A self-programming based local topography difference RVE and modeling method and application thereof

Abstract

The application discloses a local morphology difference RVE model based on self-programming and a modeling method and application thereof. The method removes surface residual stress of a target alloy, obtains a crystallographic data set of the target alloy surface through a scanning electron microscope, obtains a feature distribution map of the target alloy through visual analysis, analyzes the obtained crystallographic data to obtain a data set A containing phase numbers and a data set B containing phase positions, numbers and Euler angles, calls a damask library to read the data set B, outputs grain numbers of a region of interest in the data set B into a data set C in an ASCII format, reads the data set A and the data set C, obtains a data set D with local morphology difference through differential editing of the region of interest, and the local morphology difference RVE model is obtained. The model is used for crystal plastic finite element simulation of local mechanical response, a comparable local morphology difference RVE model is constructed, and technical support is provided for exploring the influence of grain morphology on material deformation behavior.

Description

Technical Field

[0001] The present invention relates to an alloy crystal plasticity model, in particular to a local morphology difference RVE based on self-programming and a modeling method and application thereof, belonging to the technical field of alloy numerical simulation analysis. Background Art

[0002] Crystal plasticity is a theory that describes the deformation mechanism of crystals based on the framework of continuum mechanics. It is an important theoretical basis for finite element analysis of crystal plasticity. Crystal plasticity assumes that crystalline materials undergo elastic deformation and rotation, while plastic deformation is caused by the movement of dislocations along the crystal slip system. It relates the deformation mechanism of the material to the macroscopic mechanical response. The finite element analysis method is a numerical technique for solving partial differential equations. It discretizes the study area into multiple grids and solves the approximate solution at each grid point. Finite element analysis of crystal plasticity combines the above two theories and is widely used to simulate the deformation behavior of crystalline materials, predict and analyze data such as microscopic strain, stress values ​​and damage-sensitive areas at the grain scale under different loading conditions, such as compression, tension and cyclic loading. At the same time, the finite element method of crystal plasticity can also be combined with other multiscale simulation methods, such as molecular dynamics simulation and macroscopic mechanics simulation, to achieve multiscale simulation and analysis from micro to macro.

[0003] An accurate crystal plasticity model and a representative volume element (RVE) are key to crystal plasticity modeling. Domestic and international scholars have conducted extensive research on how to establish a realistic RVE. Currently, the main methods include the Voronoi method, open source software methods (Neper, DREAM3D), and microstructure image-based methods. The former two methods mostly output regular polygonal grain morphologies, resulting in poor grain reproducibility. Microstructure image-based methods, on the other hand, can more accurately reflect the characteristic information of crystals in the microstructure, such as crystal size, distribution, and grain boundary state, and are therefore gradually becoming more widely adopted.

[0004] Chinese patent (CN112712860A) discloses a grain finite element modeling method based on the structure of a real metallographic image. The patent extracts the grain position and shape diagram by grayscale binarization processing on the metallographic photo. However, the image processing process in the patent is too cumbersome, and the distortion of the obtained results is relatively serious. Chinese patent (CN118395786A) discloses a crystal plasticity finite element modeling method and application based on EBSD data. On the basis of the original, the patent derives the grain boundary contour based on EBSD data, and then further stretches and assigns material properties in ABAQUS software. However, although this method can obtain more crystal structure information in the alloy, its modifiability is poor, and it is impossible to perform effective differential editing on the local part of the alloy, and it is difficult to accurately and truly reflect the impact of the local morphology differences of the alloy on the mechanical response. Summary of the Invention

[0005] To address the challenges of the prior art, the first objective of the present invention is to provide a self-programming RVE modeling method for local morphological differences. This method, based on a crystallographic dataset of alloy surfaces acquired through EBSD, combines self-programming scripts with open-source software to achieve localized differential editing of the alloys, constructing comparable RVE models with morphological differences.

[0006] A second objective of the present invention is to provide a self-programming RVE model for local morphology differences. This model, derived from the aforementioned method, significantly reduces operating costs while ensuring that the subject's microstructural information is derived from real materials, providing technical support for exploring the influence of grain morphology on material deformation behavior.

[0007] A third objective of the present invention is to provide an application of a self-programmed local morphology differential (RVE) modeling method for crystal plasticity finite element analysis of the local micromechanical response of alloys. This method, with its superior differential editing properties, enables accurate simulation of alloy microdeformation. Using this model for local crystal plasticity finite element analysis of alloys, it accurately reproduces the grain-scale micromechanical response of alloys with varying morphologies under varying strains.

