Sand soil hydraulic-mechanical coupling settlement monitoring method and system

Abstract

This invention discloses a hydraulic-mechanical coupled settlement monitoring method and system for sandy soil. The method includes the following steps: deploying a three-dimensional fiber optic network with boundary protection; applying coupled loads; simultaneously calculating the strain and temperature fields; inverting the seepage field and gradation parameters based on the strain field; calculating and predicting settlement using coupled multi-field information; and triggering early warning of seepage failure based on the seepage anomaly index. The system includes a model box, a three-dimensional fiber optic network, a loading and water supply system, a demodulator, and a data processing platform. This invention achieves simultaneous multi-field sensing within sandy soil, non-destructive gradation identification, intelligent settlement prediction, and early warning of seepage failure.

Description

Technical Field

[0001] This invention relates to the field of intelligent monitoring technology in geotechnical engineering, and in particular to a hydraulic-mechanical coupled settlement monitoring method and system for sandy soil. Background Technology

[0002] In geotechnical engineering, the settlement and deformation of sand under the coupled effects of hydraulic (seepage) and mechanical (load) processes is a key issue in the safety assessment of projects such as roadbeds, embankments, and foundations. Its deformation mechanism is complex, involving multiple couplings of seepage, effective stress, and soil deformation.

[0003] Traditional settlement monitoring methods primarily rely on point sensors, such as dial gauges, LVDT displacement gauges, and pore water pressure gauges. These technologies are typically based on principles such as Terzaghi's one-dimensional consolidation theory and Darcy's law for measurement and calculation. However, traditional methods have significant limitations: they monitor only one dimension, failing to simultaneously acquire continuous strain, seepage, and temperature field information within the soil; they struggle to capture the microstructural response caused by differences in soil gradation; they lack sufficient early warning capabilities for seepage failures such as piping; and they are mostly "monitor-record" models, lacking the intelligent inversion capability from monitoring data to key soil parameters.

[0004] With the development of sensing technology, distributed optical fiber sensing technology, especially Brillouin optical time-domain analysis (BOTDA / BOTDR) technology, has shown potential in geotechnical engineering monitoring due to its ability to monitor strain and temperature over long distances, in a distributed manner, and with high precision. Existing technologies have already attempted to use distributed optical fibers for geotechnical structure monitoring. For example, prior art document CN116840223A discloses a "Model Test Device and Test Method for Geogrid-Reinforced Hidden Karst Collapse Subgrade." This technical solution utilizes a distributed optical fiber sensing network (combined with FBG optical fiber) to monitor soil deformation and geogrid strain, used to study the mechanical response of reinforced subgrades under collapse conditions. It reveals the application of distributed optical fibers in monitoring deformation and reinforcement internal forces in model tests. However, this existing technology still has shortcomings: its monitoring targets are concentrated on structural deformation and mechanical response, failing to achieve synchronous distributed sensing of seepage and temperature fields during hydraulic-mechanical coupled settlement of sandy soil, lacking the ability to intelligently invert soil gradation parameters from optical fiber data, and not constructing a real-time prediction and seepage failure early warning model for coupled settlement. Furthermore, its fiber optic deployment scheme does not adequately consider the protection of boundary effects in model tests.

[0005] Therefore, existing technologies are still unable to meet the needs of refined monitoring for the entire process of hydraulic-mechanical coupled settlement of sandy soil, including multi-field synchronous sensing, micro-parameter identification, intelligent prediction, and early warning. Summary of the Invention

[0006] The main objective of this invention is to provide a method and system for monitoring hydraulic-mechanical coupled settlement of sandy soil, aiming to solve the problem that existing technologies cannot meet the needs of refined monitoring for the entire process of hydraulic-mechanical coupled settlement of sandy soil, including multi-field synchronous sensing, micro-parameter identification, intelligent prediction, and early warning.

[0007] To achieve the above objectives, the present invention provides a hydraulic-mechanical coupled settlement monitoring method for sandy soil, the method comprising the following steps:

[0008] Step 1: Deploy a three-dimensional fiber optic sensing network structure in the model box and fill it with sand. The three-dimensional fiber optic sensing network structure includes multiple layers of horizontal sensing fibers spaced apart along the height direction of the model and multiple rows of vertical sensing fibers spaced apart along the length direction of the model. The placement of all fibers maintains a preset distance from the boundary of the model box.

[0009] Step 2: Apply loads and / or water heads to the sand sample to simulate hydraulic-mechanical coupling effects;

[0010] Step 3: The Brillouin frequency shift signals of all sensing fibers are synchronously acquired by a distributed fiber demodulator, and the three-dimensional strain field and temperature field inside the sand sample are obtained after processing.

[0011] Step 4: Based on the strain field data, invert the seepage field parameters inside the sand sample;

[0012] Step 5: Extract the feature vector of the strain field and intelligently invert the gradation parameters of the sand based on the pre-established mapping model;

[0013] Step 6: Based on the strain field, the inverted seepage field, and the gradation parameters, calculate and predict the settlement of the sand sample;

[0014] Step 7: Analyze the temperature field data and trigger an early warning of seepage failure based on the temperature gradient or seepage anomaly index.

