A coupled monitoring system and method for seepage pressure and deformation in deep-buried layered water-rich surrounding rock tunnels

Abstract

This invention provides a pressure-deformation coupled monitoring system and method for deeply buried layered water-rich surrounding rock tunnels, belonging to the field of tunnel engineering safety monitoring technology. It includes a multi-source sensing module, a coupling analysis module, a risk early warning module, and a visualization platform. The multi-source sensing module integrates a distributed fiber optic sensor array, a three-dimensional pressure monitoring unit, a data fusion module, and an edge computing terminal. The method achieves minute-level stability assessment of deeply buried layered water-rich surrounding rock tunnels through synchronous multi-source data acquisition, pressure-deformation coupled modeling, plastic zone expansion prediction, and dynamic risk early warning. This invention overcomes the technical shortcomings of traditional monitoring methods, such as fragmented data, inability to capture spatiotemporal coupling effects, and delayed response. It reduces the strain measurement error of anisotropic rock masses from 25% to 8%, can predict plastic zone expansion 30 minutes in advance, and has an early warning accuracy rate of over 90%. It is suitable for layered rock tunnel projects with water pressure > 2 MPa and burial depth > 500 m.

Description

Technical Field

[0001] This invention relates to the field of tunnel engineering safety monitoring technology, specifically to a seepage pressure-deformation coupled monitoring system and method for deeply buried layered water-rich surrounding rock tunnels, which is particularly suitable for dynamic stability assessment of tunnels with high water pressure and strong anisotropy in layered rock masses. Background Technology

[0002] As infrastructure construction in my country's transportation, water conservancy, and energy sectors expands into deeper rock masses, the scale and number of deeply buried layered water-rich surrounding rock tunnels continue to grow. Deeply buried layered water-rich surrounding rock is characterized by high stress, strong anisotropy, and high water permeability. During tunnel construction and operation, the seepage pressure in the surrounding rock dynamically changes with the development of rock fissures and groundwater migration. Furthermore, the deformation of the surrounding rock further alters the distribution of seepage channels, creating a strong spatiotemporal coupling effect.

[0003] Traditional tunnel monitoring methods have many technical bottlenecks: Fragmented data acquisition: Point strain gauges and single-point piezometers are often used to monitor surrounding rock deformation and seepage pressure separately, which makes it impossible to achieve synchronous acquisition and spatiotemporal matching of strain field and seepage pressure field data, and makes it difficult to reflect the coupling law between the two. Limitations of monitoring scope: Point-based monitoring is sparsely distributed, and for anisotropic rock masses with well-developed bedding, it cannot capture non-uniform deformation and seepage pressure anomalies such as bedding plane slippage and local plastic zone expansion, resulting in many monitoring blind spots. The lack of a coupling model: Existing monitoring and analysis methods mostly use deformation and seepage pressure as independent indicators to assess the stability of the surrounding rock. No quantitative model of the seepage pressure-deformation coupling has been established, and it is impossible to predict the dynamic expansion trend of the plastic zone of the surrounding rock. Insufficient response timeliness: Relying on manual data analysis, the cycle from data collection to risk identification can take several hours or even days. For sudden changes in seepage pressure and accelerated deformation in deeply buried tunnels, timely early warning cannot be achieved, which can easily lead to engineering disasters such as instability of support structure and sudden water and mud inrush.

[0004] While existing research has made breakthroughs in the scope of distributed monitoring systems, it has not yet solved core problems such as dynamic coupling modeling of seepage pressure field and deformation field, low accuracy of anisotropy monitoring of layered rock mass, and poor dynamic adaptability of early warning thresholds. These issues make it difficult to meet the safety monitoring needs of deeply buried layered water-rich surrounding rock tunnels. Summary of the Invention

[0005] To address the aforementioned deficiencies in existing technologies, the present invention aims to provide a seepage pressure-deformation coupling monitoring system and method for deeply buried layered water-rich surrounding rock tunnels. This system enables synchronous monitoring, coupled analysis, and dynamic early warning of the surrounding rock strain field and seepage pressure field, solving the technical problems of fragmented monitoring data, inability to capture spatiotemporal coupling effects, and delayed response in traditional monitoring methods.

