A three-dimensional modeling system and method for actual wall of plastic concrete anti-seepage wall

Abstract

The present invention relates to the technical field of three-dimensional model analysis, and in particular to a system and method for three-dimensional modeling of an actual plastic concrete cut-off wall. The method comprises the following steps: scanning the plastic concrete cut-off wall using geological radar equipment to obtain plastic concrete cut-off wall scanning data; gridding the plastic concrete cut-off wall scanning data to obtain plastic concrete cut-off wall grid data; obtaining wall structure characteristic data based on the plastic concrete cut-off wall grid data; modeling the wall structure characteristic data to obtain a plastic concrete cut-off wall material property model; obtaining a plastic concrete cut-off wall fusion model based on the plastic concrete cut-off wall grid data and the plastic concrete cut-off wall material property model; and analyzing the plastic concrete cut-off wall fusion model to obtain a plastic concrete cut-off wall simulation model. The present invention can identify stress concentration areas and improve the safety performance of plastic concrete cut-off walls.

Description

Technical Field

[0001] The present invention relates to the technical field of three-dimensional model analysis, and in particular to a system and method for three-dimensional modeling of an actual wall of a plastic concrete anti-seepage wall. Background Art

[0002] Plastic concrete cut-off walls, as important structures in water conservancy and civil engineering, are widely used in buildings such as dams, dikes, and subway tunnels to prevent water leakage and ensure the long-term safety of the structure. However, due to the material properties of plastic concrete cut-off walls and their complex mechanical environment, they are prone to stress concentration, cracks, and structural deformation under long-term loads. To ensure the safety of cut-off walls, the stress distribution and dynamic variation characteristics of plastic concrete cut-off walls under different working conditions must be considered during structural design and analysis. Traditional structural design methods mainly rely on empirical formulas and linear static analysis, which makes it difficult to accurately reflect stress concentration areas and potential failure risks under complex stress fields. Summary of the Invention

[0003] In order to solve the above technical problems, the present invention proposes a three-dimensional modeling system and method for the actual wall of a plastic concrete anti-seepage wall to solve at least one of the above technical problems.

[0004] The present application provides a method for three-dimensional modeling of an actual wall of a plastic concrete anti-seepage wall, the method comprising:

[0005] S1. Scan the plastic concrete cut-off wall using geological radar equipment to obtain scanning data of the plastic concrete cut-off wall;

[0006] S2. Gridding the scanned data of the plastic concrete cut-off wall to obtain gridded data of the plastic concrete cut-off wall; extracting wall structural features from the gridded data of the plastic concrete cut-off wall to obtain wall structural feature data; and modeling wall material properties based on the wall structural feature data to obtain a material property model of the plastic concrete cut-off wall;

[0007] S3, fusing the gridded data of the plastic concrete cut-off wall and the material property model of the plastic concrete cut-off wall to obtain a fusion model of the plastic concrete cut-off wall;

[0008] S4. Perform stress analysis on the plastic concrete cut-off wall fusion model to obtain a plastic concrete cut-off wall simulation model.

[0009] The scanning data of the plastic concrete anti-seepage wall generated in the present invention greatly improves the accuracy of wall modeling, lays a good foundation for subsequent processing, reduces modeling errors, and improves the accurate reproduction of the wall geometry. The grid processing converts the scanning data into an operational three-dimensional grid model, which can accurately describe the surface details of the wall. Material property modeling of the wall structure characteristic data can be combined with the physical properties of the material (such as density, compressive strength, impermeability, etc.), which increases the physical reality of the model, so that the material properties can be more accurately reflected in subsequent simulations and analyses, especially for multi-material hybrid structures, which can better analyze the performance differences between materials. The generated fusion model not only contains geometric information, but also combines material properties, so that subsequent stress analysis can be based on the real structure and material interaction, improving the accuracy and reliability of the simulation results. Stress analysis is performed on the fusion model to effectively simulate the stress state of the wall under actual working conditions and predict potential stress concentration areas or structural weaknesses. The simulation results can provide data support for optimized design, maintenance management and safety assessment.

[0010] Optionally, the plastic concrete cut-off wall scanning data includes first plastic concrete cut-off wall scanning data and second plastic concrete cut-off wall scanning data, and S1 includes:

[0011] Scanning the plastic concrete cut-off wall at a first angle using a geological radar device to obtain first plastic concrete cut-off wall scanning data;

[0012] The plastic concrete cut-off wall is scanned at a second angle by geological radar equipment to obtain second plastic concrete cut-off wall scanning data, wherein the angle threshold data corresponding to the first angle scan and the angle threshold data corresponding to the second angle scan are in a vertical relationship.

[0013] In the present invention, by scanning at different angles (perpendicular relationships), it is ensured that every surface of the plastic concrete anti-seepage wall can be captured by the geological radar equipment, thereby reducing the shadow, obstruction or blind area problems caused by a single angle. The scanning data of the first angle and the second angle complement each other and provide complete three-dimensional surface information of the wall. Due to the reflection characteristics of scanning at different angles, scanning at a single angle may cause certain surface features (such as cracks, textures, etc.) to be lost or blurred. By scanning at different angles with a vertical relationship, all the detailed features of the wall surface are captured from different directions, making the surface features clearer and more accurate.

[0014] Optionally, S2 includes:

[0015] Performing point cloud meshing processing on the scanning data of the plastic concrete anti-seepage wall to obtain mesh data of the plastic concrete anti-seepage wall;

[0016] Optimize the mesh data of the plastic concrete cut-off wall to obtain simplified mesh data of the plastic concrete cut-off wall;

[0017] Performing mesh smoothing processing on the simplified mesh data of the plastic concrete cut-off wall to obtain mesh smoothing data of the plastic concrete cut-off wall;

[0018] Calculate the mesh normal of the plastic concrete anti-seepage wall mesh smoothing data to obtain the wall structure characteristic data;

[0019] Initialize the material mapping model according to the wall structure characteristic data to obtain the material mapping model;

[0020] Physical material parameter data is obtained, and finite element modeling is performed on the material mapping model according to the physical material parameter data to obtain a material property model of the plastic concrete cut-off wall.

[0021] In the present invention, a large amount of three-dimensional point cloud data is converted into an ordered grid structure through gridding of point cloud data, and irregular point cloud data is organized into a structured grid. The grid optimization step reduces redundant information by reducing the number of grid faces and nodes while maintaining the geometric accuracy of the model. By simplifying the grid, the computational complexity is greatly reduced, and the subsequent simulation and calculation speed are accelerated. The simplified grid is smoothed to eliminate surface irregularities caused by simplification, making the surface flatter and smoother, which helps to improve the visual effect and physical consistency of the model and ensure the stability of the surface normal. By calculating the normals of the smoothed grid, the orientation of each grid face or vertex is accurately determined, which is crucial for subsequent lighting calculations, material mapping and physical simulation, and can help better capture the structural characteristics of the wall. The material mapping model is initialized according to the wall structural feature data, and by combining the physical material properties with the geometric model, the material properties of the wall can correspond to its geometric characteristics. Finite element modeling makes the wall model not only geometrically accurate but also physically realistic.

[0022] Optionally, S3 includes:

[0023] According to the mesh data of the plastic concrete cut-off wall and the material property model of the plastic concrete cut-off wall, mesh material parameters are matched to obtain initial fusion model data;

[0024] Perform material parameter interpolation processing on the initial fusion model data to obtain material interpolation model data;

[0025] The material interpolation model data is optimized to obtain the fusion model of plastic concrete cut-off wall.

[0026] The present invention ensures that each part of the model can reflect its true physical properties through precise material parameter matching, making the simulation analysis results more reliable and helping to accurately identify stress concentration areas, deformation conditions and material failure risks during structural analysis. Material parameter interpolation processing allows the material properties to be smoothly distributed between different nodes and regions in the grid. In particular, when the material parameters are unevenly distributed, the material properties are adjusted according to the local stress field, temperature gradient or other physical conditions to ensure that the material properties in the model transition smoothly and consistently. After the interpolation processing is completed, the material interpolation model is optimized to further improve its accuracy and computational efficiency. The optimization process involves locally correcting the material distribution, removing redundant nodes or adjusting the parameters of the transition area to ensure that the computational complexity is reduced without affecting the simulation results. The present invention ensures that the geometric model and material model of the plastic concrete cut-off wall can be closely combined and generate a physically accurate three-dimensional fusion model. Since both the geometric structure and material properties are accurately modeled, the stress and deformation conditions in actual engineering can be better simulated in subsequent structural analysis and simulation.

[0027] Optionally, S4 includes:

[0028] The stress validity analysis of the plastic concrete cut-off wall fusion model was performed to obtain the first plastic concrete cut-off wall simulation model;

[0029] Perform stress multi-scale analysis on the fusion model of plastic concrete cut-off wall to obtain the second simulation model of plastic concrete cut-off wall;

[0030] The safety factors of the first plastic concrete cut-off wall simulation model and the second plastic concrete cut-off wall simulation model are fused and screened to obtain a concrete cut-off wall simulation model.

[0031] In the present invention, a stress validity analysis is performed on the fusion model of the plastic concrete anti-seepage wall. By identifying the stress distribution of the wall under actual working conditions, the stress concentration area and its possible influence are analyzed to ensure that the stress analysis results are true and reliable, and to avoid erroneous conclusions caused by errors or calculation deviations in the preliminary simulation. Through multi-scale analysis, the stress state of the plastic concrete anti-seepage wall is analyzed in detail at different scales (such as global, local, and micro). Global stress analysis can provide overall mechanical behavior, while local and micro stress analysis can capture more subtle stress changes and local stress concentration areas. Based on the stress validity analysis and multi-scale analysis, the safety factors in each simulation model are fused and screened. The stress analysis results of different scales are combined to perform a unified safety factor evaluation to ensure that the safety factor of each area is fully verified. Through the screening process, the most unsafe areas are identified and the design is further optimized.

