Water-rich sand layer permeation grouting model test modeling analysis method, device and system

Abstract

The invention provides a water-rich sand layer permeation grouting model test modeling analysis method, device and system, and belongs to the technical field of grouting model tests.The method comprises the steps that a water-rich sand layer permeation grouting visual simulation device is built; obtaining three-dimensional structure information in the simulation device in the water-rich sand layer permeation grouting test process, and generating time sequence voxel data; extracting features in the voxel data by using a convolutional neural network, and reconstructing a three-dimensional model for the voxel data of each time sequence according to the extracted features; and based on the three-dimensional model of each time sequence, performing quantitative analysis on the slurry diffusion process, and calculating test parameters.

Description

Technical Field

[0001] The present invention belongs to the technical field of grouting model tests, and particularly relates to a modeling analysis method, device and system for permeability grouting model tests in water-rich sand layers. Background Art

[0002] The statements in this part only provide background technical information related to the present invention and do not necessarily constitute prior art.

[0003] In the field of engineering construction, water-rich sand layers bring great challenges to construction due to their complex characteristics. Water-rich sand layers usually have poor cementation ability, low strength and strong water permeability. The characteristics of water-rich sand layers vary greatly in different regions, which easily cause geological disasters such as water inrush, sand gushing, quicksand, surrounding rock instability, tunnel collapse, surface subsidence and groundwater level decline during the construction of tunnel engineering, foundation reinforcement, etc., seriously threatening engineering safety.

[0004] The development of existing grouting technology theories has deficiencies in water-rich sand layers. The diffusion mode of slurry in water-rich sand layers is affected by various factors and cannot be clearly defined at present, resulting in difficult accurate prediction of the reinforcement effect and the existing grouting theories being difficult to be directly applied to actual projects. Existing grouting simulation test devices also have limitations, with poor intuitive observation effects, mostly having the "black box problem", making it difficult to directly observe the slurry diffusion process and the movement law of the wetting front, which is not conducive to understanding the physical phenomena during the grouting process.

[0005] In view of this, engineering practice urgently needs a more efficient and accurate research method to study the grouting process in water-rich sand layers. With the continuous development of civil engineering construction, the requirements for grouting technology in water-rich sand layers are getting higher and higher. Accurately mastering the diffusion of slurry in water-rich sand layers is crucial for determining reasonable grouting parameters and ensuring the construction safety and quality of projects such as tunnels. Summary of the Invention

[0006] To overcome the deficiencies of the above-mentioned existing technologies, the present invention provides a modeling analysis method for permeability grouting model tests in water-rich sand layers, which establishes a three-dimensional model for quantitative analysis and restores the whole grouting process in the test device, providing a powerful tool for the research of grouting technology in water-rich sand layers.

[0007] To achieve the above object, one or more embodiments of the present invention provide the following technical solutions:

[0008] In the first aspect, a modeling analysis method for permeability grouting model tests in water-rich sand layers is disclosed, including:

[0009] Construct a visual simulation device for permeability grouting in water-rich sand layers;

[0010] Obtain the three-dimensional structure information inside the simulation device during the permeability grouting test in water-rich sand layers to generate time-series voxel data;

[0011] Extract features from voxel data using a convolutional neural network, and reconstruct a three-dimensional model for the voxel data of each time series according to the extracted features;

[0012] Based on the three-dimensional models of each time series, quantitatively analyze the slurry diffusion process and calculate the test parameters.

[0013] As a further technical solution, obtain the three-dimensional structural information inside the simulation device during the permeation grouting test in the water-rich sand layer, including:

[0014] Before the sand layer is filled and the slurry is not injected, perform an initial scan;

[0015] Select a suitable scanning mode, set the scanning parameters to be the same as those determined in the debugging stage, and record the initial state of the water-rich sand layer;

[0016] Start grouting while performing continuous scanning;

[0017] Set the scanning time interval, and scan the sand layer and the grouting situation in the test box once at a fixed time interval.

[0018] As a further technical solution, the process of generating time series voxel data is as follows:

[0019] Denoise the two-dimensional image data obtained by scanning;

[0020] Perform image enhancement operations on the denoised two-dimensional image data;

[0021] According to the differences in the X-ray absorption degrees of the water-rich sand layer, the slurry, and the test box, use the threshold segmentation method to segment the enhanced image;

[0022] According to the test box size and the scanning resolution, divide the three-dimensional space of the scanning area into small cubic voxels. For each voxel, assign corresponding attribute values according to its position in the image and the image segmentation result;

[0023] Store all voxel data in the form of a three-dimensional array, where the value of each element is the attribute value of the corresponding voxel.

[0024] As a further technical solution, use a convolutional neural network to extract features from voxel data, where the convolutional neural network includes:

[0025] The first convolutional layer, the second convolutional layer, the third convolutional layer, the max pooling layer, and the fully connected layer;

[0026] The first convolutional layer is used to perform preliminary feature extraction on the voxel data;

[0027] The second convolutional layer is used to extract more complex features from the voxel data;

[0028] The third convolutional layer can capture more subtle and higher-level features in the voxel data;

[0029] After each convolutional layer, a max pooling layer is connected. The max pooling layer is used to downsample the feature map output by the connected convolutional layer;

[0030] After the max pooling layer, a fully connected layer is connected. The fully connected layer is used to integrate the extracted features for the final classification or regression task.

[0031] As a further technical solution, a three-dimensional model is reconstructed for the voxel data of each time series according to the extracted features. The specific process is as follows:

[0032] The three-dimensional space is divided into small cubes, i.e., voxels. For each small cube, its intersection with the isosurface is determined according to the attribute values of its vertices;

[0033] Then, according to a predefined lookup table, the topological structure of the isosurface on this small cube is determined;

[0034] Finally, the coordinates of the points on the isosurface are calculated by interpolation, and the isosurfaces on all small cubes are spliced together to obtain the three-dimensional model;

[0035] This model can intuitively display the diffusion process of the slurry in the water-rich sand layer for subsequent visualization and quantitative analysis.

[0036] As a further technical solution, test parameters are calculated, including calculating the diffusion volume, coverage area, and diffusion speed parameters of the slurry.

[0037] In a second aspect, a modeling analysis system for water-rich sand layer permeation grouting model tests is disclosed, including:

[0038] A time-series voxel data generation module, configured to: obtain the three-dimensional structure information inside the simulation device during the water-rich sand layer permeation grouting test and generate time-series voxel data;

[0039] A three-dimensional model reconstruction module, configured to: use a convolutional neural network to extract features from the voxel data and reconstruct a three-dimensional model for the voxel data of each time series according to the extracted features;

[0040] A quantitative analysis module, configured to: based on the three-dimensional model of each time series, perform quantitative analysis on the slurry diffusion process and calculate test parameters.