[0006] To achieve the above-mentioned technical features, the objective of this invention is as follows: A coupled monitoring system for seepage pressure and deformation in a deeply buried, layered, water-rich surrounding rock tunnel, comprising: Multi-source sensing module to collect real-time data on surrounding rock strain, seepage pressure and temperature; The coupling analysis module performs coupling analysis on the data collected by the multi-source sensing module, based on an improved LSTM neural network, using the strain gradient tensor. Δε ij With osmotic gradient ΔP As input, output the predicted rate of plastic zone expansion; The risk warning module compares the predicted value of the plastic zone expansion rate output by the coupling analysis module with a set threshold. When the set threshold is exceeded, a warning signal of the corresponding level is triggered. The visualization platform generates a 3D holographic model of the tunnel and dynamically renders a coupled cloud map of the seepage pressure field and deformation field.

[0007] Preferably, the multi-source sensing module includes: The distributed fiber optic sensing array includes multiple sets of distributed fiber optic strain sensors arranged along the tunnel axis at the interface between the lining and the surrounding rock. The distributed fiber optic strain sensors include FBG strain sensors and BOTDR temperature compensation fiber optics. The three-dimensional pressure monitoring unit includes multiple sets of pressure boreholes drilled radially along the tunnel. Three-dimensional pressure sensors are embedded in the pressure boreholes and arranged in the normal, tangential and perpendicular directions of the bedding plane. The data fusion module uses a data fusion acquisition instrument that simultaneously connects a distributed fiber optic sensor array and a three-dimensional seepage pressure monitoring unit. The data fusion acquisition instrument integrates a multi-channel acquisition card and synchronizes the timestamps to align strain and seepage pressure data. The edge computing terminal uses a computing terminal computer connected to a data fusion module, and has a built-in seepage pressure-deformation coupling analysis algorithm to output the equivalent plastic strain cloud map of the surrounding rock in real time.

[0008] Preferably, the measurement range of the FBG strain sensor is ±5000. με ; The spatial resolution of the BOTDR temperature-compensated fiber is 0.5m.

[0009] Preferably, the three-dimensional osmotic pressure sensor has a range of 0-10 MPa and an accuracy of 0.1%FS; The sampling frequency of the data fusion acquisition instrument is ≥100Hz.

[0010] Preferably, the distributed optical fiber sensing array adopts a layered topology structure, with denser fiber deployment in sections where the inclination angle changes. The fiber spacing is dynamically adjusted with the layering density, and the fiber density increases by 50% for every 5 fibers / m increase in layering density.

[0011] Preferably, the three-dimensional osmotic pressure monitoring unit is networked according to spatial vectors, the perpendicularity deviation between the axis of the normal osmotic pressure sensor and the bedding plane is ≤5°, and the spacing between the tangential osmotic pressure sensors is 1.2-1.5 times the average spacing between bedding planes.

[0012] Preferably, the predicted rate of plastic zone expansion includes plastic strain. ε and osmotic pressure change rate ; Furthermore, in the aforementioned risk warning module, when predicting plastic strain... ε >0.15 or osmotic pressure mutation rate When the pressure is greater than 0.5 MPa / min, a level 3 warning signal is triggered.

[0013] Preferably, in the coupling analysis module, the training data for the improved LSTM neural network includes bedding control parameters: bedding dip angle. θ Stratification density D Rock mass anisotropy coefficient k The network structure is an input layer - a bidirectional LSTM layer - an output layer.

[0014] Preferably, the risk warning module has a built-in dynamic threshold adjustment mechanism, and the warning threshold adjusts with burial depth. H and water pressure P 0 is corrected according to the following formula: ; In the formula: To correct the plastic strain threshold, To adjust the threshold for osmotic pressure mutation rate, D This represents the bedding density.

