In-situ testing device and method for water characteristics of subgrade filler unsaturated soil

Abstract

The application discloses a kind of roadbed filler unsaturated soil water characteristic in-situ testing device and method, belong to geotechnical engineering parameter test field.Device includes intelligent scanning monitoring module, high-precision closed-loop control module and central computing inversion module.The application carries out non-invasive high-precision layered scanning to infiltration soil column by ray tomography technology, and accurately maintains constant water head using PID algorithm, while central computing inversion module real-time acquisition moisture profile data, combined with Richards equation analytical solution model dynamic inversion unsaturated hydraulic diffusivity and convection velocity, and then automatically deduce soil water characteristic curve and permeability coefficient function.The application solves the problem of long time consumption and data dispersion of traditional testing method, shortens the testing period of several months to several hours, realizes the full-automatic, fast, continuous and accurate acquisition of unsaturated hydraulic parameters of roadbed filler.

Description

Technical Field

[0001] This invention belongs to the field of seepage parameter testing of roadbed fill, specifically relating to an in-situ testing device and method for the characteristics of unsaturated soil water in roadbed fill. Background Technology

[0002] The hydraulic properties of unsaturated soils, especially the soil-water characteristic curve (SWRC) and the unsaturated permeability coefficient function, are core parameters for analyzing the water migration patterns and stability evolution of railway and highway subgrades under the coupled effects of rainfall infiltration and long-term train loads. Accurately and quickly obtaining these parameters is of crucial engineering significance for preventing subgrade damage such as frost heave, mudslides, and subsidence.

[0003] However, existing methods for testing the hydraulic parameters of unsaturated soil have significant drawbacks. Currently, the most mainstream testing methods are the pressure plate method and the filter paper method. These methods are based on axis translation technology or the principle of balanced suction measurement, requiring long periods of suction equilibration of the test sample. The testing cycle for a complete soil-water characteristic curve often takes several days or even months. This lengthy testing cycle not only consumes a large amount of manpower, material resources, and financial resources, but more seriously, the data points obtained are discrete equilibrium points, failing to reflect the continuous hydraulic response of the subgrade filler during dynamic infiltration, resulting in a lack of accurate prediction of the changes in hydraulic characteristics over time in engineering projects.

[0004] While existing technologies have attempted to indirectly derive hydraulic parameters using the horizontal soil column infiltration method, this approach requires destructive sampling of the soil column to determine the moisture content distribution. This method is cumbersome, has limited accuracy, and cannot achieve in-situ, continuous, and automated monitoring. Therefore, there is an urgent need for a testing device and method that can quickly, conveniently, continuously, and automatically acquire a complete set of hydraulic parameters for unsaturated soil. This would fundamentally solve the problems of long testing cycles and discrete data in traditional testing techniques, providing an efficient and reliable technical means for unsaturated seepage analysis in roadbed engineering. Summary of the Invention

[0005] The purpose of this invention is to overcome the above-mentioned defects in the prior art and provide an in-situ testing device and method for the hydraulic characteristics of unsaturated soil in roadbed fill, aiming to solve the problems of long testing cycle, discrete data and inability to continuously and automatically acquire hydraulic parameters of unsaturated soil in the prior art.

[0006] To achieve the above objectives, the present invention provides an in-situ testing device for the characteristics of unsaturated soil and water in roadbed fill, comprising an intelligent scanning monitoring module, a high-precision closed-loop control module, and a central calculation and inversion module.

[0007] The intelligent scanning and monitoring module includes an X-ray source, a detector array, a vertical linear module, a rotating platform, and a multi-channel pulse amplitude analyzer. The high-precision closed-loop control module includes a non-contact liquid level sensor, a precision peristaltic pump, an electronic balance, and a water supply tank. The central computing and inversion module includes a high-performance edge computing terminal and a human-machine interface.

[0008] The rotating platform is fixed to the system base and is used to support the soil column containing roadbed filler. The vertical linear module is vertically mounted on one side of the rotating platform, and the X-ray source and the detector array are respectively mounted on opposite sides of the vertical linear module. This structure allows the X-ray source and detector to move up and down synchronously along the depth direction of the soil column, achieving layered scanning. The input terminal of the multichannel pulse amplitude analyzer is electrically connected to the output terminal of the detector array, and its output terminal is electrically connected to the input terminal of the high-performance edge computing terminal, used to convert the photon signals received by the detector into digitized photon counting data.

[0009] The non-contact liquid level sensor is suspended directly above the soil column by a bracket to monitor the water depth at the top of the soil column in real time. The water supply tank is placed on the electronic balance, which monitors the weight change of the tank in real time to calculate the infiltration flow rate. The inlet of the precision peristaltic pump is connected to the water supply tank via a hose, and its outlet is connected to the top of the soil column via a hose to supply water to the soil column to maintain a constant water head.

[0010] The control output terminals of the high-performance edge computing terminal are electrically connected to the control terminals of the vertical linear module, the rotating platform, the X-ray source, and the precision peristaltic pump, respectively, to achieve precise coordination of scanning motion and head control. Simultaneously, the high-performance edge computing terminal is also electrically connected to the signal output terminals of the non-contact liquid level sensor and the electronic balance, respectively, for collecting real-time liquid level and flow rate data. The human-machine interface is bidirectionally electrically connected to the high-performance edge computing terminal for displaying data and receiving operation commands.

