Method for quickly generating simulation sample library

Abstract

The invention relates to the technical field of computer-aided engineering, and discloses a rapid simulation sample library generation method, which is executed by a computer system and comprises the following steps: modeling scheme design: dividing a modeling range into a core area and a transition area, extracting coordinates and geometric parameters of the core area and the transition area, and abstracting simulation model parameters; obtaining a core area model file: modeling a core area through pre-processing software, exporting an input file, and converting the input file into an FLAC3D readable. Flaac3d file through an interface program; and defining a generation class: creating simulation scheme parameters, an initialization function and model resetting method, a model generation method, a constitutive model and parameter assignment method and a boundary condition loading method. According to the method, the modeling range is divided into the core area and the transition area, so that the problems that a traditional modeling mode only depends on FLAC3D built-in tool modeling, complex grids need to be manually defined one by one, the modeling efficiency is low, and the precision is difficult to guarantee are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of computer-aided engineering, and in particular to a method for quickly generating a simulation sample library. Background Art

[0002] CAE (Computer-Aided Engineering) is a numerical analysis method used to solve complex engineering problems and is widely used in the automotive, aviation, and aerospace industries. Artificial intelligence (AI), a technical discipline that simulates, extends, and expands human intelligence, has demonstrated significant application potential. With the development of AI technology, its application in CAE has become a research hotspot, significantly improving the efficiency and accuracy of simulation analysis and revolutionizing engineering simulation analysis. The combination of AI and CAE will not only improve simulation efficiency but also enhance the efficiency and quality of product design, while also opening up new areas of research and application.

[0003] FLAC3D is a finite-difference software specifically developed for geotechnical engineering. It excels in simulating post-yield plastic flow, large deformation, and instability in geotechnical materials, and is widely used in engineering scenarios such as tunnels, underground power plants, and chamber clusters. However, FLAC3D suffers from a major drawback: its limited pre-processing capabilities. FLAC3D performs poorly in building complex simulation models, such as those involving closely intersecting tunnels, combined surface and underground mining, underground power plants for pumped-storage power plants, and chamber clusters for high-level radioactive waste disposal. Previous research has addressed this challenge, and numerous approaches have been proposed, such as modeling with Rhino-Griddle, ANSYS, Abaqus, or Midas GTS-NX, and then importing the model into FLAC3D. This approach only allows for single-shot modeling and analysis of a specific problem, and is unable to rapidly model and calculate a large number of simulation samples, which would be essential for integrating FLAC3D simulation software with AI technology. Summary of the Invention

[0004] In response to the shortcomings of the existing technology, the present invention provides a method for quickly generating a simulation sample library, which solves the problem that the traditional modeling method only relies on the built-in tools of FLAC3D for modeling, is limited by its weak pre-processing function, and is prone to modeling complex structures such as upper and lower intersecting tunnels and chamber groups, which is difficult and time-consuming, has low modeling efficiency and is difficult to ensure accuracy.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: A method for quickly generating a simulation sample library, executed by a computer system, comprises the following steps: Modeling scheme design: Divide the modeling range into the core area and the transition area, extract the coordinates and geometric parameters of the two, and abstract the simulation model parameters; Obtain core area model file: Model the core area using pre-processing software and export the input file, which is then converted into a .flac3d file readable by FLAC3D using an interface program. Define the generation class: create a class that includes simulation scheme parameters, initialization function and model reset method, model generation method, constitutive model and parameter assignment method, boundary condition loading method, monitoring scheme calling method and simulation sample library saving method; Write example code: Use the "python-reset-state false" command to reset the model. Use "zone import" to import the core zone and "zone create" to create the transition zone. Use the "zone cmodel" and "zone property" commands to assign constitutive parameters. Use the "zone history" command to record the monitoring results. Main program execution: import the itasca module, call the generation class instance method, and execute model reset, generation, parameter assignment, boundary loading, monitoring and sample library saving in sequence.

[0006] By adopting the above technical solution, the modeling range is divided into a core area and a transition area. The core area is modeled using Abaqus and Midas GTS-NX, and the transition area is quickly generated through FLAC3D built-in commands, thereby achieving accurate modeling of complex geometric areas and parametric expansion of simple areas, realizing batch rapid modeling, and then quickly calculating to obtain a large number of simulation samples, realizing the combination of FLAC3D simulation software and AI technology, thereby improving the traditional modeling method that only relies on FLAC3D built-in tools for modeling, which is limited by its weak pre-processing function and is prone to modeling difficulties, time-consuming, low modeling efficiency and difficulty in ensuring accuracy for complex structures such as upper and lower crossing tunnels and chamber groups.

[0007] Preferably, in the modeling scheme design step, the core area and the transition area are regular hexahedrons or parallelepipeds, the union of the two is the complete modeling range and the intersection is an empty set, and the basis for determining whether the geometric shape is complex is whether it can be directly modeled through the "zone creat" command of FLAC3D.

[0008] Preferably, the pre-processing software includes Abaqus, Midas GTS-NX, COMSOL, LS-DYNA and ANSYS, the input file is in .inp format, .msh format, .k format and .fpn format, and the interface program is used to convert the grid nodes and grid, unit grouping and other information of the input file into the grid definition format of FLAC3D.

[0009] Preferably, the simulation scheme parameters are determined by the research object, including simulation model parameters, physical and mechanical parameters, boundary condition parameters, excavation method and support parameters. The simulation model parameters include chamber shape, chamber size, burial depth, vertical spacing of chambers, horizontal spacing of chambers, and number of chambers. The physical and mechanical parameters include density, elastic modulus, Poisson's ratio, bulk modulus, shear modulus, cohesion, internal friction angle, tensile strength, shear dilation angle, permeability coefficient, thermal conductivity coefficient, thermal expansion coefficient, specific heat capacity, heat release power, thermal stress coefficient, ground temperature, temperature limit, effective stress coefficient, and ground stress values in three directions. The excavation methods include full-section excavation, upper and lower step excavation, and pilot tunnel construction method.