[0015] Optionally, in step 1, the safety distance is 50mm; the specific deployment parameters of the three-dimensional fiber optic sensing network are as follows: five layers of horizontal optical fibers are deployed at equal intervals along the z-direction, eight columns of vertical optical fibers are deployed at equal intervals along the x-direction, and three optical fibers are deployed along the y-direction to form a 5×8×3 three-dimensional monitoring grid; wherein, the horizontal layer spacing is 150mm and the vertical column spacing is 200mm.

[0016] Optionally, in step 3, temperature compensation is performed using a dual-fiber compensation method. The true strain ε is obtained by calculating the Brillouin frequency shift difference between the stressed fiber and the reference fiber. The specific formula for calculating the true strain ε is as follows:

[0017] ;

[0018] in, For the Brillouin frequency shift of the stressed optical fiber, For the Brillouin frequency shift of the reference fiber, C ε denoted as the strain coefficient of the optical fiber.

[0019] Optionally, in step 5, the characteristic vector of the strain field includes the mean strain, strain standard deviation, mean strain gradient, strain curvature, and vertical strain gradient; the mapping model is a model established based on support vector machine or multiple regression analysis.

[0020] Optionally, in step 6, a consolidation equation that considers the seepage hydrodynamic gradient term is used for settlement calculation. The modified consolidation equation for one-dimensional vertical consolidation is expressed as: ;

[0021] Where p is the excess pore water pressure, t is time, and c is... v σ is the consolidation coefficient. v Let be the total vertical stress, and j be the seepage force.

[0022] Optionally, in step 7, the formula for calculating the seepage anomaly index η is: ;in This represents the magnitude of the local temperature gradient. The background average temperature gradient magnitude.

[0023] Furthermore, to achieve the above objectives, the present invention also provides a monitoring system for the method described above, the system comprising:

[0024] The model box used to hold the sand sample has a filter layer at the bottom;

[0025] The three-dimensional fiber optic sensing network structure consists of multiple tightly packed single-mode optical fibers arranged in a three-dimensional grid within the sand sample area above the filter layer, and their arrangement is maintained at a preset distance from each boundary of the model box.

[0026] The head control subsystem, connected to the model box, is used to control the inlet water head;

[0027] The loading subsystem is used to apply a vertical load to the surface of the sand sample. The loading subsystem includes a hydraulic jack located directly above the model box and a rigid loading plate located between the hydraulic jack and the surface of the sand sample.

[0028] The distributed fiber optic demodulator, located outside the model box, is connected to all sensing fiber optic paths in the three-dimensional fiber optic sensing network via fiber optic cables led out from the holes on the side wall of the model box.

[0029] The data processing and inversion platform is electrically connected to the distributed optical fiber demodulator and is used to perform signal processing, multi-field parameter inversion, settlement prediction and early warning.

[0030] Optionally, the layout parameters of the three-dimensional fiber optic sensing network are as follows: it includes 5 layers of horizontal optical fibers and 8 columns of vertical optical fibers, with 3 optical fibers arranged in the y-direction in each layer and column; the distance from the fiber endpoints to the side wall of the model box, the top surface of the filter layer, and the bottom surface of the loading plate is 50mm.

[0031] Optionally, the filter layer comprises, from bottom to top, a porous plate layer, a coarse gravel layer, a medium gravel layer, a fine gravel layer, and a geotextile layer.

[0032] Optionally, the porous plate layer has a thickness of 8 mm, the coarse gravel layer has a thickness of 30 mm, the medium gravel layer has a thickness of 30 mm, the fine gravel layer has a thickness of 30 mm, and the geotextile layer has a thickness of 2 mm.

[0033] In addition, to achieve the above objectives, the present invention also provides the application of any of the above methods in geotechnical engineering model tests, dam seepage stability monitoring, or roadbed settlement monitoring.

[0034] Beneficial effects:

[0035] (1) Boundary protection and high survival rate: By setting a three-dimensional fiber optic deployment scheme with a safe distance, construction damage, boundary friction and local stress concentration are effectively avoided, which significantly improves the survival rate of optical fibers and data reliability in complex test environments.

[0036] (2) Multi-physics synchronous sensing: It breaks through the limitations of traditional point monitoring and can synchronously and distributedly acquire information on the strain field, seepage field (inversion) and temperature field of the entire space inside the sand sample, and comprehensively capture the soil response mechanism under hydraulic-mechanical coupling.

[0037] (3) Non-destructive identification of gradation parameters: The optical fiber strain response characteristics and sand gradation parameters (C) were established. u C c The quantitative mapping relationship between the parameters can be used to achieve in-situ, real-time identification of soil microstructure parameters without sampling, providing a new means for engineering property assessment.

[0038] (4) Intelligent inversion and prediction: By integrating physical mechanism models (such as modified consolidation equations) and data-driven methods (such as machine learning), real-time inversion and advanced prediction of key soil parameters (permeability coefficient, compression modulus, gradation) and settlement development process from distributed optical fiber data are realized.

[0039] (5) Early warning of seepage damage: The innovative use of temperature field anomalies (thermal convection effect) caused by seepage to identify priority seepage channels can provide early warning of seepage damage such as piping and contact loss, thus improving the safety of testing and engineering.