[0041] In a third aspect, a modeling analysis device for water-rich sand layer permeation grouting model tests is disclosed, including: a water-rich sand layer grouting simulation test device and a computer;

[0042] The grouting simulation test device for water-rich sand layer includes: a grouting system, a water pressure control system, a visual water-rich sand layer simulation system and a data acquisition system;

[0043] The grouting system is directly connected to the water-rich sand layer simulation grouting device to meet different grouting test requirements;

[0044] The water pressure control system is directly connected to the water-rich sand layer simulation grouting device to provide pore water pressure for it;

[0045] The data acquisition system is used to collect the structural information inside the water-rich sand layer during the grouting process;

[0046] The data acquisition system communicates with a computer, and the computer is configured to:

[0047] Receive the three-dimensional structural information inside the water-rich sand layer during the grouting process collected by the data acquisition system and generate time-series voxel data;

[0048] Use a convolutional neural network to extract the features in the voxel data, and reconstruct a three-dimensional model for the voxel data of each time series according to the extracted features;

[0049] Based on the three-dimensional models of each time series, quantitatively analyze the slurry diffusion process and calculate the test parameters.

[0050] The above device can accurately simulate the actual working conditions, precisely control the test parameters, and ensure the reliability and accuracy of the test research.

[0051] The above one or more technical solutions have the following beneficial effects:

[0052] The technical solution of the present invention uses an X-ray scanner to scan and record the grouting simulation test device for the water-rich sand layer, and can collect the three-dimensional structural information inside the water-rich sand layer during the grouting process, ensuring a comprehensive monitoring of the sand layer structure, slurry distribution and the interaction between the two. Compared with the traditional local observation or indirect measurement methods, the data obtained in this way is more complete and accurate, and can reflect the dynamic changes of the micro and macro structures inside the sand layer during the grouting process.

[0053] The convolutional neural network (CNN) architecture constructed by the technical solution of the present invention can automatically extract the features in the voxel data. Compared with the traditional manual feature extraction methods, CNN has stronger adaptability and accuracy. It can identify complex slurry penetration paths, accurately determine the boundary range of slurry diffusion and other information, overcoming the subjectivity and errors that may exist in manual analysis. By learning a large amount of voxel data, CNN can discover the hidden patterns and rules in the data, provide a more reliable basis for three-dimensional model reconstruction, and thus generate a three-dimensional model that more conforms to the actual situation and more accurately reflects the diffusion state of the slurry in the sand layer.

[0054] The technical solution of the present invention combines X-ray scanning technology with convolutional neural network and applies it to the experimental study of permeation grouting model in water-rich sand layer, which is an innovative research method. This interdisciplinary technology integration breaks the limitations of traditional grouting research methods and provides new ideas and means for studying the grouting problem in water-rich sand layer. Through the complementary advantages of the two technologies, a comprehensive innovation from experimental data collection, processing to analysis is achieved, which can promote the research progress in the field of grouting technology for water-rich sand layer.

[0055] Advantages of additional aspects of the present invention will be partly given in the following description, partly will become obvious from the following description, or will be learned through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The accompanying drawings forming a part of this specification are used to provide a further understanding of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation to the present invention.

[0057] Figure 1 It is a schematic diagram of the overall process of the method of the embodiment of the present invention.

[0058] Figure 2 It is a schematic diagram of the structure of the modeling analysis device for the permeation grouting model test in water-rich sand layer.

[0059] Figure 3 It is a schematic diagram of the convolutional neural network structure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0060] It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0061] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present invention.

[0062] Without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0063] Embodiment 1

[0064] This embodiment discloses a method for modeling and analyzing the permeation grouting model test in water-rich sand layer, which is implemented based on X-ray scanning and convolutional neural network, and includes:

[0065] Step 1: Build a visualization simulation device for permeation grouting in water-rich sand layer. For the convenience of direct observation during the grouting test, the test device uses transparent materials.

[0066] Specifically, refer to the appendix Figure 2 As shown, the visualized simulation device for permeation grouting in water-rich sand layer constructed by the technical solution of this embodiment includes a grouting system, a water pressure control system, a visualized water-rich sand layer simulation system, and a data acquisition system.

[0067] In this embodiment, the grouting system is composed of a slurry storage tank, a slurry conveying pipe, a grouting pump, and a grouting pipe. The inlet of the grouting pump is connected to the slurry storage tank, and one end of the grouting pipe is connected to the outlet of the grouting pump; the entire grouting system is directly connected to the visualized water-rich sand layer simulation system through the slurry conveying pipe, which can meet the requirements of different grouting tests such as double-fluid and single-fluid. The slurry storage tank should be equipped with a well-sealed lid to prevent slurry volatilization and impurity entry. Place the slurry storage tank on a stable bracket to ensure that its height is convenient for connection with the grouting pump and that it will not affect the grouting stability due to vibration or shaking during the test. The grouting pump is a plunger grouting pump, with a flow rate adjustment range of 0.1 L / min - 10 L / min and a pressure range of 0 MPa - 5 MPa. Connect the grouting pump to the outlet of the slurry storage tank through a corrosion-resistant high-pressure rubber hose. Install a valve on the slurry conveying pipe to control the grouting flow rate and start / stop the grouting pump. The slurry conveying pipe is used to convey the slurry from the slurry storage tank to the grouting pump. The grouting pipe is a high-strength rubber grouting pipe with an inner diameter of 15 mm, and its length is determined according to the size of the test box, generally 1 m - 3 m. Connect one end of the grouting pipe tightly to the outlet of the grouting pump, and insert the other end into the grouting test box of the visualized water-rich sand layer simulation system through the grouting hole reserved in the test box. The insertion depth can be determined according to the test design, such as 1 / 2 or 2 / 3 of the depth of insertion into the sand layer.

[0068] In this embodiment, the water pressure control system is composed of a high-head water storage tank, a water diversion pipe, and a flow control valve, and is connected to the test box in the visualized water-rich sand layer simulation system to provide pore water pressure. The material of the high-head water storage tank should be plastic or stainless steel. An injection port and a water level monitoring device, such as a float-type water level gauge, are installed on the top of the water tank. A drain port and an interface connected to the water diversion pipe are installed at the bottom. The water storage tank should be placed on a bracket 1 m - 2 m higher than the test box, and the pore water pressure is controlled by adjusting the height of the bracket. Flow control valves are installed at both ends of the water diversion pipe connecting the high-head water storage tank and the test box.