[0032] Optionally, the stress effectiveness analysis includes:

[0033] Obtaining plastic concrete cut-off wall constraint data;

[0034] Applying load to the plastic concrete cut-off wall fusion model according to the plastic concrete cut-off wall constraint data to obtain load application model data;

[0035] Performing static stress processing and dynamic stress processing on the load application model data to obtain static stress distribution data and dynamic stress distribution data;

[0036] Extract stress concentration areas based on static stress distribution data and dynamic stress distribution data to obtain stress concentration area data;

[0037] The plastic concrete cut-off wall fusion model is annotated according to the stress concentration area data, the static stress distribution data, and the dynamic stress distribution data to obtain a first plastic concrete cut-off wall simulation model.

[0038] The present invention obtains the constraint data of the plastic concrete cut-off wall to ensure that the model can truly reflect the constraint conditions (such as support, fixed boundary conditions, foundation conditions, etc.) in the actual project in the simulation analysis. By applying actual loads (such as gravity, wind load, water pressure, etc.) to the plastic concrete cut-off wall fusion model, load application model data is generated to ensure that the stress analysis results can truly reflect the performance of the plastic concrete cut-off wall under different load conditions. Static stress analysis (such as stress distribution under long-term fixed load) and dynamic stress analysis (such as stress changes under transient load or cyclic load) are performed on the load application model to ensure a comprehensive understanding of the stress distribution of the plastic concrete cut-off wall under static and dynamic working conditions. According to the static and dynamic stress distribution data, stress concentration areas are extracted, and areas with highly concentrated or sudden stress changes in the wall can be identified. These areas are often potential risk points for structural failure or damage. Based on the stress concentration areas, static stress distribution data and dynamic stress distribution data, the plastic concrete cut-off wall fusion model is annotated. The stress distribution is visualized by annotation, and the stress concentration areas are highlighted, so that designers can intuitively understand the stress conditions in the structure.

[0039] Optionally, the stress multi-scale analysis includes:

[0040] Initialize the stress field of the plastic concrete cut-off wall fusion model to obtain initial stress field data;

[0041] Perform stress field stratification processing according to the initial stress field data to obtain stress field stratification data;

[0042] Perform multi-scale coupling analysis based on stress field layered data to obtain multi-scale coupled stress data;

[0043] The stress model is fused according to the multi-scale coupled stress data and the plastic concrete cut-off wall fusion model to obtain a second plastic concrete cut-off wall simulation model;

[0044] The stress models include:

[0045] Extract macro-micro stress data based on multi-scale coupled stress data to obtain macro-micro stress data;

[0046] Extract stress concentration areas based on macro and micro stress data to obtain stress concentration area data;

[0047] The stress concentration area data is used to mark the plastic concrete cut-off wall fusion model and obtain a preliminary stress fusion model.

[0048] Based on the macro-micro stress data and the preliminary stress fusion model, hierarchical stress fusion is performed to obtain the second plastic concrete cut-off wall simulation model;

[0049] The macro-micro stress data includes macro-stress data and micro-stress data, and the stress concentration area extraction includes:

[0050] Perform graph neural network encoding on the macro stress data and the micro stress data to obtain macro stress characteristic map data and micro stress characteristic map data;

[0051] Performing distance calculation on the macroscopic stress characteristic map data and the microscopic stress characteristic map data to obtain macroscopic stress map characteristic clustering data and microscopic stress map characteristic clustering data respectively;

[0052] Dynamic stress threshold control is performed based on the macroscopic stress map characteristic clustering data and the microscopic stress map characteristic clustering data, and macroscopic stress threshold matrix data and microscopic stress threshold matrix data are obtained respectively;

[0053] Using the macro stress threshold matrix data to locate the macro stress hotspot of the macro stress map feature clustering data, and obtaining macro stress hotspot location map data, and using the micro stress threshold matrix data to locate the micro stress hotspot of the micro stress map feature clustering data, and obtaining micro stress hotspot location map data;

[0054] Perform multimodal stress layered mapping based on the macroscopic stress hotspot location map data and the microscopic stress hotspot location map data to obtain layered stress mapping matrix data;

[0055] The concentrated area is synthesized according to the layered stress mapping matrix data to obtain the stress concentration area data.

[0056] In the present invention, the stress field of the plastic concrete cut-off wall fusion model is initialized to generate initial stress field data. Stress field initialization can provide basic data for subsequent stress stratification and multi-scale analysis, ensuring that the initial state of the stress field is based on accurate model geometry and material properties. The initial stress field data is processed in layers, and the entire stress field is divided into different layers according to stress gradient or stress distribution characteristics, so that the overall stress field can be refined into more manageable layered data, especially for the analysis of complex structures and uneven stress fields. By performing multi-scale coupling analysis on the layered stress field, stress data of various scales (such as macroscopic and microscopic levels) are integrated. Multi-scale coupling analysis can capture stress changes at different levels from overall structural stress to local details. On the basis of multi-scale coupling analysis, the multi-scale coupling stress data is fused with the plastic concrete cut-off wall fusion model for stress model fusion to ensure that stress information of different scales is integrated in the overall model, thereby achieving more accurate global stress field simulation. Through stress field stratification, multi-scale coupling analysis and stress model fusion, the stress distribution of plastic concrete cut-off walls can be comprehensively covered, not only focusing on the macroscopic stress state, but also going deep into the local stress concentration area, comprehensively improving the accuracy and depth of stress analysis.

[0057] The present invention combines macroscopic and microscopic stress data to accurately identify stress concentration areas at different scales, ensuring a significant improvement in the accuracy of the model, being able to more comprehensively display the stress characteristics of the plastic concrete anti-seepage wall, and enhancing the accuracy of stress concentration area identification. The dynamic threshold control mechanism adaptively adjusts the identification criteria according to real-time data, allowing the model to flexibly respond to different stress distribution conditions. Through multimodal stress hierarchical mapping, the macroscopic and microscopic stress data are finely divided and displayed, and the stress data of each level are mapped into a multimodal matrix, so that the model can display in detail the distribution of stress at different spatial and temporal levels. By using the stress concentration area data to mark the plastic concrete anti-seepage wall fusion model, and then performing hierarchical stress fusion based on the model, the simulation model is closer to the actual situation. Hierarchical stress fusion provides a more realistic simulation effect on stress concentration areas of different scales. The hierarchical mapping data is integrated through concentrated area synthesis to generate a high-precision stress concentration area model. The synthesis step integrates the multi-level stress information so that the final generated model can fully display the regional distribution and characteristics of stress concentration.

[0058] Optionally, the stress field layering process includes:

[0059] Perform material property stratification based on initial stress field data to obtain stress field material stratification data;

[0060] Perform stress gradient stratification according to the initial stress field data to obtain stress field gradient stratification data;

[0061] Cross-labeling is performed based on the stress field material layering data and the stress field gradient layering data to obtain the stress field layering data.

[0062] In the present invention, material properties are layered according to the initial stress field data to ensure that the stress analysis of each layer can reflect the difference in material properties, taking into account the different mechanical properties of the material (such as elastic modulus, density, strength, etc.), so that the stress field distribution can more truly reflect the different reactions of the material. According to the initial stress field data, the stress gradient is layered, and the area where the stress gradient is larger or changes faster is identified, and the stress concentration area and the stress change slowly area are distinguished, so that subsequent stress field analysis can be focused on key positions. Cross-labeling is performed according to the material property layered data and the stress gradient layered data to generate more refined stress field layered data, combine the material property with the stress gradient, and mark out the area where the stress change is significant and the material property difference is obvious. The combination of material property layering and stress gradient layering enables the method to carry out a deeper stress field analysis on complex structures. Different materials may show different stress behaviors in the structure, and the rapid change of stress gradient may also indicate potential structural failure risks. Through cross-labeling, these complex stress distribution characteristics can be captured.

[0063] Optionally, the stress gradient layering includes:

[0064] Perform gradient calculation based on the initial stress field data to obtain stress field gradient data;

[0065] Cluster calculation is performed based on stress field gradient data to obtain stress field gradient cluster characteristic data;

[0066] The initial stress field data is partitioned into adjacent regions according to the stress field gradient clustering feature data to obtain the initial stress field gradient layered data;

[0067] Calculate the local gradient change rate based on the initial stress field gradient stratification data to obtain the gradient stratification local change rate data;

[0068] The gradient sensitive areas of the initial stress field gradient stratification data are classified according to the gradient stratification local change rate data to obtain the stress field gradient stratification data.

[0069] The stress gradient calculation in the present invention ensures that the complex changes in the stress field can be captured, so that the local stress behavior of the structure can be accurately reflected. Cluster calculation helps to more clearly show the characteristics of the wall stress distribution by grouping areas with similar stress gradients. The initial stress field data is partitioned into adjacent areas based on the stress field gradient clustering feature data, and the structure is divided into multiple adjacent areas. The local stress differences in the structure are identified and summarized into specific areas. By calculating the local gradient change rate of the initial stress field gradient stratification data, the dynamic changes of the stress gradient in each area can be captured, revealing the local change characteristics in the stress distribution, especially identifying areas where the stress gradient suddenly changes or transitions more slowly. The initial stress field gradient stratification data is classified into gradient sensitive areas based on the gradient stratification local change rate data to ensure that sensitive areas can be labeled and identified.