[0008] In order to achieve the above technical objectives, the present invention provides a local morphology difference RVE modeling method based on self-programming, comprising:

[0009] Step S1, removing residual stress from the surface of the target alloy, obtaining a surface crystallographic data set related to the target alloy through a scanning electron microscope, and then obtaining a characteristic distribution map of the target alloy through visual analysis;

[0010] Step S2, performing source data analysis on the target alloy surface crystallography related files obtained in step S1 to obtain a data set A containing phase numbers and a data set B containing phase positions, numbers, and Euler angles;

[0011] Step S3, calling the damask library to read the dataset B obtained in step S2, and outputting the grain numbers of the region of interest therein as dataset C in ASCII format;

[0012] Step S4: read the dataset A and the dataset C, and obtain the dataset D with local morphological differences by differentially editing the region of interest, that is,

[0013] The differential editing method is to replace or add specific phase or grain numbers in the region of interest.

[0014] Differential editing is one of the important innovations of the model of the present invention. It keeps the main information from the experimental data unchanged and only changes the local grain or grain boundary state. The RVE constructed by this method not only has morphological differences but also has strong comparability.

[0015] As a preferred solution, the method for removing the surface residual stress is: using 80-3000 mesh sandpaper to grind and polish step by step, and then performing rough polishing and sub-ion polishing in sequence.

[0016] As a preferred solution, the target alloy surface crystallography-related dataset is obtained by a SEM electron microscope equipped with an EBSD probe, which consists of .cpr, .crc and .ctf files containing alloy surface grain morphology, grain orientation and grain position.

[0017] As a preferred solution, the target alloy characteristic distribution diagram includes a pole figure, an inverse pole figure and a phase distribution diagram.

[0018] As a preferred solution, the target alloy characteristic distribution map is obtained by visualization using AZtecCrystal analysis software.

[0019] As a preferred solution, before analyzing the source data in step S2, it is also necessary to load the target alloy surface crystallography-related dataset using DREAM3D software.

[0020] As a preferred solution, the process of source data analysis is: performing noise reduction processing on the original data and extracting characteristic data such as grains, phases and Euler angles.

[0021] As a preferred solution, the process of calling the damask library to read the data set B is: calling damask.Grid.load_DREAM3D to read the data set B through a self-programmed Python script.

[0022] As a preferred solution, the differential editing process is performed by reading dataset A and dataset C through a self-programmed Matlab script.

[0023] As a preferred solution, the dataset D can also be visually edited. The visual editing process is: calling damask.Grid.load_DREAM3D through a self-programmed Python script to read the dataset D, and then using AZtecCrystal analysis software to obtain visualization.

[0024] The present invention also provides a local morphology difference RVE model based on self-programming, which is obtained by any of the above-mentioned modeling methods.

[0025] The present invention also provides an application of a local morphology difference RVE model based on self-programming, which is used for crystal plasticity finite element analysis of local micromechanical response of alloys.

[0026] Compared with the prior art, the beneficial technical effects of the technical solution of the present invention are:

[0027] 1) The differentiated model provided by the present invention is based on the alloy surface crystallography-related dataset collected by EBSD. By combining self-programming scripts with open source software, it achieves local differentiated editing of the alloy and constructs a comparable RVE model with morphological differences. On the one hand, this model greatly reduces operating costs, and on the other hand, it ensures that the main microstructure information is derived from real materials, providing technical support for exploring the influence of grain morphology on material deformation behavior.

[0028] 2) The technical solution provided by the present invention is based on the excellent differentiated editing characteristics of the above-mentioned model, which can realize the accurate simulation of the microscopic deformation of alloy materials. The model is used for the crystal plasticity finite element analysis of the local micromechanical response of the alloy, and can realize the accurate reproduction of the micromechanical response of the grain scale of alloys with different morphologies under different strain amounts. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is a characteristic distribution diagram of the target alloy provided in Example 1 of the present invention;

[0030] in, Figure 1 (a) is the IPF diagram, Figure 1 (b) is the phase distribution diagram, Figure 1 (c) is the tungsten phase pole diagram, Figure 1 (d) is the pole figure of the binder phase;

[0031] Figure 2 The three RVE models and their corresponding pole figures provided in Example 1 of the present invention;