[0015] Another aspect of the present invention provides a method for coupled monitoring of seepage pressure and deformation in deeply buried layered water-rich surrounding rock tunnels. The method employs a coupled seepage pressure and deformation monitoring system for deeply buried layered water-rich surrounding rock tunnels and includes the following steps: S1. Multi-source data acquisition: Surrounding rock strain data is acquired through the distributed optical fiber sensor array of the multi-source sensing module, surrounding rock seepage pressure data is acquired through the three-dimensional seepage pressure monitoring unit, and temperature compensation data is acquired at the same time. The data fusion module is used to synchronize and align the acquired strain and seepage pressure data with timestamps to obtain a synchronous monitoring dataset. S2. Coupled Model Construction and Prediction: The strain gradient tensor in the synchronous monitoring dataset will be used for... Δε ij With osmotic gradient ΔP The input is fed into the improved LSTM neural network of the coupling analysis module, and the predicted value of the expansion rate of the plastic zone of the surrounding rock is obtained through neural network operation. At the same time, the edge computing terminal outputs the equivalent plastic strain cloud map of the surrounding rock based on the synchronous monitoring dataset. S3. Dynamic Risk Warning: The risk warning module is based on the actual burial depth of the tunnel. H and water pressure P 0, combined with bedding density D The corrected strain threshold and the pressure change rate threshold are obtained through the threshold correction formula. The predicted value of plastic strain obtained in step S2 is compared with the corrected strain threshold, and the actual monitored pressure change rate is compared with the corrected pressure change rate threshold. If any indicator exceeds the threshold, the corresponding level of warning signal is triggered. S4. Visualization: The synchronous monitoring dataset, the predicted value of the plastic zone expansion rate, the plastic strain cloud map and the early warning signal are integrated into the three-dimensional holographic model of the tunnel through the visualization platform, so as to realize the dynamic rendering and visualization of the coupling state of the seepage pressure field and the deformation field.

[0016] The present invention has the following beneficial effects: 1. This invention improves monitoring accuracy and range: By using a layered topology distributed optical fiber sensor array and a vector network three-dimensional seepage pressure monitoring unit, it adapts to the anisotropic characteristics of layered rock masses, reduces the strain measurement error of anisotropic rock masses from the traditional 25% to 8%, and realizes full-domain synchronous monitoring of strain field and seepage pressure field, eliminating monitoring blind spots.

[0017] 2. This invention can achieve precise prediction of coupling: Based on the improved LSTM neural network, a pressure-deformation coupling model is constructed and bedding control parameters are incorporated. It can predict the expansion trend of the plastic zone of the surrounding rock 30 minutes in advance, which is 10 times more efficient than traditional manual analysis.

[0018] 3. This invention can realize dynamic intelligent early warning: the early warning threshold can be dynamically adjusted according to the burial depth, water pressure and bedding density, the early warning accuracy rate is over 90%, and the stability of the surrounding rock can be dynamically assessed at the minute level, solving the defect of slow response of traditional monitoring.

[0019] 4. This invention enables efficient and convenient visual management: The three-dimensional holographic visualization platform integrates and displays multi-source monitoring information, providing intuitive and comprehensive data support for on-site construction decisions and reducing the probability of engineering disasters. Attached Figure Description

[0020] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0021] Figure 1 This is a cross-sectional view of the present invention.

[0022] Figure 2 This is a longitudinal section view of the present invention.

[0023] Figure 3 This is a system architecture diagram of the present invention.

[0024] Figure 4This is a flowchart of the LSTM network training process.

[0025] In the diagram: 1-Distributed fiber optic strain sensor; 1-1FBG strain sensor; 1-2BOTDR temperature-compensated fiber optic cable; 2-Data fusion acquisition instrument; 3-Pyrostatic borehole; 4-Three-dimensional pyrostatic sensor; 4-1Vertical direction; 4-2Normal direction; 4-3Tangential direction; 5-Computer terminal. 6-Multi-source sensing module; 7-Coupling analysis module; 8-Risk early warning module; 9-Visualization platform. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0027] Example 1: Please see Figures 1-2 This embodiment provides a seepage pressure-deformation coupled monitoring system for deeply buried layered water-rich surrounding rock tunnels, including: a multi-source sensing module for real-time acquisition of surrounding rock strain, seepage pressure, and temperature data; and a coupling analysis module for performing coupling analysis on the data acquired by the multi-source sensing module, based on an improved LSTM neural network, using the strain gradient tensor... Δε ij With osmotic gradient ΔP The system takes input as input and outputs a predicted value for the plastic zone expansion rate. A risk warning module compares the predicted plastic zone expansion rate output by the coupling analysis module with a set threshold; when the set threshold is reached, a warning signal of the corresponding level is triggered. A visualization platform generates a 3D holographic model of the tunnel and dynamically renders a coupled cloud map of the pressure and deformation fields. Through this monitoring system, real-time synchronous acquisition of surrounding rock strain and pressure field data is achieved. Based on an improved LSTM neural network, a pressure-deformation coupling model is constructed to dynamically predict the expansion trend of the surrounding rock's plastic zone.