[0011] In the technical solution of this invention, the intelligent scanning and monitoring module, the high-precision closed-loop control module, and the central computing and inversion module work together to form a complete automated testing closed loop. The intelligent scanning and monitoring module is responsible for non-invasively acquiring spatiotemporal distribution data of water transport within the soil column; the high-precision closed-loop control module ensures the stability and consistency of boundary conditions, providing a high-quality data foundation for parameter inversion; the central computing and inversion module, as the core brain, receives and processes data, and quickly inverts key hydraulic parameters through built-in physical models and optimization algorithms. All three are indispensable, jointly solving the core technical problems of long cycles and discrete data in traditional methods.

[0012] Furthermore, the high-precision closed-loop control module also includes an ambient temperature and humidity sensor. This sensor is positioned around the soil column, and its signal output is electrically connected to the high-performance edge computing terminal. This design enables the monitoring of ambient temperature and relative humidity during the experiment, allowing the central computing inversion module to correct for moisture loss due to evaporation, thereby further improving testing accuracy.

[0013] Furthermore, the central computing inversion module also includes a multi-functional motion controller. The high-performance edge computing terminal is electrically connected to the control terminals of the vertical linear module and the rotating platform via the multi-functional motion controller. This design separates motion control logic from core computing tasks, improving the system's real-time performance and stability.

[0014] Furthermore, the high-performance edge computing terminal incorporates image reconstruction algorithms and analytical solution fitting algorithms. The high-performance edge computing terminal is configured to perform the following core operations: First, it receives raw counting data from the multichannel pulse amplitude analyzer and calculates and outputs a volumetric water content profile in real time. The volumetric moisture content profile Substituting the data into the analytical solution model of the Richards equation, and performing dynamic fitting using the nonlinear least squares method, the unsaturated hydraulic diffusivity is inverted. Unsaturated convection velocity Using formulas The shape parameters of the soil-water characteristic curve are automatically derived; based on the shape parameters of the soil-water characteristic curve... Generate complete soil-water characteristic curves and unsaturated permeability coefficient functions The Richards equations are the fundamental governing equations describing unsaturated seepage flow, derived from numerous continuous experimental studies in both the time and spatial domains. By performing optimal fitting on the data, unique and highly accurate identification can be achieved. and This fundamentally avoids the cumbersome process of indirectly inferring parameters by measuring discrete suction equilibrium points, as is the traditional method. It enables rapid, continuous, and direct parameter inversion, which can significantly shorten the traditional testing cycle of several months to several hours. At the same time, due to the use of continuous data in the entire time and space domain, the inversion results are more representative and reliable.

[0015] Furthermore, the high-performance edge computing terminal is also configured to perform image subtraction operations. During the infiltration process, the reconstructed wet soil attenuation coefficient map is then used. Subtract the pre-collected and stored dry soil attenuation coefficient baseline map The attenuation component caused by moisture was obtained. Then, the moisture attenuation component is divided by the pre-calibrated pure water attenuation coefficient. To calculate the volumetric water content of each pixel. The above technical solution, based on Beer-Lambert's law and the linear superposition of ray attenuation, cleverly removes the contribution of the soil skeleton to ray attenuation, retaining only the contribution of water. This enables non-invasive, high-precision, pixel-level quantitative calculation of volumetric water content, providing high-quality input data for subsequent parameter inversion.

[0016] This invention also provides an in-situ testing method for the characteristics of unsaturated soil water in roadbed fill, applied to the aforementioned apparatus, comprising the following steps: S1: Test preparation and system calibration, including acquiring dry soil baseline maps and calibrating the pure water attenuation coefficient; S2: Constant head establishment and dynamic infiltration control, which maintains a constant head at the top of the soil column through a PID closed-loop control algorithm and records the instantaneous infiltration flow rate; S3: Tomographic scanning and volumetric water content acquisition. Attenuation coefficient maps of wet soil at different depths and times are obtained through X-ray tomography, and real-time volumetric water content profiles are generated through image subtraction processing. ; S4: Fitting the analytical solution of the Richards equation and identifying parameters, using the volumetric water content profile. The optimal unsaturated hydraulic diffusivity is dynamically derived by inputting the analytical solution model of the Richards equations using the nonlinear least squares method. Unsaturated convection velocity ; S5: Generation of soil-water characteristic curves and permeability coefficient function, based on the inverted unsaturated hydraulic diffusivity. Unsaturated convection velocity Calculate the shape parameters of the soil-water characteristic curve. and saturated permeability coefficient And generate complete soil-water characteristic curves SWRC and unsaturated permeability coefficient functions. ; S6: Aperture distribution derivation and test report output.

[0017] The above testing method deeply integrates physical experiments, digital twins (real-time scanning mapping), and intelligent inversion algorithms, achieving a leap from "discrete point measurement" to "continuous field inversion," which is a concrete manifestation of the overall technical concept of this invention at the method level.

[0018] Furthermore, the analytical solution model of the Richards equation used in step S4 is as follows:

[0019]

[0020] in, This represents the volumetric water content at depth z and time t predicted by the model. This indicates the initial volumetric water content of the soil column; This indicates the volumetric water content corresponding to the top boundary of the soil column. This represents unsaturated convection velocity; The unsaturated hydraulic diffusivity is represented by z; the depth is represented by t; and the complementary error function is represented by erfc. The nonlinear least squares method minimizes the objective function: To identify the optimal and ;in, This represents the sum of squared residuals from the fit; This represents the measured average volumetric water content at the i-th depth and the j-th time. This represents the calculated model prediction value; and This represents the parameter to be identified. The analytical solution model is applicable to one-dimensional vertical infiltration problems under constant head boundary conditions. It is concise in form, computationally efficient, and highly suitable for real-time, dynamic parameter inversion.

