A prediction method and system for the soil arching effect of the soil between embankment piles under coupling action

Abstract

The present invention relates to the technical field of geotechnical engineering, and particularly to a method and system for predicting the soil arching effect between embankment piles under the coupling action. The method includes the following steps: Layering the test soil in the model box and presetting data monitoring devices to establish a multi-dimensional data monitoring array; Controlling the hydraulic gradient of the model box through an adjustable water pressure water pump to obtain the pore water pressure distribution; Applying static loads to the model box using a hydraulic servo loader to obtain the stress distribution of the test soil; Vibration the model box based on a hydraulic vibrator to obtain vibration displacement information; Triggering the soil arch phenomenon, conducting multi-condition comparative tests on the model box, obtaining soil arch shape evolution data, and establishing a quantitative correlation model; Combining machine learning algorithms to construct a soil arching effect prediction model to achieve the prediction of the soil arching effect. The present invention studies the mechanical mechanism of the soil arching effect under the coupling action of multiple factors, providing a scientific basis for the design and construction of geotechnical engineering.

Description

Technical Field

[0001] The invention relates to the technical field of geotechnical engineering, and in particular to a method and system for predicting soil arching effect between embankment piles under coupling action. Background Art

[0002] The traditional sliding door experiment is a classic method to study the soil arch effect. By moving the sliding door, the formation and development process of the soil arch inside the soil can be intuitively observed. However, current research is mostly limited to simple static conditions and cannot meet the complex and changeable environmental requirements in actual engineering.

[0003] In actual geotechnical engineering, soil is often in an environment of groundwater seepage and various vibrations. Seepage will change the effective stress state of the soil, affect the cohesion and friction between soil particles, and thus significantly change the mechanical properties of the soil arch. Vibrations, such as those caused by traffic loads and earthquakes, will cause soil particles to rearrange, interfere with the stability of the soil arch, and even lead to the destruction and remodeling of the soil arch. However, there are many gaps in the research on the movable door soil arch effect under the coupling of seepage and vibration.

[0004] There are many challenges in simulating a real vibration environment under indoor conditions. Vibration waves have complex spectral characteristics, including rich frequency components, amplitude changes, and phase differences. At the same time, site soil conditions have an important influence on the propagation and attenuation of vibration waves. How to accurately simulate the interaction between site soil and vibration waves under indoor conditions is one of the key issues. In addition, the processing of indoor test boundary conditions is also crucial. Unreasonable boundary conditions will cause vibration wave reflections and interfere with the accuracy of experimental results.

[0005] In view of this, the present invention proposes a method and system for predicting the soil arch effect between embankment piles under coupling, which breaks through the limitations of traditional single-factor experiments and realizes the quantitative characterization of the soil arch effect under multi-field coupling for the first time. The present invention has the characteristics of precise and controllable parameters, comprehensive working condition simulation and three-dimensional testing dimensions, providing a scientific basis for the optimal design of embankments in complex environments. Summary of the invention

[0006] In view of the defects in the prior art, the present invention provides a method and system for predicting the soil arching effect between embankment piles under coupling action.

[0007] To achieve the above object, in a first aspect, the present invention provides a method for predicting the soil arch effect between embankment piles under coupling action. The method includes the following steps: Layering the test soil in the model box and presetting a data monitoring device inside the model box to establish a multi-dimensional data monitoring array; Regulating the hydraulic gradient of the model box through an adjustable water pressure water pump, and obtaining the pore water pressure distribution by using the multi-dimensional data monitoring array; Applying a static load to the model box by using a hydraulic servo loader, and obtaining the stress distribution of the test soil according to the multi-dimensional data monitoring array; Vibration is performed on the model box based on a hydraulic vibrator, and vibration displacement information is obtained according to the multi-dimensional data monitoring array; Triggering the soil arch phenomenon through a movable block, performing comparative tests on the model box under multiple working conditions, obtaining the soil arch shape evolution data of the soil arch phenomenon, and establishing a quantitative correlation model; Based on the soil arch shape evolution data, a soil arch effect prediction model is constructed by combining a machine learning algorithm to realize the prediction of the soil arch effect. By constructing a multi-dimensional data monitoring array and multi-field coupling action, the present invention improves the scientificity and engineering applicability of the research on the soil arch effect, regulates the hydraulic gradient to simulate seepage, combines the application of static load and vibration load, realizes the synchronous observation of the pore water pressure field and the stress field under complex environmental conditions, and effectively reveals the coupling action mechanism of seepage-loading-vibration; Comparative tests are carried out under multiple working conditions to obtain the soil arch shape evolution data and establish a quantitative correlation model, breaking through the limitations of traditional empirical formulas; A soil arch effect prediction model is constructed based on a machine learning algorithm, integrating multi-source monitoring quantization parameters and the spatio-temporal evolution law of the soil arch shape. It can not only dynamically characterize the evolution process of the soil arch effect, but also predict the soil bearing characteristics under different working conditions, providing an innovative method with both theoretical depth and engineering accuracy for the stability evaluation of subgrade engineering, and significantly improving the reliability of the prediction of the soil arch effect.

[0008] Optionally, establishing the multi-dimensional data monitoring array includes: The data monitoring device includes a pore water pressure gauge, a soil pressure sensor, a displacement sensor and an image acquisition device to establish the multi-dimensional data monitoring array. By integrating a pore water pressure gauge, a soil pressure sensor, a displacement sensor and an image acquisition device to construct a multi-dimensional monitoring array, the present invention realizes the all-round dynamic synchronous acquisition of pore water pressure, soil stress, displacement field and soil arch shape, and the deep fusion of multi-source data accurately reveals the spatio-temporal evolution law of the soil arch effect under coupling action, overcomes the limitations of traditional single-parameter monitoring, combines image vision analysis technology, quantitatively analyzes the internal mechanical response of the soil arch structure, intuitively captures the dynamic change characteristics of the soil arch shape, and provides multi-dimensional and high-precision data support for constructing a refined numerical model and a machine learning prediction algorithm.

[0009] Optionally, the hydraulic gradient of the model box is regulated by an adjustable water pressure pump, and the pore water pressure distribution is obtained by using the multi-dimensional data monitoring array, including: using the adjustable water pressure pump and combining with a precision flow control valve to adjust the water delivery flow rates of the water tanks on both sides of the model box, so as to realize the regulation of the hydraulic gradient; based on the regulation of the hydraulic gradient, enabling the test soil mass in the model box to reach a steady seepage state; in the steady seepage state, obtaining the pore water pressure distribution of the test soil mass based on the multi-dimensional data monitoring array. The present invention realizes the precise dynamic regulation of the hydraulic gradient and simulates the soil seepage by the coordinated adjustment of the adjustable water pressure pump and the precision flow control valve for the water delivery flow rates of the water tanks on both sides of the model box; establishes a steady seepage state to ensure the reliability of the observation of the mechanical response of the test soil mass in a real seepage environment; obtains the pore water pressure distribution in real time based on the multi-dimensional data monitoring array, quantitatively analyzes the evolution law of the seepage path and the distribution of the pore water pressure field, improves the accuracy of the seepage test, provides key parameter support for the evolution analysis of the soil arch effect under complex hydraulic conditions, enhances the engineering applicability, and provides a scientific basis for the subgrade anti-seepage design and disaster prevention and control.

