A subsurface reconstruction method for bearing steel considering phase distribution characteristics
Through the combination of electron probe and electron backscattering diffraction pattern testing combined with image processing and probability density function fitting, the three-stage reconstruction of the secondary surface of the bearing steel is achieved, solving the problems of poor reconstruction performance and insufficient generalization ability in the existing technology. , improves the accuracy and adaptability of reconstruction.
Patent Information
- Application Number
- CN202410582121.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-05-11
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2044-05-11
AI Technical Summary
The existing subsurface reconstruction methods cannot fully reproduce the volume and distribution characteristics of the metal matrix, carbides and inclusions of bearing steel, resulting in poor reconstruction performance and insufficient generalization ability.
The image data is obtained by using electron probe test and electron backscattering diffraction pattern test, and through image processing and probability density function fitting, an irregular phase real shape database and Voronoi diagram are constructed to realize the three-stage reconstruction of the secondary surface of the bearing steel, including Accurate reconstruction of substrate, irregular phase and regular phase.
It improves the accuracy and authenticity of the secondary surface reconstruction of bearing steel, enhances the generalization ability of the reconstruction scheme, and adapts to the secondary surface reconstruction requirements of bearing steel of different components and forms.
Smart Images

Figure CN118518634B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of material sub-surface numerical reconstruction, and in particular relates to a bearing steel sub-surface reconstruction method taking phase distribution characteristics into consideration. Background Art
[0002] As one of the important components of aircraft engines, the main shaft bearing plays a vital role in the thrust-to-weight ratio and other performance indicators and safe operation of aircraft engines. In recent years, as aircraft engine technology develops towards high thrust and high efficiency, the main shaft bearings of aircraft engines are exposed to high temperature, high speed, heavy load and strong corrosive working conditions, and the working environment is becoming increasingly harsh. According to the statistical results of aircraft bearing failure modes analyzed in recent years, contact fatigue spalling is the most common, accounting for more than 1 / 3 of the total failure. According to the starting position of fatigue cracks, rolling contact fatigue is divided into subsurface origin type and surface origin type. Traditional bearing life prediction models such as the LP model generally consider the subsurface of the main load-bearing area of the bearing as the origin of fatigue damage. From the perspective of bearing steel materials, studying the failure behavior and performance characterization of bearings is a hot topic in promoting high-end bearing research. In order to further comprehensively study the failure modes of the above two origins, digital reconstruction of the surface and subsurface combined with finite element simulation analysis has been proven feasible in related research in recent years.
[0003] Subsurface reconstruction technology has developed a little slower than surface numerical reconstruction technology, and many scholars do not consider subsurface reconstruction indicators comprehensively. For example, for the molybdenum-based high-speed alloy steel 8Cr4Mo4V, which is the main bearing material of an aircraft engine, the matrix phase of the material is -Fe phase, and contains strong carbide-forming elements such as Cr, Mo, and V. The second reinforcement phase is mostly Cr, Mo, V and other alloy carbides, which are evenly dispersed in the material matrix in the form of fine particles, improving the material's resistance to deformation. Therefore, a perfect subsurface reconstruction needs to consider the material matrix, carbides, inclusions and other components at the same time, and it is very necessary to include perfect reconstruction parameters for subsurface reconstruction. There is a good foundation for related research on the influence of a single subsurface indicator on bearing life, but there are the following problems in conducting subsurface reconstruction research: the metal phase composition and distribution of specific materials are not targeted, and the measured parameters are rarely well combined in the reconstruction method; the reconstruction scheme is not versatile, and it is impossible to reconstruct multiple materials in one reconstruction scheme. With the continuous iterative development of materials science and technology and metallurgical technology, the composition, heat treatment preparation process, etc. of high-end bearing steel are undergoing rapid changes. The sub-surface reconstruction method directly affects the simulation research and analysis of high-end bearings, and further affects the performance characterization and quality evaluation of high-end bearings.
[0004] In summary, since the existing subsurface reconstruction methods cannot well reproduce the volume and distribution probability characteristics of metal matrix crystals, carbides, inclusions, the indicators considered during reconstruction are not comprehensive, and the generalization ability of the reconstruction method is poor, it is necessary to propose a new reconstruction method for the subsurface of bearing steel. Summary of the invention
[0005] The purpose of the present invention is to solve the problem that the existing sub-surface reconstruction methods do not consider comprehensive indicators, resulting in poor reconstruction performance and poor generalization ability of the reconstruction methods, and propose a bearing steel sub-surface reconstruction method that takes phase distribution characteristics into consideration.