[0028] Further, see Figures 1-2The multi-source sensing module includes: a distributed optical fiber sensor array, comprising multiple sets of distributed optical fiber strain sensors 1 arranged along the tunnel axis at the interface between the lining and the surrounding rock, each distributed optical fiber strain sensor 1 containing an FBG strain sensor 1-1 and a BOTDR temperature-compensated optical fiber 1-2; a three-dimensional pressure monitoring unit, comprising multiple sets of pressure boreholes 3 drilled radially along the tunnel, with three-dimensional pressure sensors 4 embedded inside each borehole 3, arranged according to the bedding plane normal, tangential, and perpendicular directions; a data fusion module, employing a data fusion acquisition instrument 2 that simultaneously connects the distributed optical fiber sensor array and the three-dimensional pressure monitoring unit, the data fusion acquisition instrument 2 integrating a multi-channel acquisition card and synchronizing timestamps to align strain and pressure data; and an edge computing terminal, employing a computing terminal computer 5 connected to the data fusion module, with a built-in pressure-deformation coupling analysis algorithm to output a real-time equivalent plastic strain cloud map of the surrounding rock. Through the above-mentioned multi-source sensing module, real-time acquisition of surrounding rock strain and pressure field data is possible, and a pressure-deformation coupling model is constructed based on an improved LSTM neural network to dynamically predict the expansion trend of the plastic zone in the surrounding rock.

[0029] Furthermore, the measurement range of the FBG strain sensor 1-1 is ±5000. με The spatial resolution of the BOTDR temperature-compensated optical fibers 1-2 is 0.5m. This measurement range ensures that subsequent measurement accuracy is within acceptable limits.

[0030] Furthermore, the three-dimensional osmotic pressure sensor 4 has a range of 0-10 MPa and an accuracy of 0.1%FS; The sampling frequency of the data fusion acquisition instrument 2 is ≥100Hz.

[0031] Furthermore, the distributed optical fiber sensing array adopts a layered topology structure, with denser fiber deployment in sections where the inclination angle changes. The fiber spacing is dynamically adjusted with the layering density, and the fiber density increases by 50% for every 5 fibers / m increase in layering density.

[0032] Furthermore, the three-dimensional osmotic pressure monitoring unit is networked according to spatial vectors, the perpendicularity deviation between the axis of the normal osmotic pressure sensor and the bedding plane is ≤5°, and the spacing between the tangential osmotic pressure sensors is 1.2-1.5 times the average spacing between bedding planes.

[0033] Furthermore, the predicted rate of plastic zone expansion includes plastic strain. ε and osmotic pressure change rate ; Furthermore, in the risk warning module, when predicting plastic strain... ε >0.15 or osmotic pressure mutation rate When the pressure is greater than 0.5 MPa / min, a level 3 warning signal is triggered.

[0034] Furthermore, in the coupling analysis module, the training data for the improved LSTM neural network includes bedding control parameters: bedding dip angle. θ Stratification density D Rock mass anisotropy coefficient k The network structure is an input layer - a bidirectional LSTM layer - an output layer.

[0035] Furthermore, the risk warning module has a built-in dynamic threshold adjustment mechanism, and the warning threshold changes with the burial depth. H and water pressure P 0 is corrected according to the following formula: ; In the formula: To correct the plastic strain threshold, To adjust the threshold for osmotic pressure mutation rate, D Stratification density.

