A method for evaluating the size uniformity and distribution uniformity of the second phase

Abstract

The present invention belongs to the field of evaluation of material microstructure, and particularly relates to a method for evaluating the size uniformity and distribution uniformity of the second phase, which is applicable to evaluating the size and distribution uniformity of the internal second phase of materials with different compositions after being processed by different processes. First, the image recognition method is applied to the microstructural characterization photos of the material to obtain the contour information and centroid coordinates of the second phase. The area of each second phase and the shortest distance between the second phases are subjected to deviation normalization using the deviation normalization method, and the data is normalized to the interval [0, 1]. Finally, the size uniformity U1 and the distribution uniformity U2 are obtained. By quantitatively statistically analyzing the uniformity of the second phase, the relationship between the distribution of the second phase and hot working processes such as heat treatment can be established, guiding the optimization of relevant processes to obtain a second phase with uniform size and good distribution, thereby improving the mechanical properties of the actual material. The logic of the present invention is clear, the operation is relatively easy to implement, and the applicable range is relatively wide.

Description

Technical Field

[0001] The present invention belongs to the field of evaluation of the microstructure of materials, and particularly relates to a method for evaluating the size uniformity and distribution uniformity of the second phase, which is applicable to evaluating the distribution uniformity and size uniformity of the second phase in the microstructure of materials. Background Art

[0002] The microstructure of materials significantly affects their mechanical properties. To improve the strength, plasticity and toughness of materials, the microstructures of many engineering applications contain multiple constituent phases. For example: GCr15 bearing steel widely used in bearings of various mechanical equipment, its microstructure usually consists of carbides and a martensite / bainite matrix; duplex stainless steel widely used in many fields such as natural gas and nuclear industry, its microstructure consists of two phases, ferrite and austenite; particle-reinforced metal matrix composites consist of a metal matrix and particle reinforcement phases dispersed therein.

[0003] In multiphase materials, due to the different intrinsic properties of different constituent phases, the properties of materials are non-uniform at the microscale. This non-uniform property will lead to differences in micro stress and strain, as well as micro-region electrochemical properties, etc., which will in turn affect the macroscopic mechanical properties and corrosion resistance of materials. Therefore, a reasonable evaluation of the size and distribution law of the second phase in materials is of great significance for analyzing and adjusting the mechanical and chemical properties of materials.

[0004] Recently, the research of the applicant has shown that through the method of image recognition, the distribution characteristic information of the second phase in the microstructure of materials (such as: coordinates, size, distribution distance between each other, etc.) can be quickly determined, and then the size uniformity and distribution uniformity information of the second phase can be obtained. Finally, the overall distribution situation is quantitatively evaluated to guide the composition design and heat treatment process to obtain a second phase structure with a better distribution state. This method is of great significance for evaluating the overall distribution state of the second phase and has guiding significance for other materials. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for evaluating the size uniformity and distribution uniformity of the second phase. Through the method of image recognition, the distribution characteristic information of the second phase in the microstructure of materials (such as: coordinates, size, distribution spacing between each other, etc.) can be quickly determined, and then the size uniformity and distribution uniformity information of the second phase can be obtained. Finally, the overall distribution situation is quantitatively evaluated to guide the composition design and heat treatment process to obtain a second phase structure with a better distribution state.

[0006] The technical solution of the present invention:

[0007] A method for evaluating the size uniformity and distribution uniformity of the second phase, comprising the following steps:

[0008] (1) First, obtain a microscopic tissue photo where there is an obvious contrast difference between the second phase and the matrix, and call the OpenCV library to perform grayscale processing on the image; remove the noise of the redundant interference information in the image through Gaussian Blur, and perform binary threshold segmentation on the denoised image.

[0009] (2) According to the contrast difference between the matrix and the second phase, identify the contour information of each second phase, and obtain the area S of each second phase n , and finally, normalize the area of each second phase through the method of deviation normalization to obtain the size uniformity U1. Table 1 shows the meanings and units of each parameter.

[0010] S max =max{S1, S2, S3…S n}

[0011] S min =min{S1, S2, S3…S n}

[0012]

[0013] S′ mean =max{S′1, S′2, S′3...S′ n}

[0014]

[0015] U1=1 - σ S , U1∈(0, 1)

[0016] Table 1 Meanings and units of each parameter

[0017]

[0018] (3) Based on the contour information of each second phase identified in the above steps, construct the minimum circumscribed circle of each second phase to obtain the center coordinates as its centroid; calculate the distance between the centroid coordinates of a single second phase and those of all other second phases Find the minimum value and record it as Repeat the above process until the shortest distances between all second phases are found Table 2 shows the meanings and units of each parameter.