[0006] The technical solution adopted by the present invention to solve the above technical problems is:
[0007] A bearing steel subsurface reconstruction method considering phase distribution characteristics, the method specifically comprising the following steps:
[0008] Step 1: Use bearing steel to prepare bearing samples, and then perform electron probe test and electron backscatter diffraction pattern test on the samples respectively;
[0009] An image group I1 taken from the sample surface to the sub-surface is obtained by electron probe testing; an image group I2 taken from the sample surface to the sub-surface is obtained by electron backscatter diffraction pattern testing;
[0010] Step 2: Process the image group I1 to obtain an overall image of the shooting area, and then segment the overall image of the shooting area to obtain separate image groups A of each phase in the sample except the matrix;
[0011] The image group I2 is processed, and then a distribution image group B with austenite as the matrix is obtained according to the processed images;
[0012] Step 3: Obtain the volume and distribution depth of each phase according to the image group A; extract the volume and distribution depth of each metal parent austenite in the matrix according to the image group B;
[0013] Step 4: According to the volume size and distribution depth of each phase, the two-dimensional probability density function of each phase is obtained respectively; according to the volume size and distribution depth of each metal parent austenite in the matrix, the relationship curve between the number of metal parent austenite grains and the depth is fitted;
[0014] Step 5: constructing an irregular phase true shape database based on the image group A obtained in step 2;
[0015] Step 6: construct a Voronoi diagram based on the relationship curve between the number of grains and the depth, and store the horizontal and vertical coordinate data of the endpoints of each Voronoi line in the Voronoi diagram;
[0016] Step 7, reconstructing the irregular phase according to the two-dimensional probability density function of step 4, the real shape database of the irregular phase of step 5 and the Voronoi diagram of step 6, and obtaining the point and line coordinate data of the reconstructed irregular phase;
[0017] Step 8: reconstruct the regular phase according to the two-dimensional probability density function of step 4, and obtain the point and line coordinate data of the reconstructed regular phase;
[0018] Step 9: According to the Voronoi diagram of step 6, the point and line coordinate data of the irregular phase and the point and line coordinate data of the regular phase are cut off, and the intersection points of the Voronoi diagram and each phase are obtained according to the line cut off results, and the point and line coordinate data of the matrix grain boundary line are obtained according to the intersection points;
[0019] Step 10: Use the point and line coordinate data of the matrix grain boundary line, the point and line coordinate data of the irregular phase, and the point and line coordinate data of the regular phase as pre-processing data for finite element analysis to complete the reconstruction of the bearing steel subsurface.
[0020] The beneficial effects of the present invention are:
[0021] The present invention identifies the subsurface structure of bearing steel through electron probe testing and electron backscatter diffraction pattern testing, collects the numerical characteristics of matrix grains, carbides, and inclusions in a more systematic manner, and applies the representation form of probability density function to a reconstruction method with randomness; by establishing an irregular phase true shape database, the true shape of the irregular phase is reconstructed during the reconstruction process. The method of the present invention improves the accuracy and authenticity of the bearing steel subsurface reconstruction scheme and ensures the reconstruction performance.
[0022] Moreover, through the three-stage reconstruction scheme of "matrix-irregular phase-regular phase", it can effectively adapt to the subsurface reconstruction requirements of bearing steels with different compositions and morphologies. Compared with existing methods, it has better application space and development prospects, and has good generalization ability. BRIEF DESCRIPTION OF THE DRAWINGS
[0023] Figure 1 It is the overall flow chart of the method of the present invention;
[0024] Figure 2 is the processing flow chart of step 4;
[0025] Figure 3 is the process flow chart of step five;
[0026] Figure 4 is a process flow chart of step six, step seven and step eight;
[0027] Figure 5 It is a schematic diagram of the classification of the lines of the Voronoi diagram;
[0028] Figure 6 It is a line processing flow chart of the Voronoi diagram. DETAILED DESCRIPTION
[0029] Specific implementation method 1: Combination Figure 1 The present embodiment is described as follows. The present embodiment describes a bearing steel subsurface reconstruction method taking into account phase distribution characteristics, and the method specifically comprises the following steps:
[0030] Step 1: Use bearing steel to prepare bearing samples, and then perform electron probe microscopy (EPMA) and electron backscatter diffraction pattern (EBSD) tests on the samples respectively;
[0031] An image group I1 taken from the sample surface to the sub-surface is obtained by electron probe testing; an image group I2 taken from the sample surface to the sub-surface is obtained by electron backscatter diffraction pattern testing;
[0032] Step 2: Process the image group I1 to obtain the overall image of the shooting area, and then segment the overall image of the shooting area to obtain a separate image group A of each phase (regular phase and irregular phase) in the sample except the matrix, that is, each image in the image group A can only include one single phase;
[0033] The image group I2 is processed, and then a distribution image group B based on austenite (including retained austenite) is obtained according to the processed image;
[0034] Step 3: Obtain the volume and distribution depth of each phase according to the image group A; extract the volume and distribution depth of each metal parent austenite in the matrix according to the image group B;