[0010] Optionally, a static load is applied to the model box by using a hydraulic servo actuator, and the stress distribution of the test soil mass is obtained according to the multi-dimensional data monitoring array, including: based on the hydraulic servo actuator, applying a graded vertical static load to the model box through hydraulic adjustment as the static load; controlling the hydraulic servo actuator to maintain the graded vertical static load until the test soil mass reaches a deformation stable state; in the deformation stable state, obtaining the stress distribution of the test soil mass by using the multi-dimensional data monitoring array. The present invention realizes the application and stable maintenance of the graded vertical static load on the model box through the high-precision dynamic control ability of the hydraulic servo actuator, accurately simulates the progressive characteristics of the soil mass under load in actual engineering, and the graded loading strategy effectively avoids the soil disturbance error caused by instantaneous loading, ensures the reliability of the data acquisition of the internal stress of the soil mass, and the multi-dimensional data monitoring array synchronously captures the stress field distribution of the soil mass under different load levels, reveals the evolution law of the stress transfer path during the bearing process of the soil arch structure, quantitatively characterizes the dynamic response characteristics of the soil mass, and provides key mechanical parameters for revealing the formation mechanism of the soil arch effect and its stability determination through refined load simulation and multi-dimensional dynamic coupling analysis, enhancing the matching degree between the test and the actual engineering conditions.

[0011] Optionally, the model box is vibrated based on a hydraulic vibrator, and vibration displacement information is obtained according to the multi-dimensional data monitoring array, including: establishing a vibration tabletop, with the model box located above the vibration tabletop; using the hydraulic vibrator to apply a vibration load to the vibration tabletop to vibrate the vibration tabletop, where the vibration load includes a horizontal vibration load and a vertical vibration load; realizing the vibration of the model box based on the vibration of the vibration tabletop, and obtaining the vibration displacement information of the model box according to the multi-dimensional data monitoring array. Through the synergistic effect of the hydraulic vibrator and the vibration tabletop, the present invention simulates the influence of complex dynamic vibrations such as traffic loads or seismic waves on the soil body, adopts a horizontal and vertical composite vibration load loading mode, restores the multi-dimensional vibration coupling effect scenario in actual engineering, improves the test condition coverage, the multi-dimensional data monitoring array captures the spatio-temporal evolution characteristics of the soil body vibration displacement field in real time, combines high-frequency sampling and image analysis techniques, quantitatively analyzes the response of the soil arch structure under dynamic loads, provides a high-precision experimental basis for evaluating the evolution law of the soil arch effect under cyclic loads and the long-term dynamic stability of the subgrade, and has important theoretical support value for seismic design and disaster warning in traffic engineering.

[0012] Optionally, the soil arch phenomenon is induced by a movable block, and multi-condition comparative tests are carried out on the model box to obtain the soil arch morphology evolution data of the soil arch phenomenon and establish a quantitative correlation model, including: the coupling includes seepage, loading, and vibration; by moving the movable block in the vertical direction, the soil between the piles settles, thereby inducing the soil arch phenomenon; respectively changing the hydraulic gradient magnitude of the seepage, the static load application rate of the loading, the frequency of the vibration, and the coupling sequence, thereby constructing the multi-conditions; under the multi-conditions, obtaining the soil arch morphology evolution data based on the multi-dimensional data monitoring array; establishing a quantitative correlation model between the soil arch effect and the seepage, the loading, and the vibration according to the soil arch morphology evolution data. The present invention makes the soil between the piles settle by the movable block to induce the soil arch phenomenon, combines variable-parameter multi-condition tests with multi-factor coupling of seepage-loading-vibration, reveals the dynamic evolution law of the soil arch effect under the interaction of complex environments, quantitatively analyzes the differential influence mechanism of single factors and multi-field coupling on the soil arch morphology by independently or jointly regulating the hydraulic gradient, static load rate, vibration frequency, and coupling sequence, captures the parameters of the soil arch effect in real time based on the multi-dimensional monitoring array, constructs a quantitative correlation model, breaks through the limitations of traditional single-factor empirical models, provides a universal theoretical framework and parametric design basis for evaluating the soil arch effect of subgrade soil in different engineering scenarios, and improves the adaptability to complex working conditions.

[0013] Optionally, the quantitative correlation model satisfies the following relationship:

[0014] ;

[0015] Among them is the soil arch height, is the least squares fitting coefficient, is the seepage force parameter, is the load parameter, is the vibration parameter. By introducing multiple parameters such as seepage force, load, and vibration, and with the help of the least squares fitting coefficient, the present invention accurately quantifies the relationship between various factors and the soil arch height, providing a scientific basis for a deeper understanding of the soil arch effect and helping to more accurately predict the evolution of the soil arch shape.

[0016] Optionally, constructing a soil arch effect prediction model based on the soil arch shape evolution data in combination with a machine learning algorithm to achieve the prediction of the soil arch effect, including: introducing the random forest algorithm as the machine learning algorithm to construct the soil arch effect prediction model; dividing the soil arch shape evolution data into a training set, a validation set, and a test set, training, tuning parameters, and evaluating the soil arch effect prediction model; under the multiple working conditions, predicting the evolution law of the soil arch effect according to the soil arch effect prediction model. The present invention uses the random forest algorithm to construct a soil arch effect prediction model, which can handle high-dimensional data and effectively cope with the complex and variable characteristics of the soil arch shape evolution data; reasonably dividing the data into a training set, a validation set, and a test set can comprehensively evaluate the model performance, avoid overfitting, ensure the generalization ability of the model, optimize the model parameters through the training, tuning parameter, and evaluation processes, improve the prediction accuracy, and accurately predict the evolution law of the soil arch effect under multiple working conditions, providing a reliable basis for engineering practice and ensuring the safety and stability of the engineering structure.

[0017] Optionally, the soil arch effect prediction model satisfies the following relationship:

[0018] ;

[0019] Among them, is the soil arch effect prediction result, is the number of decision trees, is the traversal count flag, is the prediction result of a single decision tree, is the seepage velocity, is the pressure of the static load, is the vibration frequency, is the amplitude, is the test soil body parameter. Based on the random forest algorithm, the present invention comprehensively considers multiple factors such as seepage velocity, static load pressure, vibration frequency, and soil body parameters, and through the integrated prediction of multiple decision trees, can more comprehensively and accurately reflect the soil arch effect, effectively improving the reliability and stability of the soil arch effect prediction result, and providing strong support for engineering practice.

