless steel storage tanks are installed in areas close to the tunnel construction site with convenient transportation, flat terrain, and good drainage. These tanks store different types of grouting materials, such as low-temperature early-strength cement grout and quick-setting chemical grout. The storage tanks are equipped with level sensors, stirring motors, and other equipment. The stirring frequency and duration are precisely set according to the characteristics of different grouting materials (e.g., the initial setting time of low-temperature early-strength cement grout and the chemical reactivity of quick-setting chemical grout) to prevent grout sedimentation and stratification. An intelligent pipeline system is laid connecting the storage tanks to the tunnel grouting points. High-performance polyurethane insulated pipes are selected, with a tightly wrapped outer layer of heat-tracing cable. The cable spacing is scientifically set according to the pipeline's heat dissipation. High-precision flow meters and pressure sensors are installed at key pipeline nodes (such as tees and valves), working in conjunction with electric regulating valves to achieve precise monitoring and control of grout flow and pressure, ensuring accurate grout preparation and transportation.

[0097] Grouting equipment placement: The high-pressure grouting pump set with adjustable flow and pressure functions is hoisted to the designated location inside the tunnel and connected to a stable and reliable power supply line and grouting pipeline. The pump set undergoes comprehensive commissioning, testing the flow adjustment range and pressure output stability under different operating conditions. Based on data obtained from preliminary geological surveys regarding the pore structure and fracture development of the soft rock, multiple sets of scientifically sound grouting parameter templates are pre-set for quick recall during construction, ensuring the efficiency and quality of the grouting operation.

[0098] Step 5: Operate the support system, monitor and analyze the condition of the surrounding rock and support structure, and carry out grouting reinforcement of the surrounding rock as appropriate.

[0099] 1. Excavation Stage Support: During tunnel excavation, the advance is strictly in accordance with the predetermined advance rate of 0.8-1.2 meters. After each section of excavation is completed, the construction personnel quickly hoist the pre-fabricated modular rigid outer protective temperature-regulating layer 1 to the working face and assemble and install it. Then, the elastic components and modular steel arch frame 6 are installed, so that the support structure can quickly play its role, resist the pressure of the surrounding rock and external impact, and adapt to the initial deformation of the surrounding rock.

[0100] 2. Real-time Monitoring and Analysis: The monitoring and data transmission system operates continuously around the clock. Sensors capture data such as surrounding rock deformation, temperature, and contact pressure in real time, transmitting them to the data processing center within seconds via a high-efficiency wireless transmission link. The data processing center's cloud computing platform, equipped with specialized software, receives the data in real time, employs advanced filtering algorithms to remove noise interference, and then conducts comprehensive and in-depth data analysis. Real-time deformation-time curves, temperature distribution cloud maps, and pressure contour maps are generated, visually displaying the tunnel's condition. Based on these visualized charts, engineers can accurately assess the working status of the support structure and the stability of the surrounding rock, providing a scientific basis for construction decisions.

[0101] 3. Grouting Reinforcement: Grouting begins when monitoring data shows that the deformation of the surrounding rock reaches the grouting timing predicted by the machine learning model. The intelligent pipeline system quickly and accurately mixes different proportions of grout according to a preset program, and the high-pressure grouting pump unit simultaneously and rapidly adjusts the flow and pressure parameters to initiate the grouting process. During grouting, operators closely monitor the feedback data from the flow meter and pressure sensor, dynamically adjusting the grouting parameters based on data changes to ensure that the grout evenly diffuses and fills the pores and fissures of the soft rock, effectively reinforcing the surrounding rock and improving the overall stability of the tunnel.

[0102] To make the technical solution provided by this invention clearer, two examples are given to illustrate the invention:

[0103] Example 1

[0104] In a railway tunnel project, the soft rock support system and method for shallow buried tunnels in high-altitude permafrost areas proposed in this invention were used to assist in the initial support construction of the tunnel.

[0105] I. Project Overview

[0106] A railway tunnel is located in a high-altitude permafrost region at an elevation of approximately 4,800 meters. The geological conditions of the shallow-buried section are extremely complex, traversing a layer mainly composed of silty clay and gravel, with an average rock strength of only 30 MPa, classifying it as typical soft rock. The average annual temperature in this area is -8°C, with a diurnal temperature range of up to 35°C. Permafrost is widely distributed, with a freeze-thaw layer thickness ranging from 3 to 6 meters. Groundwater flow is complex and variable, posing a significant challenge to tunnel construction.