[0070] Optionally, the present invention further provides a three-dimensional modeling system for an actual wall of a plastic concrete cut-off wall, for executing the above-mentioned three-dimensional modeling method for an actual wall of a plastic concrete cut-off wall. The three-dimensional modeling system for an actual wall of a plastic concrete cut-off wall comprises:

[0071] The plastic concrete cut-off wall scanning module is used to scan the plastic concrete cut-off wall using geological radar equipment to obtain scanning data of the plastic concrete cut-off wall;

[0072] The plastic concrete cut-off wall material property model construction module is used to grid the plastic concrete cut-off wall scan data to obtain plastic concrete cut-off wall grid data; extract wall structure features from the plastic concrete cut-off wall grid data to obtain wall structure feature data; and perform wall material property modeling on the wall structure feature data to obtain a plastic concrete cut-off wall material property model;

[0073] The plastic concrete cut-off wall model fusion module is used to fuse the plastic concrete cut-off wall grid data and the plastic concrete cut-off wall material property model to obtain a plastic concrete cut-off wall fusion model;

[0074] The plastic concrete cut-off wall stress analysis module is used to perform stress analysis on the plastic concrete cut-off wall fusion model to obtain a plastic concrete cut-off wall simulation model.

[0075] The objects of the present invention are:

[0076] 1. This invention uses geological radar equipment to perform high-precision scanning of plastic concrete cut-off walls, capturing the wall's full range of geometric morphology and surface details. Geological radar scanning not only rapidly acquires large amounts of accurate point cloud data but also reduces the human error inherent in traditional measurement methods. Especially for plastic concrete cut-off walls with complex geometries, geological radar scanning can effectively detect the internal structure of the wall (e.g., cracks, voids, material density variations, etc.), ensuring comprehensive and high-precision data.

[0077] 2. The scanned point cloud data is meshed and converted into a structured geometric model. The meshing step converts the irregular point cloud data into a regular mesh model. Structural feature extraction is then performed on the meshed data to capture key geometric features of the wall, such as cracks, holes, and corners.

[0078] 3. By modeling the material properties of the wall structure using characteristic data, the present invention accurately describes the material properties of the plastic concrete cut-off wall (such as density, elastic modulus, and compressive strength). This material property modeling, combined with a physical properties database, enables the model to not only reflect the geometric form but also accurately predict the mechanical performance of the wall under different working conditions through the precise description of physical parameters.

[0079] 4. Stress analysis provides a multi-level structural mechanics assessment, enabling designers to fully understand the stress state of the wall. By analyzing the wall's performance under different loads, potential failure areas can be effectively identified, helping engineers optimize the design phase and thus improve the safety and stability of the structure. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] Other features, objects and advantages of the present application will become more apparent upon reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0081] Figure 1 A flowchart showing the steps of a method for three-dimensional modeling of an actual wall of a plastic concrete anti-seepage wall according to an embodiment is shown;

[0082] Figure 2 A flowchart showing the steps of a method for constructing a material property model for a plastic concrete cut-off wall according to an embodiment is shown;

[0083] Figure 3 A flowchart showing the steps of a method for merging a plastic concrete cut-off wall model according to an embodiment is shown;

[0084] Figure 4 A flowchart showing a method for analyzing stress of a plastic concrete cut-off wall according to an embodiment of the present invention is provided;

[0085] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0086] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. It is obvious that the embodiments described are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts are within the scope of protection of the present invention.

[0087] Furthermore, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor and / or microcontroller approaches.

[0088] It should be understood that although the terms "first," "second," and the like may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used solely to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the exemplary embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the listed associated items.

[0089] A section of plastic concrete cut-off wall at a water conservancy project was inspected. The wall was 10 meters long and 0.5 meters thick, with no reinforcement. Cracks and porosity were expected. Geological survey radar (GPR) was used to scan the wall, identifying internal structural features and reconstructing a 3D model.

[0090] Plan the detection path, the path design is 10 meters long, the spacing is 0.5 meters, and the number of scanning points is a total of 20 scanning points. The path diagram shows that the scanning path is divided longitudinally along the wall, and each point is scanned once.

[0091] Scan parameter settings: Geological radar equipment uses high-frequency GPR (such as 1 GHz frequency), which is suitable for detecting smaller cracks. The scanning speed is 0.5 m / s. The sampling frequency is 100 samples per scanning point. The reflected signal is recorded at each scanning point. As shown in the following table:

[0092] Scanning depth calculation table 2

[0093]

[0094]

[0095] Calculating depth based on reflection time. Data analysis indicates that the propagation speed of electromagnetic waves in plastic concrete is 0.1 m / ns. Depth (m) = time (ns) × 0.1 / 2. Multiply the time of the reflected signal by the propagation speed (m / ns), then divide by 2 (for both upward and downward reflections). This is shown in the following table:

[0096] Reflection signal data table 1

[0097]

[0098]

[0099] By calculating the gradient (rate of change) of the reflected signal intensity, areas with significant intensity changes are identified. Areas with larger gradients often indicate stress concentration or internal defects. Calculate the difference in intensity between adjacent scanning points to obtain the gradient value. Set a gradient threshold. When the gradient exceeds the threshold, it is considered a stress concentration area. Alternatively, set an intensity threshold (such as 0.5) to extract reflected signals with intensities greater than this value. The stress concentration area extraction results show that the depths corresponding to the stress concentration areas (such as cracks, pores, etc.) are: scanning point 3 (depth 0.70m, intensity 0.60), scanning point 4 (depth 0.80m, intensity 0.55), scanning point 7 (depth 1.10m, intensity 0.70), and scanning point 8 (depth 1.20m, intensity 0.80).

[0100] Based on the extraction results, 3D point cloud data was generated. The point cloud coordinates (x, y, z) are: Point 1: (0, 0, 0.50), Point 2: (0, 0.5, 0.60), Point 3: (0, 1.0, 0.70), Point 4: (0, 1.5, 0.80), Point 5: (0, 2.0, 1.10), and Point 6: (0, 2.5, 1.20). The generated point cloud data was imported into 3D visualization software (such as AutoCAD, 3D Studio Max, or Revit). The 3D model of the plastic concrete cut-off wall was displayed in the software, highlighting cracks and porosity. Stress contours and cross-sectional diagrams were generated to help engineers understand the wall's internal structure and potential problems.

[0101] See also Figures 1 to 4 The present application provides a three-dimensional modeling method for an actual wall of a plastic concrete cut-off wall, the method comprising:

[0102] S1. Scan the plastic concrete cut-off wall using geological radar equipment to obtain scanning data of the plastic concrete cut-off wall;

[0103] Specifically, a high-precision geological radar device is aimed at the exposed surface of the plastic concrete cut-off wall to perform a full-scale scan, generating high-density point cloud data. The scanning device needs to be calibrated to ensure that the scanning range covers the entire wall structure.

[0104] Specifically, before the cut-off wall is fully poured, a geological radar device is used to scan it from an accessible location. Scanning is performed regularly during the construction process to obtain progressive data of the wall and obtain scanning data of the plastic concrete cut-off wall.

[0105] Specifically, a geological radar device is selected for scanning, which includes a geological radar. Geological radar is a detection technology that can penetrate soil and plastic concrete. It can scan even when the surface of the wall is covered by soil, and detects the internal structure of the wall by emitting electromagnetic waves.

[0106] S2. Gridding the scanned data of the plastic concrete cut-off wall to obtain gridded data of the plastic concrete cut-off wall; extracting wall structural features from the gridded data of the plastic concrete cut-off wall to obtain wall structural feature data; and modeling wall material properties based on the wall structural feature data to obtain a material property model of the plastic concrete cut-off wall;

[0107] Specifically, the Delaunay triangulation algorithm is used to mesh the point cloud data to generate a three-dimensional mesh model. By optimizing the size and distribution of the triangles, the generated mesh is ensured to accurately reflect the actual shape of the wall. The gradient analysis algorithm and edge detection technology are used to identify the structural features of the wall, such as corners, edges, and irregular shapes. The extracted feature point set is used to define the geometric shape of the wall. Based on the extracted structural feature data, material property parameters in finite element analysis (such as elastic modulus, density, tensile strength, etc.) are used for material property modeling. The material property model uses multi-physical quantity modeling technology, combined with the physical parameters of the wall and experimental data, to form a material property model of the plastic concrete anti-seepage wall.

[0108] S3, fusing the gridded data of the plastic concrete cut-off wall and the material property model of the plastic concrete cut-off wall to obtain a fusion model of the plastic concrete cut-off wall;

[0109] Specifically, the meshed data is fused with the material property model. A weighted fusion algorithm is used to combine the geometric model and material properties to generate a complete 3D fused model. During the fusion process, the material parameters of each mesh element are adjusted to account for the influence of the geometric structure on the material properties. An adaptive adjustment algorithm is introduced to automatically adjust the mesh density and material property parameters in the fused model, ensuring accuracy and stability under varying stress conditions.

[0110] S4. Perform stress analysis on the plastic concrete cut-off wall fusion model to obtain a plastic concrete cut-off wall simulation model.