[0019]

[0020] Table 2 Meanings and units of each parameter

[0021]

[0022] (4) Perform range normalization on all the shortest distances between the second phases, and normalize the data to the interval [0, 1]; then calculate the standard deviation of these processed data and define it as the distribution uniformity U2. Table 3 shows the meanings and units of each parameter.

[0023]

[0024]

[0025]

[0026]

[0027] L′ mean =mean{L′1, L′2, L′3…L′ n}

[0028]

[0029] U2=1 - σ L , U2∈(0, 1)

[0030] Table 3 Meanings and units of each parameter

[0031]

[0032] The method for evaluating the size uniformity and distribution uniformity of the second phase uses the obvious contrast difference between the second phase and the matrix in the picture to determine the area and position occupied by the second phase.

[0033] The method for evaluating the size uniformity and distribution uniformity of the second phase evaluates its size uniformity based on the area uniformity of the second phase.

[0034] The method for evaluating the size uniformity and distribution uniformity of the second phase uses the center coordinates of the minimum circumscribed circle of the second phase as its centroid.

[0035] The method for evaluating the size uniformity and distribution uniformity of the second phase evaluates the distribution uniformity of the second phase based on the distribution between the centroid coordinates of the second phase.

[0036] The method for evaluating the size uniformity and distribution uniformity of the second phase: both the size uniformity and distribution uniformity of the second phase are between 0 and 1, and the larger the value, the higher the uniformity.

[0037] The design concept and principle of the present invention are as follows:

[0038] When there is a second phase in the microstructure of a material that is significantly different from the matrix, the size uniformity and distribution uniformity of the second phase significantly affect the material properties, and there is a lack of effective evaluation methods for their size uniformity and distribution uniformity. This invention develops a new quantitative evaluation method. Its design concept and basic principle are as follows: First, apply the image recognition method to the microstructural characterization photos of the material to obtain the contour information and centroid coordinates of the second phase. Use the deviation normalization method to perform deviation normalization processing on the area of each second phase and the shortest distance between second phases, and normalize the data to the interval [0, 1]. Then, quantitatively evaluate the size and distribution uniformity of the second phase respectively. Among them, the size uniformity of the second phase is evaluated by the area uniformity of the second phase, and the distribution uniformity of the second phase is evaluated by the distance between the centroid coordinates of the second phase. This invention adopts an image recognition method based on OpenCV, which can batch-recognize and count the second phase information in the microstructural photos of the material, laying a foundation for evaluating the size uniformity and distribution uniformity of the second phase. Finally, use the numerical values of size uniformity and distribution uniformity to quantitatively evaluate the second phase uniformity of the material microstructure, which can guide the heat treatment process to obtain a better distributed second phase structure.

[0039] The advantages and beneficial effects of this invention are as follows:

[0040] 1. This invention uses the image recognition method. By performing Gaussian blur, non-linear stretching, threshold segmentation, etc. on the microstructural photos of the material, it quantitatively describes the main second phase structures (such as: number, area ratio, size, distribution uniformity) that affect the mechanical or chemical properties of the material, establishing a quantitative evaluation method different from the previous qualitative observation only by the naked eye. It is beneficial to judge the overall distribution state of the second phase and then guide the composition design or hot working process to achieve a better second phase distribution state. This method provides a method for quantitatively evaluating the second phase distribution state, which is beneficial to establishing the corresponding relationship between the heat treatment process - second phase distribution state evaluation - service performance.

[0041] 2. This invention has universality and can not only be applied to evaluate carbides and second phases in steel, but also be applicable to quantitatively evaluate the uniformity of other material microstructures and even the distribution of inclusions.

[0042] 3. The method of this invention is of great significance for evaluating the overall distribution state of the second phase and has guiding significance for other materials. Description of the Drawings

[0043] Figure 1 SEM photos and processed images of the GCr15 bearing steel microstructure magnified 20,000 times. Among them, (a) SEM photo of Area.1; (b) SEM photo of Area.2; (c) Processed image of Area.1; (d) Processed image of Area.2.

[0044] Figure 2 It is the distribution frequency of carbide area and the minimum distance between any two carbides. Among them, (a) shows the carbide area distribution. The abscissa "Size of carbides" represents the normalized carbide area (unit: 1), and the ordinate "Frequency" represents the distribution frequency (unit: %). (b) shows the minimum distance distribution between two carbides. The abscissa "Mindistance of each carbides" represents the normalized minimum distance between two carbides (unit: 1), and the ordinate "Frequency" represents the distribution frequency (unit: %).

[0045] Figure 3 It is the SEM photo of the microstructure of 2205 duplex stainless steel magnified 1000 times and the processed image. Among them, (a) is the SEM photo of Area.1; (b) is the SEM photo of Area.2; (c) is the processed image of Area.1; (d) is the processed image of Area.2.