[0107] II. Implementation Process

[0108] (1) Preliminary data preparation and model training

[0109] Numerical simulation calculations: Extensive geological survey data from surrounding high-altitude permafrost areas were collected. FLAC3D software was used to construct three-dimensional tunnel models encompassing various combinations of parameters such as rock strength (25-35 MPa), porosity (15%-25%), and elastic modulus (5-15 GPa), as well as multiple tunnel cross-sectional dimensions (area 80-120 square meters, width 10-14 meters, height 8-12 meters, burial depth 50-100 meters). Seasonal factors such as high summer temperatures, extreme winter cold, and sudden temperature changes in spring and autumn were considered. Combined with different surrounding rock types (soft rock, fractured rock, etc.) and complex groundwater flow conditions, the tunnel excavation and support process was simulated. Detailed records were kept of the mechanical response of the temperature-adaptive and stable composite arch frame and surrounding rock deformation data under various working conditions. The simulation period covered two years of construction and three years of operation, accumulating a massive amount of data.

[0110] Database Improvement and Model Training: A database was established using SQL Server. Simulated data was entered and accurately labeled with geological, tunnel, structural parameters, and grouting timing information. After normalization, the input data was divided into 70%, 20%, and 10% datasets. An LSTM-Random Forest model was built. LSTM processed the surrounding rock temperature field data using its gating mechanism, and the data was then concatenated with other datasets for training the Random Forest. After multiple adjustments to hyperparameters such as the LSTM hidden layers (set to 3 layers), the number of random forest trees (set to 150), and 10-fold cross-validation, the model achieved an accuracy of 92% on the test set, completing the training.

[0111] (2) Construction of support system

[0112] Assembling a temperature-controlled and stable composite arch frame structure:

[0113] Modular rigid outer protective temperature-regulating layer 1 fabrication and installation: Select Q355D steel according to the model, whose low-temperature impact toughness reaches 40J / cm. 2 The polyurethane foam has a thermal conductivity of 0.022 W / (m•K) and a closed-cell rate of 92%. The paraffin phase change material has a phase change temperature of -8℃ to -4℃ and contains 12% graphite filler.

[0114] Elastomer component installation: According to the model parameters, nickel-titanium wire and rubber are composited in a professional factory, with a pre-tightening force of 120kN and an elastic modulus of 850MPa.

[0115] Modular steel arch frame 6: Select I22a I-beams to process the steel arch frame, with a spacing of 0.7 meters, a cross-section length of 220 mm, a height of 220 mm, an error of ±1.5 mm, and hot-dip galvanized treatment.

[0116] Deployment of monitoring and data transmission systems:

[0117] Sensor network layout: A set of sensors is arranged every 1 meter in the tunnel axial entrance and exit and complex areas, and every 4 meters in normal sections; sensors are fixed at 45° intervals at the arch crown, arch waist and arch foot of the cross section with ring brackets, and anchored with holes drilled to a depth of 35 cm to ensure good contact.

[0118] Data transmission and processing system setup: Sensors are equipped with LoRa modules, and repeaters are installed every 40 meters to ensure data transmission. A data center is built at the tunnel entrance, equipped with a server cluster of 10-core processors, 20GB of memory, and 15TB of storage devices. Python data analysis software and a large visualization screen are installed to process and display data in real time.

[0119] Grouting system installation:

[0120] Material storage and conveying system installation: Four stainless steel storage tanks will be built 600 meters from the tunnel, with mixing parameters set according to the slurry characteristics. Polyurethane insulated pipes will be laid and wrapped with heat tracing cables (spaced 0.45 meters apart). Monitoring and control components will be installed at key points to ensure the smooth preparation and conveying of the slurry.

[0121] Grouting equipment in place: hoist the high-pressure grouting pump set to the tunnel, adjust the flow rate to 0-120L / min and the pressure to 0-6MPa, and preset 6 sets of grouting parameters.

[0122] (3) Operation of the support system

[0123] Excavation stage support: Excavate at a rate of 0.8-1.2 meters per section, and quickly install the support structure after each section is excavated. First, hoist and splice the rigid outer protective temperature-regulating layer, then install the elastic components and steel arch frame to enable it to take effect quickly, resist the pressure and impact of the surrounding rock, and adapt to deformation.

[0124] Real-time monitoring and analysis: The monitoring system runs continuously, with sensors transmitting data to the processing center in seconds. The software filters and analyzes the data, creating charts to display the tunnel conditions and assisting in decision-making.

[0125] Grouting reinforcement: When the monitoring reaches the grouting time, the intelligent system allocates grout, the grouting pump set adjusts parameters to start grouting, and the operator adjusts according to feedback to ensure that the grout fills and reinforces the surrounding rock, ensuring the safety and stability of the tunnel.

[0126] III. Implementation Results

[0127] During construction, the maximum deformation of the tunnel was controlled within 12 mm, the support structure was stable, and no safety accidents occurred, effectively verifying the feasibility and reliability of the support system and method in the construction of railway tunnels in high-altitude permafrost areas.