[0035] Step 4: According to the volume size and distribution depth of each phase, the two-dimensional probability density function of each phase is obtained respectively; according to the volume size and distribution depth of each metal parent austenite in the matrix, the relationship curve between the number of metal parent austenite grains and the depth is fitted;
[0036] Step 5: construct an irregular phase true shape database (CICD) based on the image group A obtained in step 2;
[0037] Step 6: Construct a Voronoi diagram based on the relationship curve between the number of grains and the depth, construct a Voronoi diagram to reconstruct the image of austenite in the metal matrix, and store the horizontal and vertical coordinate data of the endpoints of each Voronoi line in the Voronoi diagram (Voronoi line Figure 5 ) as indicated by the blue and red lines in ;
[0038] Step 7, reconstructing the irregular phase according to the two-dimensional probability density function of step 4, the real shape database of the irregular phase of step 5 and the Voronoi diagram of step 6, and obtaining the point and line coordinate data of the reconstructed irregular phase;
[0039] Step 8: reconstruct the regular phase according to the two-dimensional probability density function of step 4, and obtain the point and line coordinate data of the reconstructed regular phase;
[0040] Step 9: According to the Voronoi diagram of step 6, the point and line coordinate data of the irregular phase and the point and line coordinate data of the regular phase are cut off, and the intersection points of the Voronoi diagram and each phase are obtained according to the line cut off results, and the point and line coordinate data of the matrix grain boundary line are obtained according to the intersection points;
[0041] Step 10: Use the point and line coordinate data of the matrix grain boundary line, the point and line coordinate data of the irregular phase, and the point and line coordinate data of the regular phase as pre-processing data for finite element analysis to complete the reconstruction of the bearing steel subsurface.
[0042] The output objects and their methods are executable Python scripts of the large-scale general-purpose finite element analysis software ABAQUS, or the parametric design language APDL of ANASYS.
[0043] It should be noted that: when using an electron probe test, the same magnification must be maintained, and continuous and clear images must be taken from the metal surface to the sub-surface. The final imaging depth should ensure that the depth of the maximum stress area is covered (the imaging depth is generally selected as 500-700 microns), and the continuous shooting of images should maintain a small horizontal swing to ensure that the number of characteristic phase samples is sufficient. The phase to be extracted should have a gray depth difference visible to the naked eye from the metal matrix or other main phase components in the lens; the test using electron backscatter diffraction patterns is relatively expensive, sample preparation is difficult, and the test speed is slow, so continuous shooting is not required, but it should still be ensured that multiple images are taken one by one at intervals along the same position on the surface in the depth direction to reflect the changes in the crystal numerical characteristics of the matrix.
[0044] The irregular phase in the present invention includes irregular carbides, and the regular phase includes regular carbides and inclusions.
[0045] Specific implementation method 2: This implementation method is different from the specific implementation method 1 in that the specific process of step 2 is as follows:
[0046] Image Group I1:
[0047] Step 1: crop, splice, adjust contrast, denoise and sharpen the edges of the image group I1. The processing can improve the distinction between different phases. The processing is completed using the open source image processing software ImageJ, which completes the image processing in batches.
[0048] Step 2: Use the gray threshold segmentation method to segment the processed image to obtain separate image groups of each phase;
[0049] Image Group I2
[0050] Step (1), removing zero resolution points and denoising the image group I2;
[0051] The accuracy of the mother crystal reconstruction can be improved through processing. The processed images are mainly used to verify the effective components of all phases, provide the basis for the constitutive relationship for subsequent simulation verification, and ensure the consistency of the phase composition obtained by the electron probe; extract the distribution law of the matrix crystal dominated by austenite or martensite, and provide a numerical basis for the generation of the matrix in the subsequent numerical reconstruction.
[0052] Step (2), reconstructing the parent phase grains of the martensite in the image processed in step (1) to obtain a distribution image group with austenite as the matrix (including residual austenite).
[0053] The other steps and parameters are the same as those in the first embodiment.
[0054] In step (2), the EBSD image post-processing software Aztechcrystal or the EBSD image post-processing software EDAXOIM Analysis is used. The version of the EBSD image post-processing software Aztechcrystal is required to be no less than 2.1, and the version of the EBSD image post-processing software EDAXOIM Analysis is required to be no less than 8.6.
[0055] Specific implementation method three: Combination Figure 2 This embodiment is different from the first or second embodiment in that the specific process of step 4 is as follows:
[0056] Set the depth D to be reconstructed MAX and width W MAX ;
[0057] For any image in the image group B, according to the same distribution depth D in the matrix of the image C The volume size of each metal parent austenite is used to calculate the average austenite grain diameter L C , and then according to the average grain diameter L C Calculate the average number of grains N C :
[0058]
[0059] Then, the number of grains N is fitted according to each image in image group B. C With depth D CThe relationship curve of
[0060] For the regular phase in image group A, record the maximum volume V of the regular phase. MAX , with the VD scatter plot point density as the Z axis, a two-dimensional function H(v,d) that conforms to the VD (volume-depth) scatter plot is fitted;
[0061] According to D MAX and V MAX Integrate and normalize the two-dimensional function H(v,d) to obtain the two-dimensional probability density function P(v,d);
[0062] Similarly, the two-dimensional probability density function of the irregular phase is obtained.