[0010] Preferably, in the model generation method, the "zone reflect" command is used to mirror the transition zone model, the "zone attach" command is used to bind the units on the contact surface of the core zone and the transition zone, and the grid size is adjusted so that the difference between the two grid sizes is less than a preset threshold.

[0011] Preferably, in the constitutive model and parameter assignment method, the constitutive model includes an elastic-plastic constitutive model, a strain softening / hardening model, a nonlinear constitutive model, and a joint and interface model. The elastic-plastic constitutive model includes a Mohr-Coulomb model and a Drucker-Prager model. The strain softening / hardening model includes a strain softening model and a bilinear strain softening / hardening model. The nonlinear constitutive model is a Hoek-Brown model. The joint and interface model includes a bilinear Mohr-Coulomb joint model. The grid of the specified area is encrypted by the "zone densify" command, and the corresponding density, elastic modulus, Poisson's ratio, bulk modulus, shear modulus, cohesion, internal friction angle, tensile strength, shear dilatancy angle, permeability coefficient, thermal conductivity coefficient, thermal expansion coefficient, specific heat capacity, heat release power, thermal stress coefficient, ground temperature, temperature limit, effective stress coefficient, and ground stress values in three directions are assigned to different areas by the "zone property" command.

[0012] Preferably, in the boundary condition loading method, the "zone face" command is used to set stress or displacement conditions for the model surface, and the "zone gridpoint" command is used to set stress or displacement conditions for the model grid nodes.

[0013] Preferably, in the monitoring scheme calling method, the monitored simulation result evaluation indicators include characteristic point displacement, number of plastic zone units and unit stress difference, and the monitoring variables are recorded through the "zone history" and "fish history" commands.

[0014] Preferably, in the simulation sample library saving method, the open function is used to open the file in text or CSV format, the simulation scheme parameters and the corresponding simulation result evaluation indicators are written into the file through the write method of the file object, and the file is closed through the close function.

[0015] Preferably, the definition generation class also includes Plot setting and saving methods, creating a post-processing view through the "plot create" command, setting view parameters using the "plot view" and "plot legend" commands, and saving the view as an image file through the "plotexport" command.

[0016] The present invention provides a method for quickly generating a simulation sample library. It has the following beneficial effects: 1. In the present invention, the modeling range is divided into a core area and a transition area. The core area is modeled using Abaqus and Midas GTS-NX, and the transition area is quickly generated using FLAC3D built-in commands, thereby achieving accurate modeling of complex geometric areas and parametric expansion of simple areas, realizing batch rapid modeling, and then quickly calculating and obtaining a large number of simulation samples, realizing the combination of FLAC3D simulation software and AI technology, thereby improving the traditional modeling method that only relies on FLAC3D built-in tools for modeling, which is limited by its weak pre-processing function and is prone to problems such as high modeling difficulty, long time consumption, low modeling efficiency and difficulty in ensuring accuracy for complex structures such as upper and lower crossing tunnels and chamber groups.

[0017] 2. In the present invention, by defining a simulation sample library rapid generation class that includes a model reset method, a model generation method, a constitutive model and parameter assignment method, a boundary condition loading method, a monitoring scheme calling method, and a simulation sample library saving method, and combining the main program to automatically call the instance method, the entire process from modeling to result storage is realized without human intervention in batch calculations. Traditional models mostly require manual parameter adjustment and repeated operations. Due to the lack of systematic code encapsulation and process integration, it is easy to lead to problems such as long sample library construction cycle and poor data consistency.

[0018] 3. In the present invention, the simulation scheme parameters of chamber size, physical and mechanical parameters and the simulation result evaluation indicators of characteristic point displacement and number of plastic zone units are structured and stored in the local sample library, thereby providing a standardized training data set for the artificial intelligence algorithm, thereby improving the problems of traditional methods such as inconsistent data format, insufficient sample size, and lack of automated data collection and storage mechanism, which easily lead to difficulty in obtaining AI algorithm training data and weak model generalization ability.

[0019] 4. In the present invention, the contact surface elements of the core area and the transition area are bound by "zone attach", the grid of the key area is encrypted by "zonedensify", and multiple types of boundary conditions are applied by "zone face apply", thereby ensuring the continuity of mechanical transmission between different areas, the calculation accuracy of key parts and the realistic simulation of complex boundary conditions. This improves the traditional method that mostly uses idealized boundaries to simplify calculations, which easily leads to large deviations between simulation results and actual engineering behaviors due to large differences in grid size or distorted boundary conditions.