[0040] (6) High system integration: The system integrates a dedicated model box, a three-dimensional fiber optic network, a high-precision loading and water supply system, a distributed demodulation and intelligent analysis platform, forming a complete solution for monitoring hydraulic-mechanical coupling settlement of sandy soil. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0042] Figure 1 This is a schematic diagram of the overall structure of the monitoring system of the present invention.

[0043] Figure 2 A front view (xz plane) showing the layout of the optical fiber and filter layer inside the model box.

[0044] Figure 3 for Figure 2 Top view (xy plane) of the fiber optic cable layout inside the model box.

[0045] Figure 4 This is a flowchart of the monitoring method of the present invention.

[0046] Explanation of icon numbers:

[0047] 1-Hydraulic jack, 2-Rigid loading plate, 3-Model box, 4-Seepage outlet pipe, 5-Distributed fiber optic demodulator, 6-Data processing and inversion platform, 7-Outlet hole, 8-Bottom drainage pipe, 9-Three-dimensional fiber optic sensor network structure, 10-Filter layer, 11-Water head control subsystem; 12-Seepage inlet pipe; 1001-Porous plate layer; 1002-Coarse gravel layer; 1003-Medium gravel layer; 1004-Fine gravel layer; 1005-Geotextile; Sand sample area-A.

[0048] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0049] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0050] The monitoring system provided by this invention is an integrated physical simulation and intelligent sensing platform, designed to achieve multi-field, distributed, and intelligent monitoring of the settlement and deformation process of sand under hydraulic-mechanical coupling. For example... Figure 1As shown, the system consists of a physical model component, a sensor network component, a load excitation component, and a data acquisition and processing component. The components work together through a specific structural distribution and connection relationship.

[0051] The physical model component includes a model box 3 and a filter layer 10, wherein the model box 3 serves as the basic physical container and observation window for the entire system. Preferably, the model box 3 is made of acrylic material, and therefore has a transparent structure. In this embodiment, the effective internal dimensions of the model box 3 are 1500mm in length (x-direction), 300mm in width (y-direction), and 800mm in height (z-direction), with a wall thickness of 10mm. The side wall of the model box 3 has multiple fiber optic lead-out holes 7 with rubber sealing rings, preferably with a diameter of 10mm, for leak-free lead-out of the internal sensing fibers to the outside.

[0052] Furthermore, a specially designed filter layer 10 with a thickness of (80, 100] mm is laid at the bottom of the model box 3. This filter layer 10 is precisely layered from bottom to top, consisting of a porous plate layer 1001 (thickness [5, 8] mm, made of PVC material, pore size 5 mm, porosity 30%), a coarse gravel layer 1002 (thickness [25, 30] mm) with a particle size of 20-40 mm, and a medium gravel layer 1003 (thickness [25, 30] mm) with a particle size of 10-20 mm. The multi-layer filter structure 10 consists of a fine gravel layer 1004 (thickness [25,30] mm), a fine gravel layer 1004 (thickness [25,30] mm), and a non-woven geotextile layer 1005 (thickness (0~2] mm). Preferably, the multi-layer filter structure 10 consists of, from bottom to top, a porous plate layer 1001 (thickness 8 mm, made of PVC material, pore size 5 mm, open porosity 30%), a coarse gravel layer 1002 (thickness 30 mm), a coarse gravel layer 1002 (thickness 30 mm), a medium gravel layer 1003 (thickness 10-20 mm), and a coarse gravel layer 1004 (thickness [25,30] mm), a coarse gravel layer 1002 (thickness 30 mm), and a medium gravel layer 1003 (thickness 30 mm). The filter layer 10 consists of a 30mm thick layer of fine gravel (5-10mm particle size) 1004 and a 2mm thick layer of non-woven geotextile 1005 (200g / m² non-woven geotextile). The function of this filter layer 10 is to ensure uniform drainage and prevent sand sample loss during the experiment. Simultaneously, its top surface (geotextile layer) provides a flat and stable reference starting surface for sand sample filling and subsequent fiber optic network deployment. The filter layer 10 is directly connected to the drainage pipe 8 at the bottom of the model box 3, ensuring smooth drainage of seepage water.

[0053] The sensing network component, a three-dimensional fiber optic sensing network structure 9, is the core sensing component of this system, responsible for capturing subtle changes in the internal physical field. In this embodiment, the network structure is composed of multiple tightly packed single-mode optical fibers with an outer diameter of 0.9 mm, and its spatial configuration follows the principles of "boundary protection" and "uniform coverage." "Boundary protection" means maintaining a safe distance of 50 mm between all fiber endpoints and the side walls of the model box, the top surface of the bottom filter layer, and the bottom surface of the top loading plate.

[0054] Furthermore, such as Figure 2-3 In the embodiment shown, five layers (z-direction) of horizontal optical fibers and eight columns (x-direction) of vertical optical fibers are laid in the sand sample area A inside the model box 3 and above the filter layer 10. Three optical fibers are also laid at each position in the y-direction, forming a three-dimensional grid containing 120 spatial monitoring nodes in total, consisting of 5 (layers) × 8 (columns) × 3 (rows). Figure 2-3 The intersection shown in the image represents the space monitoring node.