[0069] In this embodiment, the visualized water-rich sand layer simulation system consists of a cubic transparent permeation grouting test box and an internally filled sand layer. This system is connected to the grouting system and the water pressure control system. The box body of the grouting test box is made of transparent plexiglass. A grouting hole is reserved at the bottom of one side for inserting a grouting pipe; a water pressure inlet is reserved at the bottom of the other side for connecting a water inlet pipe; an exhaust hole is reserved at the top to discharge the air in the sand layer. Reinforcing rings are used to reinforce around each hole to prevent the hole edge from cracking due to pressure or vibration during the test. The internal sand layer is made of high-quality transparent glass sand with as high transparency as possible for easy observation during the test. The particle size of the glass sand can be selected according to the test requirements. The method of layered filling is adopted, with the filling thickness of each layer being 5 cm - 10 cm. A small vibrator is used to vibrate each layer of the sand layer to make the sand layer dense and uniform. During the filling process, the density of the sand layer can be checked by inserting thin rods or probes into the sand layer to ensure that the porosity of the sand layer meets the test requirements.

[0070] In this embodiment, the data acquisition system mainly consists of an X-ray scanner. The scanned images are transmitted to the computer supporting the X-ray scanner. What is directly obtained by the X-ray scanner is a two-dimensional image. Three-dimensional structure information is generated in the computer based on the two-dimensional image, which can reflect the internal structure information of the water-rich sand layer during the grouting process. The X-ray scanner should cover the entire area of the water-rich sand layer and be able to scan the entire test box comprehensively.

[0071] Step 2: Use the X-ray scanner to scan the grouting test process of the visualized simulation device, obtain the three-dimensional structure information inside the model during the test process, reveal and record information such as the penetration change process, diffusion range, path of the grout, and the interaction with the sand layer particles, and generate time-series voxel data.

[0072] When using the X-ray scanner to scan the grouting test process of the visualized simulation device to obtain the three-dimensional structure information inside the model during the test process, the X-ray scanner should be debugged first. Appropriate parameters such as resolution, scanning range, and exposure time should be selected to make the scanned image as consistent as possible with the actual size and density distribution of the standard object. The specific steps are as follows:

[0073] Select the resolution: According to the particle size of the sand layer particles and the requirements for observing the details of the grout penetration in the test, select an appropriate scanning resolution. If studying the relationship between the microscopic pore structure of the sand layer and the grout penetration, select a resolution of 0.2 mm; if mainly focusing on the macroscopic grout diffusion range and the overall change of the sand layer, a resolution of 0.5 mm can be selected.

[0074] Determine the scanning range: Measure the actual size of the test chamber, and set the scanning range to be slightly larger than the size of the test chamber, exceeding it by 5 cm - 10 cm in each direction to ensure complete acquisition of information on the sand layer and the grouting area. Enter the scanning range parameters in the scanning software, and check whether the scanning range covers the entire test area through the pre-scanning and image preview functions.

[0075] Adjust the exposure time: Conduct pre-scanning tests, and scan the ungrouted sand layer in the test chamber using different exposure times (such as 1 s, 2 s, 3 s, etc.) to obtain multiple sets of images. Analyze the contrast and clarity of the images, and select the exposure time that makes the sand layer and the test chamber structure clearly distinguishable and has rich gray levels. At the same time, use a radiation dosimeter to measure the radiation dose at different exposure times, and try to select a lower exposure time to reduce the impact of radiation on the test personnel and equipment under the premise of meeting the image quality requirements.

[0076] The scanning of the test device using the X-ray scanner includes: Before the sand layer is filled and the slurry is not injected, start the X-ray scanner for initial scanning. Select a suitable scanning mode (such as continuous scanning or layer-by-layer scanning) in the scanning control software, and set the scanning parameters (resolution, scanning range, exposure time, etc.) to be consistent with the parameters determined in the debugging stage, and record the initial state of the water-rich sand layer. Start the grouting pump to start grouting, and at the same time start the X-ray scanner for continuous scanning, and store the scanned data in the form of two-dimensional image data. Set the scanning time interval, and scan the sand layer and the grouting situation in the test chamber at a fixed time interval. The time interval can be determined according to the fineness of the test analysis.

[0077] Convert the X-ray scan data into voxel data in the computer, and perform denoising and enhancement processing on the two-dimensional image data obtained by X-ray scanning to improve the image quality. Select a suitable method for image segmentation according to the differences in X-ray absorption of the water-rich sand layer, slurry, and test chamber.

[0078] Divide the three-dimensional space of the scanning area into small cubic voxels, and assign corresponding attribute values to each voxel according to the image segmentation results. Specifically, the image segmentation result is based on the gray threshold to judge whether this part of the image is the sand layer, the slurry, or others. The main steps include:

[0079] Image denoising: Use the median filtering algorithm to perform denoising processing on the two-dimensional image data obtained by scanning. Select a 3×3 pixel filtering window. For each pixel in the image, sort the pixel values in its surrounding 3×3 neighborhood, and take the median value as the new value of this pixel. The median filtering formula is: g(x,y) = med{f(x - i,y - j)}, (i,j ∈ [-1,1]), where g(x,y) is the pixel value of the denoised image, f(x,y) is the pixel value of the original image, and med represents the median operation.

[0080] Image enhancement: Perform image enhancement operations using histogram equalization. Calculate the grayscale histogram of the image and count the frequency of each grayscale level. Calculate the cumulative distribution function (CDF) based on the grayscale histogram, normalize the CDF to obtain the mapping function. Transform the grayscale values of the original image through the mapping function to make the grayscale distribution of the image more uniform and improve the image contrast. The calculation formula for the pixel values of the enhanced image is: g(x,y) = T(f(x,y)), where T is the mapping function, f(x,y) is the pixel value of the original image, and g(x,y) is the pixel value of the enhanced image.

[0081] Image segmentation: According to the differences in the X-ray absorption degrees of the water-rich sand layer, slurry, and test box, adopt the threshold segmentation method. By analyzing the grayscale histogram of the image, select an appropriate threshold. For example, after observation, it is found that the grayscale value of the sand layer is low, the grayscale value of the slurry is high, and the grayscale value of the test box is between the two. Select a grayscale threshold T, and determine the pixels with grayscale values less than T in the image as the sand layer, and the pixels with grayscale values greater than T as the slurry or test box part. In the segmented image, the pixel values of the sand layer part are set to 0, and the pixel values of the slurry and test box parts are set to 1.