[0020] Second aspect, the present invention provides a prediction system for the soil arch effect between embankment piles under coupling action. The system uses a prediction method for the soil arch effect between embankment piles under coupling action provided by the present invention. The system includes: a model box main body for layered landfill of the test soil mass; a seepage module including the adjustable water pressure water pump for regulating the hydraulic gradient of the model box and conducting seepage on the test soil mass; a loading module including a hydraulic servo loader for applying static load to the model box to realize loading of the test soil mass; a vibration module including a hydraulic vibrator for vibrating the model box to complete vibration of the test soil mass; a movable door module including a movable block for inducing the soil arch phenomenon; a data acquisition module including a multi-dimensional data monitoring array for collecting multi-dimensional data under multiple working conditions; and a data processing module for constructing the quantitative correlation model and the soil arch effect prediction model to realize prediction of the soil arch effect. Each module of the system provided by the present invention has clear division of labor and operates collaboratively, comprehensively considering the coupling action of multiple factors, providing a complete and efficient solution for in-depth exploration of the soil arch effect, being able to more accurately reveal the law of the soil arch effect, providing a reliable basis for engineering practice, and helping to optimize the design of the soil between embankment piles. Description of the Drawings

[0021] Figure 1 It is a flowchart of a prediction method for the soil arch effect between embankment piles under coupling action according to an embodiment of the present invention;

[0022] Figure 2 It is a framework diagram of a prediction system for the soil arch effect between embankment piles under coupling action according to an embodiment of the present invention;

[0023] Figure 3 It is a layout design diagram of the soil arch effect experimental device according to an embodiment of the present invention;

[0024] Figure 4 It is a front view structure diagram of the soil arch effect experimental device according to an embodiment of the present invention;

[0025] Figure 5 It is a schematic diagram of the inner structure of the model box according to an embodiment of the present invention;

[0026] Figure 6 It is a connection schematic diagram of the seepage module according to an embodiment of the present invention;

[0027] Figure 7 It is a schematic diagram of the loading module according to an embodiment of the present invention;

[0028] Figure 8 It is a schematic diagram of the vibration module and the movable door module according to an embodiment of the present invention;

[0029] Figure 9 Schematic diagram of the arrangement of the image acquisition device according to an embodiment of the present invention;

[0030] Description of reference numerals: Main body of the model box A1, seepage module A2, loading module A3, vibration module A4, movable door module A5, data acquisition module A6, data processing module A7, vibration system 1, vibration damping support 11, vibration damping rubber 111, hydraulic vibrator 12, hydraulic pump 121, hydraulic cylinder 122, control device 13, vibration table 14, connecting bolt 15, loading system 2, hydraulic servo loading device 21, fixed frame 22, steel support bracket 221, reinforcement cross beam 222, displacement sensor 231, earth pressure sensor 232, pore water pressure gauge 233, loading control system 24, loading rod 25, loading rod 251, loading plate 252, friction reducing rubber 253, seepage system 3, water tank 31, water inlet faucet 311, water delivery pipe 32, adjustable water pressure water pump 33, precision flow control valve 34, image acquisition device 4, high-speed camera 41, LED lamp 42, image acquisition and analysis system 43, machine learning platform 5, computer terminal 51, model box 6, sensor embedding groove 61, horizontal seepage hole 62, filter screen 63, plexiglass 64, sand collection device 65, bolt connection hole 66, movable door system 7, movable block 71, moving mechanism 72, movable door displacement sensor 73, multi-parameter synchronous acquisition system 8, integrated acquisition system 84. Specific embodiments

[0031] The specific embodiments of the present invention will be described in detail below. It should be noted that the embodiments described here are only for illustrative purposes and are not used to limit the present invention. In the following description, a large number of specific details are set forth in order to provide a thorough understanding of the present invention. However, it will be apparent to those of ordinary skill in the art that: the present invention does not have to be practiced with these specific details. In other instances, well-known circuits, software, or methods have not been specifically described in order to avoid obscuring the present invention.

[0032] Throughout the specification, the reference to "an embodiment", "embodiment", "an example" or "example" means that a particular feature, structure, or characteristic described in connection with the embodiment or example is included in at least one embodiment of the present invention. Thus, the phrases "in an embodiment", "in the embodiment", "an example" or "example" appearing throughout the specification do not necessarily all refer to the same embodiment or example. In addition, the specific features, structures, or characteristics may be combined in any suitable combination and / or sub-combination in one or more embodiments or examples. In addition, those of ordinary skill in the art should understand that the drawings provided herein are for illustrative purposes only and are not necessarily drawn to scale.

[0033] Please refer to Figure 1, an embodiment of the present invention provides a method for predicting the soil arch effect between piles in an embankment under coupling action, and the method includes the following steps:

[0034] S1. Layer the test soil mass and fill it into the model box 6, and preset a data monitoring device inside the model box 6 to establish a multi-dimensional data monitoring array.

[0035] In this embodiment, layer and compact the test soil mass in the model box 6; preset monitoring points of the data monitoring device inside the model box 6, reserve a sensor embedding groove 61 at the bottom of the model box 6, and the data monitoring device includes a pore water pressure gauge 233, a soil pressure sensor 232, and a displacement sensor 231; in addition, an image acquisition device 4 is arranged on the front side of the model box 6 to collect images of the evolution of the soil arch during the test process; the pore water pressure gauge 233, the soil pressure sensor 232, the displacement sensor 231, and the image acquisition device 4 serve as the multi-dimensional data monitoring array.

[0036] Select good embankment fillers and similar materials as fillers for the test soil mass, proportion them according to their target relative density and water content. First, dry the experimental materials, and then uniformly proportion them according to their set water content. After the proportioning is completed, seal each layer of soil sample. Since it is under vibration conditions, materials with a relatively large relative density need to be designed for filling. Fill the soil mass materials treated according to the expected target water content into the model box 6 layer by layer, and control the thickness of each layer to be 10 cm to ensure the uniformity of soil compaction. And start the image acquisition device 4 to work for image acquisition. It should be noted that during the compaction process, professional compaction equipment needs to be used to compact each layer of soil. During the compaction process, control the compaction energy and the number of compaction times to ensure that the soil reaches the predetermined density. After each layer is compacted, ensure that it meets the requirements of the experimental design.

[0037] Specifically, the model box 6 has characteristics such as high strength and fatigue resistance under complex conditions. Simulate the soil arch effect of the pile-supported embankment in the model box 6, and deeply explore the development and evolution of the soil arch under seepage, loading, and vibration conditions, especially simulate the soil arch effect under the real earthquake environment.

[0038] It should be noted that the model box 6 is an aluminum soil box with a size of 0.6 m (length) × 0.6 m (width) × 0.8 m (height).

[0039] The multi-dimensional data monitoring array can simultaneously monitor multiple parameters such as the internal stress, pore water pressure, settlement, displacement, and seepage flow of the soil mass, and obtain more comprehensive experimental data. Transmit the collected data to the machine learning platform 5 and the computer terminal 51 for analysis, enrich its training model database, and combine digital image correlation technology to obtain the deformation information of the soil between piles, and more deeply study the formation and failure mechanism of the soil arch effect. The introduction of the machine learning platform 5 can deeply mine the experimental data and establish a soil arch effect prediction model.

[0040] S2. Regulate the hydraulic gradient of the model box 6 through an adjustable water pressure pump 33, and obtain the pore water pressure distribution by using the multi-dimensional data monitoring array.