[0055] Five horizontal optical fibers are laid out at equal intervals (150mm) along the height direction (z), with coordinates z=150, 300, 450, 600, and 750mm. Eight vertical optical fibers are laid out at equal intervals (200mm) along the length direction (x), with coordinates x=50, 250, 450, 650, 850, 1050, 1250, and 1450mm. All fiber endpoints maintain a safe distance of 50mm from the sidewalls of the model box 3, the top surface of the filter layer 10, and the bottom surface of the top loading plate 2. This design effectively avoids the impact of boundary friction, local stress concentration, and construction damage on the reliability of the optical fibers and data. After the two ends of a total of 39 sensing optical fibers (15 horizontal + 24 vertical) are laid out according to the required path within the sand sample, they converge and pass through the fiber optic exit holes 7 on the sidewalls of the model box 3, ultimately connecting to the external distributed optical fiber demodulator 5. This centralized optical path connection is the key channel for transmitting the sensing signal.

[0056] The load excitation components include a head control subsystem 11 and a loading subsystem. The head control subsystem 11 consists of a servo motor, a water storage tank, and precision valves. The water storage tank is connected to the lower left side wall of the model box 3 via a seepage inlet pipe 12, allowing water to be introduced into the tank. The water level in the water storage tank is precisely controlled by the servo motor, enabling simulation of static conditions, stepped or linearly varying seepage head. The hydraulic gradient adjustment range covers 0 to critical conditions; preferably, the head adjustment range is 0~700mm with an accuracy of ±1mm. After passing through the sand sample and the filter layer 10, the water is discharged through the seepage outlet pipe 4 and the bottom drain pipe 8, forming a controlled water circulation path.

[0057] The loading subsystem consists of a hydraulic jack 1 and a rigid loading plate 2. The rigid loading plate 2 is placed on the top surface of the sand sample and covers its core area. The hydraulic jack 1 is located directly above the loading plate and is used to apply graded vertical loads with a maximum loading capacity of 150 kPa. This system provides a top-down mechanical excitation interface.

[0058] The data acquisition and processing components include a distributed fiber optic demodulator 5 and a data processing and inversion platform 6. The distributed fiber optic demodulator 5 is directly optically connected to the internal sensor network structure 9 via optical fiber, used to transmit probe light pulses into the optical fiber and receive backscattered Brillouin signals. Its spatial resolution reaches 0.1m, strain resolution 1µε, and temperature resolution 0.1℃. It can synchronously demodulate the entire data path of all 39 optical fibers at a sampling frequency of 1Hz, realizing the primary conversion from optical signals to digital signals containing strain and temperature information. The distributed fiber optic demodulator 5 preferably adopts a Brillouin Optical Time Domain Analysis (BOTDA) type demodulator.

[0059] The data processing and inversion platform 6 is electrically connected to the distributed fiber optic demodulator 5 via a data cable (such as a network cable) to receive the raw data uploaded by the distributed fiber optic demodulator 5. The data processing and inversion platform 6 is preferably a high-performance computer workstation. This platform integrates a dedicated software system and serves as the "brain" of the entire system, sequentially executing the following core functions:

[0060] (1) Signal preprocessing: Temperature compensation is performed on the original Brillouin frequency shift signal (e.g., using a dual-fiber compensation algorithm) to calculate the true three-dimensional strain field distribution and temperature field distribution inside the sand sample. Among them, the dual-fiber compensation method is used to eliminate the influence of temperature.

[0061] (2) Intelligent inversion of multiple field parameters: Based on strain field data, combined with physical models such as Darcy's law and Kozeny-Carman formula, the seepage field parameters (such as permeability coefficient and hydraulic gradient) are inverted; at the same time, feature vectors are extracted from the spatial distribution characteristics of the strain field and input into a pre-trained machine learning model (such as support vector machine) to realize the sand gradation parameters (non-uniformity coefficient C). u curvature coefficient C c Lossless identification of ).

[0062] (3) Settlement calculation and prediction: The strain, gradation, seepage and other information obtained by inversion are substituted into the consolidation theory equation considering seepage hydrodynamic correction, and the settlement is calculated in real time using the layered summation method. The settlement development trend can be predicted and parameters updated online based on hyperbolic models.

[0063] (4) Multi-level early warning and output: The platform calculates the maximum strain value and seepage anomaly index in real time and makes automatic judgments based on the preset four-level early warning thresholds of blue, yellow, orange, and red. Once an early warning is triggered, the platform can immediately issue an alarm message and output instructions to prompt operators or the automatic control system to adjust the head or load. Finally, the platform can output a comprehensive report that includes the settlement process curve, gradation identification results, multi-field evolution cloud map, and early warning records.