[0063] The other steps and parameters are the same as those in the first or second embodiment.
[0064] The frequency graph is curve-fitted, and the fitted curve is adjusted according to the reconstruction requirement depth. The probability density function is used to characterize the numerical characteristics of the regular phase and the irregular phase; the curve of the fitted frequency is used to characterize the number of matrix grains generated along the depth direction.
[0065] Specific implementation method four: Combination Figure 3 This embodiment is different from the first to third embodiments in that the step five is as follows:
[0066] (1) In the open source image processing software ImageJ, the image of the area where the irregular carbide is located is identified and cut out from the image group A by using the grayscale threshold segmentation method;
[0067] (2) Convert the image captured in (1) into an RGB array;
[0068] (3) Binarize the RGB array, extract the edge points of the binary-processed RGB array, record the extracted edge points as feature boundaries, and calculate the feature volume according to the number of grids inside the feature boundaries;
[0069] (4) After unitizing each feature volume, the edge line of the feature after volume unitization is obtained, and the horizontal and vertical coordinate data of the two end points of each feature edge line are stored in the database (that is, the line coordinate data is stored in the database), and the real shape database of the irregular phase is obtained.
[0070] The other steps and parameters are the same as those in Specific Embodiments 1 to 3.
[0071] Specific implementation method five: Combination Figure 4 This embodiment is different from the first to fourth embodiments in that the specific process of step 6 is as follows:
[0072] Step 61, initializing depth d=0, initializing grain row number i=1;
[0073] Step 6.2: Consider the shape of the grain as a square. C -N C The curve gets the number of grains n in the i-th row i (That is, the number of equally spaced square units in the i-th row is n i ), then the side length of the square is l i for:
[0074]
[0075] Randomly distribute points in each square unit as Delaunay points;
[0076] Step 63: Set depth d = d + l i , determine whether d≥D MAX ;
[0077] If it is satisfied, make a Voronoi diagram based on all Delaunay points obtained in each iteration;
[0078] If not satisfied, set the number of grain rows i=i+1 and return to step 62.
[0079] The other steps and parameters are the same as those in Specific Embodiments 1 to 4.
[0080] Specific implementation method six: Combination Figure 4 This embodiment is different from the first to fifth embodiments in that the specific process of step 7 is as follows:
[0081] Step 7.1 Initialize the target volume fraction V of the irregular phase G , extraction line segment ratio t, relaxation factor g and allowable volume error ω;
[0082] Step 72, initialize the number of iterations j = 1;
[0083] Step 7.3: randomly select the position of the Voronoi line from the Voronoi diagram in step 6 according to the extraction line segment ratio t to generate irregular carbides, randomly extract carbide shapes from the irregular phase true shape database constructed in step 5, and generate irregular carbides with different volumes and depths at the positions of the selected Voronoi lines according to the two-dimensional probability density function and the randomly extracted carbide shapes;
[0084] Step 74: Calculate the total volume V of the irregular carbides generated in step 72 j ;
[0085] Step 75: Perform overlap detection on the irregular carbides generated in step 72 (the carbides generated in each iteration will be overlapped with the carbides retained in all previous iterations), and then delete the overlapping carbides;
[0086] Step 76: Determine whether V is satisfied G ∈[V j -ω,V j +ω];
[0087] If not, let j = j + 1, let t = t + g (V j -V G ), then return to step seven-three;
[0088] If satisfied, the process ends and stores the endpoint horizontal and vertical coordinate data of the generated boundary line of the irregular carbide.
[0089] The other steps and parameters are the same as those in Specific Implementation Methods 1 to 5.
[0090] Specific implementation method seven: Combination Figure 4 This embodiment is different from the first to sixth embodiments in that the specific process of step eight is as follows:
[0091] Step 81: Initialize the target volume fraction V of the regular phase G , number of points n and allowable volume error ω;
[0092] Step 82, initialize the number of iterations k = 1;
[0093] Step 83: Generate regular carbides and inclusions of different volumes at random points according to the number of points n and the two-dimensional probability density function (i.e., the regular phase includes carbides and inclusions), and the shapes of carbides and inclusions are ellipses;
[0094] Step 84: Calculate the total volume V of the regular carbides and inclusions generated in step 83 k ;
[0095] Step 85: Perform overlap detection on the regular carbides and inclusions generated in step 84, and then delete the carbides and inclusions that overlap with each other;
[0096] Step 86: Determine whether V is satisfied G ∈[V k -ω,V k +ω];
[0097] If not, set k=k+1 and return to step 83;
[0098] If satisfied, the process ends and the horizontal and vertical coordinate data of the endpoints of the generated regular carbide and inclusion boundary lines are stored.