[0021] Furthermore, in step S5, the shape parameters of the soil-water characteristic curve... The calculation formula is:

[0022] Where δ represents the shape parameter of the soil-water characteristic curve; This represents the obtained optimal unsaturated convection velocity; This represents the obtained optimal unsaturated hydraulic diffusivity; Indicates the specific gravity of water; The complete soil-water characteristic curve SWRC and unsaturated permeability coefficient function Generated using the following exponential model:

[0023]

[0024] Where θ(ψ) represents the volumetric water content under suction force ψ; Indicates the residual volumetric moisture content; Expressed as the absolute value of suction force; Indicates the saturated volumetric water content; The scheme establishes inversion parameters ( , ) and macroscopic hydraulic properties (SWRC, The physical connection between the two is based on the similarity solution and the assumption of the exponential soil-water characteristic curve in the theory of unsaturated seepage. This allows a simple soil-water characteristic curve shape parameter δ to simultaneously characterize water holding capacity and water conductivity, greatly simplifying the parameter acquisition process.

[0025] Furthermore, step S6 also includes deriving the pore size distribution based on the Laplace capillary equation, specifically as follows: Establish the relationship between suction force and equivalent aperture: Where ψ represents soil suction; σ represents liquid surface tension; and α represents the contact angle. Indicates the equivalent aperture; Generate the equivalent aperture cumulative frequency distribution function:

[0026] in, Indicates less than or equal to the equivalent aperture. The cumulative frequency; Represents residual saturation; σ represents liquid surface tension; δ represents the shape parameter of the soil-water characteristic curve; Indicates the equivalent aperture; And calculate the critical aperture:

[0027] in, This represents the critical aperture corresponding to the peak value of the aperture frequency distribution curve. Indicates the surface tension of a liquid; The above technical solution represents the shape parameters of the soil-water characteristic curve; it is based on the intrinsic relationship between the soil-water characteristic curve and the pore size distribution, and obtains the shape parameters of the soil-water characteristic curve through inversion. Furthermore, the pore size distribution curve reflecting the microscopic pore structure of the soil was derived, providing more in-depth microscopic information for the interpretation of the hydraulic and mechanical behavior of the subgrade soil.

[0028] Furthermore, the acquisition of the dry soil baseline map in step S1 specifically includes: moving the X-ray source and detector array layer by layer along the Z-axis of the soil column using a vertical straight-line module, and performing multi-angle projection acquisition at each depth layer using a rotating platform. The high-performance edge computing terminal then reconstructs the dry state attenuation coefficient baseline map of each depth layer based on a filtered back-projection algorithm. The above technical solution ensures that the benchmark for subsequent image subtraction operations is three-dimensional and high-precision, rather than a simple uniform value, thereby improving the accuracy of water content field reconstruction.

[0029] The present invention has the following beneficial effects: 1. The in-situ testing device for unsaturated soil-water characteristics of roadbed fill material in this invention integrates self-testing, dynamic scanning, PID head control, automatic image processing, parameter inversion and report generation functions, realizing intelligent management and control of the entire process, reducing the operation threshold and avoiding human error.

[0030] 2. The in-situ testing method for unsaturated soil-water characteristics of roadbed fill material in this invention utilizes a high-precision closed-loop controlled real-time infiltration experiment combined with a physical model-based intelligent inversion algorithm. This reduces the time-consuming process of traditional methods (which can take months) to just a few hours, significantly improving testing efficiency and saving time and labor costs. It employs X-ray tomography for non-invasive, high-resolution three-dimensional scanning and uses digital twin technology to reconstruct the water content field in real time, avoiding human sampling disturbances. Furthermore, by utilizing continuous data inversion across the entire spatiotemporal domain, the results are more statistically representative and reliable than traditional discrete-point measurements. This testing method can not only quickly generate complete soil-water characteristic curves and unsaturated permeability coefficient functions, but also derive pore size distribution and saturated permeability coefficient, achieving a comprehensive and continuous characterization of the hydraulic properties of unsaturated soil. Attached Figure Description

[0031] Figure 1 This is a schematic diagram of the overall structure of the in-situ testing device for the characteristics of unsaturated soil water in roadbed fill material according to the present invention.

[0032] Figure 2 This is a flowchart of the in-situ testing method for the water characteristics of unsaturated soil filler in the roadbed according to the present invention.

[0033] Among them, 11. X-ray source; 12. Detector array; 13. Vertical linear module; 14. Rotating platform; 15. Multichannel pulse amplitude analyzer; 21. Non-contact liquid level sensor; 22. Precision peristaltic pump; 23. Electronic balance; 24. Water supply tank; 25. Ambient temperature and humidity sensor; 31. High-performance edge computing terminal; 32. Multifunctional motion controller; 33. Human-computer interaction interface. Detailed Implementation

[0034] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0035] Example 1

[0036] like Figure 1As shown in the figure, this embodiment provides an in-situ rapid testing device for the characteristics of unsaturated soil and water in roadbed fill. The device mainly consists of three parts: an intelligent scanning and monitoring module, a high-precision closed-loop control module, and a central calculation and inversion module.