[0064] Furthermore, this invention also provides a method for monitoring hydraulic-mechanical coupled settlement of sandy soil based on distributed optical fiber multi-field coupling sensing, such as... Figure 4 As shown, the specific workflow is as follows:

[0065] Step 1: Constructing a physical model of an integrated three-dimensional sensing network: Specifically, within the transparent model box 3, a filter layer 10, composed of a porous plate, coarse gravel, medium gravel, fine gravel, and geotextile, is laid sequentially from bottom to top. Subsequently, a three-dimensional fiber optic sensing network structure is deployed within the sand sample filling area above the filter layer 10. This network structure is composed of multiple tightly packed single-mode optical fibers arranged according to preset spatial coordinates, specifically including: N layers (e.g., 5 layers) of horizontal sensing fibers laid at equal intervals (e.g., 150mm) along the model height direction (Z-axis), M columns (e.g., 8 columns) of vertical sensing fibers laid at equal intervals (e.g., 200mm) along the model length direction (X-axis), and K fibers (e.g., 3 fibers) evenly deployed along the width direction (Y-axis), thereby forming an N×M×K three-dimensional monitoring grid. Preferably, the laying paths and endpoints of all optical fibers maintain a preset safe distance (e.g., 50mm) from the inner wall of the model box, the top surface of the filter layer, and the future loading surface (e.g., the bottom surface of the loading plate) to prevent boundary effects from damaging the optical fibers or affecting measurement accuracy. After the optical fibers are laid, the sand sample is filled in layers and compacted to ensure tight coupling between the sand sample and the optical fiber. All sensing optical fibers are finally exited through the sealed outlet holes 7 on the side wall of the model box and connected to the external demodulation equipment.

[0066] Step 2, Applying Hydraulic-Mechanical Coupled Loads: Specifically, for the consolidated and stabilized sand sample, coupled loads simulating actual working conditions are applied through an independent loading system and a head control subsystem. The mechanical load is applied in stages via a rigid loading plate 2 and a hydraulic device, for example, increasing gradually from 0 kPa to 150 kPa. The hydraulic load is achieved through precise control of the inlet head, allowing for stepped increases or constant seepage to simulate different hydraulic gradients (from steady seepage to critical states). Load application can be performed individually, alternately, or simultaneously to reproduce various hydraulic-mechanical coupled action scenarios.

[0067] Step 3: Synchronously acquire multiple raw signals and perform decoupling processing: Specifically, start the distributed fiber demodulator 5 (such as BOTDA type) to synchronously excite and acquire the backscattered Brillouin light signals along the path of each sensing fiber in all three-dimensional network structures 9 at high frequency, and obtain the raw Brillouin frequency shift distribution data. Subsequently, perform temperature compensation processing on the raw signals to separate strain and temperature effects. The preferred method is to use the "dual fiber compensation method": that is, while laying out the stress sensing fiber, lay out a reference fiber in a free and relaxed state, and calculate the true strain ε using the following formula.

[0068] ;

[0069] in, For the Brillouin frequency shift of the stressed optical fiber, For the Brillouin frequency shift of the reference fiber, C ε denoted as the strain coefficient of the optical fiber.

[0070] This processing yields precise spatiotemporal distribution data of the three-dimensional strain and temperature fields within the sand sample, specifically representing a set of discrete but dense spatial points over time. Its core is a high-resolution spatiotemporal function based on three-dimensional sensor grid coordinates.

[0071] In terms of spatial dimension, it is a collection of physical quantities at discrete spatial points. Its spatial architecture is determined by the high spatial resolution (0.1 meters) of the three-dimensional fiber optic sensing network structure 9 and the distributed fiber optic demodulator 5.

[0072] The strain field data (ε) can be specifically represented as follows:

[0073] ε={ε ijk |i=1toM,j=1toK,k=1toN};

[0074] Where (i, j, k) are the indices of the 3D spatial mesh nodes, corresponding to their positions in the X (length), Y (width), and Z (height) directions, respectively, and ε ijk This is the normal strain value calculated from the fiber after temperature compensation at this node. By combining measurements with horizontal and vertical fibers, it can be further decomposed into components in different directions. Since the fiber itself is a continuous sensing medium, the data between adjacent nodes can be considered as a quasi-continuous distribution. Therefore, the entire strain field can be effectively regarded as a spatial scalar field (or tensor field) function defined on the sand sample region Ω: ε(x, y, z).

[0075] The temperature field data (ΔT) can be specifically represented as follows:

[0076] ΔT={ΔT ijk |i=1toM,j=1toK,k=1toN};

[0077] ΔT ijk This refers to the temperature change (in degrees Celsius, °C) at the corresponding spatial node, obtained directly from the Brillouin frequency shift or through a reference fiber. Similarly, the entire temperature field can be considered as a spatial scalar field function defined on the sand sample region Ω: ΔT(x, y, z).

[0078] The specific representation of the time dimension means that the aforementioned spatial field data evolves over time, forming a time series. This is determined by the sampling frequency of the demodulator (e.g., 1 Hz). Therefore, the complete spatiotemporal distribution data is a four-dimensional data volume, which can be represented as:

[0079] ε(x, y, z, t) and ΔT(x, y, z, t), where t represents a continuous time variable starting from the beginning of the experiment. In actual computer processing and storage, it is represented as a data sequence indexed by timestamps: {ε ijk(tn) ΔT ijk(tn) |n=0, 1, 2, ...}; where t n This represents the nth sampling time. For example, in the case of sampling once per second, t... n It can be 0s, 1s, 2s, ...

[0080] Therefore, the precise spatiotemporal distribution data of the three-dimensional strain field (ε) and temperature field (ΔT) obtained in step 3 is specifically a high-density database of strain and temperature values ​​based on a preset three-dimensional coordinate grid and with timestamps. This transforms the invisible mechanical and thermal states within the sand sample into digital information that can be processed, analyzed, and visualized in three dimensions by a computer. It forms the data foundation for the entire method to achieve "multi-field coupled sensing" and "intelligent inversion."