[0046] Figure 4 It is the distribution frequency of the area of each austenite and the minimum distance between any two austenites. (a) shows the austenite area distribution. The abscissa "Size of austenite" represents the normalized austenite area (unit: 1), and the ordinate "Frequency" represents the normalized austenite area distribution rule (unit: %). (b) shows the minimum distance distribution between each austenite. The abscissa "Min distance of each austenite" represents the normalized minimum distance between two austenite grains (unit: 1), and the ordinate "Frequency" represents the normalized austenite-austenite spacing distribution frequency (unit: %). Specific implementation mode

[0047] In the specific implementation process, the present invention first applies the image recognition method to the microstructure characterization photos of materials to obtain the contour information and centroid coordinates of the second phase. The deviation normalization method is used to perform deviation normalization processing on the area of each second phase and the shortest distance between second phases, and the data is normalized to the interval [0, 1]. Finally, the size uniformity U1 and the distribution uniformity U2 are obtained. By quantitatively counting the uniformity of the second phase, the relationship between the distribution of the second phase and heat treatment and other hot working processes can be established, guiding the optimization of relevant processes to obtain a second phase structure with uniform size and good distribution, thereby improving the mechanical properties of the actual material.

[0048] In the present invention, OpenCV (Open Source Computer Vision) is an open-source computer vision algorithm library written in C / C++, aiming to take advantage of multi-core. It provides interfaces for C++, C, Python, and Java and supports all mainstream operating system platforms, including Windows, Linux, Mac OS, iOS, and Android.

[0049] In order to better demonstrate the efficiency and superiority of the present invention in quantitatively evaluating the size and distribution uniformity of the second phase and considering the feasibility of actual operation, in the subsequent implementation schemes, the present invention will be applied to GCr15 bearing steel and 2205 duplex stainless steel respectively for illustration.

[0050] The following further details the present invention through examples:

[0051] Example 1

[0052] In this example, the automatic recognition method for evaluating the carbide size and distribution uniformity is as follows:

[0053] (1) Obtain SEM photos of the microstructure of GCr15 bearing steel under standard heat treatment process magnified 20,000 times;

[0054] (2) Successively perform grayscale processing, Gaussian blur, and binary threshold segmentation on the SEM photos of the microstructures in two regions Area.1 and Area.2 with different carbide distribution characteristics to improve the contrast difference between the carbide and the matrix, so as to obtain more comprehensive carbide information and make the carbide contour easy to identify;

[0055] (3) Identify each carbide contour based on the contrast difference between the matrix and the carbide and calculate the area. Finally, obtain the size uniformity U1 through the mathematical processing method of deviation normalization;

[0056] S max = max{S1, S2, S3…S n}

[0057] S min = min{S1, S2, S3…S n}

[0058]

[0059] S′ mean = max{S′1, S′2, S′3…S′ n}

[0060]

[0061] U1 = 1 - σS , U1 ∈ (0, 1)

[0062] (4) Based on the carbide contour information identified in the above steps, construct the minimum circumscribed circle for each carbide to obtain the center coordinates as the centroid of the carbide. Calculate the distances between the centroid coordinates of a certain fixed carbide and those of all other carbides Find the minimum value and record it as Repeat the above process until the shortest distances between all carbides are found

[0063]

[0064] (5) Perform deviation normalization on the shortest distances between all carbides, and normalize the data to the interval [0, 1]. Then, calculate the standard deviation of these data and define the distribution uniformity U2. The results are shown in Table 4 below: It can be seen that the size uniformity of the carbides in Area.1 and Area.2 does not differ much, but the distribution uniformity differs significantly. Thus, this method can quantitatively represent the differences in carbide size and distribution uniformity

[0065]

[0066]

[0067]

[0068]

[0069] L′ mean = mean{L′1, L′2, L′3…L′ n}

[0070]

[0071] U2 = 1 - σ L , U2 ∈ (0, 1)

[0072] Table 4 Differences in carbide distribution characteristics in Area.1 and Area.2

[0073]

[0074] As Figure 1 shown, it can be seen from the SEM micrograph of the GCr15 bearing steel microstructure magnified 20,000 times and the processed image that the recognition of carbide contours and the number of carbides is relatively accurate

[0075] As Figure 2As shown, from the carbide area and the minimum spacing distribution frequency between any two carbides, it can be seen that the carbide size and distribution uniformity are well consistent with the distribution frequency diagram.