[0037] The intelligent scanning and monitoring module is the core of this device for achieving non-invasive, high-precision monitoring. It includes a radiation source 11, a detector array 12, a vertical linear module 13, a rotating platform 14, and a multi-channel pulse amplitude analyzer 15. The rotating platform 14 is fixed to the base of the entire device and driven by a high-precision stepper motor, enabling 360-degree precise rotation. It is used to support and precisely rotate a transparent soil column (such as a PVC or acrylic tube) containing roadbed filler material to obtain complete projection data required for tomographic imaging. The vertical linear module 13 is a ball screw slide rail system driven by a servo motor, vertically positioned on one side of the rotating platform 14. The radiation source 11 and detector array 12 are respectively mounted on opposite sides of the vertical linear module 13. Driven by this module, they can perform micron-level precise lifting and lowering positioning along the Z-axis (depth direction) of the soil column, achieving dynamic scanning of cross-sections at different depths. The input of the multichannel pulse amplitude analyzer 15 is electrically connected to the output of the detector array 12, and its output is electrically connected to the input of the central calculation and inversion module. It is responsible for rapidly converting the analog photon signal received by the detector into digital photon counting data.

[0038] The high-precision closed-loop control module is responsible for maintaining stable and accurate water supply boundary conditions and monitoring the infiltration process. It includes a non-contact level sensor 21, a precision peristaltic pump 22, an electronic balance 23, a water supply tank 24, and an environmental temperature and humidity sensor 25. The non-contact level laser sensor 21 is suspended directly above the soil column by a bracket, with its probe aimed at the water surface at the top of the soil column, for real-time, non-contact monitoring of water depth. The water supply tank 24 contains purified water and is placed on the high-precision electronic balance 23. The electronic balance 23 monitors the mass change of the water tank in real time and sends the data to the central calculation and inversion module for accurate calculation of instantaneous infiltration flow rate. The inlet of the precision peristaltic pump 22 is submerged in the water supply tank 24 via a hose, and its outlet is connected to the top of the soil column via a hose to supply water to the soil column. The environmental temperature and humidity sensor 25 is arranged around the soil column to monitor environmental conditions for subsequent evaporation correction.

[0039] The central computing inversion module is the "brain" of the device, responsible for coordinating control, data acquisition, image reconstruction, and intelligent computing. It includes a high-performance edge computing terminal 31, a multi-functional motion controller 32, and a human-machine interface 33. The high-performance edge computing terminal 31 is the core processor of the system, with built-in image reconstruction and analytical solution fitting algorithms. Its control output is electrically connected to the control terminals of the vertical linear module 13 and the rotating platform 14 via the multi-functional motion controller 32 to precisely control the scanning motion. Simultaneously, it directly controls the start and stop of the X-ray source 11 and the rotation speed of the precision peristaltic pump 22. Its signal input is electrically connected to a multi-channel pulse amplitude analyzer 15, a non-contact liquid level sensor 21, an electronic balance 23, and an ambient temperature and humidity sensor 25 to collect all experimental data in real time. The human-machine interface 33 is bidirectionally electrically connected to the high-performance edge computing terminal 31, providing the operator with a graphical monitoring and operating environment.

[0040] like Figure 2 As shown, the testing method using this device includes the following specific steps: Step 1: Experiment Preparation and System Calibration This embodiment specifically includes the following steps: S11. Equipment Self-Check and System Initialization: First, the central computing inversion module is started. The high-performance edge computing terminal 31 performs connectivity checks and status detection on the X-ray source 11, detector array 12, multichannel pulse amplitude analyzer 15, vertical linear module 13, rotating platform 14, non-contact liquid level sensor 21, precision peristaltic pump 22, electronic balance 23, and ambient temperature and humidity sensor 25. If any hardware module experiences a disconnection, zero-point drift, abnormal stroke, or unstable signal, the system will issue a prompt on the human-machine interface 33 and pause the test. When all modules are working normally, the system establishes the time reference, data sampling frequency, scanning layer sequence, and storage directory for this test.

[0041] S12. Soil Sample Preparation: After sampling the subgrade fill material to be tested, the sample is dried, sieved, and weighed. It is preferable to dry it at 105℃ to constant weight, and then sieve it through a 1mm sieve to reduce the impact of large particles and local agglomeration on the uniformity of ray attenuation. Subsequently, it is filled into the soil column in layers according to the predetermined dry density. Each layer is quantitatively filled and compacted and leveled to improve the consistency of material distribution along the height direction.

[0042] S13. Sample Placement: The filled soil column is vertically fixed on the rotating platform 14, and the central axis of the soil column is aligned with the X-ray scanning center using the support clamps. Subsequently, the vertical linear module 13 and the rotating platform 14 are driven by the multi-functional motion controller 32 for no-load test run to confirm that the eccentricity of the soil column during rotation is within the allowable range, thus avoiding geometric distortion in the subsequent reconstructed images.

[0043] S14. Input initial parameters: Enter the geometric and boundary parameters of this test through the human-computer interaction interface 33, including the inner diameter of the soil column, the height of the soil column, the layer scanning interval, the initial volumetric water content, the set water head height, the scanning time interval, and the test termination conditions.

[0044] S15, Pure Water Attenuation Coefficient Calibration: A pure water phantom of known dimensions is placed on the rotating platform 14, and projection acquisition is automatically performed by the intelligent scanning and monitoring module. For the ray attenuation process in a homogeneous medium, the Beer-Lambert law is satisfied:

[0045] in, This represents the intensity of photons before they are incident on the sample, and is controlled by the X-ray source. This represents the photon intensity after passing through the sample, which is measured by the detector array. denoted by , where represents the linear attenuation coefficient when the ray passes through pure water; x represents the total path length of the ray through the sample.