[0081] Step 4: Invert the seepage field parameters based on the strain field.

[0082] Specifically, the seepage field is inverted based on the three-dimensional strain field data obtained by decoupling and combined with the basic principles of soil mechanics.

[0083] First, the volumetric strain ε at each point is calculated based on the strain data. v Then, the change in void ratio (e) can be deduced.

[0084] ;

[0085] Δe is the change in porosity. ε is the conversion factor. v ε is the volumetric strain, calculated from three-dimensional fiber strain data. v =ε x +ε y +ε z .

[0086] Then, Darcy's law is combined with the Kozeny-Carman formula, which reflects the structural properties of the soil, to perform seepage field inversion.

[0087] The permeability coefficient was inverted based on Darcy's law and fiber optic strain data. Under steady-state seepage conditions, the relationship between the hydraulic gradient and pore water pressure is as follows:

[0088]

[0089] i: hydraulic gradient; γ w : Specific weight of water; du / dz: Rate of change of pore water pressure u along depth z, i.e., pore water pressure gradient, in kPa / m. It describes how quickly pore water pressure changes with depth.

[0090] The permeability coefficient, expressed using the Kozeny-Carman formula, is as follows:

[0091] ;

[0092] Where k is the permeability coefficient, C is the shape factor, and γ w Let be the specific weight of water, μ be the dynamic viscosity of water, M be the shape factor related to gradation, and d be the specific weight of water. 10 The effective particle size.

[0093] Furthermore, the permeability coefficient (k) and hydraulic gradient (i) of each point inside the sand sample can be indirectly and distributedly calculated from the strain field data, thereby reconstructing the seepage field.

[0094] Step 5: Intelligent inversion of sand gradation parameters.

[0095] Specifically, firstly, a physical relationship between the fiber strain response and soil gradation parameters is established through a theoretical model. This mainly includes...

[0096] Fiber-to-soil interface mechanical model: A hyperbolic model is used to describe the shear stress transfer at the interface. The specific formula is as follows:

[0097] In the formula, τ is the interfacial shear stress; γ is the shear strain; and a and b are fitting parameters that are related to the soil particle size.

[0098] And establish the relationship between the interfacial friction angle and the average particle size d of the soil. 50 The relationship is expressed by the following formula:

[0099] In the formula, Φ is the internal friction angle of the sand; d fiber d is the diameter of the optical fiber (typically 0.25 mm); 50 The average particle size is denoted as .

[0100] Strain transfer coefficient model: Define the fiber strain transfer coefficient α (the ratio of fiber strain to the actual soil strain), and clarify that it is a gradation parameter (uniformity coefficient C). u curvature coefficient C c Relative density D r and effective stress σ v ′ The function, the specific formula is as follows

[0101] .

[0102] From the three-dimensional strain field ε(x,y,z) obtained in step 3, the eigenvector X characterizes its spatial distribution and variation law.

[0103] ;in, σ represents the average strain within the monitoring area. ε G represents the standard deviation of strain, reflecting the degree of strain dispersion. ε κ represents the average strain gradient, reflecting the drastic degree of strain change. ε Strain curvature reflects the local bending characteristics of strain distribution. This represents the vertical strain gradient, reflecting the change in strain along the depth.

[0104] Subsequently, these feature vectors X are input into a pre-established mapping model based on a machine learning algorithm. This mapping model (e.g., trained using a support vector machine (SVM) or neural network algorithm) can learn and establish the strain field characteristics and key gradation parameters of sand (uniformity coefficient C). u curvature coefficient C c The complex nonlinear relationship between these parameters. This model enables in-situ, non-destructive intelligent identification of sand gradation parameters without the need for sampling.

[0105] Preferably, the kernel function used is a radial basis function (RBF):

[0106] .

[0107] X i X j The feature vectors of the i-th and j-th samples, where σ is the width parameter of the RBF kernel function, also known as bandwidth.

[0108] Step 6: Couple multi-field information for settlement calculation and real-time prediction.

[0109] First, the outputs from steps 3, 4, and 5 are integrated as inputs to the settlement calculation model: that is, the strain field (ε) obtained in step 3 is used to obtain the inverted seepage field parameters (k,i) from step 4, and combined with the load conditions, the effective stress (σ) of each soil layer is determined.v = (Known load + Soil self-weight) - (Pore water pressure obtained from fiber strain inversion)), obtain the sand gradation parameters (uniformity coefficient C) from step 5 using intelligent inversion. u curvature coefficient C c ) and the known relative density D r Calculate or update the compression modulus E of each soil layer using the following formula. s ,

[0110] ;

[0111] In the formula, K and n are experimental constants (for quartz sand, K≈300, n≈0.5), p a Reference stress, usually taken as standard atmospheric pressure. D r Relative density. This is the core parameter describing the compaction state of sandy soil. It is a dimensionless number and its value ranges from 0 to 1.

[0112] Secondly, the settlement was calculated using the consolidation theory that takes into account seepage force correction and the corrected layered summation method.