[0020] 5. In the present invention, by defining the parametric abstract modeling range and simulation scheme of chamber spacing, support parameters, etc. as adjustable variables, combined with the modular encapsulation of Python scripts and FLAC3D commands, rapid adaptation to different engineering scenarios such as mining, tunnel construction, and high-level radioactive waste disposal is achieved, thereby improving the traditional method of independently developing codes for a single project. Due to the lack of parametric design and universal interfaces, it is easy to cause repeated development and poor compatibility when applied across projects. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 Schematic diagram of the method steps of the present invention. DETAILED DESCRIPTION

[0022] The following will clearly and completely describe the technical solution of the present invention in conjunction with the accompanying drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0023] Please see the attached Figure 1 The embodiment of the present invention provides a method for quickly generating a simulation sample library, which is executed by a computer system and includes the following steps: Modeling scheme design: Divide the modeling range into the core area and the transition area, extract the coordinates and geometric parameters of the two, and abstract the simulation model parameters; Obtain core area model file: Model the core area using pre-processing software and export the input file, which is then converted into a .flac3d file readable by FLAC3D using an interface program. Define the generation class: create a class that includes simulation scheme parameters, initialization function and model reset method, model generation method, constitutive model and parameter assignment method, boundary condition loading method, monitoring scheme calling method and simulation sample library saving method; Write example code: Use the "python-reset-state false" command to reset the model. Use "zone import" to import the core zone and "zone create" to create the transition zone. Use the "zone cmodel" and "zone property" commands to assign constitutive parameters. Use the "zone history" command to record the monitoring results. Main program execution: import the itasca module, call the generation class instance method, and execute model reset, generation, parameter assignment, boundary loading, monitoring and sample library saving in sequence.

[0024] Specifically, the modeling scope is divided into a core area (complex geometry, fixed size) and a transition area (simple geometry, variable size). The core area is modeled with the help of external professional software to solve the problem of insufficient FLAC3D pre-processing capabilities, and the transition area is directly modeled using FLAC3D built-in commands, forming an efficient modeling mode of multi-software collaboration; by extracting coordinates and geometric parameters and abstracting them as simulation model parameters, standardized input data is provided for subsequent automated modeling, so that the size change of the transition area can be achieved through parameter adjustment, meeting the needs of batch generation of different simulation models, and then quickly building a diversified sample library; by fixing the size of the core area and modeling it independently, the geometric accuracy of the key areas of research focus is ensured; the parametric design of the transition area supports flexible adjustment of the modeling scope. The combination of the two ensures the accuracy of the model while achieving adaptive expansion of complex engineering scenarios (such as cross tunnels and chamber groups); FLAC3D has shortcomings in establishing complex models, such as poor performance in dealing with complex structures such as cross tunnels at close distances above and below. By using Abaqus, Midas The core area model is constructed by using software with powerful pre-processing functions such as GTS-NX, which can accurately model complex areas and ensure the accuracy and precision of the model; after the pre-processing software builds the model, it exports the input file in a specific format, and then converts it into a .flac3d file readable by FLAC3D through the interface program, so that the model data between different software can be circulated, allowing FLAC3D to read and process complex models created by other software, providing a basis for subsequent calculation and analysis; the obtained core area .flac3d file is the key part of building a complete simulation model, and combined with the transition area model subsequently created by the FLAC3D command, they together constitute various models in the simulation sample library, laying the foundation for the rapid generation of simulation sample libraries; by defining a generation class, the simulation scheme parameters, initialization functions and multiple instance methods are integrated into one class to form a complete set of simulation operations, among which the simulation scheme parameters include It covers multiple parameters such as simulation model, physical mechanics, boundary conditions, etc., and is the basic data for the entire simulation; the initialization function abstracts the initialization parameters from the modeling scheme design and instance method execution code, provides initial settings for subsequent operations, and ensures the consistency and accuracy of parameters in each link; various instance methods have different functions, and the model reset method ensures that the completed calculation results are not lost when the model is updated; the model generation method is responsible for importing or establishing the core area and transition area models, and generating a model with a complete modeling range; the constitutive model and parameter assignment method set the constitutive relationship and physical mechanics parameters of the model; the boundary condition loading method sets the boundary conditions such as stress, displacement, and temperature for the model; the monitoring scheme call method obtains the simulation result evaluation index; the save method saves the simulation sample library data locally. These methods work together to realize the automated operation from model construction to result saving, improve modeling efficiency and accuracy, and reduce manual intervention;The definition of the generated class makes the code well structured and modular, which is convenient for reuse in different projects or studies. When a new simulation study is needed, it is only necessary to adjust the parameters and methods in the class according to specific needs to quickly build a new simulation model, which has promoted the research on the combination of artificial intelligence algorithms and FLAC3D software and provided strong support for solving geotechnical engineering problems. The "python-reset-state false" command is used to reset the model to ensure that the completed calculation results will not be lost when the model is updated. This function is of great significance for continuous multi-group simulation calculations. It can effectively reduce repeated calculations, improve modeling efficiency, and ensure the continuity and efficiency of research. With the help of "zone import" to import the core area model and "zone creat" to create the transition area model, a complete simulation model covering the core area and transition area can be quickly built. This process is based on the previous modeling scheme design and core area model acquisition steps, realizing the transformation from design to actual model construction, and providing a basic model for subsequent simulation analysis. Using "zone cmodel" and "zone The "property" command is used to assign constitutive models and parameters, which can accurately set the constitutive relationship of the model (such as elastic constitutive model, Mohr-Coulomb constitutive model) and the corresponding physical and mechanical parameters (such as density, elastic modulus, Poisson's ratio, etc.); by accurately setting these parameters, the simulation model can be more consistent with the actual geotechnical engineering situation, and the accuracy and reliability of the simulation results can be improved; the calculation results of specific positions of the simulation model are monitored through the "zone history" command to obtain simulation result evaluation indicators such as characteristic point displacement, number of plastic zone units, unit stress difference, etc.; these data are the key basis for evaluating simulation effects and studying geotechnical engineering problems, and provide important support for subsequent analysis and decision-making; importing the itasca module is the basis for using the relevant functions of the FLAC3D software, and provides support for the subsequent calling of the FLAC3D command; calling the generation class instance method, the model reset, generation, parameter assignment, boundary loading, monitoring and sample library storage methods contained in the definition generation class are connected in series, so that each link is executed in the established order, ensuring the consistency and integrity of the entire simulation process; through this This approach avoids manual operation in the middle, improves modeling efficiency, and reduces errors that may be caused by manual intervention; executing these example methods in a fixed order forms a standardized simulation process; in different studies or projects, as long as the same settings and parameters are followed, similar simulation sample libraries can be repeatedly generated, ensuring the repeatability of research results and facilitating comparison and verification between different studies; this is crucial for promoting the application of artificial intelligence algorithms in geotechnical engineering, and provides a stable and reliable data set for algorithm training; by automating each link, the entire process from model construction, parameter setting, calculation and analysis to result storage is quickly completed, thereby quickly obtaining a simulation sample library;These sample libraries contain simulation scheme parameters and simulation result evaluation indicators, providing a rich training data set for artificial intelligence algorithms, helping to explore the general laws of research problems, promote the application of artificial intelligence technology in geotechnical engineering problems, and further enhance the intelligent level of geotechnical engineering.