[0082] The threshold segmentation formula is: where g(x,y) is the pixel value of the segmented image, f(x,y) is the pixel value of the original image, and T is the threshold.

[0083] Voxel data generation: Divide the three-dimensional space of the scanned area into small cubic voxels according to the test box size and scanning resolution. In each small three-dimensional space, whether it is defined as a sand layer, slurry, or test box body is determined by the result of image segmentation.

[0084] If the side length of the test box is 50 cm and the scanning resolution is 0.5 mm, then 1000 voxels can be divided in each direction. For each voxel, assign corresponding attribute values according to its position in the image and the result of image segmentation. The corresponding attribute values are the key data describing its characteristics and can indicate the material category of the voxel. Store all voxel data in the computer's memory in the form of a three-dimensional array. The dimensions of the array are (number of voxels x, number of voxels y, number of voxels z), where the value of each element is the attribute value of the corresponding voxel, the grayscale value, material category, etc. corresponding to the voxel.

[0085] Step 3: Construct a convolutional neural network (CNN) architecture suitable for voxel data, extract the features in the voxel data, identify the penetration path of the slurry in the sand layer, determine information such as the boundary range of slurry diffusion, etc. Finally, reconstruct the three-dimensional model for the voxel data of each time series based on the extracted features.

[0086] Constructing a convolutional neural network (CNN) architecture suitable for voxel data includes designing network layers, selecting activation functions, determining loss functions, and optimization algorithms, as follows:

[0087] In the first step, design the network layers. Refer to the appendix Figure 3 As shown, that is, design the convolutional layer, pooling layer, and fully connected layer of the convolutional neural network. The main steps are as follows:

[0088] (1) Design the convolutional layer

[0089] Design 3 cascaded convolutional layers. The first convolutional layer uses 32 convolutional kernels of size 3×3×3, with a stride of 1 and a padding method of'same' (i.e., keeping the image size unchanged after the convolution operation). Its function is to perform preliminary feature extraction on voxel data in three-dimensional space. The convolution operation formula is: where V o (i,j,k) is the output voxel value, W(m,n,p) is the convolutional kernel weight, V i (i,j,k) is the input voxel value, b is the bias term, and i,j,k are the coordinates of the voxel in three-dimensional space. m, n, p are used to locate the elements in the convolutional kernel. m can represent the index of the element in the x dimension of the convolutional kernel, and its value range depends on the size of the convolutional kernel in this dimension; n is the y dimension index, and p is the z dimension index, jointly determining the coordinates of each element in the convolutional kernel.

[0090] The second convolutional layer uses 64 convolutional kernels of 3×3×3, with a stride of 1 and a padding method of'same'. This layer further extracts more complex features and enhances the network's ability to understand data features.

[0091] The third convolutional layer uses 128 convolutional kernels of 3×3×3, with a stride of 1 and a padding method of'same'. By increasing the number of convolutional kernels, the network can capture more subtle and advanced features in voxel data, providing richer information for subsequent classification and regression tasks. After each convolutional layer, a ReLU activation function is connected to introduce non-linearity and enhance the network's expressive ability. The ReLU activation function formula is: f(x) = max(0, x), where x is the input value and f(x) is the output value.

[0092] (2) Configure the pooling layer

[0093] Connect a max-pooling layer after each convolutional layer. The pooling window size is 2×2×2, and the stride is 2. The role of the max-pooling layer is to downsample the feature map output by the convolutional layer, reduce the data volume, and at the same time retain the main feature information. The max-pooling operation formula is: V p (i,j,k) = max{V c(2i + m, 2j + n, 2k + p)}, (m, n, p ∈ [0, 1]), where V p (i, j, k) is the output voxel value after pooling, V c (i, j, k) is the output voxel value of the convolutional layer, and i, j, k are the coordinates of the voxel in the three-dimensional space.

[0094] (3) Add fully connected layers

[0095] Connect two cascaded fully connected layers after the pooling layer. The first fully connected layer has 256 neurons, and the second fully connected layer has 128 neurons. The fully connected layer integrates the features extracted by the previous network layers for the final classification or regression task. The calculation method of the fully connected layer is: y = Wx + b, where y is the output vector, W is the weight matrix, x is the input vector, and b is the bias vector.

[0096] In the second step, select the activation function. Select the ReLU activation function in both the convolutional layer and the fully connected layer to introduce non-linearity and enhance the expression ability of the network. The ReLU function outputs the same as the input when the input is greater than 0 and outputs 0 when the input is less than 0. The ReLU activation function is simple to calculate, has a fast convergence speed, can effectively alleviate the vanishing gradient problem, enabling the network to learn the features in the data faster. And it performs better in deep networks, improving the training efficiency and performance of the network.

[0097] In the third step, determine the loss function and the optimization algorithm. Select a suitable loss function according to the task objective (such as the accuracy of identifying the slurry penetration path, the error of predicting the diffusion boundary range, etc.). Such as the cross-entropy loss function, mean squared error loss function, etc. At the same time, select a suitable optimization algorithm, such as Stochastic Gradient Descent (SGD), Adagrad, Adadelta, Adam, etc., to optimize and train the network parameters to minimize the loss function and improve the network performance. Specifically as follows:

[0098] (1) Select the loss function:

[0099] For the task of identifying the slurry penetration path (a binary classification problem, such as determining whether a voxel belongs to the penetration path), select the binary cross-entropy loss function. The formula is: where L is the loss value, y is the true label (0 or 1), is the predicted probability. This loss function can measure the difference between the prediction result and the true label, and optimize the network parameters by minimizing the loss function to improve the prediction accuracy.

[0100] For the task of determining the slurry diffusion boundary range (a regression problem), select the mean squared error loss function. The formula is: where L is the loss value, y iis the true value, is the predicted value, and is the number of samples. The mean squared error loss function can measure the average error between the predicted value and the true value, enabling the network to learn a more accurate prediction model.

[0101] (2) Select an optimization algorithm

[0102] Select the Adam optimization algorithm to optimize and train the network parameters. The Adam algorithm combines the advantages of the Adagrad and RMSProp algorithms, can adaptively adjust the learning rate, use different learning rates for different parameters, and converge to a better solution faster during training. The update formula of the Adam algorithm is:

[0103] m t = β 1 m t-1 +(1 - β 1 )g t

[0104] v t = β 2 v t-1 +(1 - β 2 )g t 2

[0105]

[0106] where m t and v t are the first-order moment estimate and the second-order moment estimate respectively, β 1 and β 2 are the decay rates (usually β 1 = 0.9, β 2 = 0.999), g t is the gradient, and are the corrected first-order moment estimate and the second-order moment estimate, θ t is the parameter, α is the learning rate (initially can be set to 0.001), l is a small constant to prevent division by zero (usually the value), and is the number of iterations.