[0041] In this embodiment, an adjustable water pressure pump 33 is used to regulate the water delivery flow rate of the water tanks 31 on both sides of the model box 6 in combination with a precision flow control valve 34 to regulate the hydraulic gradient, providing good support for stable horizontal seepage; based on the regulation of the hydraulic gradient, the test soil body in the model box 6 reaches a stable seepage state; in the stable seepage state, the water pressure value is monitored based on the pore water pressure gauge 233 in the multi-dimensional data monitoring array, and the pore water pressure distribution of the test soil body is obtained.

[0042] Specifically, the adjustable water pressure pump 33 can accurately regulate the water flow rate. Cooperating with the precision flow control valve 34, it can simulate various complex seepage conditions, such as constant seepage, non-constant seepage, etc. The pore water pressure gauges 233 are evenly arranged at different positions in the soil body to monitor the changes in pore water pressure and seepage velocity in real time. A pair of left and right water tanks 31 are used for temporarily storing water, and a number of seepage holes are arranged on the inner side thereof for the water outlets of horizontal seepage to provide a stable horizontal seepage environment.

[0043] S3. Apply a static load to the model box 6 by using a hydraulic servo actuator 21, and obtain the stress distribution of the test soil body according to the multi-dimensional data monitoring array.

[0044] In this embodiment, a graded vertical static load is applied to the model box 6 through hydraulic regulation based on the hydraulic servo actuator 21 as the static load; the graded vertical static load is maintained by controlling the hydraulic servo actuator 21 until the test soil body reaches a deformation stable state; in the deformation stable state, the stress distribution of the test soil body is obtained by using the earth pressure sensor 232 in the multi-dimensional data monitoring array.

[0045] S4. Vibrate the model box 6 based on a hydraulic vibrator 12, and obtain vibration displacement information according to the multi-dimensional data monitoring array.

[0046] In this embodiment, a vibration table 14 is established, and the model box 6 is located on the vibration table 14; a vibration load is applied to the vibration table 14 by using the hydraulic vibrator 12 to vibrate the vibration table 14. The vibration load includes a horizontal vibration load and a vertical vibration load; the vibration of the model box 6 is realized based on the vibration of the vibration table 14, and the vibration displacement information of the model box 6 is obtained according to the displacement sensor 231 in the multi-dimensional data monitoring array.

[0047] Specifically, different vibration modes with different amplitudes and frequencies are realized through frequency conversion, while reducing the impact on the surrounding environment and the ground. The frequency range is 0.1 Hz - 50 Hz, and the vertical acceleration is 1.0g.

[0048] In an alternative embodiment, the vibrating table 14 is discretized into a set of point clouds and the Monte Carlo random field is used to stochastically model the random vibration of each point on the vibrating table 14 in the three directions of (X, Y, Z); the Monte Carlo random field limits the vibration amplitude of each point on the vibrating table 14 within a certain range to prevent damage to the platform structure, and at the same time the vibration frequency should be limited within the designed frequency range of the platform; the vibrations between adjacent points on the vibrating table 14 should have a certain correlation to simulate the overall movement of the vibrating device, and the Gaussian random field is used to model this correlation.

[0049] Furthermore, the vibrating table 14 is rectangular. The grid division algorithm is used to uniformly generate points of the grid and store the point cloud coordinates (x, y, z). The state variable of each point at each time step is its displacement in the three directions of X, Y, Z (dx, dy, dz). The energy function is defined to describe the vibration state of the point cloud. The Metropolis–Hastings Algorithm is used to sample from the multi-scale conditional random field, and at each time step, the displacement of each point is updated according to the sampling result of the multi-scale conditional random field and the displacement is converted into a control signal and sent to the actuator hydraulic vibrator 12. The finite element analysis software is used to simulate the vibration of the vibrating table 14 to verify the effectiveness of the multi-scale conditional random field control.

[0050] In an alternative embodiment, sensors are arranged on the vibrating table 14 in a grid pattern, and the arrangement density needs to satisfy the Nyquist Sampling Theorem (Nyquist). The time series data is synchronously collected to generate the point cloud. The sensor coordinates and the vibration amplitude are combined into a spatio-temporal point cloud. The platform is divided into finite elements using the finite element simplified model. The vibration characteristics of each element are represented by interpolation of the adjacent sensor points. Subsequently, the Monte Carlo random field is constructed and the first covariance function is defined, and then the random phase spectrum Φ(ω)-U(0, 2π) is generated according to the target power spectral density.

[0051] The spatio-temporal point cloud satisfies the following relationship:

[0052]

[0053] where is the spatio-temporal point cloud, is the sensor coordinate, is the vibration amplitude, is the traversal count flag, is the total quantity.

[0054] The first covariance function satisfies the following relationship:

[0055]

[0056] where is the first covariance, is the variance, are the sensor coordinates, is the correlation length.

[0057] The excitation signal generated by the multi-scale conditional random field is input to the control device 13 and fed back to the hydraulic vibrator 12 for feedback correction. The sensor collects the point cloud vibration data in real time, and the excitation signal is corrected by proportional-integral-derivative control (PID) or adaptive filtering (such as the least mean square algorithm, LMS).

[0058] Specifically, the surface of the vibration table 14 is discretized into a three-dimensional point cloud plane, and a high-density dot matrix is generated through grid division technology. Each point represents an independent vibration unit. The Monte Carlo random field generates random vibration parameters for each point, which can simulate the multi-dimensional random vibration environment of the random vibration of the vibration table 14, and introduce a Gaussian process or a Markov random field to describe the spatial correlation between point clouds to ensure the uniform distribution of vibration energy. The energy function of the vibration energy distribution satisfies the following relationship:

[0059]

[0060] where is the energy, is the frequency, is the total number of points, is the th point in the point cloud at frequency under the vibration amplitude, is the target spectrum, is the frequency weight, is the regularization parameter, is the point and the point the vibration covariance between, is the point and the point the target vibration covariance between.

[0061] The energy function is used to simultaneously constrain the spectral distribution and spatial correlation of vibration energy, and is applicable to multi-objective optimization. The vibration states of each point are fed back to the control device 13 in real time through sensors arranged on the vibration device, and the driving signal of the vibration device is dynamically adjusted. The vibration generated by the random parameters of Monte Carlo can realize the quasi-simulation of multi-direction and multi-frequency composite vibration of the vibration table 14, which can actually cope with complex working conditions. Its technical steps include: discretizing the vibration table 14, discretizing the surface of the vibration table 14 into multiple points using the uniform grid division method; point cloud dynamics modeling, establishing a dynamics model between discrete points using the spring-damper model, connecting springs and dampers between adjacent discrete points to establish a spring-damper system, and the spring stiffness and damping coefficient need to be determined according to the material properties and structural parameters of the vibration table 14; Monte Carlo random field generation, assuming that the parameters (internal friction angle, cohesion, permeability coefficient, etc.) of the soil body follow a Gaussian distribution using the Gaussian random field, generating the random field of the soil body, and simulating the heterogeneity of the soil body; closed-loop feedback control, using a PID controller to control the movement of the vibration table 14, and achieving precise control of the vibration table 14 through closed-loop feedback control.