[0099] The other steps and parameters are the same as those in Specific Embodiments 1 to 6.
[0100] Specific implementation method eight: Combination Figure 5 This embodiment is described below. The difference between this embodiment and any one of the specific embodiments 1 to 7 is that the lines of the Voronoi diagram include the following five categories:
[0101] Category 1: Both ends of the line are on a phase diagram line B n The interior of the line is B. n Inside;
[0102] Category 2: One of the endpoints of the line is on a phase diagram line B n The interior of the line, that is, the line only penetrates the phase diagram line B n Does not cross the phase line B n , and the line crosses or penetrates other phase diagram lines;
[0103] Category 3: Both end points of the line are not inside any phase diagram line, and the line completely crosses at least one phase diagram line;
[0104] Category 4: One of the endpoints of the line is on a phase diagram line B n The interior of the line, that is, the line only penetrates the phase diagram line B n Does not cross the phase line B n , and the line does not cross any other phase diagram line;
[0105] Category 5: Both end points of the line are not inside any phase diagram line, but the two end points can be on the edge of any phase diagram line, and the line does not cross any phase diagram line, that is, the line as a whole is outside any phase diagram line.
[0106] The other steps and parameters are the same as those in Specific Embodiments 1 to 7.
[0107] Specific implementation method 9: This implementation method is different from any one of specific implementation methods 1 to 8 in that, in step 9, the processing flow for any line in the Voronoi diagram is as follows:
[0108] Step 91: Determine whether the line belongs to category 1 (according to the Voronoi diagram in step 6, the point and line coordinate data of the irregular phase and the point and line coordinate data of the regular phase can determine the category of the line);
[0109] If the line belongs to category 1, delete the line;
[0110] If the line does not belong to category 1, execute step ninety-two;
[0111] Step 92: Determine whether the line belongs to category 2;
[0112] If the line belongs to category 2, the line is split based on the phase diagram line and step ninety-five is executed;
[0113] If the line does not belong to category 2, execute step nine-three;
[0114] Step 93: Determine whether the line belongs to category 3;
[0115] If the line belongs to category 3, the line is split based on the phase diagram line and step ninety-five is executed;
[0116] If the line does not belong to category 3, execute step 94;
[0117] Step 94: Determine whether the line belongs to category 4;
[0118] If the line belongs to category 4, the line is split based on the phase diagram line and step ninety-five is executed;
[0119] If the line does not belong to Category 4, then the line is Category 5 and no processing is required for the line;
[0120] Step 95: If there is a second type of line in the split result, return to step 92 for the second type of line;
[0121] If there is a third type of line in the split result, return to step nine-three for the third type of line;
[0122] If there is a Category 4 line in the splitting result, return to execute step 94 for the Category 4 line.
[0123] The other steps and parameters are the same as those in Specific Embodiments 1 to 8.
[0124] Similarly, after processing each line in the Voronoi diagram respectively using the method of this embodiment, the intersection points of the Voronoi diagram lines and each phase can be obtained, and the final reconstruction data of the matrix grain boundary line can be obtained by saving the intersection point and line coordinate data.
[0125] Example
[0126] The technical solution of the present invention is further described in detail below. The present invention is a bearing steel subsurface reconstruction method considering multiple phase distribution characteristics. The entire implementation process of the method of the present invention is as follows: Figure 1 As shown, the specific implementation steps are as follows:
[0127] Step 1: Get the original image
[0128] The samples were prepared using the bearing steel to be tested, and then subjected to electron probe microscopy (EPMA) and electron backscatter diffraction pattern (EBSD) testing. The former obtained a group of continuous and clear images taken from the sample surface to the subsurface; the latter obtained several images taken at intervals from the metal surface to the subsurface.
[0129] Step 2: Raw image processing
[0130] According to the continuous image group obtained by the electron probe test (EPMA) in step 1, the open source image processing software imagej is used to write scripts for secondary development, batch cropping, splicing, optimization, and enhancement to obtain the overall image of the shooting area. The grayscale threshold segmentation method is used to segment the image and separate the image group A of each phase of the sample except the matrix;
[0131] Based on the multiple interval images obtained by the electron backscatter diffraction pattern test (EBSD) in step 1, the EBSD image post-processing software Aztechcrystal of Aztech Company is used for optimization and enhancement. The parent phase grains of the martensite in the image are reconstructed to obtain the distribution image B with austenite as the matrix (including retained austenite).
[0132] Step 3: Obtaining raw data
[0133] Based on the images in the image group A obtained in step 2, the volume size and distribution depth of regular and irregular phases such as inclusions and carbide structures are extracted according to categories, and the proportion of each category in the entire image is extracted.