[0107] Extract the features in the voxel data, identify the penetration path of the slurry in the sand layer, determine the boundary range of the slurry diffusion, etc. Among them, according to the characteristics of the voxel data, the material type of each small three-dimensional space can be identified. Some are slurry and some are sand layers. Based on this, the diffusion boundary can be determined. According to the changes in the voxel data in time series, the penetration path can be determined.

[0108] Finally, reconstruct the three-dimensional model for the voxel data of each time series according to the extracted features. The specific steps include:

[0109] (1) Data preprocessing: Normalize the generated voxel data so that the data is within a specific range (e.g., 0 - 1) to improve the efficiency and stability of network training. A simple linear normalization method is used, and the formula is: where V n (i,j,k) is the normalized voxel value, V(i,j,k) is the original voxel value, V min and V max are the minimum and maximum values in the voxel data respectively. According to the data volume and actual situation, divide the preprocessed voxel data into a training set, a validation set, and a test set, and the ratio can be set as 70%:15%:15%.

[0110] (2) Network training: Input the training set into the constructed convolutional neural network for multiple iterative trainings. During the training process, adjust the network parameters (such as the learning rate, the number of convolutional kernels, etc.) according to the loss value and accuracy rate and other indicators of the validation set to prevent overfitting. The training process continues until the performance indicators of the network on the validation set reach stability or meet the predetermined stopping conditions (such as reaching the maximum number of training epochs of 100 or the loss value of the validation set no longer decreases).

[0111] (3) Feature extraction and model reconstruction

[0112] After training is completed, input the voxel data collected from the test set or actual experiment into the trained convolutional neural network, and extract the features in the voxel data through the forward propagation process of the network. These features can be connected regions representing the slurry penetration path, contour features of the diffusion boundary, etc.

[0113] Utilize the extracted features to carry out the reconstruction of the three-dimensional model for the voxel data of each time series through appropriate algorithms (such as voxel-based modeling algorithms, Marching Cubes algorithms, etc.), and generate a three-dimensional model that can intuitively display the diffusion process of the slurry in the water-rich sand layer for subsequent visualization and quantitative analysis.

[0114] Taking the Marching Cubes algorithm as an example, the basic idea of this algorithm is to construct an isosurface on the boundary of the voxel by analyzing the attribute values of each voxel in the voxel data (in this case, the voxel related to the slurry determined after classification or regression). The specific steps are as follows:

[0115] First, divide the three-dimensional space into small cubes (i.e., voxels). For each small cube, determine its intersection with the isosurface (here, a threshold isosurface representing the slurry boundary can be set) according to the attribute values of its vertices (0 or 1, indicating whether it belongs to the slurry region).

[0116] Then, according to the pre-defined lookup table, determine the topological structure of the isosurface on this small cube (i.e., which edges form the isosurface). The lookup table is pre-computed based on all possible combinations of vertex attribute values (a total of 256 cases).

[0117] Finally, calculate the coordinates of the points on the isosurface through interpolation, and splice the isosurfaces on all small cubes to obtain the three-dimensional model. For example, for a small cube with vertex attribute values of (0, 0, 0, 0, 1, 1, 1, 1) (representing the 8 vertices of the small cube in a certain order), according to the lookup table, its isosurface is composed of three edges. Calculate the coordinates of the isosurface points on these three edges through linear interpolation, and then connect these isosurface points to form the isosurface part on this small cube. Repeat this process for all small cubes to finally obtain the entire three-dimensional model, which can intuitively display the diffusion process of the slurry in the water-rich sand layer for subsequent visualization and quantitative analysis.

[0118] Step 4: Based on the three-dimensional model for each time series, conduct a quantitative analysis of the slurry diffusion process, and calculate parameters such as the diffusion volume, coverage area, and diffusion speed of the slurry.

[0119] Based on the established three-dimensional model of the time series, the above parameters can be calculated to conduct an analysis and research on the slurry diffusion. These data can be accurate and are used for quantitative comparative analysis of the grouting effects under different grouting parameters, providing a basis for the optimization of grouting parameters.

[0120] Calculate the diffusion volume. Count the number of voxels in the slurry diffusion area and multiply it by the volume of each voxel to obtain the diffusion volume of the slurry. For the three-dimensional models of different time series, calculate their slurry diffusion volumes respectively to analyze the variation law of the slurry diffusion volume with time. The specific calculation process is as follows:

[0121] (1) Basis for voxel counting and volume calculation: First, calculate the volume of a single voxel according to the voxel size determined when generating the voxel data before. Assume the scanning resolution is r (unit: cm), then the volume V of a single voxel v = r 3 (unit: cm 3 ).

[0122] (2) Determine the voxels in the slurry diffusion area: In the three-dimensional model, determine the voxels belonging to the slurry diffusion area according to the attribute values of the voxels. As in the previous steps, the attribute values of some voxels of the slurry and the test box are set. Here, it is necessary to further distinguish the slurry part. This can be achieved by analyzing features such as the connectivity of the voxels. For example, starting from the voxels known to be slurry near the grouting inlet, using the region growing algorithm, mark the adjacent voxels with the attribute value as slurry voxels until there are no new eligible voxels. The basic principle of the region growing algorithm is to start from one or more seed points and continuously add adjacent pixels or voxels to the growing area according to a certain similarity criterion (here, the voxel attribute value is and adjacent).

[0123] (3) Calculate the diffusion volume: Count the number of voxels marked as slurry, N. Then the diffusion volume V of the slurry = N × V v . For the three-dimensional models at different time sequences, repeat the above steps to calculate their slurry diffusion volumes respectively, so as to obtain the variation law of the slurry diffusion volume with time. For example, at time t 1 , the calculated slurry diffusion volume is V 1 , at time t 2 , the calculated V 2 . It is possible to analyze the change of V 2 - V 1 to understand the diffusion of the slurry during this period.

[0124] Calculate the coverage area. For the selected calculation cross-section (such as the sand layer surface or a specific vertical cross-section), calculate the coverage area according to the distribution of the slurry on this cross-section. It can be achieved by identifying the boundaries of the slurry voxels on the cross-section and extracting the contours, and then using polygon area calculation methods (such as Green's formula, integral method, etc.) to calculate the coverage area. For models at different time sequences, calculate the coverage area separately as well, and observe its change trend over time to evaluate the lateral diffusion range of the slurry in the sand layer. The specific calculation process is as follows:

[0125] (1) Select the calculation cross-section: According to the test requirements and research focus, select a suitable calculation cross-section. For example, the sand layer surface cross-section can reflect the lateral diffusion range of the slurry at the top of the sand layer, and a specific vertical cross-section (such as a cross-section parallel to the grouting direction) can observe the diffusion of the slurry at different depths. For the sand layer surface cross-section, its plane coordinates can be determined according to the size and coordinate system of the test box.