[0062] Specifically, the Gaussian random field assumes that the displacements of each point in the point cloud follow a multivariate Gaussian distribution in the three directions of (X, Y, Z); select an appropriate covariance function according to the actual situation; calculate the distance between any two points in the point cloud to obtain the distance matrix; use the selected covariance function and the distance matrix to construct the covariance matrix; sample from the multivariate Gaussian distribution to sample the displacements of each point in the point cloud.

[0063] Specifically, the covariance function satisfies the following relationship:

[0064]

[0065] Among them, is the covariance between point and point , is the variance, is point and point The distance between, is the correlation length.

[0066] It should be noted that the variance is used to control the overall displacement amplitude; the correlation length is used to control the attenuation speed of the correlation. The larger the correlation length, the stronger the correlation and the more similar the displacements of adjacent points.

[0067] S5. Trigger the soil arching phenomenon through the movable block 71, conduct multi-condition comparative tests on the model box 6, obtain the soil arch shape evolution data of the soil arching phenomenon, and establish a quantitative correlation model.

[0068] In this embodiment, coupling includes seepage, loading, and vibration; by moving the movable block 71 in the vertical direction, settlement of the soil between the piles is generated, thereby triggering the soil arch phenomenon; the magnitude of the hydraulic gradient of seepage, the application rate of the static load of loading, the frequency of vibration, and the order of coupling are respectively changed to construct multiple working conditions; under multiple working conditions, the evolution data of the soil arch morphology is obtained based on the multi-dimensional data monitoring array; a quantitative correlation model of the soil arch effect with seepage, loading, and vibration is established according to the evolution data of the soil arch morphology.

[0069] According to the experimental design, select the seepage hydraulic gradient, vibration frequency and amplitude, and the magnitude of the loading value and its time to turn on and debug the device. At the same time, the experimental device adopts a modular design and can be flexibly configured according to different experimental requirements. After the experiment is completed, weigh and analyze the sand and water collected by the sand collection device 65, and clean and organize the experimental device for the next use.

[0070] Furthermore, through trigger signal synchronization, a unified trigger signal is used to start the data acquisition of all sensors. This method can ensure that the starting time of data acquisition is consistent. Analyze the acquired evolution data of the soil arch morphology, transmit the sensor data to the data processing module for preprocessing, and align the data of different sensors using timestamp synchronization. Analyze the data through machine learning algorithms and export the analysis results into various formats of files.

[0071] Specifically, the quantitative correlation model satisfies the following relationship:

[0072]

[0073] where is the soil arch height, is the least squares fitting coefficient, is the seepage force parameter, is the load parameter, is the vibration parameter.

[0074] S6. Based on the evolution data of the soil arch morphology, combine machine learning algorithms to construct a soil arch effect prediction model to realize the prediction of the soil arch effect.

[0075] In this embodiment, the random forest algorithm is introduced as the machine learning algorithm to construct a soil arch effect prediction model; the evolution data of the soil arch morphology is divided into a training set, a validation set, and a test set, and the soil arch effect prediction model is trained, tuned, and evaluated; under multiple working conditions, the evolution law of the soil arch effect is predicted according to the soil arch effect prediction model.

[0076] Specifically, a machine learning platform 5 is constructed to train the collected data and predict the development and failure mechanism of soil arches. The random forest algorithm is selected as the machine learning algorithm, and the collected soil arch morphology evolution data is divided into a training set, a validation set, and a test set. Among them, 70% of the data is divided into the training set for training the machine learning model; 15% is the validation set for adjusting the model hyperparameters to prevent overfitting; 15% is the test set for evaluating the generalization ability of the model. Then, the training set data is used to train the machine learning model, and the test set data is used to evaluate the performance of the trained model. The minimum mean square error is selected as the evaluation index to measure the prediction accuracy of the model and optimize the model. Finally, the machine learning model is trained with a large amount of training data to obtain the soil arch effect prediction model.

[0077] Furthermore, the machine learning platform 5 transmits the collected data to the computer terminal 51 to enrich the database, develops appropriate machine learning algorithms for training, trains the model with the training data, and conducts model verification and evaluation. The machine learning platform 5 supports deploying the trained model to the experimental device or the cloud to achieve real-time prediction and analysis of the development of soil arches. The machine learning platform 5 can achieve more accurate and efficient analysis and prediction of the soil arch effect, optimize the embankment design, improve the stability and safety of the embankment, reduce the engineering cost. At the same time, machine learning can also be used for the intelligent control and data acquisition of the experimental device to improve the experimental efficiency and data quality.

[0078] The machine learning platform 5 is based on a large amount of accumulated data such as seepage failure, vehicle load, and vibration parameters. The machine learning platform 5 uses this massive data for training, continuously optimizes the model parameters, and improves the prediction accuracy and reliability. By learning the historical vehicle load, seepage failure parameters, and vibration parameters, the model can discover potential trends and thus make more accurate experimental simulations. The machine learning platform 5 can also input set parameters to control the vibration system 1, the loading system 2, and the seepage system 3. The data acquisition and analysis system can obtain various seepage hydraulic gradients and directions, loading time and magnitude, and vibration frequency and amplitude data of the model under seismic action and automatically save them for subsequent query, comparison, and in-depth research.

[0079] Specifically, the minimum mean square error satisfies the following relationship:

[0080]

[0081] Where is the minimum mean square error, is the total number of samples, is the traversal counting flag, is the predicted soil arch effect coefficient, is the experimental soil arch effect coefficient.

[0082] Specifically, the soil arching effect prediction model satisfies the following relationship:

[0083]

[0084] Wherein, is the prediction result of the soil arching effect, is the number of decision trees, is the traversal counting flag, is the prediction result of a single decision tree, is the seepage velocity, is the pressure of the static load, is the vibration frequency, is the amplitude, is the test soil body parameter.

[0085] In an alternative embodiment, through the differential settlement of the soil between piles, the embankment loading, cyclic load and dynamic regulation of pore water pressure are simulated; during the test, the contact pressure between the pile and the soil, the stress redistribution of the soil between piles, the pile top displacement and the evolution data of pore water pressure are synchronously collected, the data of all sensors are synchronously recorded using a multi-dimensional data monitoring array, and the data processing module is used to preprocess the data to ensure the time synchronization of all sensor data. Based on the processed sensor data, the three-dimensional stress field, displacement field and pore water pressure field are reconstructed, and a professional three-dimensional visualization software is used to visualize the reconstructed three-dimensional stress field, displacement field and pore water pressure field and analyze the three-dimensional shape of the soil arching effect.

[0086] Please refer to Figure 2 , in an alternative embodiment, the present invention provides a prediction system for the soil arching effect of the soil between piles in an embankment under coupling action. The system uses a prediction method for the soil arching effect of the soil between piles in an embankment under coupling action provided by the present invention. The system includes a model box main body A1, a seepage module A2, a loading module A3, a vibration module A4, a movable door module A5, a data acquisition module A6 and a data processing module A7.