[0111] Specifically, a stress analysis algorithm based on finite element analysis was used to perform stress simulation on the fused 3D model. By dividing the mesh into cells and applying boundary conditions (such as load, temperature, and humidity), the stress distribution of the plastic concrete cutoff wall under different operating conditions was calculated. Simulation software was used to visualize the stress distribution and deformation, generating stress contours and displacement diagrams to illustrate the cutoff wall's response under different stress conditions.

[0112] Optionally, the plastic concrete cut-off wall scanning data includes first plastic concrete cut-off wall scanning data and second plastic concrete cut-off wall scanning data, and S1 includes:

[0113] Scanning the plastic concrete cut-off wall at a first angle using a geological radar device to obtain first plastic concrete cut-off wall scanning data;

[0114] Specifically, adjust the geological radar equipment to the preset first angle (45 degrees or 60 degrees from top to bottom) to ensure that the equipment is stably placed and aligned with the surface of the plastic concrete cut-off wall. The angle threshold data should be accurately set according to the height and width of the specific wall to ensure that sufficient surface area is covered. Start scanning and obtain high-density point cloud data from the wall surface through a point cloud acquisition algorithm (such as a three-dimensional scanning algorithm based on LIDAR). The point cloud data format is a three-dimensional coordinate point set (X, Y, Z), and each point represents a measurement point on the wall surface. Perform preliminary cleaning and noise reduction on the first plastic concrete cut-off wall scanning data to remove noise points and unnecessary redundant data to ensure the accuracy and integrity of the data.

[0115] The plastic concrete cut-off wall is scanned at a second angle by geological radar equipment to obtain second plastic concrete cut-off wall scanning data, wherein the angle threshold data corresponding to the first angle scan and the angle threshold data corresponding to the second angle scan are in a vertical relationship.

[0116] Specifically, rotate the geological radar device 90 degrees (45 degrees or 30 degrees from bottom to top) so that it is perpendicular to the first angle. Make sure that the scanning device is aimed at the same wall area, but collect data from an angle perpendicular to the first scan. Repeat the scanning operation and use the same point cloud acquisition algorithm to obtain the second plastic concrete cut-off wall scanning data. Since the scanning angle is perpendicular to the first angle, these point cloud data can capture the geometric details of the wall at different angles and supplement the areas that the first scan cannot cover. The second plastic concrete cut-off wall scanning data is denoised, cleaned, and format converted to ensure the same accuracy and format as the first scan data.

[0117] Based on the preset angle threshold, the point cloud data of the first angle and the second angle are spatially calibrated using an angle conversion algorithm (such as a rotation matrix and a quaternion algorithm). This step ensures that the scan data at different angles are aligned in the same coordinate system. The two sets of point cloud data are accurately aligned using the ICP (Iterative Closest Point) algorithm. The ICP algorithm minimizes the difference between the two sets of point clouds by iteratively matching the closest points, ensuring the best match between the two sets of data in three-dimensional space. Using a weighted average algorithm or a voxel-based point cloud fusion algorithm, the point cloud data of the first angle and the second angle are fused into a complete three-dimensional point cloud model.

[0118] Optionally, S2 includes:

[0119] S21, performing point cloud meshing processing on the scanning data of the plastic concrete cut-off wall to obtain meshed data of the plastic concrete cut-off wall;

[0120] Specifically, the point cloud data of the plastic concrete cut-off wall is converted into a 3D mesh model using either Delaunay triangulation or Poisson surface reconstruction algorithms. Delaunay triangulation is suitable for surfaces with good regularity, while Poisson reconstruction is more suitable for complex, irregular surfaces. Initial meshing data is generated by connecting the nearest neighbor points in the point cloud one by one to form triangular facets. The resulting meshed model intuitively reflects the geometry of the plastic concrete wall. The number of generated triangular facets depends on the density of the point cloud.

[0121] S22. Optimizing the mesh data of the plastic concrete cut-off wall to obtain simplified mesh data of the plastic concrete cut-off wall;

[0122] Specifically, the Quadric Error Metrics (QEM) algorithm is used for mesh simplification. QEM removes unnecessary mesh points by minimizing the error after vertex merging, reducing mesh complexity while maintaining geometric features. During the optimization process, the error matrix is ​​first calculated for each vertex, and then vertex pairs with the smallest error are gradually merged until the mesh reaches the desired level of simplification. The resulting simplified mesh contains fewer triangles while retaining essential geometric details.

[0123] S23, performing mesh smoothing processing on the simplified mesh data of the plastic concrete cut-off wall to obtain mesh smoothing data of the plastic concrete cut-off wall;

[0124] Specifically, the simplified mesh data is smoothed using the Laplacian smoothing algorithm. Laplacian smoothing reduces surface irregularities by iteratively shifting each vertex to the geometric center of its neighboring vertices. The smoothing process requires setting the number of iterations and the shift factor. The resulting mesh data has a smoother surface while maintaining its original structural features.

[0125] S24, performing mesh normal calculation on the mesh smoothing data of the plastic concrete anti-seepage wall to obtain wall structure characteristic data;

[0126] Specifically, a normal calculation algorithm based on geometric normal vectors is used to calculate the normal vector for each mesh patch. For each triangle, the normal direction is calculated using a cross product based on the order of its vertices. All triangles are traversed, each normal vector is calculated, and these normal vectors are assigned to the corresponding vertices to generate complete wall structural feature data. This normal data is used for subsequent lighting calculations and material mapping.

[0127] S25. Initialize the material mapping model according to the wall structure characteristic data to obtain the material mapping model;

[0128] Specifically, a UV mapping algorithm is used to generate the initial state of the material mapping model based on the wall's structural characteristics. UV mapping projects the 3D mesh coordinates onto a 2D texture coordinate system, determining the texture coordinates corresponding to each vertex of the mesh. The mesh is first flattened, and then, based on the flattened mesh coordinates, UV texture coordinates are assigned to each vertex to ensure that the material is correctly mapped to the wall surface.

[0129] S26. Obtain physical material parameter data, and perform finite element modeling on the material mapping model according to the physical material parameter data to obtain a material property model of the plastic concrete cut-off wall.

[0130] Specifically, the physical parameters of plastic concrete, such as elastic modulus, Poisson's ratio, density and compressive strength, are obtained. These parameters are provided by experimental data or industry standards. Finite element analysis (FEA) software, such as ANSYS or ABAQUS, is used to combine the material mapping model with the physical material parameters to construct a material property model of the plastic concrete cut-off wall. In finite element modeling, mesh nodes are regarded as calculation nodes for mechanical analysis, and the physical material parameters determine the behavior of these nodes under mechanical stress and deformation. Boundary conditions (such as fixed constraints and load application) are set, and then the stress, strain and displacement responses of the material property model of the plastic concrete cut-off wall are solved to generate the mechanical performance data of the wall under different working conditions.

[0131] Optionally, S3 includes:

[0132] S31, performing mesh material parameter matching based on the mesh data of the plastic concrete cut-off wall and the material property model of the plastic concrete cut-off wall to obtain initial fusion model data;

[0133] Specifically, the meshed data of the plastic concrete cut-off wall is matched to the physical parameters of the material property model using a nearest neighbor matching algorithm or a bilinear interpolation algorithm. Each mesh node is compared with the corresponding point in the material property model, and material parameters (such as elastic modulus and density) are assigned to the mesh node.

[0134] The system traverses all mesh nodes and, based on their geometric positions, searches for the nearest corresponding node in the material property model. Using a nearest neighbor algorithm, the parameter values ​​from the material property model are assigned to the mesh nodes, generating the initial fused model data. For complex surfaces, a bilinear interpolation algorithm is used to ensure smooth transitions between adjacent nodes and continuity of physical parameters.

[0135] S32, performing material parameter interpolation processing on the initial fusion model data to obtain material interpolation model data;

[0136] Specifically, the material parameters of the initial fusion model data are interpolated using a cubic spline interpolation algorithm or a Kriging interpolation method. The interpolation algorithm infers the material properties of unknown nodes based on the material property values ​​of known nodes, ensuring the overall smoothness of the model.

[0137] First, all mesh nodes in the model are calibrated. Using the known material property values ​​of the nodes as constraints, cubic spline interpolation is used to generate continuous material parameter distribution curves on the mesh surface. If the material parameters exhibit significant spatial variation, Kriging interpolation is employed. This interpolation method accounts for spatial autocorrelation and better reflects the variability of material properties across different regions. Once interpolation is complete, the material interpolation model data is generated.

[0138] S33. Optimize the material interpolation model data to obtain a plastic concrete cut-off wall fusion model.

[0139] Specifically, the material interpolation model is optimized using a gradient descent method or a multi-objective optimization algorithm. The optimization goal is to minimize the error in the material property distribution while ensuring the physical consistency and accuracy of the model.

[0140] An error function is defined, which calculates the error value for each mesh node based on the difference between the material interpolation model and the actual physical material parameters. By calculating the gradient of the error function, the material parameters of each mesh node are adjusted to gradually approach the actual value. The gradient descent method can effectively reduce local errors during the interpolation process. If the model needs to consider multiple material parameters (such as density, elastic modulus, compressive strength, etc.), a multi-objective optimization algorithm is used to comprehensively adjust the distribution of each material parameter to achieve optimal performance under different working conditions.

[0141] Optionally, S4 includes:

[0142] S41, performing stress validity analysis on the plastic concrete cut-off wall fusion model to obtain a first plastic concrete cut-off wall simulation model;

[0143] Specifically, a statics analysis method within finite element analysis (FEA) was used to conduct a stress validity analysis on the fusion model of a plastic concrete cut-off wall. The purpose of this analysis was to verify whether the model met the rationality of stress distribution and structural safety under the design conditions.