[0032] Figure 2 (a) and Figure 2 (d) RVE model and pole figure containing grain boundaries and twin boundaries in the binder phase; Figure 2 (b) and Figure 2 (e) RVE model and pole figure without grain boundaries in the binder phase; Figure 2 (c) and Figure 2 (f) RVE model and pole figure of tungsten particles added to the binder phase;

[0033] Figure 3 The strain ε obtained by CPFEM simulation when the compression strain of the alloys with different morphologies is 25% in Example 1 of the present invention is VM Distribution cloud map; Figure 3 (a) is the strain ε of GB model VMDistribution cloud map, Figure 3 (b) is the strain ε of the NGB model VM Distribution cloud map, Figure 3 (c) is the strain ε of the NGB-A model VM Distribution cloud map;

[0034] Figure 4 is the ε of the tungsten phase and the binder phase of the three alloys with different morphologies under different strains in Example 1 of the present invention. VM Distribution curve graph;

[0035] Figure 4 (a) is the strain ε when the strain is 6% VM Distribution curve graph, Figure 4 (b) is the strain ε when the strain is 12% VM Distribution curve graph, Figure 4 (c) is the ε when the strain is 25% VM Distribution curve graph;

[0036] Figure 5 The ε inside the tungsten phase and the binder phase of the three alloys with different morphologies under different strains in Example 1 of the present invention VM strain line graph;

[0037] Figure 5 (a) ε inside the tungsten phase and the binder phase VM Line graph of maximum strain value, Figure 5 (b) ε inside the tungsten phase and the binder phase VM Line graph of the strain mean values. DETAILED DESCRIPTION

[0038] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] Example 1

[0040] The present invention provides a local morphology difference RVE modeling method based on self-programming, comprising:

[0041] Step S1: Cut the tungsten alloy sample prepared by liquid phase sintering, grind and polish it with 80#, 200#, 400#, 800#, 1500# and 3000# sandpaper, then perform rough polishing, and then obtain a high-flatness EBSD test sample through sub-ion polishing. In the SEM electron microscope equipped with an EBSD probe, a 450μm×450μm area is selected, the scanning step length is 0.5μm, and the detection phases are selected as pure tungsten phase and bonding phase. After the detection is completed, the data file is exported to obtain .cpr, .crc and .ctf files, and the AZtecCrystal analysis software is used to visualize the obtained results. Figure 1 The IPF diagram, phase distribution diagram, and pole figure are shown.

[0042] Step S2: Load the .ctf file using DREAM3D software. Filter and remove noise from the source data, and output A.txt (containing the phase number) and B.dream3d (containing information such as the phase position, number, and Euler angles).

[0043] Step S3: Write a Python script that calls the damask library and uses damask.Grid.load_DREAM3D to read the data in B.dream3d. The script then outputs the grain numbers of the region of interest in ASCII format to the file C.txt. This file contains the numbers of the different tungsten and binder phase grains in the selected region.

[0044] Step S4: Write a Matlab script to read A.txt and C.txt, replace all the random grain numbers representing the binder phase with a single grain number, output the grain number to get the D.txt file, write a Python script, call the damask library, use damask.Grid.load_ASCII to read the data in D.txt respectively, and use Paraview software to visualize the D.vti file, as shown in the following example: Figure 2 (b) is shown and named as NGB.

[0045] Example 2

[0046] This embodiment is identical to embodiment 1, except that: Step S4: Write a Matlab script to read A.txt and C.txt, replace all random grain numbers representing the binder phase with a single grain number, select an appropriate binder phase region position, set a suitable diameter, set the position as the grain number of the tungsten phase, output a D.txt file, write a Python script, call the damask library, use damask.Grid.load ASCII to read the data in D.txt respectively, and use Paraview software to visualize the D.vti file, as shown in FIG. Figure 2(C) and named NGB-A.

[0047] Comparative Example 1

[0048] This comparative example is exactly the same as Example 1, except that: Step S4: Write a Matlab script to read A.txt and C.txt without deletion or addition, output the grain numbers to obtain D.txt file, write a Python script to call the damask library, use damask.Grid.load_ASCII to read the data in D.txt respectively, and use Paraview software to visualize the D.vti file, as shown in FIG. Figure 2 (A) and named GB.