[0076] Example 2

[0077] In this embodiment, the automatic identification method for evaluating the uniformity of austenite size and distribution in dual-phase steel is as follows:

[0078] (1) A 1000-fold magnified SEM photograph of the microstructure of 2205 duplex stainless steel under a standard heat treatment process was obtained, and there was an obvious contrast difference between the second phase austenite and the matrix ferrite, which falls within the scope of application of the present invention;

[0079] (2) The microstructure SEM photos of Area.1 and Area.2 with different austenite distribution characteristics were processed with grayscale, Gaussian blur and binary threshold segmentation in turn to improve the contrast difference between austenite and matrix ferrite, so as to obtain more comprehensive austenite distribution information and make its contour easy to identify;

[0080] (3) According to the contrast difference between the matrix and austenite, the contours of each austenite are identified and their areas are calculated. Finally, the dimensional uniformity U1 is obtained through a mathematical treatment method of deviation standardization;

[0081] S max =max{S1, S2, S3...S n}

[0082] S min =min{S1, S2, S3…S n}

[0083]

[0084] S′ mean =max{S′1, S′2, S′3…S′ n}

[0085]

[0086] U1=1-σ S , U1∈(0,1)

[0087] (4) Based on the austenite contour information identified in the above steps, construct the minimum circumscribed circle of each austenite to obtain the coordinates of the center of the circle as the centroid of each austenite. Calculate the distance between a fixed austenite and the centroid coordinates of all other austenites Find the minimum value and record it as Repeat the above process until the shortest distance between all austenites is found.

[0088]

[0089] (5) Perform deviation normalization on the shortest distance between all austenites, and normalize the data to the interval [0, 1]. Calculate the standard deviation of these data, and define the distribution uniformity U2. The results are shown in Table 5 below. It can be seen that this method can quantitatively represent the differences in austenite size and distribution uniformity.

[0090]

[0091]

[0092]

[0093]

[0094] L′ mean =mean{L′1, L′2, L′3…L′ n}

[0095]

[0096] U2=1 - σ L , U2∈(0, 1)

[0097] Table 5 Differences in austenite distribution characteristics in Area.1 and Area.2

[0098]

[0099] As Figure 3 shown, from the SEM micrographs of the 2205 duplex stainless steel microstructure magnified 1000 times and the processed images, it can be seen that this method is relatively accurate in identifying austenite grain contours and the number of grains.

[0100] As Figure 4 shown, from the distribution frequencies of the areas of each austenite and the minimum distance between any two austenites, it can be seen that the size and distribution uniformity of austenite grains are in good agreement with the distribution frequency diagram.

[0101] The results of the examples show that the present invention is applicable to evaluating the problems of the size and distribution uniformity of internal second phases after materials are subjected to different composition designs and different process treatments. This method has clear logic, is relatively easy to implement, and has a wide range of applications.

Claims

1. A method for evaluating the dimensional uniformity and distribution uniformity of the second phase, characterized in that, Including the following steps: (1) First, obtain a microstructural photograph in which there is an obvious contrast difference between the second phase and the matrix, and call the OpenCV library to perform gray-scale processing on the picture; remove the noise of the redundant interference information in the picture through Gaussian blur (GaussianBlur), and perform binary threshold segmentation on the picture after noise reduction processing; (2) Identify the contour information of each second phase according to the contrast difference between the matrix and the second phase, and obtain the area S of each second phase n , and finally normalize the area of each second phase by the method of deviation standardization to obtain the size uniformity U1; S max = max{S1, S2, S3…S n} S min = min{S1, S2, S3... S n} S′ meaan = max{S′1, S′2, S′3... S′ n} U1 = 1 - σ S , U1 ∈ (0, 1) (3) Based on the contour information of each second phase identified in the above steps, construct the minimum circumscribed circle of each second phase to obtain the center coordinates as its centroid; Calculate the distance between the centroid coordinates of a single second phase and those of all other second phases Let n represent the nth second phase particle, find the minimum value and record it as Repeat the above process until the shortest distances between all second phases are found (4) Perform deviation standardization processing on the shortest distance between all second phases, and normalize the data to the interval [0,1]; then calculate the standard deviation of these processed data and define it as the distribution uniformity U2; L′ mean = mean{L′1, L′2, L′3...L′ n} U2 = 1 - σ L , where U2 ∈ (0, 1).

2. The method for evaluating the dimensional uniformity and distribution uniformity of the second phase according to claim 1, wherein, Utilize the obvious contrast difference between the second phase and the matrix in the picture to determine the area and position occupied by the second phase.

3. The method for evaluating the size uniformity and distribution uniformity of the second phase according to claim 1, characterized in that, Evaluate the size uniformity of the second phase by the area uniformity of the second phase.

4. The method for evaluating the size uniformity and distribution uniformity of the second phase according to claim 1, characterized in that, Take the center coordinates of the minimum circumscribed circle of the second phase as its centroid.

5. The method for evaluating the size uniformity and distribution uniformity of the second phase according to claim 1, wherein Evaluate the distribution uniformity of the second phase by the distribution between the centroid coordinates of the second phase.

6. The method for evaluating the dimensional uniformity and distribution uniformity of the second phase according to claim 1, characterized in that, Both the size uniformity and the distribution uniformity of the second phase are between 0 and 1, and the larger the value, the higher the uniformity.

Images