[0144] Apply realistic boundary conditions to the plastic concrete cutoff wall, such as loads, temperature fluctuations, and humidity. Using FEA software (such as ANSYS or ABAQUS), calculate the stress values ​​at each grid node and generate a stress contour map. Verify the calculated stress distribution for validity, determining whether stress concentration or uneven stress distribution exists. If the stress distribution meets design expectations, generate the first simulation model of the plastic concrete cutoff wall.

[0145] S42, performing stress multi-scale analysis on the plastic concrete cut-off wall fusion model to obtain a second plastic concrete cut-off wall simulation model;

[0146] Specifically, a hierarchical multi-scale analysis method was used to analyze the structure of the plastic concrete cut-off wall at different scales (such as macro, meso, and micro). This method can more carefully analyze the stress distribution and deformation behavior at different scales within the wall.

[0147] The fusion model of the plastic concrete cutoff wall is divided into different scale levels, with corresponding material parameters and boundary conditions set for each scale. Independent stress analysis is performed on each scale model to calculate the stress distribution at each scale. For example, the macroscale focuses on the structural stress distribution of the entire wall, while the microscale focuses on subtle stress variations in local materials. The stress analysis results at each scale are integrated to generate a second simulation model of the plastic concrete cutoff wall. Multi-scale analysis helps identify local structural weaknesses and potential failure risks.

[0148] S43: Fusing and screening the safety factors of the first plastic concrete cut-off wall simulation model and the second plastic concrete cut-off wall simulation model to obtain a concrete cut-off wall simulation model.

[0149] Specifically, based on the stress distribution results of the first and second simulation models, a safety factor calculation method, such as the Von Mises stress criterion or the Mohr-Coulomb criterion, is used to calculate the safety factor of the wall in each area.

[0150] The stress data and corresponding safety factors from the first and second simulation models are fused using a weighted average method or a decision tree-based fusion algorithm. The weighted average method assigns weights to each model based on its contribution to the stress distribution, integrating safety factors at different scales to generate a global safety factor distribution. Based on this fused safety factor data, threshold filtering conditions are set to identify areas with safety factors below the set threshold and mark them as potential risk areas. These areas are then analyzed and reinforced.

[0151] Optionally, the stress effectiveness analysis includes:

[0152] Obtaining plastic concrete cut-off wall constraint data;

[0153] Specifically, physical constraint data for the plastic concrete cutoff wall is obtained through on-site surveys and design drawings. Constraint data includes fixed support points, boundary conditions, structural contact surfaces, and the connection between the wall and the foundation. Common constraints include fixed constraints (zero displacement) and slip constraints (limiting displacement in a specific direction). These conditions ensure that the wall conforms to actual working conditions during calculations.

[0154] Applying load to the plastic concrete cut-off wall fusion model according to the plastic concrete cut-off wall constraint data to obtain load application model data;

[0155] Specifically, based on the constraint data of the plastic concrete cutoff wall, static and dynamic load application algorithms were used to apply loads to the wall model. Load types included deadweight, seismic loads, wind loads, and hydrostatic pressure. Finite element analysis software was used to gradually increase the load intensity based on actual working conditions, generating load application model data. When applying loads to each grid node, the influence of the constraints was considered to ensure a reasonable distribution of the loads.

[0156] Performing static stress processing and dynamic stress processing on the load application model data to obtain static stress distribution data and dynamic stress distribution data;

[0157] Specifically, a linear static analysis method is used to perform static stress processing on the load-applied model data. This method assumes that the material remains elastic under load and calculates the stress distribution using equilibrium equations. Using a statics solver, the static stress distribution of the plastic concrete cut-off wall under load is calculated. Static stress distribution data is generated to reflect the stress changes in the wall under static loads, enabling the identification of weak areas under static conditions.

[0158] Dynamic stress analysis is performed on the load-applied model data using either the time-domain integration method or modal analysis. The time-domain integration method is used to analyze transient responses, while modal analysis is used for frequency-domain analysis. Finite element software calculates the stress response of the wall under dynamic loads, taking into account transient loads such as earthquakes or wind loads, and generates dynamic stress distribution data. This data demonstrates the magnitude and frequency characteristics of stress changes in the wall under dynamic loads.

[0159] Extract stress concentration areas based on static stress distribution data and dynamic stress distribution data to obtain stress concentration area data;

[0160] Specifically, a stress concentration extraction algorithm based on stress gradient analysis is used to identify stress concentration areas in static and dynamic stress distribution data. Local gradient analysis is performed on the static and dynamic stress distribution data to identify stress concentrations. Stress concentration areas are locations where stress is significantly higher than the surrounding area. These areas occur at geometric discontinuities, edges, or areas with dramatic load variations. The extracted stress concentration area data is used for subsequent structural assessment and annotation.

[0161] The plastic concrete cut-off wall fusion model is annotated according to the stress concentration area data, the static stress distribution data, and the dynamic stress distribution data to obtain a first plastic concrete cut-off wall simulation model.

[0162] Specifically, the fused model of the plastic concrete cutoff wall was annotated based on stress concentration area data, static stress distribution data, and dynamic stress distribution data. This annotation process involved visualizing stress concentration areas and high-risk regions and labeling these areas with important stress values ​​or safety factors. In the simulation model, stress concentration areas were annotated using colors or symbols to highlight potential structural weaknesses. The annotated simulation model of the first plastic concrete cutoff wall will be used for further safety assessment and design optimization.

[0163] Optionally, the stress multi-scale analysis includes:

[0164] Initialize the stress field of the plastic concrete cut-off wall fusion model to obtain initial stress field data;

[0165] Specifically, a finite element mesh-based initial stress field generation algorithm is employed. Load conditions, boundary conditions, and material properties are input into the fusion model of the plastic concrete cutoff wall to initialize the wall's stress field data. Based on the physical parameters of the plastic concrete material (such as elastic modulus and density) and its stress response, the initial stress value at each mesh node is calculated. A static solver is used to generate initial stress field data, which reflects the overall stress distribution under the initial operating conditions.

[0166] Perform stress field stratification processing according to the initial stress field data to obtain stress field stratification data;

[0167] Specifically, a grid-based hierarchical algorithm is used to stratify stress field data at different scales. This process divides the stress field into different stress levels based on the internal structure of the material, local areas of the wall, and the overall structure.

[0168] The stress field data of the plastic concrete cutoff wall is divided into different levels based on region and stress gradient, with each level representing a different stress region of the wall. Layered processing is performed step by step from the microscopic level (such as the internal stress field of the material) to the macroscopic level (such as the overall stress field of the wall) to generate layered stress field data.

[0169] Perform multi-scale coupling analysis based on stress field layered data to obtain multi-scale coupled stress data;

[0170] Specifically, a multiscale coupled analysis algorithm, such as multiscale finite element analysis (MS-FEA) or multi-level reduced-order modeling (ROM), is used to couple the stress fields of each layer. This algorithm combines the stress behaviors at the micro and macro scales to calculate the coupled stress distribution.

[0171] At each stress level, the influence of the local stress field on the global stress field is calculated. A coupled stress analysis algorithm simulates the interaction of stresses from the microscopic to the macroscopic level, generating multiscale coupled stress data. Iterative calculations are performed between each level to ensure the proper transfer of the mutual influence of the local and global stress fields. By gradually adjusting the stress distribution at each scale, accurate multiscale stress data is obtained.

[0172] The stress model is fused according to the multi-scale coupled stress data and the plastic concrete cut-off wall fusion model to obtain a second plastic concrete cut-off wall simulation model;

[0173] Specifically, a stress field fusion algorithm, such as a weighted average-based fusion algorithm or a layered composite algorithm, is used to fuse the multi-scale coupled stress data with the plastic concrete cut-off wall fusion model.

[0174] Multiscale coupled stress data is mapped to the overall wall model, and a layer-by-layer fusion approach is used to ensure that the impact of microscopic stress data on the macroscopic stress field is properly reflected. Stress data at different levels is weighted and integrated to eliminate boundary effects in multiscale analysis, generating a simulation model of a second plasticity concrete cut-off wall.

[0175] The stress models include:

[0176] Extract macro-micro stress data based on multi-scale coupled stress data to obtain macro-micro stress data;

[0177] Specifically, the finite element analysis software was used to perform numerical simulation of the plastic concrete cut-off wall, and the material parameters and loading conditions were input. Meshing was performed to ensure that the model could accurately reflect the geometric characteristics and stress state of the wall. The simulation was run to obtain the stress data of each mesh unit, including macroscopic and microscopic stress distribution. Macroscopic stress data (depth-stress): depth 0.0m: stress value 20MPa, depth 1.0m: stress value 30MPa, depth 2.0m: stress value 40MPa, depth 3.0m: stress value 45MPa, depth 4.0m: stress value 50MPa. Microscopic stress data (depth-stress): depth 0.5m: stress value 35MPa, depth 1.5m: stress value 42MPa, depth 2.5m: stress value 55MPa (stress concentration), depth 3.5m: stress value 48MPa, depth 4.5m: stress value 52MPa.