[0025] In the modeling scheme design step, the core area and transition area are regular hexahedrons or parallelepipeds, the union of the two is the complete modeling range and the intersection is an empty set. The complexity of the geometric shape is determined by whether it can be directly modeled through the "zone creat" command of FLAC3D.

[0026] Specifically, the core area and transition area are set as regular hexahedron or parallelepiped, which unifies the regional shape and makes the model structure more regular and orderly. This not only reduces the complexity of modeling and facilitates the subsequent meshing, parameter setting and calculation analysis of the model, but also improves the model processing efficiency, enhances the calculation stability, and reduces the calculation error caused by irregular shape. The union of the two is stipulated as the complete modeling range and the intersection is an empty set, which clearly defines the modeling boundary and regional division. It ensures that the model covers the entire range required for the study and avoids the calculation conflict caused by regional overlap, ensuring that each region has clear independence and uniqueness in the model, which is conducive to accurately controlling and analyzing the impact of different regions on the overall model. Whether it can pass the "zone The "creat" command is used to directly model the geometric shape to determine its complexity, thereby dividing the core area and transition area. For complex areas (core areas) that cannot be directly modeled with this command, external professional pre-processing software is used to model them, giving full play to its powerful complex modeling capabilities. Simple areas (transition areas) are quickly modeled using FLAC3D's own commands. This division of labor improves modeling efficiency, ensures model accuracy, and effectively solves the problem of FLAC3D's weak pre-processing function.

[0027] The pre-processing software includes Abaqus, Midas GTS-NX, COMSOL, LS-DYNA and ANSYS. The input files are in .inp format, .msh format, .k format and .fpn format. The interface program is used to convert the mesh nodes, mesh, cell grouping and other information of the input files into the FLAC3D mesh definition format.

[0028] Specifically, FLAC3D has limitations when building complex models, while pre-processing software such as Abaqus, Midas GTS-NX, and ANSYS have powerful modeling capabilities and can handle complex structures such as closely spaced intersecting tunnels and high-level radioactive waste disposal chambers, accurately constructing core area models and ensuring the accuracy and completeness of the models. .inp, .msh, .k, and .fpn formats are specified as input file formats because they can effectively store key model information and facilitate data exchange between pre-processing software and interface programs. The interface program converts the mesh node and cell information in these files into FLAC3D's mesh definition format, achieving smooth transmission and compatibility of model data between different software and ensuring that FLAC3D can read and process complex models. By selecting appropriate pre-processing software and standardized input file formats, and using the interface program for format conversion, an efficient modeling process is formed. This avoids the difficulties of directly building complex models in FLAC3D, improves modeling efficiency, and ensures model quality, laying a solid foundation for subsequent simulation calculations and sample library generation.

[0029] The simulation scheme parameters are determined by the research object, including simulation model parameters, physical and mechanical parameters, boundary condition parameters, excavation method and support parameters. The simulation model parameters include chamber shape, chamber size, burial depth, vertical spacing of chambers, horizontal spacing of chambers, and number of chambers. The physical and mechanical parameters include density, elastic modulus, Poisson's ratio, bulk modulus, shear modulus, cohesion, internal friction angle, tensile strength, shear dilatancy angle, permeability coefficient, thermal conductivity coefficient, thermal expansion coefficient, specific heat capacity, heat release power, thermal stress coefficient, ground temperature, temperature limit, effective stress coefficient, and ground stress values in three directions. The excavation methods include full-section excavation, upper and lower step excavation, and pilot tunnel construction method.

[0030] Specifically, clarifying the simulation scheme parameters can comprehensively and meticulously depict the simulation scenario of geotechnical engineering; simulation model parameters such as chamber size and spacing directly determine the geometric shape and layout of the model; physical and mechanical parameters such as density, elastic modulus, and Poisson's ratio reflect the basic properties of geotechnical materials; boundary condition parameters stipulate the constraints and loading conditions of the model boundary; excavation methods and support parameters simulate the actual construction process; these parameters together construct a simulation environment that is highly consistent with the actual project, making the simulation results more reliable and valuable for reference; the parameterized setting method provides great flexibility for modeling; by adjusting these parameters, different simulation models can be quickly generated, and then A variety of simulation results can be obtained; for example, by changing the spacing between chambers and the excavation method, the impact of different layouts and construction methods on project stability can be studied; this parametric modeling method meets different research needs, provides a powerful means for in-depth exploration of geotechnical engineering problems, and also provides rich and diverse sample data for the simulation sample library; the large number of simulation results generated by rich and detailed simulation scheme parameters constitute the core content of the simulation sample library; these sample library data are the basis for the training of artificial intelligence algorithms. By learning from these data and excavating the laws therein, the algorithm can improve the prediction and analysis capabilities of geotechnical engineering problems and promote the application and development of artificial intelligence technology in the field of geotechnical engineering.