[0046] S16. Dry Soil Baseline Map Acquisition: The dry soil column to be measured is placed on the rotating platform 14. The vertical linear module 13 is controlled by the multi-functional motion controller 32 to move layer by layer along the height direction of the soil column, and the rotating platform 14 is controlled to perform multi-angle projection acquisition at each depth layer. For non-homogeneous objects such as soil columns, the following projection formula is used:

[0047] in, I represents the intensity of the incident photon; I represents the intensity of the transmitted photon after passing through the soil column. denoted by , where is the local linear attenuation coefficient at position (x, y) on the cross section of the dry soil column; L represents the penetration path of the ray in the cross section of the soil column; dl represents the minute length element on the path.

[0048] The high-performance edge computing terminal 31 performs fault reconstruction based on projection values ​​from various angles to obtain the benchmark attenuation coefficient map of each depth layer under dry soil conditions. This is used to contribute to the subsequent subtraction of soil skeleton attenuation from wet soil images.

[0049] S2: Establishment of constant head and dynamic infiltration control This embodiment specifically includes the following steps: S21. Constant Head Setting: After the dry soil baseline map is acquired, water from the water supply tank 24 is delivered to the top of the soil column via a precision peristaltic pump 22, forming a water accumulation layer on the top surface of the soil column. A non-contact level sensor 21 continuously measures the liquid level at the top of the soil column and sends the measurement results to the central calculation and inversion module. When the liquid level approaches the set head value, the system switches from rapid water supply mode to constant head control mode. This step establishes a constant head infiltration boundary for the soil column, ensuring that the top of the soil column remains under stable infiltration conditions throughout the entire test.

[0050] S22, PID Closed-Loop Control: In this step, the central calculation and inversion module uses PID closed-loop control to adjust the water supply of the precision peristaltic pump 22. First, the liquid level deviation formula is used:

[0051] Where e(t) represents the liquid level error at time t; H represents the preset target head height; and h(t) represents the real-time liquid level height measured by the liquid level sensor 21 at time t. The control output formula is then used:

[0052] Where u(t) represents the control output of the peristaltic pump 22 at time t; Indicates the proportionality coefficient; Indicates the integral coefficient; The coefficient represents the differential coefficient; e(t) represents the current liquid level deviation. This represents the cumulative amount of error over time; This indicates the rate of change of error. The central calculation and inversion module continuously adjusts the pump speed according to the above two formulas to keep the liquid level at the top of the soil column within the set range.

[0053] S23, Infiltration Flow Record: While maintaining stable head control, the electronic balance 23 continuously monitors the mass change of the water supply tank 24. The central calculation and inversion module calculates the instantaneous infiltration flow using the following formula:

[0054] Where Q(t) represents the instantaneous infiltration flow rate at time t; The density of water is represented by m(t); the mass of water tank 24 at time t is represented by m(t). This indicates the rate of change of the water tank's mass over time.

[0055] S24. Dynamic Scanning: Once a stable water head forms at the top of the soil column and infiltration begins, the central calculation and inversion module controls the intelligent scanning and monitoring module to perform dynamic tracking scans. In the early stages of infiltration, the scanning frequency is prioritized to increase in the area near the top of the soil column; as the wetting front advances downwards, the scanning depth range is gradually expanded; the central calculation and inversion module can also automatically increase the scanning density in the area near the wetting front based on the water content change gradient obtained from the previous scan, thereby enabling the system to more accurately capture the water transport process.

[0056] S3: Tomographic Scanning and Volumetric Water Content Acquisition This embodiment specifically includes the following steps: S31, Photon Counting to Attenuation Coefficient: During dynamic scanning, the X-ray source 11 continuously emits a photon beam towards the soil column, and the detector array 12 collects the photon count signal after it passes through the soil column. For any rotation angle φ, any depth z, and any time t, the transmission count measured by the detector array 12 can be denoted as I(φ,z,t). The testing device at each z-axis depth and... At the rotation angle, the detector array 12 collects the photon counts after the rays pass through the soil column. According to Beer-Lambert's law, calculate the total attenuation rate of the ray as it passes through the soil column:

[0057] in Let L be the incident intensity and L be the diameter of the soil column. The attenuation rate is a linear attenuation coefficient along the ray path. The integral of the signal is then processed by the multichannel pulse amplitude analyzer 15 and transmitted as projection data to the high-performance edge computing terminal 31.

[0058] S32, Image Reconstruction: The high-performance edge computing terminal 31 preprocesses the projection data obtained in step S31 at various angles. The preprocessing includes filtering and noise reduction of the collected projection data (i.e., attenuation rate) to prepare for image reconstruction, thereby obtaining the attenuation coefficient distribution map of each depth layer under wet soil conditions. This reflects the ray attenuation capability caused by the combined effects of the wet soil skeleton and the moisture in the pores at time t and cross-sectional coordinates (x, y).

[0059] S33. Image Subtraction: To eliminate the attenuation contribution of the soil skeleton itself, the following image subtraction formula is used in this step:

[0060] in, This indicates the additional attenuation caused by changes in water content; This represents the attenuation coefficient under wet soil conditions; This represents the dry soil baseline attenuation coefficient.

[0061] S34. Calculation of volumetric water content: Based on the water attenuation coefficient, calculate the water content of each pixel:

[0062] in, This represents the volumetric water content at time t and position (x, y). Indicates the attenuation coefficient of wet soil; This represents the attenuation coefficient of pure water.