[0178] Construct a geometric model of the plastic concrete cut-off wall in finite element analysis software (such as ANSYS, ABAQUS, etc.), ensuring that the model size is consistent with the actual wall. Set the material properties of the plastic concrete, such as: elastic modulus (E): about 25GPa, Poisson's ratio (ν): about 0.2, density: about 2400kg / m 3 . Set the boundary conditions of the wall according to the actual working conditions, such as the bottom is fixed (cannot move) and the side is subject to soil pressure. Calculate the deadweight of the anti-seepage wall and apply it according to the volume and density. Apply soil pressure, hydrostatic pressure, etc. For example, the height of the anti-seepage wall is 5m, the width is 1m, and the thickness is 0.5m. The volume of the wall is: V = 5m × 1m × 0.5m = 2.5m 3 The self-weight is calculated as F 自重=V×density×g≈58860N, divide the model into a finite number of units. Use tetrahedral or hexahedral units. Perform meshing and generate nodes and elements. Record the coordinates of each node and the corresponding unit type. Start the solution process, and the software will calculate the stress, strain and other data of each node based on the input load and boundary conditions. The software automatically solves the node response and calculates the stress value of each node. The stress of each node includes: principal stress (σ1, σ2, σ3), equivalent stress (von Mises stress). After the calculation is completed, extract the stress data of each node. Macro stress records such as calculating the macro stress of the entire wall and taking the average stress value of each layer of nodes. Micro stress records such as recording the specific stress value of each node. As shown in the following table:

[0179] Stress calculation table 3:

[0180]

[0181]

[0182] Extract stress concentration areas based on macro and micro stress data to obtain stress concentration area data;

[0183] Specifically, the extracted stress data is analyzed to identify stress concentration areas. A stress threshold is set to filter out areas that exceed the threshold. The extracted stress data is sorted and a stress distribution diagram is drawn. According to material properties and engineering requirements, a stress threshold (for example, 40MPa) is set. The stress distribution data is traversed, and all nodes with stress values ​​greater than the threshold are extracted and marked as stress concentration areas. The depth and intensity of the stress concentration area are recorded. The stress concentration area is extracted as follows: stress concentration area (greater than 40MPa): depth 2.5m (stress 55MPa).

[0184] The stress concentration area data is used to mark the plastic concrete cut-off wall fusion model and obtain a preliminary stress fusion model.

[0185] Specifically, the fused model of the cutoff wall is annotated using data from stress concentration areas to clearly identify potential risk areas. The fused cutoff wall model is then imported into visualization software. Using the extracted stress concentration area data, different colors or symbols are applied to the model. A corresponding label is generated for each marked area, including information such as stress value and depth. For example, in the 3D model, the area at a depth of 2.5m is labeled as a "stress concentration area" and labeled with a stress value of 55MPa.

[0186] Based on the macro-micro stress data and the preliminary stress fusion model, hierarchical stress fusion is performed to obtain the second plastic concrete cut-off wall simulation model;

[0187] Specifically, the extracted macro-micro stress data are combined with the preliminary stress fusion model to perform hierarchical stress analysis. The model is divided into several layers, such as the upper layer, the middle layer and the lower layer, to ensure that each layer can reflect different stress states. According to the macro-micro stress data, the stress values ​​of each layer are summarized and the average stress of each layer is calculated. The stress of each layer is weighted to ensure that the influence of each layer is reflected in the overall model. The stress data of all layers are integrated to obtain the second plastic concrete anti-seepage wall simulation model. Upper layer (0.0m-1.0m): average stress 25MPa, middle layer (1.0m-3.0m): average stress 40MPa, lower layer (3.0m-4.5m): average stress 50MPa, calculate the weighting, obtain the second plastic concrete anti-seepage wall simulation model, and form a stress cloud map.

[0188] The macro-micro stress data includes macro-stress data and micro-stress data, and the stress concentration area extraction includes:

[0189] Perform graph neural network encoding on the macro stress data and the micro stress data to obtain macro stress characteristic map data and micro stress characteristic map data;

[0190] Specifically, feature maps are constructed using the collected macro- and micro-stress data. The extracted macro- and micro-stress data are organized into a two-dimensional matrix, where each matrix element corresponds to the stress value at a measurement point. The macro- and micro-stress data are then fed into an encoding module (calculated via a graph convolutional layer) to generate feature maps. These feature maps reflect the spatial distribution of stress.

[0191] Specifically, each node and its stress value are constructed as a node in the graph, and the edges between nodes can be established based on spatial position or stress relationship. Edges are established for adjacent nodes to form a graph structure. The macroscopic and microscopic feature graph data are input into the encoding module of the graph neural network. A graph node is created for each node, the stress value is used as the node feature, and edges are established to connect adjacent nodes (such as between node 1 and node 2, between node 2 and node 3). Stress data of the node (macroscopic and microscopic). Feature extraction is performed through the graph convolution layer, and convolution operation is performed using the combination of node features and the features of its adjacent nodes. After graph convolution, an activation function (such as ReLU) is used for nonlinear transformation. The features of each node will be updated according to the features of its neighbors. After multiple layers of graph convolution, macroscopic stress feature graph data and microscopic stress feature graph data are generated, and these feature graphs will reflect the spatial distribution of stress.

[0192] Performing distance calculation on the macroscopic stress characteristic map data and the microscopic stress characteristic map data to obtain macroscopic stress map characteristic clustering data and microscopic stress map characteristic clustering data respectively;

[0193] Specifically, distance calculations are performed based on the feature map data to identify feature similarities. For each point in the feature map, the distance to other points is calculated to generate a distance matrix. Distance calculations can be based on Euclidean distance. Based on the distance matrix, feature clustering is performed. Points with similar features are clustered together to form macro-stress map feature cluster data and micro-stress map feature cluster data.

[0194] Specifically, the distance between each point in the feature map is calculated to generate a distance matrix. For example, Euclidean distance is used. The distance matrix is ​​used to cluster the macroscopic and microscopic features. Clustering is performed based on a distance threshold (e.g., 10 MPa) to group similar stress values ​​together.

[0195] Dynamic stress threshold control is performed based on the macroscopic stress map characteristic clustering data and the microscopic stress map characteristic clustering data, and macroscopic stress threshold matrix data and microscopic stress threshold matrix data are obtained respectively;

[0196] Specifically, stress thresholds are set and dynamically adjusted to suit different conditions. Preliminary macro- and micro-stress thresholds are set (e.g., based on material properties). Clustering results are analyzed and thresholds are dynamically adjusted. Based on the changing trends of local stresses and the distribution of clustered data, thresholds can be reset to improve extraction accuracy. Based on the adjusted thresholds, macro- and micro-stress threshold matrix data are generated.

[0197] Specifically, the macro threshold is set to 40 MPa and the micro threshold is set to 45 MPa.

[0198] After obtaining preliminary stress data, cluster analysis was performed to identify stress hotspots. The clustering results showed that certain nodes showed a clear concentration trend in the stress distribution. Based on the cluster analysis results, the number of nodes in each cluster was counted. A larger number of clusters indicated that these nodes had a higher degree of concentration in the stress distribution. The center of gravity (mean or median) of each cluster was calculated. If the center of gravity position was concentrated in a specific area, it indicated that stress concentration existed in that area. Macroscopic stress concentration areas were identified, and the following node stress values ​​exceeded the threshold of 40MPa: node 3 (depth 2m, stress 40MPa) and node 4 (depth 3m, stress 45MPa). Microscopic stress concentration areas were identified, and the following node stress values ​​exceeded the threshold of 45MPa: node 2 (depth 1.5m, stress 50MPa) and node 3 (depth 2.5m, stress 55MPa).

[0199] Adjustments are based on the following: If a node appears multiple times in the clustering results and its stress value is close to or above the current threshold, the threshold for that node is raised. If a node appears multiple times in the clustering results and its stress value is far below the current threshold, the threshold for that node can be lowered. If the mean stress value of a layer increases significantly (e.g., greater than 10%), the threshold is raised overall. If the mean stress value of a layer decreases significantly (e.g., less than 10%), the threshold is lowered overall.

[0200] After each stress data update, monitor the clustering results and record the frequency of occurrence and stress value of each node. For each node, adjust the stress value assessment threshold based on its frequent occurrence in the cluster: If the stress value of a node exceeds 50 MPa in multiple analyses and the clustering results show that the node appears frequently, increase the macro threshold to 45 MPa. If the stress value of a node is less than 35 MPa in multiple analyses, reduce the macro threshold to 35 MPa. If the stress value of a node exceeds 55 MPa in multiple analyses and the clustering results show that the node appears frequently, increase the micro threshold to 50 MPa. If the stress value of a node is less than 40 MPa in multiple analyses, reduce the micro threshold to 40 MPa. Update the macro and micro stress thresholds to ensure that the new thresholds are used when locating the next stress hotspot.

[0201] Using the macro stress threshold matrix data to locate the macro stress hotspot of the macro stress map feature clustering data, and obtaining macro stress hotspot location map data, and using the micro stress threshold matrix data to locate the micro stress hotspot of the micro stress map feature clustering data, and obtaining micro stress hotspot location map data;

[0202] Specifically, threshold matrix data is used to locate stress hotspots. Based on the macroscopic stress threshold matrix, regions with stress values ​​exceeding the threshold are screened and marked as macroscopic stress hotspots. Similarly, microscopic stress hotspots are located using the microscopic stress threshold matrix. The located hotspots are recorded in the stress hotspot location map data, including information such as location and stress value.

[0203] Perform multimodal stress layered mapping based on the macroscopic stress hotspot location map data and the microscopic stress hotspot location map data to obtain layered stress mapping matrix data;

[0204] Specifically, threshold matrix data is used to locate stress hotspots. Based on the macroscopic stress threshold matrix, regions with stress values ​​exceeding the threshold are screened and marked as macroscopic stress hotspots. Similarly, microscopic stress hotspots are located using the microscopic stress threshold matrix. The located hotspots are recorded in the stress hotspot location map data, including information such as location and stress value.