[0113] A modified consolidation equation, incorporating the seepage hydrodynamic gradient term j, is used to describe the coupling process between pore pressure dissipation and deformation. The specific formula is as follows:

[0114] .

[0115] Where p is the excess pore water pressure, referring to the portion of pore water pressure exceeding the hydrostatic pressure caused by external loads or seepage changes (obtained through inversion using the parameters in step 4, combined with the strain field measured by distributed optical fiber, the soil constitutive model, and the effective stress principle); t is time, and c... ν The consolidation coefficient ( z: Vertical coordinate. Usually positive upwards; in a transparent model box, this represents the vertical height coordinate measured from the top surface of the filter layer. j: Permeable force, expressed as... (i is the hydraulic gradient). This represents the rate of change of seepage force along depth.

[0116] The entire sand sample is discretized into several layers along its depth. For each layer i, its vertical strain ε is used. z,i (Derived directly from fiber optic data) or from the effective stress increment Δp i With compressive modulus E s,i Calculate the strain. The total settlement S(t) at the current time t is calculated using the following two equivalent methods:

[0117] Based on strain integral: H represents the total thickness of the compressed layer (or sand sample); z represents the vertical coordinate (depth); εz (z,t) represents the vertical normal strain at depth z and time t; Δz j : The thickness of the j-th soil layer. Usually, all layers are taken as having the same thickness, i.e., Δz. j =H / m.

[0118] Based on stress-modulus: .

[0119] An empirical model is used to extrapolate the settlement-time relationship to predict the final settlement and future trend. Specifically, a hyperbolic model is used for prediction. Perform fitting. ∞ The predicted final settlement is represented by α, which is a fitting parameter related to the soil consolidation rate. Based on the monitored settlement-time data segment (t, S(t)), the parameter S is initially determined through curve fitting. ∞ And α, thus obtaining the subsidence development prediction curve.

[0120] To achieve "real-time prediction," a feedback update mechanism is established between the prediction model and real-time monitoring data from the optical fiber. Specifically, whenever new inversion settlement data S at time k+1 is obtained from the optical fiber network... k+1 fiber At that time, compare it with the model's current predicted value S. k+1 predict A comparison is made. Based on the prediction error, parameters are updated online using real-time fiber optic data.

[0121]

[0122] In the formula, The learning rate; For fiber optic inversion of sedimentation; To predict settlement for the model, update the parameters α and S of the hyperbolic model online. ∞ This allows the predicted curve to continuously approach the actual monitoring data trajectory, thereby dynamically correcting and optimizing the prediction of future settlement.

[0123] Step 6 is a dynamic closed-loop process of "calculation-prediction-assimilation". It first couples multi-source information such as strain, gradation, and seepage to calculate the current settlement through a physical model; secondly, it uses an empirical model to predict the settlement trend; finally, it continuously corrects the prediction model by assimilating real-time fiber optic data to ensure that the prediction results become more and more accurate over time, thus realizing the leap from static calculation to dynamic advanced prediction.

[0124] Step 7: Early warning of infiltration damage based on temperature field anomalies.

[0125] Specifically, the distributed temperature field obtained in step 3 is analyzed in real time. Data. Real-time calculation of the spatial distribution of temperature gradient within the monitoring area. The temperature gradient is a vector, and its magnitude is... This reflects the severity of local temperature changes. Under uniform seepage conditions, the temperature gradient distribution should be relatively uniform and gentle.

[0126] Calculate the temperature gradient distribution within the monitoring area and define a seepage anomaly index (η), for example:

[0127] ;

[0128] in This represents the magnitude of the local temperature gradient. The background average temperature gradient modulus. When the η value in a certain area continuously and significantly exceeds the empirical threshold (η) set based on soil properties and historical data. cr When a signal is detected (e.g., blue, yellow, orange, red), it indicates that a preferential seepage channel may have formed at that location, posing a risk of seepage instability. Based on this, the system triggers tiered early warning signals (e.g., blue, yellow, orange, and red) to provide a basis for timely intervention measures (such as reducing the hydraulic head), achieving early warning of seepage failure. And ε max It is the core criterion for triggering early warnings related to mechanical deformation. It, along with the exponent η, which reflects abnormal seepage, together constitute the dual criterion of "strain-seepage". max Exceeding the standard indicates excessive deformation in a localized area of ​​the sand sample, potentially foreshadowing shear band formation, structural instability, or impending failure. Exceeding the η standard indicates the presence of concentrated seepage channels within the sample, which may induce seepage failures such as piping. Exceeding either standard triggers a corresponding early warning system, enabling simultaneous and independent monitoring of both mechanical instability and seepage failure risks.

[0129] It is a key, global indicator extracted in real time from a distributed optical fiber three-dimensional strain field to assess the overall mechanical safety status of sandy soil. Its value directly determines the warning level and, in conjunction with the seepage anomaly index η, constitutes a comprehensive and intelligent safety monitoring network. The multi-level warning thresholds are shown in Table 1.

[0130]

[0131] If an early warning condition is triggered (e.g., η>0.5), corresponding response measures are implemented (e.g., lowering the head, strengthening monitoring, etc.); if no warning condition is triggered, the monitoring cycle continues until the test termination criteria are met (e.g., settlement stabilization or failure occurs). Finally, a comprehensive report including settlement curves, gradation identification results, and failure modes is output.