[0128] Example 2

[0129] In a highway tunnel project, the soft rock support system and method for shallow buried tunnels in high-altitude permafrost areas proposed in this invention were used to assist in the initial support construction of the tunnel.

[0130] I. Project Overview

[0131] This highway tunnel is located in a high-altitude permafrost region at an elevation of approximately 4,300 meters. It is 3,000 meters long in total, with a 500-meter shallow-buried section. It traverses sandy mudstone strata with a rock strength of approximately 25 MPa and an RQD value of 35-45, exhibiting well-developed joints and fissures. The region has an average annual temperature of -6°C, a diurnal temperature range of 30°C, frequent freeze-thaw cycles in the permafrost, and abundant and complex groundwater flow, posing significant challenges to tunnel construction.

[0132] II. Implementation Process

[0133] (1) Preliminary data preparation and model training

[0134] Numerical simulation calculation: Integrating regional geological data, using ANSYS software to construct tunnel models with various parameter combinations and cross-sectional dimensions, considering different seasonal temperatures (-10℃ to 5℃), surrounding rock (different weathering degrees of sandy mudstone) and groundwater conditions, simulating excavation and support, recording data, and the simulation period covering 1.5 years of construction and 2.5 years of operation to obtain a large amount of data.

[0135] Database Improvement and Model Training: A MySQL database was created to input and label data. After normalization, the dataset was divided, and an LSTM-Random Forest model was built and trained. After optimizing hyperparameters such as the LSTM hidden layers (2 layers) and the number of random forest trees (120 trees) and implementing 8-fold cross-validation, the model achieved an accuracy of 90% on the test set.

[0136] (2) Construction of support system

[0137] Assembling a temperature-controlled and stable composite arch frame structure:

[0138] Modular rigid outer protective temperature-regulating layer 1 fabrication and installation: Select Q355D steel (low-temperature yield strength 345MPa), polyurethane foam with thermal conductivity of 0.028W / (m•K) and closed-cell rate of 90%, paraffin phase change material with a phase change temperature of -6℃ to -6℃, and add 10% graphite filler. The modules are fabricated, transported to the site for assembly, and the seal is checked.

[0139] Elastomer component installation: composite nickel-titanium wire and rubber, with precise parameter control, preload force of 100kN, elastic modulus of 800MPa, and fixture installation and calibration.

[0140] Modular steel arch frame 6: Steel arch frames are fabricated using I20a I-beams, with a spacing of 0.6 meters, a cross-section length of 200 mm, a height of 200 mm, and an error of ±2 mm. After anti-corrosion treatment, the frames are hoisted, installed, and calibrated.

[0141] Deployment of monitoring and data transmission systems:

[0142] Sensor network layout: Sensors are arranged every 1.5 meters in special sections and every 3.5 meters in general sections along the axis. Sensors are installed at 50° intervals in key parts of the cross section and anchored by drilling 30 cm holes.

[0143] Data transmission and processing system setup: Sensors are equipped with ZigBee modules, repeaters are installed every 35 meters, and a data center is built at the tunnel entrance to process and display data.

[0144] Grouting system installation:

[0145] Material storage and conveying system installation: Three stainless steel storage tanks are set 400 meters from the tunnel, with mixing parameters set, and insulated pipes laid and control components installed.

[0146] Grouting equipment in place: hoist the grouting pump set for commissioning, with a flow rate of 0-100L / min and a pressure of 0-5MPa, and 5 preset parameters.

[0147] (3) Operation of the support system

[0148] Excavation stage support: Excavate and install support structures according to the advance requirements to ensure timely support.

[0149] Real-time monitoring and analysis: The monitoring system collects and analyzes data during operation, and generates charts to assist in decision-making.

[0150] Grouting reinforcement: Grouting is initiated when the grouting time is reached, and parameters are adjusted according to feedback to ensure tunnel stability.

[0151] III. Implementation Results

[0152] During construction, the tunnel deformation was controlled within 10 millimeters, and no major safety issues occurred, indicating that the support system and method are suitable for highway tunnel projects in high-altitude permafrost areas, effectively ensuring construction and operation safety.