[0205] Based on the depth of the plastic concrete cutoff wall, the entire wall is divided into multiple layers (e.g., upper, middle, and lower layers), each of which should reflect the stress state at different depths. Based on the located macroscopic and microscopic stress hotspot data, multimodal stress layered mapping is performed. The layering principle can be, for example, divided into three equal parts according to the depth for multimodal stress layered mapping. Alternatively, based on plastic concrete data query, the entire wall is divided into multiple layers (e.g., upper, middle, and lower layers) based on the depth of the plastic concrete cutoff wall. Each layer should reflect the stress state at different depths for multimodal stress layered mapping. Alternatively, soil layer data is obtained (through geological exploration or historical data query), and layered based on the soil layer data and the plastic concrete depth data in the plastic concrete cutoff wall scanning data (previously obtained through geological radar scanning) to obtain plastic concrete layered data. Multimodal stress layered mapping is then performed based on the located macroscopic and microscopic stress hotspot data and the plastic concrete layered data. Determine the stress distribution at different depth levels: first layer (0-1m): 20-30MPa, second layer (1-2m): 30-40MPa, third layer (2-3m): 40-45MPa (including hot spots).

[0206] The concentrated area is synthesized according to the layered stress mapping matrix data to obtain the stress concentration area data.

[0207] Specifically, the layered stress mapping matrix is ​​synthesized to identify stress concentration areas. Based on the layered stress mapping matrix, overlapping stress hotspots are identified and synthesized into final stress concentration area data. The synthesized stress concentration area data is recorded, including the characteristics, location, and corresponding stress value of each area.

[0208] Specifically, overlapping hotspots are identified based on the layered stress mapping matrix data, and stress concentration data is synthesized. The stress concentration data is recorded and marked as high-risk areas. For example, the synthesized stress concentration data includes a concentration area at a depth of 1.5m (45MPa) marked as a high-risk area.

[0209] Optionally, the stress field layering process includes:

[0210] Perform material property stratification based on initial stress field data to obtain stress field material stratification data;

[0211] Specifically, a physical property-based stratification algorithm divides the initial stress field data into strata based on the material's physical parameters. The algorithm divides the stress field into regions based on the material's different properties, such as elastic modulus, density, and tensile strength.

[0212] Material properties, including elastic modulus, Poisson's ratio, and compressive strength, are extracted from the initial stress field data. Using a clustering algorithm (such as K-means or hierarchical clustering), the stress field is divided into layers based on the similarity of material properties, each corresponding to a different material characteristic. The resulting stress field material layer data will label different regions and reflect the differences in physical material properties within those regions.

[0213] Perform stress gradient stratification according to the initial stress field data to obtain stress field gradient stratification data;

[0214] Specifically, a gradient stratification algorithm based on stress change rate is used to divide the initial stress field data into different levels according to the stress gradient change. This method is used to identify stress concentration and sharp change areas in the stress field.

[0215] Using a derivation algorithm, the spatial derivative of the stress distribution in the initial stress field data is calculated to obtain stress gradient data. Regions with larger gradients represent stress concentration areas, while regions with smaller gradients represent stress flattening areas. Based on the magnitude of the stress gradient, a threshold method or a multi-level stratification algorithm is used to divide the stress field into several gradient layers. Each layer represents a different stress gradient strength. This generates stress field gradient stratification data, displaying the stress change rate and concentration in each region.

[0216] Cross-labeling is performed based on the stress field material layering data and the stress field gradient layering data to obtain the stress field layering data.

[0217] Specifically, a cross-labeling algorithm is used to combine the material property layered data and the stress gradient layered data to generate composite labeled stress field layered data.

[0218] The material property layered data is cross-fused with the stress gradient layered data, and a cross-labeling algorithm based on logical operations is used to annotate the material properties and stress gradients of each mesh node. In the fused stress field layered data, each region is annotated with both its material properties and stress gradient information. Combining the two layered results, it is possible to identify weak points not only in areas of stress concentration but also in areas of special materials. This generates stress field layered data, marking the intersection areas of material properties and stress gradients. These intersection areas represent critical locations in the structure, posing a higher safety risk and requiring further analysis and reinforcement.

[0219] Optionally, the stress gradient layering includes:

[0220] Perform gradient calculation based on the initial stress field data to obtain stress field gradient data;

[0221] Specifically, a numerical differentiation algorithm (such as the finite difference method or the finite element method) is used to calculate the gradient of the initial stress field data, and the stress change rate of each grid node is obtained by derivative calculation.

[0222] Based on the stress distribution in three-dimensional space, the main direction of stress change is determined and the stress change vector at each node is calculated. Stress gradient values ​​are calculated for all mesh nodes to generate stress field gradient data. Larger gradient values ​​indicate stress concentration or abrupt changes in that area.

[0223] Cluster calculation is performed based on stress field gradient data to obtain stress field gradient cluster characteristic data;

[0224] Specifically, K-means or density-based clustering algorithm (DBSCAN) is used to perform clustering calculations on the stress field gradient data to identify characteristic regions in the stress field.

[0225] Extract the gradient value and spatial position of each node from the stress field gradient data to form a feature vector. Using the K-means or DBSCAN algorithm, the stress gradient data is divided into several categories, each representing a different stress variation characteristic area, to generate stress field gradient clustering feature data.

[0226] The initial stress field data is partitioned into adjacent regions according to the stress field gradient clustering feature data to obtain the initial stress field gradient layered data;

[0227] Specifically, according to the stress field gradient clustering feature data, the Voronoi diagram segmentation method or the proximity-based partitioning algorithm is used to segment the initial stress field data into multiple adjacent regions.

[0228] Using the Voronoi diagram segmentation method, the entire stress field is divided into multiple adjacent regions based on the cluster center. Each region contains the mesh nodes close to the cluster center. Initial stress field gradient layered data is generated, and each layered region corresponds to a stress gradient characteristic cluster region.

[0229] Calculate the local gradient change rate based on the initial stress field gradient stratification data to obtain the gradient stratification local change rate data;

[0230] Specifically, the stress gradient change rate is calculated for a local area of ​​the initial stress field gradient layered data using a local difference method or a curvature-based change rate calculation method.

[0231] Select a local area within each partition and calculate its gradient change rate, focusing particularly on areas with sharp changes in stress gradient. By calculating the derivative of the gradient curve, we obtain the local gradient change rate data. Areas with high gradient changes indicate dramatic stress changes and stress concentration.

[0232] The gradient sensitive areas of the initial stress field gradient stratification data are classified according to the gradient stratification local change rate data to obtain the stress field gradient stratification data.

[0233] Specifically, based on a support vector machine (SVM) or decision tree algorithm, sensitive regions in the stress field gradient layered data are classified based on local gradient change rate data. Sensitive features, such as gradient change rate and change amplitude, are extracted from the local change rate data. Using the SVM or decision tree algorithm, sensitive regions in the stress field gradient layered data are classified. Sensitive regions are areas with the most significant stress changes and may be at risk of structural failure. Stress field gradient layered data is generated, and each gradient-sensitive region is marked for subsequent structural optimization and safety assessment.

[0234] Specifically, a threshold value for the gradient change rate is set. This threshold value is determined through engineering experience or preliminary simulation results. For example, an area where the gradient change rate is greater than a certain threshold value is defined as a sensitive area.

[0235] Optionally, the present invention further provides a three-dimensional modeling system for an actual wall of a plastic concrete cut-off wall, for executing the above-mentioned three-dimensional modeling method for an actual wall of a plastic concrete cut-off wall. The three-dimensional modeling system for an actual wall of a plastic concrete cut-off wall comprises:

[0236] The plastic concrete cut-off wall scanning module is used to scan the plastic concrete cut-off wall using geological radar equipment to obtain scanning data of the plastic concrete cut-off wall;

[0237] The plastic concrete cut-off wall material property model construction module is used to grid the plastic concrete cut-off wall scan data to obtain plastic concrete cut-off wall grid data; extract wall structure features from the plastic concrete cut-off wall grid data to obtain wall structure feature data; and perform wall material property modeling on the wall structure feature data to obtain a plastic concrete cut-off wall material property model;

[0238] The plastic concrete cut-off wall model fusion module is used to fuse the plastic concrete cut-off wall grid data and the plastic concrete cut-off wall material property model to obtain a plastic concrete cut-off wall fusion model;

[0239] The plastic concrete cut-off wall stress analysis module is used to perform stress analysis on the plastic concrete cut-off wall fusion model to obtain a plastic concrete cut-off wall simulation model.

[0240] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, the scope of the present invention is limited by the attached application documents rather than the above description, and it is intended that all changes that fall within the meaning and scope of equivalent elements of the application documents are included in the present invention.