[0134] According to the image group B obtained in step 2, the volume size and distribution depth of each metal parent austenite in the sample matrix are extracted.
[0135] Step 4: Data analysis and fitting
[0136] According to the data obtained in step 3, Figure 2 The data processing method is as follows:
[0137] (1) Set the depth D to be reconstructed MAX and width W MAX ;
[0138] (2) If the data to be analyzed comes from EBSD, continue; if it comes from EPMA, proceed to (5);
[0139] (3) Calculate the average grain size L of the original austenite in each image C and the depth D C ;
[0140] (4) Calculate the average number of grains N C , fitting the relationship curve between the number of grains and the depth D C -N C ,in:
[0141] (5) Record the maximum volume V MAX, taking the point density of the VD scatter plot as the Z axis, fit the two-dimensional function H(v,d) that conforms to the VD scatter plot;
[0142] (6) According to D MAX and V MAX By integrating and normalizing H(v,d), we can obtain the two-dimensional probability density function P(v,d).
[0143] Step 5: Figure 3 As shown, the irregular carbide true shape database (CICD) is constructed based on several images in the image group A obtained in step 2; the specific process is as follows:
[0144] (1) Using ImageJ, the characteristic boundary recognition and segmentation of carbide inclusion defects are realized through grayscale threshold segmentation method;
[0145] (2) Force the .jpg file to be converted into an RGB array;
[0146] (3) Perform binary processing on the RGB array, extract the edge of the RGB array and record it as the feature boundary;
[0147] (4) Estimate the feature volume based on the number of grid cells inside the feature;
[0148] (5) The above characteristic volumes are unitized and recorded in the CICD.
[0149] Step 6: Voronoi matrix reconstruction
[0150] According to the volume distribution law of the parent austenite along the depth direction obtained from image B in step 4, a Voronoi diagram is constructed to reconstruct the image of austenite in the metal matrix, and the point and line coordinate data generated by the Voronoi are stored.
[0151] Step 7: Irregular Phase Reconstruction
[0152] The Voronoi diagram point and line coordinate data stored in step 6 are read, and according to the probability density function of the irregular phase along the depth direction obtained from image group A in step 4, lines in the Voronoi diagram are randomly selected to generate irregular carbides, and according to the irregular carbide true shape database (CICD) constructed in step 5, carbide shapes are randomly extracted from it and placed at the generation position.
[0153] The final volume of each carbide and inclusion is generated according to the volume probability distribution density function of the irregular phase obtained from the image group A in step 4, and the generation direction thereof is along the line direction.
[0154] Perform overlap detection and delete one of the overlapping carbides or inclusions.
[0155] The above operations in step 7 are iterated. According to the overall volume proportion of the phase in step 3, when the iteration reaches the volume error distribution range, the iteration stops.
[0156] Stores the point and line coordinate data of the Voronoi diagram, as well as the point and line coordinate data of the irregular phase.
[0157] Step 8: Rule Phase Reconstruction
[0158] The point and line coordinate data of the Voronoi diagram stored in step 7 and the point and line coordinate data of the irregular phase are read, and randomly generated within the reconstruction range according to the probability density function of the regular phase along the depth direction obtained from image group A in step 4.
[0159] The final volume of each carbide and inclusion is generated according to the volume probability distribution density function of the regular phase obtained from the image group A in step 4, and the generation direction is along the line direction.
[0160] Perform overlap detection and delete one of the rules that overlaps with other rules. If the rule generated in this step overlaps with the irregular rule generated in step 7, delete the rule generated in this step.
[0161] The above operations in step 8 are iterated. According to the overall volume proportion of the phase in step 3, when the iteration reaches the volume error distribution range, the iteration stops.
[0162] Stores the point and line coordinate data of the Voronoi diagram, the point and line coordinate data of the irregular phase, and the point and line coordinate data of the regular phase.
[0163] Step 9: Line truncation
[0164] Read all the data stored in step 8, such as Figure 5 As shown in , the lines are divided into 5 categories for operation. Figure 6 As shown, the Voronoi diagram passing through the regular and irregular phases is truncated and regenerated, the original line coordinate data is deleted, the intersection points of the Voronoi diagram and each phase are used as endpoints, and the new point and line coordinate data are saved as the final reconstruction data of the matrix grain boundary line.
[0165] The point and line coordinate data of the matrix grain boundary line, the point and line coordinate data of the irregular phase, and the point and line coordinate data of the regular phase are stored.
[0166] Step 10: Result output
[0167] All the data stored in step nine are read and output as pre-processed data for finite element analysis. The output object and method thereof are executable Python scripts of the large-scale general finite element analysis software ABAQUS, thus completing the present reconstruction method.
[0168] The above calculation examples of the present invention are only used to explain the calculation model and calculation process of the present invention in detail, and are not intended to limit the implementation methods of the present invention. For ordinary technicians in the relevant field, other different forms of changes or modifications can be made based on the above description. It is impossible to list all the implementation methods here. All obvious changes or modifications derived from the technical solution of the present invention are still within the scope of protection of the present invention.