[0036] Example 2: See Figures 3-4 This embodiment provides a method for coupled monitoring of seepage pressure and deformation in deeply buried layered water-rich surrounding rock tunnels. The method is implemented using a coupled monitoring system for seepage pressure and deformation in deeply buried layered water-rich surrounding rock tunnels, and includes the following steps: S1. Multi-source data acquisition: Surrounding rock strain data is acquired through the distributed optical fiber sensor array of the multi-source sensing module, surrounding rock seepage pressure data is acquired through the three-dimensional seepage pressure monitoring unit, and temperature compensation data is acquired at the same time. The data fusion module is used to synchronize and align the acquired strain and seepage pressure data with timestamps to obtain a synchronous monitoring dataset. S2. Coupled Model Construction and Prediction: The strain gradient tensor in the synchronous monitoring dataset will be used for... Δε ij With osmotic gradient ΔP The input is fed into the improved LSTM neural network of the coupling analysis module, and the predicted value of the expansion rate of the plastic zone of the surrounding rock is obtained through neural network operation. At the same time, the edge computing terminal outputs the equivalent plastic strain cloud map of the surrounding rock based on the synchronous monitoring dataset. S3. Dynamic Risk Warning: The risk warning module is based on the actual burial depth of the tunnel. H and water pressure P 0, combined with bedding density D The corrected strain threshold and the pressure change rate threshold are obtained through the threshold correction formula. The predicted value of plastic strain obtained in step S2 is compared with the corrected strain threshold, and the actual monitored pressure change rate is compared with the corrected pressure change rate threshold. If any indicator exceeds the threshold, the corresponding level of warning signal is triggered. S4. Visualization: The synchronous monitoring dataset, the predicted value of the plastic zone expansion rate, the plastic strain cloud map and the early warning signal are integrated into the three-dimensional holographic model of the tunnel through the visualization platform, so as to realize the dynamic rendering and visualization of the coupling state of the seepage pressure field and the deformation field.

[0037] Example 3: A pumped storage tunnel is buried at a depth of 650m, with layered sandstone at a dip angle of 40° and a water pressure of 2.5MPa.

[0038] A pressure-deformation coupled monitoring system for deeply buried layered water-rich surrounding rock tunnels addresses the problem of support structure instability caused by the dynamic coupling of seepage pressure and surrounding rock deformation. The system comprises a distributed fiber optic sensor array, a three-dimensional pressure monitoring unit, and a data fusion module. It acquires real-time and synchronous data of the surrounding rock strain and pressure fields, constructs a pressure-deformation coupled model based on an improved LSTM neural network, and dynamically predicts the expansion trend of the plastic zone in the surrounding rock. The system includes a multi-source sensing module, a coupling analysis module, a risk early warning module, and a visualization platform.

[0039] like Figure 1 As shown, a coupled monitoring system for seepage pressure and deformation in a deep-buried, layered, water-rich surrounding rock tunnel includes the following implementation steps: S1. Deploy a distributed fiber optic sensing array: Multiple sets of distributed fiber optic strain sensors 1 are deployed along the tunnel axis at the interface between the lining and surrounding rock, including FBG strain sensor 1-1 and BOTDR temperature compensation fiber optic cable 1-2. The FBG strain sensor 1-1 has a measurement range of ±5000 mm. με The spatial resolution of BOTDR temperature-compensated fiber optic cables 1-2 is 0.5m. S2. Arrange the three-dimensional osmotic pressure monitoring unit: the three-dimensional osmotic pressure sensor 4 is embedded in the radial osmotic pressure borehole 3. The three-dimensional osmotic pressure sensor 4 has a range of 0-10MPa and an accuracy of 0.1%FS. It is arranged in the normal direction 4-2, the tangential direction 4-3 and the vertical direction 4-1 of the bedding plane. S3. Operational data fusion module: integrates a multi-channel acquisition card, with a sampling frequency ≥100Hz, and synchronizes timestamps to align strain and seepage pressure data; S4. Run the edge computing terminal: Built-in seepage pressure-deformation coupling analysis algorithm, output the equivalent plastic strain cloud map of the surrounding rock in real time.

[0040] The distributed optical fiber sensing array described in step S1 adopts a layered topology structure and is densely deployed in the section where the inclination angle of the layers changes. The fiber spacing is dynamically adjusted with the layer density. Specifically, for every 5 fibers / m increase in layer density, the fiber density increases by 50%.

[0041] In the three-dimensional osmotic pressure monitoring unit described in step S2, the micro piezometers are networked according to spatial vectors, the perpendicularity deviation between the axis of the normal piezometer and the bedding plane is ≤5°, and the spacing between the tangential piezometers is 1.2-1.5 times the average spacing between bedding planes.