[0031] In the model generation method, the “zone reflect” command is used to mirror the transition zone model, and the “zone attach” command is used to bind the elements on the contact surface between the core zone and the transition zone. The mesh size is then adjusted to ensure that the difference between the two mesh sizes is less than a preset threshold.

[0032] Specifically, the “zone reflect” command is used to mirror the transition zone model, which can quickly generate a model area with a symmetrical structure. This operation significantly improves modeling efficiency, avoids the tedious process of repeatedly constructing similar areas, and at the same time ensures the symmetry of the model structure, which conforms to the geometric characteristics of many actual projects. The “zone attach” command is used to bind the units on the contact surface of the core zone and the transition zone, so that the two areas form a continuous mechanical transfer interface on the contact surface. This operation ensures that stress and displacement can be smoothly transmitted between the two areas during the simulation calculation process, avoiding calculation errors caused by interface discontinuity and improving the accuracy and reliability of the model. Adjusting the grid size of the core zone and the transition zone so that the difference is less than the preset threshold helps to optimize the grid quality of the entire model. The coordination of grid size can reduce the calculation instability caused by sudden changes in grid size, improve calculation efficiency and the accuracy of the results. At the same time, this also provides a good grid foundation for subsequent numerical calculations, ensuring the smooth progress of the calculation process.

[0033] In the constitutive model and parameter assignment method, the constitutive models include elastic-plastic constitutive model, strain softening / hardening model, nonlinear constitutive model and joint and interface model. The elastic-plastic constitutive model includes Mohr-Coulomb model and Drucker-Prager model. The strain softening / hardening model includes strain softening model and bilinear strain softening / hardening model. The nonlinear constitutive model is Hoek-Brown model. The joint and interface model includes bilinear Mohr-Coulomb joint model. The grid of the specified area is encrypted by the "zone densify" command, and the corresponding density, elastic modulus, Poisson's ratio, bulk modulus, shear modulus, cohesion, internal friction angle, tensile strength, shear dilatancy angle, permeability, thermal conductivity, thermal expansion coefficient, specific heat capacity, heat release power, thermal stress coefficient, ground temperature, temperature limit, effective stress coefficient, and ground stress values in three directions are assigned to different areas by the "zone property" command.

[0034] Specifically, the elastic constitutive model is suitable for simulating the behavior of geotechnical materials in the elastic deformation stage. The Mohr-Coulomb and Mohr-Coulomb constitutive models can take into account the plastic yield characteristics of geotechnical materials and are closer to the geotechnical mechanical behavior in actual engineering projects; the Drucker-Prager model is suitable for materials such as rocks, metals and concrete. It is mainly used to describe the yield behavior of materials under complex stress states and can consider the plastic deformation and hardening behavior of materials. The strain softening model is suitable for describing the situation where the strength of materials gradually decreases during deformation and is suitable for the failure process of materials such as rocks and concrete. The bilinear strain softening / hardening model is suitable for simulating the elastic-plastic deformation and failure process of materials. The Hoek-Brown model is suitable for describing the nonlinear strength characteristics of rocks and can consider the damage and failure process of rocks. The bilinear Mohr-Coulomb joint model is suitable for describing the mechanical behavior of joint surfaces and is mainly used to simulate the shear deformation and failure of joint surfaces. By using the "zone property" command to assign parameters such as density, cohesion, and internal friction angle to different regions, detailed settings can be made based on the actual characteristics of the geotechnical mass, enabling the model to accurately reflect the differences in mechanical properties of geotechnical materials in different regions and improve the authenticity of the simulation results. By using the "zone densify" command to densify the mesh in designated areas (such as core areas or stress concentration areas), higher computational accuracy can be achieved in critical areas, capturing more subtle mechanical responses and deformation characteristics. At the same time, relatively sparse meshes are maintained in non-critical areas, effectively reducing the amount of computation while ensuring computational accuracy, optimizing computational resource allocation, and improving overall computational efficiency. By assigning different constitutive models and parameters to different regions and combining them with mesh densification technology, geotechnical engineering scenarios with complex and changeable geological conditions can be simulated. For example, in the simulation of a high-level radioactive waste disposal chamber group, different parameters can be set for different rock layers or structural surfaces to accurately predict engineering behavior and provide a reliable basis for actual engineering decision-making.

[0035] In the boundary condition loading method, the "zone face" command is used to set stress or displacement conditions for the model surface, and the "zone gridpoint" command is used to set stress or displacement conditions for the model grid nodes.

[0036] Specifically, by setting stress boundary conditions through the "zone face" command, the external pressure (such as groundwater pressure, soil pressure, etc.) on the rock and soil can be simulated, so that the model can truly reflect the stress state in the actual engineering; using the "zonegridpoint" command to set displacement boundary conditions, the displacement of specific positions of the model can be constrained (such as fixed boundaries) to ensure the mechanical rationality of the model; reasonable setting of boundary conditions is the key to the convergence of numerical calculations; by accurately applying stress and displacement boundary conditions, unreasonable displacement or stress concentration of the model can be avoided during the calculation process, ensuring the stability and convergence of the calculation process and improving the reliability of the simulation results; combining multiple boundary condition application methods, the coupled analysis of multiple physical fields in geotechnical engineering can be realized; for example, by considering the interaction between stress field and temperature field at the same time, the complex physical phenomena in actual engineering can be more comprehensively simulated, providing a more accurate basis for engineering design and safety assessment.