[0132] This invention, through the combination of the aforementioned hardware system and software algorithm, achieves intelligent and precise monitoring and safety early warning of the entire process of sand settlement under hydraulic-mechanical coupling.

[0133] The above are merely preferred embodiments of the present invention and do not limit the patent scope of the present invention. All equivalent structural transformations made using the contents of the present invention's specification and drawings under the inventive concept of the present invention, or direct / indirect applications in other related technical fields, are included within the patent protection scope of the present invention.

Claims

1. A hydraulic-mechanical coupled settlement monitoring method for sandy soil, characterized in that, The method includes the following steps: Step 1: Deploy a three-dimensional fiber optic sensing network structure in the model box and fill it with sand. The three-dimensional fiber optic sensing network structure includes multiple layers of horizontal sensing fibers spaced apart along the height direction of the model and multiple rows of vertical sensing fibers spaced apart along the length direction of the model. The placement of all fibers maintains a preset distance from the boundary of the model box. Step 2: Apply loads and / or water heads to the sand sample to simulate hydraulic-mechanical coupling effects; Step 3: The Brillouin frequency shift signals of all sensing fibers are synchronously acquired by a distributed fiber demodulator, and the three-dimensional strain field and temperature field inside the sand sample are obtained after processing. Step 4: Based on the strain field data, invert the seepage field parameters inside the sand sample; Step 5: Extract the feature vector of the strain field and intelligently invert the gradation parameters of the sand based on the pre-established mapping model; Step 6: Based on the strain field, the inverted seepage field, and the gradation parameters, calculate and predict the settlement of the sand sample; Step 7: Analyze the temperature field data and trigger an early warning of seepage failure based on the temperature gradient or seepage anomaly index.

2. The method according to claim 1, characterized in that, In step 1, the preset distance is 50mm; the specific deployment parameters of the three-dimensional fiber optic sensing network are as follows: 5 layers of horizontal optical fibers are deployed at equal intervals along the z-direction, 8 columns of vertical optical fibers are deployed at equal intervals along the x-direction, and 3 optical fibers are deployed along the y-direction to form a 5×8×3 three-dimensional monitoring grid; wherein, the horizontal layer spacing is 150mm and the vertical column spacing is 200mm.

3. The method according to claim 1 or 2, characterized in that, In step 3, temperature compensation is performed using the dual-fiber compensation method. The true strain ε is obtained by calculating the Brillouin frequency shift difference between the stressed fiber and the reference fiber. The specific formula for calculating the true strain ε is as follows: ； in, For the Brillouin frequency shift of the stressed optical fiber, For the Brillouin frequency shift of the reference fiber, C ε denoted as the strain coefficient of the optical fiber.

4. The method according to claim 1, characterized in that, In step 5, the characteristic vector of the strain field includes the mean strain, strain standard deviation, mean strain gradient, strain curvature, and vertical strain gradient; the mapping model is a model established based on support vector machine or multiple regression analysis.

5. The method according to claim 1, characterized in that, In step 6, a consolidation equation that considers the seepage hydrodynamic gradient term is used to calculate settlement. The modified consolidation equation for one-dimensional vertical consolidation is expressed as follows: ; Where p is the excess pore water pressure, t is time, and c is... v σ is the consolidation coefficient. v Let be the total vertical stress, and j be the seepage force.

6. The method according to claim 1, characterized in that, In step 7, the formula for calculating the seepage anomaly index η is: ;in This represents the magnitude of the local temperature gradient. The background average temperature gradient magnitude.

7. A monitoring system for implementing the method according to any one of claims 1-6, characterized in that, The system includes: The model box used to hold the sand sample has a filter layer at the bottom; The three-dimensional fiber optic sensing network structure consists of multiple tightly packed single-mode optical fibers arranged in a three-dimensional grid within the sand sample area above the filter layer, and their arrangement is maintained at a preset distance from each boundary of the model box. The head control subsystem, connected to the model box, is used to control the inlet water head; The loading subsystem is used to apply a vertical load to the surface of the sand sample. The loading subsystem includes a hydraulic jack located directly above the model box and a rigid loading plate located between the hydraulic jack and the surface of the sand sample. The distributed fiber optic demodulator, located outside the model box, is connected to all sensing fiber optic paths in the three-dimensional fiber optic sensing network via fiber optic cables led out from the holes on the side wall of the model box. The data processing and inversion platform is electrically connected to the distributed optical fiber demodulator and is used to perform signal processing, multi-field parameter inversion, settlement prediction and early warning.

8. The system according to claim 7, characterized in that, The filter layer, from bottom to top, includes a porous plate layer, a coarse gravel layer, a medium gravel layer, a fine gravel layer, and a geotextile layer.

9. The system according to claim 8, characterized in that, The porous plate layer is 8mm thick, the coarse gravel layer is 30mm thick, the medium gravel layer is 30mm thick, the fine gravel layer is 30mm thick, and the geotextile layer is 2mm thick.

10. The application of the method according to any one of claims 1-6 in geotechnical engineering model tests, dam seepage stability monitoring, or roadbed settlement monitoring.

Images