[0042] Furthermore, such as Figure 2 The system disclosed is a coupled monitoring system for seepage pressure and deformation in a deep-buried, layered, water-rich surrounding rock tunnel, characterized by comprising the following modules: Multi-source sensing module 6: integrates the monitoring system to collect real-time data on surrounding rock strain, seepage pressure, and temperature. Running Coupled Analysis Module 7: Based on an improved LSTM neural network, input strain gradient tensor Δε ij With osmotic gradient ΔP Output the predicted value of the plastic zone expansion rate; Operating Risk Warning Module 8: When predicting plastic strain ε >0.15 or osmotic pressure mutation rate ΔP / Δt When the pressure is >0.5 MPa / min, a level three warning signal is triggered; Run Visualization Platform 9: Generate a 3D holographic model of the tunnel and dynamically render the coupled cloud map of the seepage pressure field and deformation field.

[0043] In the coupling analysis module 6, the training data for the improved LSTM neural network includes bedding control parameters: bedding dip angle. θ (0°~90°), bedding density D (1~20 lines / m), rock mass anisotropy coefficient k (0.2~5.0), the network structure is input layer (12 neurons) - bidirectional LSTM layer (64 units) - output layer (3 neurons).

[0044] The risk warning module 7 has a built-in dynamic threshold adjustment mechanism, and the warning threshold adjusts with burial depth. H (m) and water pressure P 0 (MPa) is corrected according to the following formula: ; In the formula: To correct the plastic strain threshold, To adjust the threshold for osmotic pressure mutation rate, D This represents the layering density. In this embodiment, the LSTM network outputs once every 10 seconds. ε The predicted value is when the monitoring point is 15m behind the working face. ε The error rate reached 0.18% (threshold correction value 0.172%), triggering a level-two warning. Radial grouting was promptly initiated, suppressing the expansion of the plastic zone and verifying a prediction accuracy of 92.3%.

[0045] The beneficial effects of the present invention using the above technical solution are as follows: Addressing the problem of instability of support structures caused by the dynamic coupling of seepage pressure and rock deformation in deeply buried layered surrounding rock, the strain measurement error of anisotropic rock mass is reduced from the traditional 25% to 8% through vector network piezometers and layered topology optical fibers, improving monitoring accuracy. Furthermore, the coupling model predicts the expansion of the plastic zone 30 minutes in advance, increasing efficiency by 10 times compared to manual analysis, thus solving the technical deficiency of traditional point monitoring in failing to capture spatiotemporal coupling effects.

[0046] The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention.

Claims

1. A coupled monitoring system for seepage pressure and deformation in a deep-buried, layered, water-rich surrounding rock tunnel, characterized in that, include: Multi-source sensing module to collect real-time data on surrounding rock strain, seepage pressure and temperature; The coupling analysis module performs coupling analysis on the data collected by the multi-source sensing module, based on an improved LSTM neural network and using the strain gradient tensor. Δε ij With osmotic gradient ΔP As input, output the predicted rate of plastic zone expansion; The risk warning module compares the predicted value of the plastic zone expansion rate output by the coupling analysis module with a set threshold. When the set threshold is exceeded, a warning signal of the corresponding level is triggered. A visualization platform generates a 3D holographic model of the tunnel and dynamically renders a coupled cloud map of the seepage pressure field and deformation field. The multi-source sensing module includes: The distributed optical fiber sensing array includes multiple sets of distributed optical fiber strain sensors (1) arranged along the tunnel axis at the interface between the lining and the surrounding rock. The distributed optical fiber strain sensor (1) includes an FBG strain sensor (1-1) and a BOTDR temperature compensation optical fiber (1-2). The three-dimensional seepage pressure monitoring unit includes multiple sets of seepage pressure boreholes (3) drilled along the radial direction of the tunnel. Three-dimensional seepage pressure sensors (4) are embedded inside the seepage pressure boreholes (3). The three-dimensional seepage pressure sensors (4) are arranged in the normal, tangential and vertical directions of the bedding plane. The data fusion module uses a data fusion acquisition instrument (2) that connects the distributed fiber optic sensor array and the three-dimensional seepage pressure monitoring unit simultaneously. The data fusion acquisition instrument (2) integrates a multi-channel acquisition card and synchronizes the timestamps to align the strain and seepage pressure data. The edge computing terminal uses a computing terminal computer (5) and connects to a data fusion module. It has a built-in seepage pressure-deformation coupling analysis algorithm and outputs the equivalent plastic strain cloud map of the surrounding rock in real time. The predicted rate of plastic zone expansion includes plastic strain. ε and osmotic pressure change rate ; Furthermore, in the aforementioned risk warning module, when predicting plastic strain... ε >0.15 or osmotic pressure mutation rate When the pressure is >0.5 MPa / min, a level three warning signal is triggered; In the coupling analysis module, the training data for the improved LSTM neural network includes bedding control parameters: bedding dip angle. θ Stratification density D Rock mass anisotropy coefficient k The network structure is an input layer - a bidirectional LSTM layer - an output layer; The risk warning module has a built-in dynamic threshold adjustment mechanism, and the warning threshold adjusts with burial depth. H and water pressure P 0 is corrected according to the following formula: ； In the formula: To correct the threshold for plastic strain, To adjust the threshold for osmotic pressure mutation rate, D This represents the bedding density.