[0087] Please refer to Figure 3 , which is a layout design diagram of the soil arching effect experimental device, showing the positions and connection relationships between the various devices.

[0088] Please refer to Figure 4 , which is a front view structural diagram of the soil arching effect experimental device.

[0089] The model box main body A1 is used to layer-fill the test soil body.

[0090] Please refer to Figure 5 , which is a schematic diagram of the inner structure of the model box 6.

[0091] Specifically, the model box 6 includes a sensor embedding groove 61, horizontal seepage holes 62, a filter screen 63, plexiglass 64, a sand collection device 65, and bolt connection holes 66; the model box 6 is used to fill the test soil mass; the horizontal seepage holes 62 on both left and right sides of the model box 6 are several water outlets for horizontal seepage; the filter screen 63 is arranged at the bottom of the model box 6, and its pore size is larger than the particle size of fine particles, which can effectively allow fine particles to pass through the filter screen 63 while preventing large-particle-size particles from entering the sand collection device 65 through the filter screen 63; the sensor embedding groove 61 is arranged around the bottom of the movable door system 7 and on the upper surface of the movable block 71 for monitoring data during the test, and can collect data in a variety of complex environments to provide strong support for the test; the plexiglass 64 is arranged on the front side of the model box 6, which is used to observe the occurrence and failure phenomena of soil arches and can also provide a good visual environment for the image acquisition device 4; the sand collection device 65 is arranged under the filter screen at the bottom of the model box 6, which is used to collect the fine particles and water lost during the test.

[0092] Fix the model box 6 on the vibration device to ensure that the bottom of the model box 6 is rigidly connected to the vibration table surface 14. Connect the water tanks 31 on both left and right sides of the model box 6 to the water storage tank to provide a stable water source. Thoroughly clean the model box 6, check whether there are damages on the box wall and whether the sealing performance is good, ensure that the transparent plexiglass 64 observation window has no scratches and is clear and transparent, to prevent errors in the photos taken by the high-speed camera 41.

[0093] Furthermore, bury the earth pressure sensor 232 on the upper surface of the movable block 71 and in the reserved groove at the bottom of the model box 6 for collecting the earth pressure values under seepage, vibration, and loading conditions. Arrange multiple pore water pressure gauges 233 in the model box 6 for pore water pressure measurement. Arrange a displacement sensor 231 at the bottom end of the movable block 71 to ensure that the displacement sensor 231 is firmly installed, the wire connection is correct, and it does not affect the normal operation of the soil mass and the movable block 71.

[0094] The seepage module A2 includes the adjustable water pressure water pump 33, which is used to regulate the hydraulic gradient of the model box 6 and conduct seepage on the test soil mass.

[0095] Please refer to Figure 6 , which is a schematic diagram of the connection of the seepage module.

[0096] Specifically, the seepage system 3 is located on the left and right sides of the model box 6 and is used to provide and regulate the water pressure during the seepage process. It includes a set of water tanks 31, a water inlet faucet 311, an adjustable water pressure pump 33, and a water delivery pipe 32. The set of water tanks 31 is fixed on the left and right sides of the model box 6 to temporarily store water and prevent the adverse effects caused by the sudden increase of the water head. The water inlet faucet 311 is arranged on the water tank 31 as its water inlet. The adjustable water pressure pump 33 is connected to the water delivery pipe 32, and the water pressure collected by the precision flow control valve 34 is fed back to the adjustable water pressure pump 33 to automatically adjust the water pressure. The water delivery pipe 32 is used to deliver water to the water tank 31.

[0097] The sand collection device 65 is arranged around the bottom of the model box 6 and is used to comprehensively collect the fine particles lost by the sand under the conditions of seepage, vibration, and loading. The sand collection device 65 is used to collect fine particles and water and can be used under various complex factors. Its stability is reliable and it is suitable for use in various environments. It should be noted that since the seepage outlet hole has a fixed particle size, a filter screen can be added at the outlet hole according to the selected fine particle size of the experiment to prevent fine particles from entering the water tank 31 and affecting the application of seepage.

[0098] The loading module A3 includes a hydraulic servo actuator 21, which is used to apply a static load to the model box 6 to realize the loading of the test soil mass.

[0099] Please refer to Figure 7 , which shows a schematic diagram of the loading module.

[0100] Specifically, the loading system 2 is located on the top of the model box 6 and is used to apply a load to the test soil mass in the model box 6. It includes a hydraulic servo actuator 21, a fixed frame 22, a displacement sensor 231, an earth pressure sensor 232, a loading control system 24, and a loading rod 25. The hydraulic servo actuator 21 can realize the application of vertical graded vertical loads through hydraulic regulation. The fixed frame 22 is composed of a steel support bracket 221 and a reinforcement cross beam 222 and is arranged on the vibration table 14 to support the loading equipment. The displacement sensor 231 and the earth pressure sensor 232 are used to monitor the data of the loading equipment during the test and timely adjust the loading equipment. The loading control system 24 is used to control the hydraulic servo actuator 21 to realize the regulation of different load magnitudes and loading times. The loading rod 25 is composed of a loading rod 251, a loading plate 252, and a friction reducing rubber 253, and distributes the pressure evenly on the loading plate 252 by transmitting the pressure applied by the hydraulic actuator.

[0101] The vertical load is applied to simulate the traffic load. The bottom of the loading frame is fixed above the vibration table 14, providing a rigid and reliable environment for its loading device. The loading frame can provide a stable working platform for the loading device in the vibration environment, ensuring the application of static load and the repeatability of the experiment. The hydraulic servo actuator 21 is fixed on the crossbeam of the loading frame and is used to apply a vertical load to the loading rod 25 to drive its movement, thereby applying a load to the inside of the model box 6. The hydraulic servo actuator 21 performs loading simulation of the traffic load according to different loading times and magnitudes set by the loading control system 24. The upper part of the loading rod 25 is fixed on the loading crossbeam, and the stable vertical load is applied through the set guiding device, and the load is evenly distributed above the soil through the loading plate 252. The earth pressure sensor 232 is arranged above the loading plate 252 to monitor the load applied by the hydraulic servo actuator 21 and the displacement generated by the loading rod 25. The area of the loading plate 252 is smaller than the internal area of the model box 6 to prevent damage to the model box 6 during the loading process. Anti-friction rubbers 253 are arranged outside the periphery of the loading plate 252 to reduce the friction inside it and prevent the soil and water from overflowing during the loading process, which may affect the progress of the experiment and cause experimental errors.

[0102] The vibration module A4 includes a hydraulic vibrator 12, which is used to vibrate the model box 6 to complete the vibration of the test soil.