[0037] In the monitoring scheme calling method, the monitored simulation result evaluation indicators include characteristic point displacement, number of plastic zone elements and element stress difference. The monitoring variables are recorded through the "zone history" and "fish history" commands.

[0038] Specifically, by monitoring evaluation indicators such as characteristic point displacement, the number of plastic zone elements, and element stress difference, key mechanical responses of geotechnical engineering during stress or excavation can be directly obtained. Characteristic point displacement reflects the degree of structural deformation, the number of plastic zone elements indicates the potential range of damage, and the element stress difference reflects the degree of stress concentration within the material. These indicators provide a quantitative basis for assessing engineering stability and safety, helping engineers to promptly identify potential risks. Using the "zone history" and "fish history" commands to record monitoring variables, monitoring content can be customized to meet specific research needs. For example, monitoring stress paths, energy release rates, or relative displacements of structural contact surfaces in specific areas can be used. This flexibility enables the model to capture complex mechanical behavior and engineering details, providing richer data support for in-depth research on geotechnical engineering problems. The large amount of simulation result evaluation indicators and customized variable data obtained through monitoring constitute an important component of the simulation sample library. This data can be used to train artificial intelligence algorithms, explore potential laws in geotechnical engineering, and establish predictive models, thereby realizing intelligent analysis and prediction of engineering behavior and promoting the digital transformation of the geotechnical engineering field.

[0039] In the simulation sample library saving method, the open function is used to open the file in text or CSV format, the simulation scheme parameters and the corresponding simulation result evaluation indicators are written to the file through the write method of the file object, and the file is closed through the close function.

[0040] Specifically, through the combination of open, write and close functions, the simulation scheme parameters (such as chamber size, physical and mechanical parameters) and corresponding evaluation indicators (such as characteristic point displacement, number of plastic zone units) are written to files in text or CSV format to ensure that the data can be preserved for a long time; this method avoids the loss of temporary data in memory and provides a reliable basis for subsequent analysis and reuse; the structured storage in CSV format makes the data readable and scalable; the data of different simulation experiments are stored in a unified format, which is convenient for building a standardized sample library and supports batch processing and cross-platform sharing; the openness of text files also allows the data to be directly read and processed by various data analysis tools; the structured stored simulation data constitutes a high-quality training set that can be directly used for training machine learning models; the correspondence between parameters and results provides a clear input-output mapping for the algorithm, helping the algorithm to learn the complex laws in geotechnical engineering, realize accurate prediction and optimization design of engineering behavior, and promote the application of artificial intelligence in the geotechnical field.

[0041] The definition of the generation class also includes Plot settings and saving methods. The post-processing view is created through the "plot create" command, the view parameters are set using the "plot view" and "plot legend" commands, and the view is saved as an image file through the "plot export" command.

[0042] Specifically, by creating a post-processing view through "plot create" and using "plot view" and "plot legend" to set parameters such as viewing angle, scale, color mapping, etc., abstract numerical calculation results (such as stress distribution, displacement field, and plastic zone) can be converted into intuitive graphic images; this visualization method allows researchers to quickly identify key features (such as stress concentration areas and deformation trends) and improve analysis efficiency; using "plot export" to save the view as an image file ensures that the simulation results can be output in a unified and standardized format; this not only facilitates the writing of research reports and papers, but also supports cross-departmental or cross-project data sharing, making the presentation of results more professional and standardized; the saved image file records the simulation results under specific parameter configurations, providing a visual traceability basis for the research process; by comparing images under different parameter settings, the sensitivity and reliability of the model can be verified, assisting in optimizing the simulation plan and ensuring the scientific nature of the research conclusions.

[0043] Step 1: Modeling scheme design 1. Determine the modeling scope based on the characteristics of the research question and divide it into a finite number of core and transition zones. These zones are typically designed as regular hexahedrons or parallelepipeds. When defining the modeling scope, set a specific location as the coordinate origin. Based on the resulting modeling scope, obtain the coordinate information for all core and transition zones. Next, compile the modeling file to provide the data.

[0044] 2. The union of all core areas and transition areas is the complete modeling range, and the intersection of any two areas is an empty set.

[0045] 3. The area with complex geometric shapes within the modeling range is defined as the core area. The geometric dimensions of this area are fixed and do not change with changes in the design plan. This area is often the core area of research focus.

[0046] 4. Define the modeling range outside the core area as the transition zone. The geometry of this area is generally not complex. By changing the geometric dimensions of the transition zone, the modeling range can be changed to obtain different simulation models.

[0047] 5. The basis for judging whether a region is complex is whether it can be successfully modeled using the "zone creat" command of the FLAC3D software. If so, the region is a simple region, otherwise it is a complex region.

[0048] 6. Abstract the simulation model parameters in the simulation scheme parameters from the geometric parameters of the core area model and the transition area model to automatically create a new simulation model.

[0049] 7. The core area simulation model was created using other simulation software. This software generally has powerful pre-processing capabilities, not only enabling the core area simulation model to be built but also successfully exporting input files. For example, Abaque's pre-processing capabilities are robust and can export .inp files; Midas's pre-processing capabilities are robust and can export .fpn files; and ANSYS's pre-processing capabilities are robust and can export .msh files. These can all be opened using Notepad or Wordpad.

[0050] 8. If there are multiple core areas with the same geometric shape, and they can overlap with each other through position translation in three-dimensional space, then only one core area simulation model needs to be built. Otherwise, multiple core area models that cannot overlap with each other need to be built.

[0051] 9. When importing the core area simulation model into FLAC3D software, you only need to ensure that its geometric dimensions are consistent with Article 3. You do not need to pay attention to the grid node coordinates, numbers, and number of grids.