[0049] By comparison Figure 2 (d) and Figure 1 The pole figures in the two phases are consistent, indicating that the established RVE model reproduces the grain orientation and distribution in the experiment. When the random grain numbers of the binder phase are replaced with single grains, the binder phase orientation becomes single. On this basis, when tungsten grains are added, the pole figures of the two phases are obtained as shown in the figure below. Figure 2 As shown in (f), the phenomenological crystal plasticity model is used to assign material parameters to the tungsten phase and the binder phase, respectively, and compress them along the horizontal direction.

[0050] Figure 3 is the local Mises strain ε of the three RVE alloys when the compressive strain is 25% VM The cloud plot shows significant strain concentration on the binder phase side near the tungsten / binder phase interface, with the softer binder phase bearing greater plastic deformation. In the GB alloy, due to the smaller binder phase grains and the presence of numerous binder phase grain boundaries and twin boundaries, dislocation motion is hindered during deformation, leading to more pronounced strain concentration.

[0051] Figure 4 The ε in the tungsten phase and binder phase of the three alloys were quantified at deformations of 6%, 12%, and 25%. VM As the deformation progresses, the ε VM The strain distribution changes from a tall and thin type to a short and fat type, and the GB alloy has the greatest degree of heterogeneity.

[0052] Figure 5 The ε in the tungsten phase and the binder phase of three alloys (GB, NGB and NGB-A) under different strains are shown. VMThe maximum (a) and average (b) of strain. The maximum value law is as follows: tungsten phase GB > NGB-A > NGB, binder phase GB > NGB-A > NGB. The average value law is as follows: tungsten phase NGB > NGB-A > GB, binder phase NGB > NGB-A > GB. The existence of the binder phase grain boundary makes the W / γ deformation heterogeneity intensified, and the binder phase bears more deformation. With the increase of tungsten content, the proportion of binder phase strain increases, and the local strain maximum value increases.

Claims

1. A local morphology difference RVE modeling method based on self-programming, characterized in that: include: Step S1, removing residual stress from the surface of the target alloy, obtaining a surface crystallographic data set related to the target alloy through a scanning electron microscope, and then obtaining a characteristic distribution map of the target alloy through visual analysis; Step S2, performing source data analysis on the target alloy surface crystallography data obtained in step S1 to obtain a data set A containing phase numbers and a data set B containing phase positions, numbers, and Euler angles; Step S3, calling the damask library to read the dataset B obtained in step S2, and outputting the grain numbers of the region of interest therein as dataset C in ASCII format; Step S4: read the dataset A and the dataset C, and obtain the dataset D with local morphological differences by differentially editing the region of interest, that is, The differential editing method is to replace or add specific phase or grain numbers in the region of interest.

2. The method for modeling local topography differences based on self-programming according to claim 1, characterized in that: The method for removing the surface residual stress is: using 80-3000 mesh sandpaper to grind and polish step by step, and then performing rough polishing and sub-ion polishing in sequence.

3. The method for modeling local topography differences based on self-programming according to claim 1, characterized in that: The target alloy surface crystallography-related data set is obtained by a SEM electron microscope equipped with an EBSD probe, which consists of .cpr, .crc and .ctf files containing alloy surface grain morphology, grain orientation and grain position.

4. The method for modeling local topography differences based on self-programming according to claim 1, characterized in that: The target alloy characteristic distribution diagram includes a pole figure, an inverse pole figure and a phase distribution diagram; the target alloy characteristic distribution diagram is obtained by visualization through AZtecCrystal analysis software.

5. The method for modeling local topography differences based on self-programming according to claim 1, characterized in that: Before the source data analysis in step S2, it is necessary to load the target alloy surface crystallography-related data set through DREAM3D software; the process of the source data analysis is: denoising the original data and extracting grain, phase and Euler angle feature data.

6. The method for modeling local topography differences based on self-programming according to claim 1, characterized in that: The process of calling the damask library to read data set B is: calling damask.Grid.load_DREAM3D to read data set B through a self-programmed Python script.

7. The method for modeling local topography differences based on self-programming according to claim 1, characterized in that: The differential editing process is performed by reading dataset A and dataset C using a self-programmed Matlab script.

8. The method for modeling local topography differences based on self-programming according to claim 1, characterized in that: The dataset D can also be visually edited. The visual editing process is as follows: calling damask.Grid.load_DREAM3D through a self-programmed Python script to read the dataset D, and then using AZtecCrystal analysis software to obtain visualization.

Images