[0241] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A three-dimensional modeling method for an actual wall of a plastic concrete anti-seepage wall, characterized in that: The method comprises: S1. Scan the plastic concrete cut-off wall using geological radar equipment to obtain scanning data of the plastic concrete cut-off wall; S2. Gridding the scanned data of the plastic concrete cut-off wall to obtain gridded data of the plastic concrete cut-off wall; extracting wall structural features from the gridded data of the plastic concrete cut-off wall to obtain wall structural feature data; and modeling wall material properties based on the wall structural feature data to obtain a material property model of the plastic concrete cut-off wall; S3, fusing the gridded data of the plastic concrete cut-off wall and the material property model of the plastic concrete cut-off wall to obtain a fusion model of the plastic concrete cut-off wall; S4. Perform stress validity analysis on the fusion model of the plastic concrete cut-off wall to obtain a first plastic concrete cut-off wall simulation model; perform stress multi-scale analysis on the fusion model of the plastic concrete cut-off wall to obtain a second plastic concrete cut-off wall simulation model; perform safety factor fusion and screening based on the first plastic concrete cut-off wall simulation model and the second plastic concrete cut-off wall simulation model to obtain a concrete cut-off wall simulation model; The stress multi-scale analysis includes: Initialize the stress field of the plastic concrete cut-off wall fusion model to obtain initial stress field data; perform stress field stratification processing based on the initial stress field data to obtain stress field stratification data; perform multi-scale coupling analysis based on the stress field stratification data to obtain multi-scale coupling stress data; perform stress model fusion based on the multi-scale coupling stress data and the plastic concrete cut-off wall fusion model to obtain a second plastic concrete cut-off wall simulation model; The stress models include: Extract macro-micro stress data based on multi-scale coupled stress data to obtain macro-micro stress data; extract stress concentration areas based on macro-micro stress data to obtain stress concentration area data; use stress concentration area data to annotate the plastic concrete cut-off wall fusion model to obtain a preliminary stress fusion model; perform hierarchical stress fusion based on the macro-micro stress data and the preliminary stress fusion model to obtain a second plastic concrete cut-off wall simulation model; The macro-micro stress data includes macro-stress data and micro-stress data, and the stress concentration area extraction includes: The macro-stress data and the micro-stress data are encoded by graph neural network and distributed to obtain macro-stress characteristic map data and micro-stress characteristic map data; distance calculation is performed on the macro-stress characteristic map data and the micro-stress characteristic map data to obtain macro-stress map characteristic clustering data and micro-stress map characteristic clustering data respectively; dynamic stress threshold control is performed based on the macro-stress map characteristic clustering data and the micro-stress map characteristic clustering data to obtain macro-stress threshold matrix data and micro-stress threshold matrix data respectively; macro-stress hotspots are located on the macro-stress map characteristic clustering data using the macro-stress threshold matrix data to obtain macro-stress hotspot location map data, and micro-stress hotspots are located on the micro-stress map characteristic clustering data using the micro-stress threshold matrix data to obtain micro-stress hotspot location map data; multi-modal stress layered mapping is performed based on the macro-stress hotspot location map data and the micro-stress hotspot location map data to obtain layered stress mapping matrix data; concentrated area synthesis is performed based on the layered stress mapping matrix data to obtain stress concentration area data.

2. The method according to claim 1, characterized in that The plastic concrete cut-off wall scanning data includes the first plastic concrete cut-off wall scanning data and the second plastic concrete cut-off wall scanning data, and S1 includes: Scanning the plastic concrete cut-off wall at a first angle using a geological radar device to obtain first plastic concrete cut-off wall scanning data; The plastic concrete cut-off wall is scanned at a second angle by geological radar equipment to obtain second plastic concrete cut-off wall scanning data, wherein the angle threshold data corresponding to the first angle scan and the angle threshold data corresponding to the second angle scan are in a vertical relationship.

3. The method according to claim 1, characterized in that S2 include: Performing point cloud meshing processing on the scanning data of the plastic concrete anti-seepage wall to obtain mesh data of the plastic concrete anti-seepage wall; Optimize the mesh data of the plastic concrete cut-off wall to obtain simplified mesh data of the plastic concrete cut-off wall; Performing mesh smoothing processing on the simplified mesh data of the plastic concrete cut-off wall to obtain mesh smoothing data of the plastic concrete cut-off wall; Calculate the mesh normal of the plastic concrete anti-seepage wall mesh smoothing data to obtain the wall structure characteristic data; Initialize the material mapping model according to the wall structure characteristic data to obtain the material mapping model; Physical material parameter data is obtained, and finite element modeling is performed on the material mapping model according to the physical material parameter data to obtain a material property model of the plastic concrete cut-off wall.

4. The method according to claim 1, wherein S3 includes: According to the mesh data of the plastic concrete cut-off wall and the material property model of the plastic concrete cut-off wall, mesh material parameters are matched to obtain initial fusion model data; Perform material parameter interpolation processing on the initial fusion model data to obtain material interpolation model data; The material interpolation model data is optimized to obtain the fusion model of plastic concrete cut-off wall.

5. The method according to claim 1, wherein The stress effectiveness analysis includes: Obtaining plastic concrete cut-off wall constraint data; Applying load to the plastic concrete cut-off wall fusion model according to the plastic concrete cut-off wall constraint data to obtain load application model data; Performing static stress processing and dynamic stress processing on the load application model data to obtain static stress distribution data and dynamic stress distribution data; Extract stress concentration areas based on static stress distribution data and dynamic stress distribution data to obtain stress concentration area data; The plastic concrete cut-off wall fusion model is annotated according to the stress concentration area data, the static stress distribution data, and the dynamic stress distribution data to obtain a first plastic concrete cut-off wall simulation model.

6. The method according to claim 1, characterized in that The stress field layering process includes: Perform material property stratification based on initial stress field data to obtain stress field material stratification data; Perform stress gradient stratification according to the initial stress field data to obtain stress field gradient stratification data; Cross-labeling is performed based on the stress field material layering data and the stress field gradient layering data to obtain the stress field layering data.

7. The method according to claim 6, characterized in that Stress gradient layering includes: Perform gradient calculation based on the initial stress field data to obtain stress field gradient data; Cluster calculation is performed based on stress field gradient data to obtain stress field gradient cluster characteristic data; The initial stress field data is partitioned into adjacent regions according to the stress field gradient clustering feature data to obtain the initial stress field gradient layered data; Calculate the local gradient change rate based on the initial stress field gradient stratification data to obtain the gradient stratification local change rate data; The gradient sensitive areas of the initial stress field gradient stratification data are classified according to the gradient stratification local change rate data to obtain the stress field gradient stratification data.

8. A three-dimensional modeling system for actual wall of plastic concrete anti-seepage wall, characterized in that: For executing the actual three-dimensional modeling method of the plastic concrete cut-off wall according to claim 1, the actual three-dimensional modeling system of the plastic concrete cut-off wall comprises: The plastic concrete cut-off wall scanning module is used to scan the plastic concrete cut-off wall using geological radar equipment to obtain scanning data of the plastic concrete cut-off wall; The plastic concrete cut-off wall material property model construction module is used to grid the plastic concrete cut-off wall scan data to obtain plastic concrete cut-off wall grid data; extract wall structure features from the plastic concrete cut-off wall grid data to obtain wall structure feature data; and perform wall material property modeling on the wall structure feature data to obtain a plastic concrete cut-off wall material property model; The plastic concrete cut-off wall model fusion module is used to fuse the plastic concrete cut-off wall grid data and the plastic concrete cut-off wall material property model to obtain a plastic concrete cut-off wall fusion model; The plastic concrete cut-off wall stress analysis module is used to perform stress validity analysis on the fusion model of the plastic concrete cut-off wall to obtain a first plastic concrete cut-off wall simulation model; perform stress multi-scale analysis on the fusion model of the plastic concrete cut-off wall to obtain a second plastic concrete cut-off wall simulation model; and perform safety factor fusion and screening based on the first and second plastic concrete cut-off wall simulation models to obtain a concrete cut-off wall simulation model. The stress multi-scale analysis includes: Initialize the stress field of the plastic concrete cut-off wall fusion model to obtain initial stress field data; perform stress field stratification processing based on the initial stress field data to obtain stress field stratification data; perform multi-scale coupling analysis based on the stress field stratification data to obtain multi-scale coupling stress data; perform stress model fusion based on the multi-scale coupling stress data and the plastic concrete cut-off wall fusion model to obtain a second plastic concrete cut-off wall simulation model; The stress models include: Extract macro-micro stress data based on multi-scale coupled stress data to obtain macro-micro stress data; extract stress concentration areas based on macro-micro stress data to obtain stress concentration area data; use stress concentration area data to annotate the plastic concrete cut-off wall fusion model to obtain a preliminary stress fusion model; perform hierarchical stress fusion based on the macro-micro stress data and the preliminary stress fusion model to obtain a second plastic concrete cut-off wall simulation model; The macro-micro stress data includes macro-stress data and micro-stress data, and the stress concentration area extraction includes: The macro-stress data and the micro-stress data are encoded by graph neural network and distributed to obtain macro-stress characteristic map data and micro-stress characteristic map data; distance calculation is performed on the macro-stress characteristic map data and the micro-stress characteristic map data to obtain macro-stress map characteristic clustering data and micro-stress map characteristic clustering data respectively; dynamic stress threshold control is performed based on the macro-stress map characteristic clustering data and the micro-stress map characteristic clustering data to obtain macro-stress threshold matrix data and micro-stress threshold matrix data respectively; macro-stress hotspots are located on the macro-stress map characteristic clustering data using the macro-stress threshold matrix data to obtain macro-stress hotspot location map data, and micro-stress hotspots are located on the micro-stress map characteristic clustering data using the micro-stress threshold matrix data to obtain micro-stress hotspot location map data; multi-modal stress layered mapping is performed based on the macro-stress hotspot location map data and the micro-stress hotspot location map data to obtain layered stress mapping matrix data; concentrated area synthesis is performed based on the layered stress mapping matrix data to obtain stress concentration area data.

Images