[0052] 10. Transition zone modeling: Use the "zone create" command in FLAC3D software to create the model. Since the element sizes at the interface between the core and transition zone models are different, the "zone attach" command can be used to bind the elements at the interface. However, the difference in mesh size between the two models should be minimized.

[0053] Step 2: Get the .flac3d file of the core area model 1. Use appropriate methods or simulation software to model the core area. These methods include but are not limited to: Rhino and Griddle coupling, ANSYS, Abaqus, Midas GTS-NX, etc. After building the grid model, export the input file.

[0054] 2. Use the interface program to convert the input file into a .flac3d file that can be read by the FLAC3D software, and use the "zone import" command to confirm the imported core zone model.

[0055] Step 3: Define the class for quick generation of the simulation sample library (class Simulation_Sample_Library:) 1. The simulation sample library quickly generates classes including: class attributes, initialization functions and instance methods.

[0056] 2. Class attributes mainly involve attributes related to the sample library, which are mainly used for reading and calling simulation results, such as simulation scheme parameters and simulation evaluation indicators, sample result storage arrays, etc.

[0057] 3. The initialization function primarily defines initialization parameters, which are abstracted from the first step (modeling scheme design) and the fourth step (writing the implementation code for the instance method). The content and number of initialization parameters to be set depend on the complexity of the research problem, such as transition zone modeling parameters, the location of fixed boundary loads, and the location of monitoring points. The initialization function can be passed a file path, and the file content can be designed as a dictionary data type.

[0058] 4. Example methods include: (1) Model reset method (Model_Reset); (2) Plot setting and saving method (Plot_Set_Save); (3) Model generation method (Model_Generation); (4) Constitutive model and parameter assignment method (Constitutive_Model_Parameter); (5) Boundary condition loading method (Boundary_Load); (6) Model initial state setting method (Initial_Set); (7) Monitoring scheme calling method (Monitor_Scheme); (8) Calculation start method (Solve_Start); (9) FISH language calling method (FISH_Language); (10) Simulation sample library saving method (Save_Simulation_Sample_Library), etc.

[0059] Step 4: Write the execution code of the instance method 1. Model reset method. This method ensures that the calculated model results are not lost after the model is updated. It includes the "python-reset-state false" command.

[0060] 2. Plot setting and saving method. This method is to set the display of post-processing results and export pictures, including "model title <s>”、"plot <s>export keyword”、"plot create <s>”、"plot <s>view keyword ...”、"plot <s>legend keyword ..." and other commands 3. Model generation method. This method is used to import or establish core zone and transition zone models. This method can generate all models within the modeling range, including "zone import s <s2>keyword”、"zonecreate keyword”、"zone reflect keyword ... <range>" and other commands.

[0061] 4. Constitutive model and parameter assignment method. This method involves setting the constitutive model and the corresponding physical and mechanical parameters. For example, setting the elastic constitutive model and its associated physical and mechanical parameters, setting the Mohr-Coulomb constitutive model and its associated physical and mechanical parameters, includes "model configure keyword," "zone cmodel keyword," "zonefluid keyword," "zone thermal keyword," "zone creep keyword," "zone dynamic keyword," "zone densify keyword..." <range>”、"zone property <density f >keyword ... <range>" and other commands.

[0062] 5. Boundary condition loading method. This method is to set the boundary conditions of the simulation model, such as setting stress boundary conditions, setting displacement boundary conditions, setting temperature boundary conditions, etc., including "zone face skin<keyword... > <range>”、"zone face apply keyword <range>”、"zone face apply-remove <keyword> <range>”、"zone gridpoint fix keyword <range>" and other commands.

[0063] 6. Model initial state setting method. This method is to initialize the simulation model, such as initial ground stress field, initial temperature field, etc., including "zone initialize keyword <range>”、"zone gridpointinitialize keyword <range>" and other commands.

[0064] 7. Monitoring scheme call method. This method is to monitor the simulation calculation results of certain positions of the simulation model. These results generally include the simulation result evaluation indicators and the content that needs to be displayed in the Plot setting and saving method, such as: the displacement, stress, temperature of a certain point, the number of plastic units within a certain range, including "zone history<name s > keyword", "fish history<name s > sym" and other commands.

[0065] 8. Calculation startup method. This method configures and switches the calculation mode and starts the calculation, including "model mechanical keyword”、"model thermal keyword”、"model fluid keyword", "model dynamic keyword", "model solve keyword...", "model cycle i<calm i2 > " and other commands.

[0066] 9. FISH language calling method. This method designs the FISH functions required for calculations. These functions are used by other related methods in this step. The specific content of this method must not only comply with the usage rules of the FISH language but also make appropriate use of relevant built-in variables based on the specific engineering problem. The FISH language calling method may contain multiple calling files, each of which contains multiple FISH functions.

[0067] 10. Simulation sample library saving method. This method saves the simulation sample library data locally. Use the open function to open a file, for example: file = open('example.txt', 'w'). Use the file object's write method to write data to the file, for example: file.write('Hello, this is a test file.\n'). Use the file object's close method to close the file, for example: file.close().

[0068] Step 5: Write the main program execution entry file 1. The main program execution entry file is the starting point of the program, organizing and calling other modules or functions.

[0069] 2. The module must use the "import itasca as it" command to import the itasca module.

[0070] 3. To illustrate the overall concept of this patent in more detail, an example of writing a main program execution entry file (main.py) is given below, as shown below: import itasca as it from fileneame import Simulation_Sample_Library def main(): #Call the function in the module and output the result. This part is just an example. # Different projects are different and require flexible adjustments.