[0063] S35. Generating Real-Time Profiles: After obtaining the pixel-by-pixel volumetric water content, the central calculation and inversion module averages the effective area of ​​the same depth section using the following formula:

[0064] in, represents the average volumetric water content of the cross section at time t and depth z; A represents the effective analytical area of ​​the cross section. This represents the volumetric water content at any point within the cross-section. This average value is used to characterize the overall water content of a layer at a certain depth and serves as direct input for fitting the analytical solution of the Richards equation.

[0065] Step 4: Fitting the analytical solution of the Richards equation and parameter identification S41. Model Fitting: The central calculation and inversion module uses a one-dimensional analytical expression for vertical transient infiltration to fit the average volumetric water content profile obtained in step S35. The formula used is:

[0066] The complementary error function is:

[0067] In the formula, θ(z,t) represents the volumetric water content at depth z and time t predicted by the model; This indicates the initial volumetric water content of the soil column; This indicates the volumetric water content corresponding to the top boundary of the soil column. This represents unsaturated convection velocity; Let θ(z,t) represent the unsaturated hydraulic diffusivity; z represent depth; t represent time; and erfc represent the complementary error function. The model treats θ(z,t) as a function of z and t, and includes two core unknown parameters: unsaturated hydraulic diffusivity. Unsaturated convection velocity In order to identify the optimal and The central computing inversion module adopts the following objective function:

[0068] in, This represents the sum of squared residuals from the fit; This represents the measured average volumetric water content at the i-th depth and the j-th time. This represents the calculated model prediction value; and This represents the parameters to be identified. The objective function reflects the overall deviation between the model-predicted profile and the measured profile from the tomographic scan. The central computation and inversion module minimizes this objective function to obtain the parameter values ​​that best match the current testing process.

[0069] S42. Dynamic Optimization: The central computational inversion module uses a nonlinear least squares method to solve the formula in S41, preferably employing the Levenberg-Marquardt algorithm. In each iteration, the system updates the formula obtained in step S35. Data is used as input, and... and Continuous corrections are performed until the sum of squared residuals reaches its minimum or the parameter variation is less than a set threshold. The final output is... and These represent the optimal unsaturated convection velocity and unsaturated hydraulic diffusivity of the soil column under unsaturated infiltration conditions in this experiment, respectively.

[0070] S43. Determine the validity of the fitting results: After obtaining... and Then, the central computational inversion module judges the validity of the fitting results. Preferably, the judgment can be made based on at least one of the following conditions: First, is the objective function value lower than the set threshold? Second, is the range of parameter changes during iteration small enough? That is:

[0071]

[0072] Where Yes represents the final output. and This is the optimal unsaturated convection velocity and unsaturated hydraulic diffusivity of the soil column under unsaturated infiltration conditions in this experiment. If not, continue to repeat steps S42 and S43 to iterate and obtain the optimal unsaturated convection velocity and unsaturated hydraulic diffusivity again. Indicates the final output and The sum of squared residuals of the objective function after that. Indicates the final output and Threshold for the sum of squared residuals of the objective function. and These represent the iterative variation amplitudes of the unsaturated convection velocity and the unsaturated hydraulic diffusivity of the soil column under unsaturated infiltration conditions in this experiment, respectively. and These represent the minimum required values ​​for the iterative variation amplitudes of the unsaturated convection velocity and the unsaturated hydraulic diffusivity of the soil column under unsaturated infiltration conditions in this experiment.

[0073] When the output is "Yes", it means that the sum of squared residuals of the objective function has fallen below the set threshold, or and The magnitude of the iterative parameter change is already small enough that the final output can be considered... and It represents the optimal unsaturated convection velocity and unsaturated hydraulic diffusivity of the soil column under unsaturated infiltration conditions in this experiment.

[0074] Furthermore, when the output is No, the objective function will iterate again.

[0075] At the same time, check whether the fitted curve and the measured moisture content curve are consistent in the main area of ​​the wetting front; If the conditions are met, proceed to the next step; if the conditions are not met, extend the scanning time or re-fit the local data.

[0076] Step 5: Generation of Soil-Water Characteristic Curves and Permeability Coefficient Function S51, SWRC shape parameter δ calculation: The central calculation and inversion module uses the following formula to calculate the shape parameter of the soil-water characteristic curve:

[0077] Where δ represents the shape parameter of the soil-water characteristic curve; This represents the optimal unsaturated convection velocity obtained in step S43; This represents the optimal unsaturated hydraulic diffusion rate obtained in step S42; It indicates the density of water.

[0078] Subsequently, the central calculation and inversion module uses the following formula to generate soil and water characteristic curves:

[0079] Where θ(ψ) represents the volumetric water content under suction force ψ; Indicates the residual volumetric moisture content; δ represents the saturated volumetric water content; δ represents the shape parameter of the soil-water characteristic curve calculated in step S51; ψ represents the soil suction; |ψ| represents the absolute value of the suction.

[0080] The central calculation and inversion module calculates θ(ψ) point by point according to different suction values, forming a complete soil-water characteristic curve.

[0081] S52, saturated permeability coefficient Calculation: The central calculation and inversion module uses the following formula to calculate the saturated permeability coefficient:

[0082] in, Indicates the saturated permeability coefficient; This represents the optimal unsaturated convection velocity; Indicates the saturated volumetric water content; This indicates the residual volumetric moisture content.