[0103] Specifically, the vibration system 1 is arranged on the lower side of the model box 6 and is used to apply a vibration effect to the model box 6. It includes a vibration table 14, a vibration damping support 11, a hydraulic vibrator 12, a control device 13, and connecting bolts 15. The vibration table 14 is used to carry the model box 6 and the loading device and is located below the model box 6. It has sufficient strength and stiffness to ensure that it does not deform during vibration. Its material is made of alloy steel to avoid affecting the test results. The connecting bolts 15 are hexagonal anti-vibration bolts made of steel plate material. They have high strength, good plasticity and toughness, and are used to connect the model box 6 and the vibration device. The connecting bolts 15 contain cushions to transmit the vibration to the model box 6 and reduce damage to itself, improving the vibration consistency between the model box 6 and the vibration table 14. The vibration table 14 is connected to the bottom vibration damping support 11 through the hydraulic vibrator 12. The lower part of the vibration damping support 11 is provided with vibration damping rubbers 111 to absorb the damage of the vibration to the ground and the surrounding environment and provide good test conditions. The hydraulic vibrators 12 are arranged at the four vertices of the vibration table 14. Different hydraulic pressures are applied to the hydraulic pumps 121 in the hydraulic vibrators 12 through the frequencies and amplitudes set by the control device 13 to drive the movement of the hydraulic cylinders 122, thereby realizing vibration. The control device 13 adjusts its vibration frequency and amplitude to realize the up and down vibration of the vibration table 14.

[0104] Specifically, the hydraulic vibrator 12 is composed of a hydraulic pump 121 and a hydraulic cylinder 122. The hydraulic pump 121, as a power source, converts mechanical energy into hydraulic energy, provides pressure oil, and drives the hydraulic cylinder 122 to work; the hydraulic cylinder 122 converts hydraulic energy into mechanical energy, and drives the vibration table surface to vibrate through the reciprocating movement of the piston. The hydraulic pipeline is used to transmit pressure oil and connect the hydraulic pump 121 and the hydraulic cylinder 122. Its material is made of high-pressure rubber hoses, and the high-pressure rubber hoses have good flexibility and fatigue resistance.

[0105] The movable door module A5 includes a movable block 71 for triggering the soil arch phenomenon.

[0106] Specifically, the movable door system 7 is located exactly in the middle of the bottom of the model box 6. Through the vertical movement of the movable block 71, differential settlement between piles is generated, thereby triggering the soil arch phenomenon.

[0107] The movable door system 7 includes a movable block 71, a moving mechanism 72, and a movable door displacement sensor 73; the movable block 71 is connected to the moving mechanism 72. Through the movement of the movable block 71, differential settlement is generated with the fixed part of the model box 6, thereby triggering the soil arch effect; the lower part of the moving mechanism 72 is fixed on the vibration table surface 14, and the upper part is connected to the movable block 71 to drive its movement.

[0108] The maximum stroke of the moving mechanism 72 is 100 mm, and the displacement rate can be adjusted within the range of 0.01 mm / s - 10 mm / s; when the movable block 71 is subjected to various external forces such as soil pressure, seepage force, and vibration, the moving mechanism 72 can stably output sufficient torque to ensure that the movable block 71 moves in a predetermined manner and will not have position deviation or unstable movement due to external force interference, thereby ensuring the reliability and repeatability of the experimental results. A displacement sensor 231 is arranged at the bottom end of the movable block 71 to detect the displacement of the movable block 71 and feed it back to the moving mechanism 72 in a timely manner, and make adjustments accurately and timely to improve its accuracy. The moving mechanism 72 is connected to the movable block 71 and directly controls its displacement. The moving mechanism 72 device has mechanisms such as real-time error monitoring and safety protection. Real-time error monitoring is used to stop the moving mechanism 72 from working when the deviation between the actual displacement and the command > 1 mm lasts for 0.5 seconds.

[0109] Please refer to Figure 8 , which shows a schematic diagram of the vibration module and the movable door module.

[0110] The data acquisition module A6 includes a multi-dimensional data monitoring array for collecting multi-dimensional data under the multi-working conditions.

[0111] Specifically, the multi-parameter synchronous acquisition system 8 includes an earth pressure sensor 232, a displacement sensor 231, a pore water pressure gauge 233, and an integrated acquisition system 84; the earth pressure sensor 232 is used to monitor the earth pressure values in its fixed area part and above the movable block 71; the displacement sensor 231 is arranged below the movable block 71 and above the loading plate 252 to monitor its displacement value; the pore water pressure gauge 233 is arranged inside the model box 6 to monitor the water pressure value under seepage conditions; the earth pressure sensor 232 is arranged above the loading plate 252 to monitor its pressure.

[0112] Please refer to Figure 9 , which is a schematic layout diagram of the image acquisition device 4.

[0113] Specifically, the image acquisition device 4 is arranged in front of the model box 6 and is used to capture images of the development process of the soil arch during the experiment; it consists of a high-speed camera 41, an LED lamp 42, and an image acquisition and analysis system 43; the high-speed camera 41 is arranged on the front side of the model box 6 and is used to collect images during the experiment and save the collected images; the LED lamp 42 is used to be arranged on the high-speed camera 41 to improve the quality of the collected images; the image acquisition system transmits the collected images to the computer terminal 51 for analysis by professional analysis software and reconstructs the occurrence and failure process of the soil arch.

[0114] The plexiglass 64 has the characteristics of being transparent, easy to image, and firm. The occurrence and failure process of the soil arch can be observed through this plexiglass 64, providing strong support for the subsequent reconstruction of the occurrence and failure process of the soil arch.

[0115] Before the experiment starts, it is necessary to focus according to the actual situation to make the target area completely presented in the image and eliminate the influence of lens distortion on the measurement results. The exposure time of the high-speed camera 41 is 50 μs - 500 μs, which is specifically adjusted according to the ambient light intensity and the reflection characteristics of the object surface to ensure that the image is clear and not overexposed or underexposed. The LED lamp 42 avoids direct strong light and shadows for the image acquisition of the high-speed camera 41, ensures uniform illumination on the object surface, and improves the image contrast. The image acquisition device 4 is used to store the collected images and transmit them to the computer terminal 51 for analysis. According to the deformation and displacement data obtained from the analysis, software such as Matrix Laboratory (abbreviated as MATLAB) is used to draw deformation cloud diagrams, displacement vector diagrams, etc. In the deformation cloud diagram, different colors represent different deformation amounts, intuitively showing the deformation distribution on the object surface; at the same time, the image acquisition measurement results are correlated with other data in the experiment (such as seepage pressure, vibration acceleration, etc.) for research and analysis of their internal relationships.

[0116] The data processing module A7 is used to construct the quantitative correlation model and the soil arch effect prediction model to realize the prediction of the soil arch effect.

[0117] Specifically, the information collected by the data acquisition module is transmitted to the data analysis module and analyzed using the computer terminal 51 of the machine learning platform 5 to deeply understand the soil arch effect of the soil between the piles of the embankment under the coupling action of seepage, loading and vibration, provide a more reliable basis for the embankment engineering design, optimize parameters such as pile body arrangement, pile spacing, and filler selection, and improve the stability and bearing capacity of the embankment.