Claims
1. A bearing steel subsurface reconstruction method considering phase distribution characteristics, characterized in that: The method specifically comprises the following steps: Step 1: Use bearing steel to prepare bearing samples, and then perform electron probe test and electron backscatter diffraction pattern test on the samples respectively; An image group I1 taken from the sample surface to the sub-surface is obtained by electron probe testing; an image group I2 taken from the sample surface to the sub-surface is obtained by electron backscatter diffraction pattern testing; Step 2: Process the image group I1 to obtain an overall image of the shooting area, and then segment the overall image of the shooting area to obtain separate image groups A of each phase in the sample except the matrix; The image group I2 is processed, and then a distribution image group B with austenite as the matrix is obtained according to the processed images; Step 3: Obtain the volume and distribution depth of each phase according to the image group A; extract the volume and distribution depth of each metal parent austenite in the matrix according to the image group B; Step 4: According to the volume size and distribution depth of each phase, the two-dimensional probability density function of each phase is obtained respectively; according to the volume size and distribution depth of each metal parent austenite in the matrix, the relationship curve between the number of metal parent austenite grains and the depth is fitted; Step 5: constructing an irregular phase true shape database based on the image group A obtained in step 2; Step 6: construct a Voronoi diagram based on the relationship curve between the number of grains and the depth, and store the horizontal and vertical coordinate data of the endpoints of each Voronoi line in the Voronoi diagram; Step 7, reconstructing the irregular phase according to the two-dimensional probability density function of step 4, the real shape database of the irregular phase of step 5 and the Voronoi diagram of step 6, and obtaining the point and line coordinate data of the reconstructed irregular phase; Step 8: reconstruct the regular phase according to the two-dimensional probability density function of step 4, and obtain the point and line coordinate data of the reconstructed regular phase; Step 9: According to the Voronoi diagram of step 6, the point and line coordinate data of the irregular phase and the point and line coordinate data of the regular phase are cut off, and the intersection points of the Voronoi diagram and each phase are obtained according to the line cut off results, and the point and line coordinate data of the matrix grain boundary line are obtained according to the intersection points; Step 10: Use the point and line coordinate data of the matrix grain boundary line, the point and line coordinate data of the irregular phase, and the point and line coordinate data of the regular phase as pre-processing data for finite element analysis to complete the reconstruction of the bearing steel subsurface.
2. A bearing steel subsurface reconstruction method considering phase distribution characteristics according to claim 1, characterized in that: The specific process of step 2 is as follows: Image Group I1: Step 1: cropping, splicing, adjusting contrast, denoising and edge sharpening the image group I1; Step 2: Use the gray threshold segmentation method to segment the processed image to obtain separate image groups of each phase; Image Group I2 Step (1), removing zero resolution points and denoising the image group I2; Step (2), reconstructing the parent phase grains of the martensite in the image processed in step (1) to obtain a distribution image group with austenite as the matrix.
3. A bearing steel subsurface reconstruction method considering phase distribution characteristics according to claim 1, characterized in that: The specific process of step 4 is as follows: Set the depth D to be reconstructed MAX and width W MAX ; For any image in the image group B, according to the same distribution depth D in the matrix of the image C The volume size of each metal parent austenite is used to calculate the average austenite grain diameter L C , and then according to the average grain diameter L C Calculate the average number of grains N C : Then, the number of grains N is fitted according to each image in image group B. C With depth D C The relationship curve of For the regular phase in image group A, record the maximum volume V of the regular phase. MAX , fit the two-dimensional function H(v,d) that conforms to the VD scatter plot; According to D MAX and V MAX Integrate and normalize the two-dimensional function H(v,d) to obtain the two-dimensional probability density function P(v,d); Similarly, the two-dimensional probability density function of the irregular phase is obtained.
4. The method for subsurface reconstruction of bearing steel considering phase distribution characteristics according to claim 1, characterized in that: The step five is specifically as follows: (1) In the open source image processing software ImageJ, the image of the area where the irregular carbide is located is identified and cut out from the image group A by using the grayscale threshold segmentation method; (2) Convert the image captured in (1) into an RGB array; (3) Binarize the RGB array, extract the edge points of the binary-processed RGB array, record the extracted edge points as feature boundaries, and calculate the feature volume according to the number of grids inside the feature boundaries; (4) After normalizing each feature volume, the edge line of the feature after the volume normalization is obtained, and the horizontal and vertical coordinate data of the two end points of each feature edge line are stored in the database to obtain the real shape database of the irregular phase.