[0083] S53. Generation of Unsaturated Permeability Coefficient Function: The central calculation and inversion module generates the unsaturated permeability coefficient function using the following formula:

[0084] Where k(ψ) represents the unsaturated permeability coefficient under suction ψ; δ represents the saturated permeability coefficient obtained in step S52; δ represents the shape parameter of the soil-water characteristic curve; ψ represents the soil suction. The central calculation and inversion module discretizes k(ψ) according to the set suction interval and automatically plots the corresponding function curve. Step 6: Aperture Distribution Derivation and Test Report Output S61. Derivation of Aperture Distribution: Establishing the relationship between suction and equivalent aperture. The central calculation and inversion module uses the Laplace capillary equation to establish the relationship between suction and equivalent aperture.

[0085] Where ψ represents soil suction; σ represents liquid surface tension; and α represents the contact angle. This represents the equivalent aperture. In this test setup and experimental scenario, the liquid is typically water; therefore, σ represents the surface tension of water at the corresponding temperature; α reflects the wetting relationship between the water and the pore wall. This represents the equivalent capillary pore size corresponding to the current suction force. The cumulative frequency distribution function of the equivalent pore size is then generated using the following formula:

[0086] in, Indicates less than or equal to the equivalent aperture. The cumulative frequency; Represents residual saturation; σ represents liquid surface tension; δ represents the shape parameter of the soil-water characteristic curve; Indicates the equivalent aperture.

[0087] The equivalent aperture frequency distribution function is generated using the following formula:

[0088] in, Indicates equivalent aperture The corresponding frequency distribution function value; σ represents the liquid surface tension; δ represents the shape parameter of the soil-water characteristic curve; Indicates the equivalent aperture; represents the residual saturation; ln(10) represents the natural logarithm constant term.

[0089] The critical aperture for the aperture frequency distribution curve is calculated using the following formula:

[0090] in, The pore size represents the critical pore size corresponding to the peak value of the pore size frequency distribution curve; σ represents the surface tension of the liquid; and δ represents the shape parameter of the soil-water characteristic curve.

[0091] The corresponding peak frequency is expressed by the following formula:

[0092] in, This represents the peak frequency distribution at the critical aperture. This indicates the residual saturation.

[0093] The corresponding suction power is calculated using the following formula:

[0094] in, δ represents the suction value corresponding to the critical aperture; δ represents the shape parameter of the soil-water characteristic curve.

[0095] S62. Output test results and generate reports: The results can be displayed through the human-computer interaction interface 33, and an electronic test report can be automatically generated for subsequent analysis of hydraulic parameters of roadbed fill and engineering evaluation.

[0096] When the preset termination conditions are met, the central calculation and inversion module automatically stops the water supply of the precision peristaltic pump 22, stops the scanning tasks of the vertical linear module 13 and the rotating platform 14, and archives all original projection data, reconstructed images, control records, moisture content profiles, and inversion results of this experiment. Preferably, the termination conditions include at least one of the following: the wetting front reaches the target depth, the test duration reaches the set upper limit, the parameter identification results are continuously stable, or the operator manually ends the test.

Claims

1. An in-situ testing device for the water characteristics of unsaturated soil in roadbed fill, characterized in that, It includes an intelligent scanning and monitoring module, a high-precision closed-loop control module, and a central computing and inversion module; The intelligent scanning monitoring module includes an X-ray source (11), a detector array (12), a vertical linear module (13), a rotating platform (14), and a multi-channel pulse amplitude analyzer (15). The high-precision closed-loop control module includes a non-contact liquid level sensor (21), a precision peristaltic pump (22), an electronic balance (23), and a water supply tank (24). The central computing inversion module includes a high-performance edge computing terminal (31) and a human-computer interaction interface (33). The rotating platform (14) is fixed on the system base and is used to support the soil column containing the roadbed filler; the vertical linear module (13) is vertically arranged on one side of the rotating platform (14), the X-ray source (11) and the detector array (12) are respectively installed on both sides of the vertical linear module (13), the input end of the multi-channel pulse amplitude analyzer (15) is electrically connected to the output end of the detector array (12), and the output end of the multi-channel pulse amplitude analyzer (15) is electrically connected to the input end of the high-performance edge computing terminal (31); The non-contact liquid level sensor (21) is suspended above the soil column by a bracket. The water supply tank (24) is placed on the electronic balance (23). The inlet of the precision peristaltic pump (22) is connected to the water supply tank (24) through a hose. The outlet of the precision peristaltic pump (22) is connected to the top of the soil column through a hose. The control output terminal of the high-performance edge computing terminal (31) is electrically connected to the control terminals of the vertical linear module (13), the rotating platform (14), the X-ray source (11), and the precision peristaltic pump (22), respectively. The high-performance edge computing terminal (31) is also electrically connected to the signal output terminals of the non-contact liquid level sensor (21) and the electronic balance (23), respectively. The human-machine interface (33) is bidirectionally electrically connected to the high-performance edge computing terminal (31).

2. The in-situ testing device for the characteristics of unsaturated soil water in roadbed fill material according to claim 1, characterized in that, The high-precision closed-loop control module also includes an ambient temperature and humidity sensor (25), which is arranged around the soil column and its signal output terminal is electrically connected to the high-performance edge computing terminal (31).

3. The in-situ testing device for the characteristics of unsaturated soil water in roadbed fill material according to claim 1 or 2, characterized in that, The central computing inversion module also includes a multi-functional motion controller (32), and the high-performance edge computing terminal (31) is electrically connected to the control terminals of the vertical linear module (13) and the rotating platform (14) through the multi-functional motion controller (32).