[0126] (2) Identification and contour extraction of slurry liquid element boundaries: On the selected cross-section, traverse the voxel data to find the voxel boundaries belonging to the slurry. The boundary voxels can be determined by judging the change in the attribute values of adjacent voxels. For example, when the attribute value of a voxel is (slurry) and the attribute value of its adjacent voxel is (sand layer), this voxel may be a boundary voxel. For these boundary voxels, a contour extraction algorithm (such as the boundary tracking algorithm) is used to obtain the contour of the slurry region. The basic idea of the boundary tracking algorithm is to start from a boundary point and sequentially find adjacent boundary points along the boundary until returning to the starting point, thus forming a closed contour.

[0127] (3) Calculate the coverage area: According to the extracted contour, use the polygon area calculation method to calculate the coverage area. For a simple polygon (such as the contour is usually a simple polygon), Green's formula can be used to calculate the area. Assume that the vertex coordinates of the polygon are (x i , y i )(i = 1, 2, …, n, where n is the number of vertices), then the area (where x n+1 = x 1 , y n+1 = y 1 ). For models at different time sequences, repeat the above process to calculate the coverage area at each moment and observe its change trend over time to evaluate the lateral diffusion range of the slurry in the sand layer.

[0128] Diffusion speed calculation: According to the change in the diffusion volume or coverage area of the slurry at different time sequences, combined with the time interval between two adjacent scans, calculate the diffusion speed of the slurry. The diffusion speed can be calculated from two perspectives: volume change and area change, reflecting the diffusion dynamic characteristics of the slurry in the water-rich sand layer from different aspects. By analyzing the diffusion speed at different stages, understand the fast and slow change process of slurry diffusion, and provide a basis for the design and optimization of grouting projects, specifically including:

[0129] (1) Diffusion speed calculation based on volume change: Select two adjacent time sequences t n-1 and t n , calculate the change in the diffusion volume of the slurry between these two moments ΔV = V n - V n-1 , where V n and V n-1 are the diffusion volumes of the slurry at t n-1 and t n respectively. Given that the time interval between two adjacent scans is Δt, the diffusion speed based on volume change

[0130] (2) Diffusion speed calculation based on area change: Similarly, select two adjacent time sequences t n-1 and tn , calculate the change in the slurry coverage area ΔA = A n - A n-1 , where A n and A n-1 are the slurry coverage areas at times t n and t n-1 respectively. Then, analyze the change in the diffusion rate based on the area change

[0131] : Analyze the change in the diffusion rate

[0132] By calculating the diffusion rate at different stages (whether based on volume or area), the process of the slurry diffusion speed change can be understood. For example, if the diffusion rate is fast at the initial stage of grouting and gradually slows down over time, it may mean that the diffusion of the slurry in the sand layer is subject to gradually increasing resistance, such as the pores in the sand layer being gradually filled. This information can provide a basis for the design and optimization of the grouting project, such as adjusting the grouting pressure, grouting volume, or the proportion of grouting materials, to achieve better grouting effects. At the same time, by comparing the diffusion rates calculated based on volume and area, the diffusion characteristics of the slurry can be comprehensively understood from different perspectives. For example, when the volume diffusion rate changes little but the area diffusion rate decreases significantly, it may indicate that the slurry is diffusing deeper into the sand layer while the lateral diffusion slows down relatively.

[0133] The technical solution of this embodiment uses plexiglass and transparent glass sand to construct a transparent test device, simulates the water-rich sand layer environment, conducts grouting tests, uses an X-ray scanner to capture the subtle changes in the slurry and generate time-series voxel data, processes the voxel data through a convolutional neural network, identifies the slurry penetration path, determines the diffusion boundary range, and establishes a three-dimensional model for quantitative analysis to restore the entire grouting process in the test device, providing a powerful tool for the research of water-rich sand layer grouting technology.

[0134] In a more specific embodiment, a method for modeling and analyzing the permeation grouting model test in a water-rich sand layer is implemented. The specific process is as follows:

[0135] (1) In a water-rich sand layer reinforcement project of a certain underground project, the method of the present invention is used for modeling and analyzing the permeation grouting model test.

[0136] First, build a visual simulation device for permeation grouting in water-rich sand layers. Select high-strength transparent acrylic material to make a cube transparent permeation grouting test box, and fill the box with a sand layer that simulates the characteristics of the water-rich sand layer in actual projects. The grouting system adopts a double-fluid grouting mode, and prepare a slurry storage tank, a high-performance grouting pump and grouting pipes to ensure that the grouting system is tightly connected to the test box. The water pressure control system consists of a large-capacity high-head water storage tank and a water diversion pipe. After debugging according to the pore water pressure requirements of actual projects, it is connected to the test box to provide a stable pore water pressure for the water-rich sand layer simulation grouting device. The data acquisition system uses an advanced X-ray scanner, and prepare a data storage device and processing software.

[0137] After starting the grouting test, first debug the X-ray scanner. According to the size of the test box and the characteristics of the sand layer, select a suitable resolution of 0.5 mm / pixel, a scanning range covering the entire test box, and an exposure time of 2 seconds, so that the scanned image highly coincides with the actual size and density distribution of the standard object. When injecting slurry into the test box body, use the X-ray scanner to continuously scan, and store the scanned data in the computer in the form of two-dimensional image data. Subsequently, perform denoising and enhancement processing on the two-dimensional image data, using median filtering denoising and histogram equalization enhancement methods to improve the image quality. According to the differences in X-ray absorption of the water-rich sand layer, slurry and test box, use the threshold segmentation method for image segmentation. Divide the three-dimensional space of the scanned area into small cubic voxels, each voxel with a side length of 0.5 mm, and assign corresponding attribute values to each voxel according to the image segmentation results, such as sand layer, slurry or test box wall, etc.