5. The method for subsurface reconstruction of bearing steel considering phase distribution characteristics according to claim 3, characterized in that: The specific process of step six is as follows: Step 61, initializing depth d=0, initializing grain row number i=1; Step 6.2: Consider the shape of the grain as a square. C -N C The curve gets the number of grains n in the i-th row i , then the side length of the square is l i for: Randomly distribute points in each square unit as Delaunay points; Step 63: Set depth d = d + l i , determine whether d≥D MAX ; If it is satisfied, make a Voronoi diagram based on all Delaunay points obtained in each iteration; If not satisfied, set the number of grain rows i=i+1 and return to step 62.
6. A bearing steel subsurface reconstruction method considering phase distribution characteristics according to claim 5, characterized in that: The specific process of step seven is as follows: Step 7.1 Initialize the target volume fraction V of the irregular phase G , extraction line segment ratio t, relaxation factor g and allowable volume error ω; Step 72, initialize the number of iterations j = 1; Step 7.3: randomly select the position of the Voronoi line from the Voronoi diagram in step 6 according to the extraction line segment ratio t to generate irregular carbides, randomly extract carbide shapes from the irregular phase true shape database constructed in step 5, and generate irregular carbides with different volumes and depths at the positions of the selected Voronoi lines according to the two-dimensional probability density function and the randomly extracted carbide shapes; Step 74: Calculate the total volume V of the irregular carbides generated in step 72 j ; Step 75: Perform overlap detection on the irregular carbides generated in step 72, and then delete the overlapping carbides; Step 76: Determine whether V is satisfied G ∈[V j -ω,V j +ω]; If not, let j = j + 1, let t = t + g (V j -V G ), then return to step seven-three; If satisfied, the process ends and stores the endpoint horizontal and vertical coordinate data of the generated boundary line of the irregular carbide.
7. A bearing steel subsurface reconstruction method considering phase distribution characteristics according to claim 5, characterized in that: The specific process of step eight is as follows: Step 81: Initialize the target volume fraction V of the regular phase G , number of points n and allowable volume error ω; Step 82, initialize the number of iterations k = 1; Step 83: Generate regular carbides and inclusions of different volumes at random points according to the number of points n and the two-dimensional probability density function, and the shape of the carbides and inclusions is an ellipse; Step 84: Calculate the total volume V of the regular carbides and inclusions generated in step 83 k ; Step 85: Perform overlap detection on the regular carbides and inclusions generated in step 84, and then delete the carbides and inclusions that overlap with each other; Step 86: Determine whether V is satisfied G ∈[V k -ω,V k +ω]; If not, set k=k+1 and return to step 83; If satisfied, the process ends and the horizontal and vertical coordinate data of the endpoints of the generated regular carbide and inclusion boundary lines are stored.
8. The method for subsurface reconstruction of bearing steel considering phase distribution characteristics according to claim 5, characterized in that: The lines of the Voronoi diagram include the following 5 categories: Category 1: Both ends of the line are on a phase diagram line B n The interior of the line is B. n Inside; Category 2: One of the endpoints of the line is on a phase diagram line B n The interior of the line, that is, the line only penetrates the phase diagram line B n Does not cross the phase line B n , and the line crosses or penetrates other phase diagram lines; Category 3: Both end points of the line are not inside any phase diagram line, and the line completely crosses at least one phase diagram line; Category 4: One of the endpoints of the line is on a phase diagram line B n The interior of the line, that is, the line only penetrates the phase diagram line B n Does not cross the phase line B n , and the line does not cross any other phase diagram line; Category 5: Both end points of the line are not inside any phase diagram line, and the line does not cross any phase diagram line, that is, the line as a whole is outside any phase diagram line.
9. A bearing steel subsurface reconstruction method considering phase distribution characteristics according to claim 8, characterized in that: In step nine, the processing flow for any line in the Voronoi diagram is as follows: Step 91: Determine whether the line belongs to category 1; If the line belongs to category 1, delete the line; If the line does not belong to category 1, execute step ninety-two; Step 92: Determine whether the line belongs to category 2; If the line belongs to category 2, the line is split based on the phase diagram line and step ninety-five is executed; If the line does not belong to category 2, execute step nine-three; Step 93: Determine whether the line belongs to category 3; If the line belongs to category 3, the line is split based on the phase diagram line and step ninety-five is executed; If the line does not belong to category 3, execute step 94; Step 94: Determine whether the line belongs to category 4; If the line belongs to category 4, the line is split based on the phase diagram line and step ninety-five is executed; If the line does not belong to Category 4, then the line is Category 5 and no processing is required for the line; Step 95: If there is a second type of line in the split result, return to step 92 for the second type of line; If there is a third type of line in the split result, return to step nine-three for the third type of line; If there is a Category 4 line in the splitting result, return to execute step 94 for the Category 4 line.
Citation Information
Patent Citations
Stress intensity factor-based Method for estimating fatigue damage degree of subsurface of bearing roller path
CN103105467A
Spatial target three-dimensional reconstruction evaluation method based on reconstruction model integrity, and application
CN108876908A