4. The in-situ testing device for the characteristics of unsaturated soil water in roadbed fill material according to claim 1, characterized in that, The high-performance edge computing terminal (31) has built-in image reconstruction algorithm and analytical solution fitting algorithm. The high-performance edge computing terminal (31) is configured to perform the following operations: receive raw counting data from the multichannel pulse amplitude analyzer (15) and calculate and output volumetric water content profile in real time. The volumetric moisture content profile Substituting the data into the analytical solution model of the Richards equation, and performing dynamic fitting using the nonlinear least squares method, the unsaturated hydraulic diffusivity is inverted. Unsaturated convection velocity ; Using formula The shape parameters of the soil-water characteristic curve are automatically derived; based on the shape parameters of the soil-water characteristic curve... Generate complete soil-water characteristic curves and unsaturated permeability coefficient functions .

5. The in-situ testing device for the water characteristics of unsaturated soil in roadbed fill material according to claim 4, characterized in that, The high-performance edge computing terminal (31) is also configured to perform an image subtraction operation: during the infiltration process, the reconstructed wet soil attenuation coefficient map is used to perform an image subtraction operation. Subtract the pre-collected and stored dry soil attenuation coefficient baseline map The attenuation component caused by moisture was obtained. Then, the moisture attenuation component is divided by the pre-calibrated pure water attenuation coefficient. To calculate the volumetric water content of each pixel. .

6. A method for in-situ testing of the water characteristics of unsaturated soil in roadbed fill, applied to the apparatus described in any one of claims 1 to 5, characterized in that, Includes the following steps: S1: Test preparation and system calibration, including acquiring dry soil baseline maps and calibrating the pure water attenuation coefficient; S2: Constant head establishment and dynamic infiltration control, which maintains a constant head at the top of the soil column through a PID closed-loop control algorithm and records the instantaneous infiltration flow rate; S3: Tomographic scanning and volumetric water content acquisition. Attenuation coefficient maps of wet soil at different depths and times are obtained through X-ray tomography, and real-time volumetric water content profiles are generated through image subtraction processing. ; S4: Fitting the analytical solution of the Richards equation and identifying parameters, using the volumetric water content profile. The optimal unsaturated hydraulic diffusivity is dynamically derived by inputting the analytical solution model of the Richards equations using the nonlinear least squares method. Unsaturated convection velocity ; S5: Generation of soil-water characteristic curves and permeability coefficient function, based on the inverted unsaturated hydraulic diffusivity. Unsaturated convection velocity Calculate the shape parameters of the soil-water characteristic curve. and saturated permeability coefficient And generate complete soil-water characteristic curves SWRC and unsaturated permeability coefficient functions. ; S6: Aperture distribution derivation and test report output.

7. The in-situ testing method for the water characteristics of unsaturated soil in roadbed fill material according to claim 6, characterized in that, The analytical solution model of the Richards equation used in step S4 is as follows: in, This represents the volumetric water content at depth z and time t predicted by the model. This indicates the initial volumetric water content of the soil column; This indicates the volumetric water content corresponding to the top boundary of the soil column. This represents unsaturated convection velocity; The unsaturated hydraulic diffusivity is represented by z; the depth is represented by t; and the complementary error function is represented by erfc. The nonlinear least squares method minimizes the objective function: To identify the optimal and ;in, This represents the sum of squared residuals from the fit; This represents the measured average volumetric water content at the i-th depth and the j-th time. This represents the calculated model prediction value; and This indicates the parameter to be identified.

8. The in-situ testing method for the water characteristics of unsaturated soil in roadbed fill material according to claim 6, characterized in that, In step S5, the shape parameters of the soil-water characteristic curve The calculation formula is: Where δ represents the shape parameter of the soil-water characteristic curve; This represents the obtained optimal unsaturated convection velocity; This represents the obtained optimal unsaturated hydraulic diffusivity; Indicates the specific gravity of water; The complete soil-water characteristic curve SWRC and unsaturated permeability coefficient function Generated using the following exponential model: Where θ(ψ) represents the volumetric water content under suction force ψ; Indicates the residual volumetric moisture content; Expressed as the absolute value of suction force; Indicates the saturated volumetric water content; .

9. The in-situ testing method for the water characteristics of unsaturated soil in roadbed fill material according to claim 6, characterized in that, Step S6 further includes deriving the pore size distribution based on the Laplace capillary equation, specifically as follows: Establish the relationship between suction force and equivalent aperture: Where ψ represents soil suction; σ represents liquid surface tension; and α represents the contact angle. Indicates the equivalent aperture; Generate the equivalent aperture cumulative frequency distribution function: in, Indicates less than or equal to the equivalent aperture. The cumulative frequency; Represents residual saturation; σ represents liquid surface tension; δ represents the shape parameter of the soil-water characteristic curve; Indicates the equivalent aperture; And calculate the critical aperture: in, This represents the critical aperture corresponding to the peak value of the aperture frequency distribution curve. Indicates the surface tension of a liquid; This represents the shape parameters of the soil-water characteristic curve.

10. The in-situ testing method for the water characteristics of unsaturated soil in roadbed fill material according to claim 6, characterized in that, The step S1 of acquiring the dry soil reference map specifically includes: using a vertical straight module (13) to drive the X-ray source (11) and detector array (12) to move layer by layer along the Z-axis of the soil column, and using a rotating platform (14) to acquire multi-angle projections at each depth layer. The high-performance edge computing terminal (31) then reconstructs the dry state attenuation coefficient reference map of each depth layer based on the filtered back projection algorithm. .

Images