[0138] Then construct a convolutional neural network architecture suitable for voxel data. In terms of network layer design, design multiple convolutional layers, with a convolutional kernel size of 3×3 for each layer, configure a max pooling layer, and add a fully connected layer at the end of the network. Select the ReLU function as the activation function to introduce non-linearity factors in the convolutional layer and the fully connected layer. Select the mean squared error loss function as the loss function, and use the Adam algorithm as the optimization algorithm to optimize and train the network parameters. In the feature extraction and three-dimensional model reconstruction stage, linearly normalize the generated voxel data so that its range is between 0 and 1. Divide the voxel data into a 70% training set, a 15% validation set, and a 15% test set according to the data volume. Input the training set into the constructed convolutional neural network for 500 iterations of training. After training is completed, input the test set into the trained convolutional neural network, extract the features in the voxel data, and use the voxel-based modeling algorithm to reconstruct the three-dimensional model for the voxel data of each time series, generating a three-dimensional model that intuitively shows the diffusion process of the slurry in the water-rich sand layer.

[0139] Finally, quantitative analysis is performed based on the three-dimensional models at each time series. The number of voxels within the slurry diffusion region is counted and multiplied by the volume of each voxel to obtain the diffusion volume of the slurry. For the calculation cross-section on the sand layer surface, the boundaries of the slurry voxels on the cross-section are identified and the contours are extracted, and the Green's formula is used to calculate the covered area. According to the changes in the slurry diffusion volume and covered area at different time series, combined with the time interval between two adjacent scans, the diffusion velocity of the slurry is calculated from two perspectives of volume change and area change, providing an accurate basis for the design and optimization of the grouting project.

[0140] (2) In the foundation reinforcement project of a certain water conservancy project, the method of the present invention is used for the seepage grouting model test of a water-rich sand layer.

[0141] When building the visualization simulation device, a cube-shaped transparent seepage grouting test box is made of special transparent tempered glass to ensure that it can withstand high water pressure and grouting pressure. The grouting system selects the single-fluid grouting mode, and a large-capacity slurry storage tank, a high-pressure grouting pump, and durable grouting pipes are prepared and connected to the test box. The water pressure control system consists of a large high-head water storage tank and a strong water diversion pipe, which is adjusted according to the actual pore water pressure parameters of the project and then connected to the test box. The data acquisition system uses a high-precision X-ray scanner and is equipped with a professional data processing workstation.

[0142] During the grouting test, the X-ray scanner is debugged, and the resolution is selected to be 0.4 mm / pixel, the scanning range completely covers the test box, and the exposure time is 1.5 seconds. After starting to inject the slurry, the X-ray scanner continuously scans, and the data is stored in the form of two-dimensional images. The two-dimensional image data is processed by Gaussian filtering for denoising and adaptive histogram equalization for enhancement. According to the X-ray absorption differences of the water-rich sand layer, the slurry, and the test box, the region growing method is used for image segmentation. The three-dimensional space is divided into voxels with a side length of 0.4 mm, and corresponding attribute values are assigned to the voxels.

[0143] When constructing the convolutional neural network architecture, a deep convolutional layer is designed with a convolutional kernel size of 5×5, an average pooling layer is configured, and multiple fully connected layers are added. The activation function selects the LeakyReLU function. The loss function selects the cross-entropy loss function, and the optimization algorithm uses the Adagrad algorithm to optimize and train the network parameters. In the feature extraction and three-dimensional model reconstruction section, the voxel data is normalized to the range between -1 and 1. The voxel data is divided into a 60% training set, a 20% validation set, and a 20% test set. The training set is input into the network for 800 iterations of training. After the training is completed, the test set is input into the network, features are extracted, and the Marching Cubes algorithm is used to reconstruct the three-dimensional model of the voxel data at each time series.

[0144] Quantitative analysis is carried out based on the reconstructed three-dimensional model. The number of voxels in the slurry diffusion area is counted, multiplied by the volume of each voxel, and the slurry diffusion volume is calculated. For a specific vertical section, the voxel boundary of the slurry is identified by an edge detection algorithm and the contour is extracted, and the covered area is calculated by the integral method. According to the changes in the diffusion volume and covered area at different time sequences, combined with the time interval between two adjacent scans, the slurry diffusion speed is calculated, providing a scientific decision-making basis for the foundation reinforcement of the water conservancy project.

[0145] Example Two

[0146] The purpose of this embodiment is to provide a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the above method are implemented.

[0147] Example Three

[0148] The purpose of this embodiment is to provide a computer-readable storage medium.

[0149] A computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of the above method are executed.

[0150] Example Four

[0151] The purpose of this embodiment is to provide a modeling analysis system for the permeation grouting model test of a water-rich sand layer, including:

[0152] A time-sequence voxel data generation module, configured to: obtain the three-dimensional structure information inside the simulation device during the permeation grouting test of the water-rich sand layer, and generate time-sequence voxel data;

[0153] A three-dimensional model reconstruction module, configured to: use a convolutional neural network to extract features from the voxel data, and reconstruct a three-dimensional model for the voxel data of each time sequence according to the extracted features;

[0154] A quantitative analysis module, configured to: based on the three-dimensional model of each time sequence, quantitatively analyze the slurry diffusion process and calculate the test parameters.

[0155] Example Five

[0156] The purpose of this embodiment is to provide a computer program product containing instructions, which when running on a computer, enables the computer to execute the methods and functions involved in any one of the above embodiments.

[0157] Example Six

[0158] The purpose of this embodiment is to provide a modeling analysis device for the permeation grouting model test of a water-rich sand layer, including: a water-rich sand layer grouting simulation test device and a computer;

[0159] The grouting simulation test device for water-rich sand layer includes: a grouting system, a water pressure control system, a visual water-rich sand layer simulation system, and a data acquisition system;

[0160] The grouting system is directly connected to the water-rich sand layer simulation grouting device to meet different grouting test requirements;

[0161] The water pressure control system is directly connected to the water-rich sand layer simulation grouting device to provide pore water pressure for it;

[0162] The data acquisition system is used to collect the internal structural information of the water-rich sand layer during the grouting process;

[0163] The data acquisition system communicates with a computer, and the computer is configured to:

[0164] Receive the three-dimensional structural information inside the water-rich sand layer during the grouting process collected by the data acquisition system, and generate time-series voxel data;

[0165] Use a convolutional neural network to extract features from the voxel data, and reconstruct a three-dimensional model for the voxel data of each time series according to the extracted features;

[0166] Based on the three-dimensional models of each time series, quantitatively analyze the slurry diffusion process and calculate the test parameters.