[0153] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A construction method of a shallow-buried tunnel soft rock support system in a high-altitude permafrost region, characterized in that: the support system comprises a warm and stable protection composite arch structure, which comprises a rigid outer protection temperature regulation layer, an elastomer component and an arch; the rigid outer protection temperature regulation layer is a modular structure located on the radial outer side of the tunnel, and comprises a radial outer steel protection layer and a radial inner thermal insulation layer and temperature regulation layer; the elastomer component is located between the rigid outer protection temperature regulation layer and the arch, and comprises an elastic member and a buffer layer, the two ends of the elastic member abut against the inner side of the rigid outer protection temperature regulation layer and the outer side of the arch respectively, and the buffer layer fills the space between the rigid outer protection temperature regulation layer and the arch; the arch is a modular steel arch located on the radial inner side of the tunnel; the construction method of the support system comprises the following steps: obtaining the geological parameters and tunnel geometric parameters of the tunnel to be built; inputting the obtained geological parameters and tunnel geometric parameters into a pre-trained machine learning model to output warm and stable protection composite arch structure parameters and grouting timing; based on the output warm and stable protection composite arch structure parameters, assembling the warm and stable protection composite arch structure, and during the tunnel excavation process, after completing a section of excavation, sequentially installing the modular rigid outer protection temperature regulation layer, the elastomer component and the modular steel arch; deploying a sensor network along the axial and circumferential directions of the tunnel, and transmitting sensor data in real time to a data processing center through a wireless transmission module; starting grouting when the monitoring data shows that the surrounding rock deformation reaches the predicted grouting timing.

2. The construction method of a soft rock support system of a shallow-buried tunnel in an alpine permafrost region according to claim 1, characterized in that, The thermal insulation layer uses polyurethane foam, the temperature regulation layer uses phase change thermal insulation material, the elastic member uses nickel-titanium shape memory alloy wire, and the buffer layer uses natural rubber.

3. The construction method of the soft rock support system of the shallow-buried tunnel in the high-altitude permafrost region according to claim 1, characterized in that, It also includes a monitoring and data transmission system, which comprises a sensor network and a transmission module, the sensor network comprises sensors arranged in the axial direction of the tunnel, at the entrance and exit of the tunnel, in the geological transition zone and in the fault fracture zone, and sensors arranged at the vault, both sides of the haunch and the arch foot of the same tunnel cross section, the sensor network comprises displacement sensors, temperature sensors and pressure sensors.

4. The construction method of a soft rock support system of a shallow tunnel in a high-altitude permafrost region according to claim 1, characterized in that, It also includes a grouting system, which comprises a storage tank, a stirring motor and a liquid level sensor, the storage tank is used to store grouting materials, the stirring motor is arranged in the storage tank to stir the grouting materials, and the liquid level sensor is installed in the storage tank to detect the liquid level of the grouting materials.

5. The construction method of the soft rock support system of the shallow-buried tunnel in the high-altitude permafrost region according to claim 1, characterized in that, The geological parameters include rock strength, rock bulk density, rock porosity, rock elastic modulus, rock deformation modulus, rock mass RQD value, rock water content and rock mass freeze-thaw volume expansion rate, and the tunnel geometric parameters include tunnel area, maximum tunnel width, maximum tunnel height and tunnel burial depth.

6. The construction method of a soft rock support system of a shallow tunnel in a high-altitude permafrost region according to claim 2, characterized in that, The warm and stable protection composite arch structure parameters include steel protection layer thickness, polyurethane foam thickness, phase change thermal insulation material thickness, elastic member deformation amount, elastic member elastic coefficient, buffer layer elastic modulus, arch spacing and steel arch cross section size.

7. The construction method of the soft rock support system of the shallow-buried tunnel in the high-altitude permafrost region according to claim 1, characterized in that, The training of the machine learning model comprises the following steps: A database containing different geological conditions and tunnel geometric parameters is constructed, which is generated by numerical simulation of the tunnel excavation and support process, and the numerical simulation is based on rock strength, rock bulk density, rock porosity, rock elastic modulus, rock deformation modulus, rock mass RQD value, rock water content, rock freeze-thaw volume expansion rate, tunnel surface area, maximum tunnel width, maximum tunnel height and tunnel burial depth parameters to establish a three-dimensional tunnel model; The database is normalized to map input data and output data to a preset numerical interval and divided into a training set, a validation set and a test set; The long short-term memory network is used to process the surrounding rock temperature field time series data, and after extracting the dynamic characteristics, the random forest algorithm is combined for model training to output the warm and stable support composite arch structure parameters and grouting timing.

8. The construction method of a soft rock support system of a shallow tunnel in a high-altitude permafrost region according to claim 1, characterized in that, The training of the machine learning model also includes: A sub-data set is extracted from the training set by the bootstrap sampling method to construct a random forest decision tree, and when the node is split, the feature and threshold are selected based on the principle of minimum mean square error; The number of hidden layers of the long short-term memory network and the number of trees of the random forest are adjusted by cross-validation until the prediction accuracy of the model on the test set meets the preset threshold.

9. The construction method of claim 1, wherein According to the real-time monitoring data of the pore structure and crack development degree of soft rock, the grouting flow and pressure parameters are dynamically adjusted to adapt to the grouting requirements of different regions.

Images