[0071] Sample_Library = Simulation_Sample_Library(filepath) Sample_Library.Model_Reset(parameter) Sample_Library.Plot_Set_Save(parameter) Sample_Library.Model_Generation(parameter) Sample_Library.Constitutive_Model_Parameter(parameter) Sample_Library.Boundary_Load(parameter) Sample_Library.Initial_Set(parameter) Sample_Library.Monitor_Scheme(parameter) Sample_Library.Solve_Start(parameter) Sample_Library.FISH_Language(parameter) Sample_Library.Save_Simulation_Sample_Library(parameter) if __name__ == "__main__": main() While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents. < / range> < / range> < / range> < / range> < / keyword> < / range> < / range> < / range> < / range> < / range> < / s> < / s> < / s> < / s> < / s>

Claims

1. A method for quickly generating a simulation sample library, characterized in that: The method is executed by a computer system and includes the following steps: Modeling scheme design: Divide the modeling range into the core area and the transition area, extract the coordinates and geometric parameters of the two, and abstract the simulation model parameters; Obtain core area model file: Model the core area using pre-processing software and export the input file, which is then converted into a .flac3d file readable by FLAC3D using an interface program. Define the generation class: create a class that includes simulation scheme parameters, initialization function and model reset method, model generation method, constitutive model and parameter assignment method, boundary condition loading method, monitoring scheme calling method and simulation sample library saving method; Write example code: Use the "python-reset-state false" command to reset the model, use "zoneimport" to import the core zone and "zonecreate" to create the transition zone, use the "zonecmodel" and "zoneproperty" commands to assign constitutive parameters, and use the "zonehistory" command to record the monitoring results. Main program execution: import the itasca module, call the generation class instance method, and execute model reset, generation, parameter assignment, boundary loading, monitoring and sample library saving in sequence.

2. The method for rapidly generating a simulation sample library according to claim 1, wherein: In the modeling scheme design step, the core area and transition area are regular hexahedrons or parallelepipeds, the union of the two is the complete modeling range and the intersection is an empty set. The complexity of the geometric shape is determined by whether it can be directly modeled through the "zone create" command of FLAC3D.

3. The method for rapidly generating a simulation sample library according to claim 1, wherein: The pre-processing software includes Abaqus, Midas GTS-NX, COMSOL, LS-DYNA and ANSYS. The input files are in .inp format, .msh format, .k format and .fpn format. The interface program is used to convert the grid nodes, grids, unit grouping and other information of the input files into the grid definition format of FLAC3D.

4. The method for rapidly generating a simulation sample library according to claim 1, wherein: The simulation scheme parameters are determined by the research object, including simulation model parameters, physical and mechanical parameters, boundary condition parameters, excavation method and support parameters. The simulation model parameters include chamber shape, chamber size, burial depth, vertical spacing of chambers, horizontal spacing of chambers, and number of chambers. The physical and mechanical parameters include density, elastic modulus, Poisson's ratio, bulk modulus, shear modulus, cohesion, internal friction angle, tensile strength, shear dilation angle, permeability coefficient, thermal conductivity coefficient, thermal expansion coefficient, specific heat capacity, heat release power, thermal stress coefficient, ground temperature, temperature limit, effective stress coefficient, and ground stress values in three directions. The excavation methods include full-section excavation, upper and lower step excavation, and pilot tunnel construction method.

5. The method for rapidly generating a simulation sample library according to claim 1, wherein: In the model generation method, the "zone reflect" command is used to mirror the transition zone model, the "zone attach" command is used to bind the cells on the interface between the core zone and the transition zone, and the mesh size is adjusted so that the difference between the two mesh sizes is less than a preset threshold.

6. The method for rapidly generating a simulation sample library according to claim 1, wherein: In the constitutive model and parameter assignment method, the constitutive model includes an elastic-plastic constitutive model, a strain softening / hardening model, a nonlinear constitutive model, and a joint and interface model. The elastic-plastic constitutive model includes a Mohr-Coulomb model and a Drucker-Prager model. The strain softening / hardening model includes a strain softening model and a bilinear strain softening / hardening model. The nonlinear constitutive model is a Hoek-Brown model. The joint and interface model includes a bilinear Mohr-Coulomb joint model. The grid of the specified area is encrypted by the "zone densify" command, and the corresponding density, elastic modulus, Poisson's ratio, bulk modulus, shear modulus, cohesion, internal friction angle, tensile strength, shear dilatancy angle, permeability, thermal conductivity, thermal expansion coefficient, specific heat capacity, heat release power, thermal stress coefficient, ground temperature, temperature limit, effective stress coefficient, and ground stress values in three directions are assigned to different areas by the "zone property" command.

7. The method for rapidly generating a simulation sample library according to claim 1, wherein: In the boundary condition loading method, the "zone face" command is used to set stress or displacement conditions for the model surface, and the "zone gridpoint" command is used to set stress or displacement conditions for the model grid nodes.

8. The method for rapidly generating a simulation sample library according to claim 1, wherein: In the monitoring scheme calling method, the monitored simulation result evaluation indicators include characteristic point displacement, the number of plastic zone elements and element stress difference, and the monitoring variables are recorded through the "zone history" and "fish history" commands.

9. The method for rapidly generating a simulation sample library according to claim 1, wherein: In the simulation sample library saving method, the open function is used to open the file in text or CSV format, the simulation scheme parameters and the corresponding simulation result evaluation indicators are written into the file through the write method of the file object, and the file is closed through the close function.

10. The method for rapidly generating a simulation sample library according to claim 1, wherein: The definition generation class also includes Plot setting and saving methods, creating a post-processing view through the "plot create" command, using the "plotview" and "plot legend" commands to set view parameters, and saving the view as an image file through the "plot export" command.

Images