[0118] In summary, a method and system for predicting the soil arch effect of the soil between the piles of an embankment under a coupling action proposed by the present invention can simulate the soil arch effect of the soil between the piles of the embankment under the coupling action of three factors: seepage, loading and vibration. At the same time, combined with the machine learning platform 5, after enriching its database with the collected data, it trains, simulates and predicts the occurrence and failure process of the soil arch. At the same time, it innovatively introduces Monte Carlo random field modeling to realize the random vibration of the vibration table 14, and combines seepage and loading simulations, which can more realistically and comprehensively simulate the evolution process of the soil arch effect of the soil between the piles of the embankment under complex working conditions, improve the test accuracy and reliability. The present invention can accurately control seepage, loading and vibration parameters to realize the simulation of different working conditions, and combine the image acquisition device 4 to realize the dynamic visualization of the whole process of soil arch formation - evolution - failure, provide data support for the verification of the theoretical model, provide a more reliable basis for the embankment design, thereby optimizing the embankment design, improving the engineering safety and economy, and having important engineering application value.

[0119] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered by the scope of the claims and the description of the present invention.

Claims

1. A prediction method for the soil arching effect of the soil between embankment piles under the coupling action, characterized in that, It includes the following steps: Layer the test soil into the model box, and preset a data monitoring device inside the model box to establish a multi-dimensional data monitoring array; Regulate the hydraulic gradient of the model box through an adjustable water pressure pump, and obtain the pore water pressure distribution by using the multi-dimensional data monitoring array; Apply a static load to the model box using a hydraulic servo loader, and obtain the stress distribution of the test soil according to the multi-dimensional data monitoring array; Vibrate the model box based on a hydraulic vibrator, and obtain vibration displacement information according to the multi-dimensional data monitoring array; Trigger the soil arch phenomenon through a movable block, conduct multi-condition comparison tests on the model box, obtain the soil arch shape evolution data of the soil arch phenomenon, and establish a quantitative correlation model; Construct a soil arch effect prediction model based on the soil arch shape evolution data combined with a machine learning algorithm to achieve the prediction of the soil arch effect; The step of triggering the soil arch phenomenon through a movable block, conducting multi-condition comparison tests on the model box, obtaining the soil arch shape evolution data of the soil arch phenomenon, and establishing a quantitative correlation model includes: The coupling includes seepage, loading, and vibration; Move the movable block in the vertical direction to cause the soil between the piles to settle, thereby triggering the soil arch phenomenon; Respectively change the magnitude of the hydraulic gradient of the seepage, the application rate of the static load of the loading, the frequency of the vibration, and the order of the coupling, thereby constructing the multi-conditions; Under the multi-conditions, obtain the soil arch shape evolution data based on the multi-dimensional data monitoring array; Establish a quantitative correlation model between the soil arch effect and the seepage, the loading, and the vibration according to the soil arch shape evolution data.

2. The prediction method for the soil arching effect between embankment piles under the coupling action according to claim 1, characterized in that The establishment of the multi-dimensional data monitoring array includes: The data monitoring device includes a pore water pressure gauge, a soil pressure sensor, a displacement sensor, and an image acquisition device to establish the multi-dimensional data monitoring array.

3. The prediction method of the soil arching effect between embankment piles under the coupling action according to claim 1, characterized in that The step of regulating the hydraulic gradient of the model box through an adjustable water pressure pump and obtaining the pore water pressure distribution by using the multi-dimensional data monitoring array includes: Use the adjustable water pressure pump and combine it with a precision flow control valve to regulate the water delivery flow of the water tanks on both sides of the model box to achieve the regulation of the hydraulic gradient; Based on the regulation of the hydraulic gradient, make the test soil in the model box reach a stable seepage state; Under the stable seepage state, obtain the pore water pressure distribution of the test soil based on the multi-dimensional data monitoring array.

4. The method for predicting the soil arching effect between embankment piles under the coupling action according to claim 1, wherein The step of applying a static load to the model box using a hydraulic servo loader and obtaining the stress distribution of the test soil according to the multi-dimensional data monitoring array includes: Based on the hydraulic servo loader, apply a graded vertical static load to the model box through hydraulic regulation as the static load; Control the hydraulic servo loader to maintain the graded vertical static load until the test soil reaches a deformation stable state; Under the deformation stable state, obtain the stress distribution of the test soil by using the multi-dimensional data monitoring array.

5. The prediction method of the soil arching effect between the embankment piles under the coupling action according to claim 1, characterized in that The step of vibrating the model box based on a hydraulic vibrator and obtaining vibration displacement information according to the multi-dimensional data monitoring array includes: A vibration tabletop is established, and the model box is located above the vibration tabletop; The hydraulic vibrator is used to apply a vibration load to the vibration tabletop to vibrate the vibration tabletop. The vibration load includes a horizontal vibration load and a vertical vibration load; Based on the vibration of the vibration tabletop, the vibration of the model box is realized, and the vibration displacement information of the model box is obtained according to the multi-dimensional data monitoring array.

6. The method for predicting the soil arching effect between the embankment piles under the coupling action according to claim 1, characterized in that The quantitative correlation model satisfies the following relationship: ； Among them, is the soil arch height, is the least squares fitting coefficient, is the seepage force parameter, is the load parameter, is the vibration parameter.

7. The prediction method of the soil arching effect between the embankment piles under the coupling action according to claim 1, wherein, The soil arch effect prediction model is constructed by combining the soil arch shape evolution data with a machine learning algorithm to realize the prediction of the soil arch effect, including: Introducing the random forest algorithm as the machine learning algorithm to construct the soil arch effect prediction model; Dividing the soil arch shape evolution data into a training set, a validation set and a test set, and training, tuning and evaluating the soil arch effect prediction model; Under the multi-working conditions, the evolution law of the soil arch effect is predicted according to the soil arch effect prediction model.

8. The prediction method of the soil arching effect between the embankment piles under the coupling action according to claim 7, characterized in that The soil arch effect prediction model satisfies the following relationship: ； Among them, is the prediction result of the soil arching effect, is the number of decision trees, is the traversal count flag, is the prediction result of a single decision tree, is the seepage velocity, is the pressure of the static load, is the vibration frequency, is the amplitude, are the test soil parameters.

9. A prediction system for the soil arching effect between piles in an embankment under coupling action, wherein the system uses the prediction method for the soil arching effect between piles in an embankment under coupling action according to any one of claims 1-8, characterized in that, Including: The main body of the model box is used for layered landfilling of the test soil; The seepage module includes the adjustable water pressure water pump, which is used to regulate the hydraulic gradient of the model box and conduct seepage on the test soil; The loading module includes a hydraulic servo loader, which is used to apply a static load to the model box to realize the loading of the test soil; The vibration module includes a hydraulic vibrator, which is used to vibrate the model box to complete the vibration of the test soil; The movable door module includes a movable block, which is used to trigger the soil arch phenomenon; The data acquisition module includes a multi-dimensional data monitoring array, which is used to collect multi-dimensional data under the multi-working conditions; The data processing module is used to construct the quantitative correlation model and the soil arch effect prediction model to realize the prediction of the soil arch effect.

Images