[0167] The computer can choose a high-configuration computing implementation. Among them, the processor can be selected from Intel Core i9 or AMD Ryzen9 series multi-core models, such as i9-13900K, to provide strong computing power. The memory starts from 32GB to handle large data sets and complex tasks to ensure the smoothness of the system. The hard disk uses 1TB or more NVMe SSD to accelerate data access with high read and write speeds. The graphics card uses NVIDIA GeForce RTX 40 series or AMD Radeon RX 7000 series high-performance products to assist in accelerating specific algorithms. The operating system can be selected from Windows 11 or Linux distributions to ensure stable compatibility. Equipped with a high-resolution large-screen monitor, starting from 2560×1440 for accurate analysis. An efficient heat dissipation system maintains stable operation under high load, the network interface reaches Gigabit Ethernet or higher to ensure fast data transmission, and a high-power stable power supply is adapted, taking into account compatibility, expandability, and chassis space, and flexibly adjusting and optimizing the configuration according to the actual situation.

[0168] The steps involved in the device of the above embodiments correspond to those of Method Embodiment 1. For the specific implementation manners, reference can be made to the relevant description parts of Embodiment 1. The term "computer-readable storage medium" should be understood to include a single medium or multiple media including one or more instruction sets; it should also be understood to include any medium that can store, encode, or carry an instruction set for execution by a processor and cause the processor to execute any method in the present invention.

[0169] Those skilled in the art should understand that the various modules or steps of the present invention described above can be implemented by a general-purpose computer device. Optionally, they can be implemented by program codes executable by a computing device, so that they can be stored in a storage device and executed by the computing device, or they can be separately fabricated into individual integrated circuit modules, or multiple modules or steps among them can be fabricated into a single integrated circuit module for implementation. The present invention is not limited to any specific combination of hardware and software.

[0170] Although the specific implementation manners of the present invention have been described above in conjunction with the accompanying drawings, they do not limit the protection scope of the present invention. Those skilled in the art should understand that, based on the technical solutions of the present invention, various modifications or deformations that can be made by those skilled in the art without creative efforts are still within the protection scope of the present invention.

Claims

1. Modeling and analysis method for water-rich sand layer permeation grouting model test, characterized by: include: Construct a visual simulation device for water-rich sand layer penetration grouting; Obtain the 3D structural information inside the simulation device during the water-rich sand layer permeation grouting test and generate time-series voxel data; A convolutional neural network is used to extract features from voxel data, and a three-dimensional model is reconstructed for each time series of voxel data based on the extracted features; Based on the three-dimensional model of each time series, the slurry diffusion process is quantitatively analyzed and the experimental parameters are calculated.

2. The water-rich sand layer infiltration grouting model test modeling and analysis method according to claim 1 is characterized in that: Obtain 3D structural information inside the simulation device during the water-rich sand layer penetration grouting test, including: An initial scan is performed after the sand filling is completed and before slurry is injected; Select the appropriate scanning mode, set the scanning parameters consistent with those determined during the commissioning phase, and record the initial state of the water-rich sand layer; Start grouting while performing continuous scanning; Set the scanning time interval and scan the sand layer and grouting conditions in the test box at fixed time intervals.

3. The water-rich sand layer infiltration grouting model test modeling and analysis method according to claim 1 is characterized in that: The process of generating time series voxel data is: Performing denoising on the scanned two-dimensional image data; Perform image enhancement operation on the denoised two-dimensional image data; According to the difference in X-ray absorption among water-rich sand layer, slurry and test box, the enhanced image is segmented by using threshold segmentation method. According to the size of the test box and the scanning resolution, the three-dimensional space of the scanning area is divided into small cubic voxels. For each voxel, a corresponding attribute value is assigned according to its position in the image and the image segmentation result; All voxel data are stored in the form of a three-dimensional array, where the value of each element is the attribute value of the corresponding voxel.

4. The water-rich sand layer infiltration grouting model test modeling and analysis method according to claim 1, characterized in that: A convolutional neural network is used to extract features from voxel data, wherein the convolutional neural network comprises: The first convolutional layer, the second convolutional layer, the third convolutional layer, the maximum pooling layer and the fully connected layer; The first convolutional layer is used to perform preliminary feature extraction on voxel data; The second convolutional layer is used to extract more complex features from the voxel data; The third convolutional layer can capture more subtle and high-level features in the voxel data; A maximum pooling layer is connected after each convolutional layer. The maximum pooling layer is used to downsample the feature map output by the connected convolutional layer. The fully connected layer is connected after the maximum pooling layer. The fully connected layer is used to integrate the extracted features for the final classification or regression task.

5. The water-rich sand layer penetration grouting model test modeling and analysis method according to claim 1, characterized in that: The three-dimensional model is reconstructed for each time series of voxel data based on the extracted features. The specific process is as follows: Divide the three-dimensional space into small cubes, i.e. voxels. For each small cube, determine its intersection with the isosurface according to the attribute values ​​of its vertices. Then, the topological structure of the isosurface on the small cube is determined according to a predefined lookup table; Finally, the coordinates of the points on the isosurface are calculated by interpolation, and the isosurfaces on all the small cubes are stitched together to obtain a three-dimensional model. The model can intuitively display the diffusion process of slurry in water-rich sand layers, facilitating subsequent visualization and quantitative analysis.

6. Modeling and analysis system for water-rich sand layer infiltration grouting model test, characterized by: include: The time series voxel data generation module is configured to: obtain the three-dimensional structural information inside the simulation device during the water-rich sand layer penetration grouting test, and generate time series voxel data; The three-dimensional model reconstruction module is configured to: extract features from the voxel data using a convolutional neural network, and reconstruct a three-dimensional model for each time series of voxel data according to the extracted features; The quantitative analysis module is configured to: perform quantitative analysis on the slurry diffusion process based on the three-dimensional model of each time series and calculate the test parameters.

7. Modeling and analysis device for water-rich sand layer penetration grouting model test, characterized in that: include: Water-rich sand layer grouting simulation test device and computer; The water-rich sand layer grouting simulation test device includes: grouting system, water pressure control system, visual water-rich sand layer simulation system and data acquisition system; The grouting system is directly connected to the water-rich sand layer simulation grouting device to meet different grouting test requirements; The water pressure control system is directly connected to the water-rich sand layer simulation grouting device to provide pore water pressure therefor; The data acquisition system is used to collect structural information inside the water-rich sand layer during the grouting process; The data acquisition system is in communication with a computer, and the computer is configured to: Receiving the three-dimensional structural information inside the water-rich sand layer during the grouting process collected by the data acquisition system, and generating time series voxel data; A convolutional neural network is used to extract features from voxel data, and a three-dimensional model is reconstructed for each time series of voxel data based on the extracted features; Based on the three-dimensional model of each time series, the slurry diffusion process is quantitatively analyzed and the experimental parameters are calculated.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the method described in any one of claims 1 to 5 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method described in any one of claims 1 to 5 are performed.

Images