2. The seepage pressure-deformation coupled monitoring system for a deep-buried layered water-rich surrounding rock tunnel according to claim 1, characterized in that, The measurement range of the FBG strain sensor (1-1) is ±5000. με ; The spatial resolution of the BOTDR temperature-compensated fiber (1-2) is 0.5m.

3. The seepage pressure-deformation coupled monitoring system for a deep-buried layered water-rich surrounding rock tunnel according to claim 1, characterized in that, The three-dimensional osmotic pressure sensor (4) has a range of 0-10 MPa and an accuracy of 0.1%FS; The sampling frequency of the data fusion acquisition instrument (2) is ≥100Hz.

4. The seepage pressure-deformation coupled monitoring system for a deep-buried layered water-rich surrounding rock tunnel according to claim 1, characterized in that, The distributed optical fiber sensing array adopts a layered topology structure and is densely deployed in the section where the inclination angle of the layers changes. The fiber spacing is dynamically adjusted with the layer density. For every 5 fibers / m increase in layer density, the fiber density increases by 50%.

5. The seepage pressure-deformation coupled monitoring system for a deep-buried layered water-rich surrounding rock tunnel according to claim 1, characterized in that, The three-dimensional osmotic pressure monitoring unit is networked according to spatial vectors. The perpendicularity deviation between the axis of the normal osmotic pressure sensor and the bedding plane is ≤5°, and the spacing between the tangential osmotic pressure sensors is 1.2-1.5 times the average spacing between bedding planes.

6. A method for coupled monitoring of seepage pressure and deformation in deeply buried layered water-rich surrounding rock tunnels, characterized in that, The method is implemented using the deep-buried layered water-rich surrounding rock tunnel seepage pressure-deformation coupling monitoring system as described in any one of claims 1-5, and includes the following steps: S1. Multi-source data acquisition: Surrounding rock strain data is acquired through the distributed optical fiber sensor array of the multi-source sensing module, surrounding rock seepage pressure data is acquired through the three-dimensional seepage pressure monitoring unit, and temperature compensation data is acquired at the same time. The data fusion module is used to synchronize and align the acquired strain and seepage pressure data with timestamps to obtain a synchronous monitoring dataset. S2. Coupled Model Construction and Prediction: The strain gradient tensor in the synchronous monitoring dataset will be constructed and predicted. Δε ij With osmotic gradient ΔP The input is fed into the improved LSTM neural network of the coupling analysis module, and the predicted value of the expansion rate of the plastic zone of the surrounding rock is obtained through neural network operation. At the same time, the edge computing terminal outputs the equivalent plastic strain cloud map of the surrounding rock based on the synchronous monitoring dataset. S3. Dynamic Risk Warning: The risk warning module is based on the actual burial depth of the tunnel. H and water pressure P 0, combined with bedding density D The corrected strain threshold and seepage pressure mutation rate threshold are obtained through the threshold correction formula. The predicted value of the expansion rate of the plastic zone of the surrounding rock obtained in step S2 is compared with the corrected strain threshold, and the actual monitored seepage pressure mutation rate is compared with the corrected seepage pressure mutation rate threshold. If any indicator exceeds the threshold, the corresponding level of warning signal is triggered. S4. Visualization: The synchronous monitoring dataset, the predicted value of the plastic zone expansion rate, the plastic strain cloud map and the early warning signal are integrated into the three-dimensional holographic model of the tunnel through the visualization platform, so as to realize the dynamic rendering and visualization of the coupling state of the seepage pressure field and the deformation field.

Images