Method for detecting and quantitatively evaluating non-metallic inclusions

Abstract

The invention provides a detection and rating method for non-metallic inclusions, which comprises the following steps: scanning a sample by adopting an X-ray tomography system to obtain three-dimensional gray data, and marking the non-metallic inclusions; obtaining three-dimensional morphological parameters and spatial position information of the non-metallic inclusions; counting the obtained three-dimensional morphological parameters and spatial positions to obtain the volume fraction, size distribution concentration ratio, spatial distribution uniformity and shape distribution data of the inclusions; calculating the data by adopting an evaluation model to obtain an inclusion evaluation index, and rating the nonmetallic inclusions of the steel grade on the basis of the inclusion evaluation index. According to the method, the problems of contour distortion and form misjudgment of a traditional method are solved, the objectivity of inclusion rating is effectively improved, an IEI rating result is directly output, a visual basis can be provided for steel quality control and performance pre-judgment, and industrial application is supported.

Description

Technical Field

[0001] This invention belongs to the field of steel non-metallic inclusion detection technology, specifically relating to a method for quantitative evaluation of non-metallic inclusion detection. Background Technology

[0002] Non-metallic inclusions in steel are a key factor affecting steel performance; their volume, size, spatial distribution, and morphology directly determine core indicators such as strength and toughness. Existing methods for detecting non-metallic inclusions in steel mainly include metallography, electrolysis, and ultrasonic testing. Metallography is low-cost but can only observe two-dimensional cross-sections, limiting the detection area and failing to reflect the true distribution of large inclusions, while also easily missing small inclusions. Electrolysis can statistically analyze a large number of inclusions but cannot obtain their spatial distribution. Ultrasonic testing is convenient and fast but cannot detect small inclusions or obtain microscopic morphology. X-ray computed tomography (CT) technology, however, can achieve three-dimensional non-destructive imaging. Through image processing, it can statistically analyze the morphology, size, and spatial distribution of inclusions larger than 1 μm in steel, making it an emerging method for inclusion analysis and characterization.

[0003] A search revealed that patent CN114509456A discloses a CT detection method for non-metallic inclusions in steel, which statistically analyzes basic single parameters such as the three-dimensional morphology and sphericity of inclusions. Patent CN113155872B uses CT to characterize defects in steel billets and distinguishes defect types through sphericity, but does not correlate it with a quantitative assessment of the impact of inclusions. Currently, the three-dimensional characterization of inclusions in steel is based on qualitative characterization of image morphology or characterization using multiple single parameters. However, inclusions vary in type and state across different steel grades. Some inclusions are large in size but few in number, while others are numerous but uniform in shape, often elongated and without significant hazard, making it difficult to establish unified evaluation indicators. The multi-parameter discretization of inclusion characterization cannot intuitively reflect the comprehensive hazard level of inclusions, making it difficult to support standardized evaluation of steel quality. Therefore, there is an urgent need for a method that can comprehensively and quantitatively characterize inclusions, integrating multi-dimensional indicators into unified quantitative values, and comprehensively assessing the hazard level of inclusions to overcome the limitations of existing technologies. Summary of the Invention

[0004] To address the problems existing in the prior art, this invention provides a method for detecting and rating non-metallic inclusions. It aims to perform non-destructive imaging of the internal structure of steel using X-ray computed tomography (CT) technology, and to quantitatively calculate the three-dimensional morphological parameters and spatial location of inclusions through three-dimensional reconstruction. A "Three-Dimensional Inclusion Evaluation Index (IEI)" model is established, which comprehensively quantifies and weights multiple indicators such as inclusion volume fraction, size distribution, spatial distribution uniformity, and shape distribution, thereby achieving quantitative evaluation of inclusions.

[0005] As one aspect of the present invention, a method for detecting and rating non-metallic inclusions is provided, specifically including the following steps:

[0006] S1. Non-metallic inclusion detection

[0007] S101 uses an X-ray tomography system to scan the sample, acquire three-dimensional grayscale data, and extract features based on the three-dimensional grayscale data to identify and mark non-metallic inclusions.

[0008] S102 Perform connected cell analysis on the non-metallic inclusions marked in step S101 to obtain the three-dimensional morphological parameters and spatial location information of the non-metallic inclusions;

[0009] S2, Statistics of Non-metallic Inclusions

[0010] The volume fraction of inclusions is obtained by statistically analyzing the three-dimensional morphological parameters and spatial locations obtained in step S102. Inclusion size distribution concentration Spatial distribution uniformity of inclusions Distribution of inclusion shapes data;

[0011] S3. Quantitative Evaluation of Non-metallic Inclusions

[0012] S301 An inclusion evaluation index is calculated from the data obtained in step S2 using an evaluation model. The evaluation model is...

[0013] ,

[0014] In the formula, IEI is the inclusion evaluation index. , , and These are the volume influence factor, size distribution concentration influence factor, spatial distribution uniformity influence factor, and shape distribution influence factor of inclusions, respectively.

[0015] S302 Based on the inclusion evaluation index obtained in step S201, the non-metallic inclusions of the steel grade are rated. The smaller the IEI index, the smaller the impact of the inclusions on the performance of the steel product.

[0016] As a preferred embodiment of the detection and rating method for non-metallic inclusions of the present invention, the volume fraction of the inclusions... In the formula, For reference volume fraction, Let n be the volume of the nth inclusion; n is the total number of inclusions. To analyze the total volume of the region.

[0017] As a preferred embodiment of the detection and rating method for non-metallic inclusions of the present invention, the inclusion size distribution concentration... In the formula, These are reference dimensions, typically the median size of a qualified batch. This is the dimensional influence coefficient; This is the influence coefficient of size distribution. The interquartile interval is the equivalent sphere diameter. The median of the equivalent sphere diameter. To standardize the threshold; the size of the inclusions mainly considers the equivalent sphere diameter. .

[0018] As a preferred embodiment of the detection and rating method for non-metallic inclusions of the present invention, the spatial distribution uniformity of the inclusions... In the formula, Set of nearest neighbor distances Standard deviation; Set of nearest neighbor distances The mean, The standardized threshold for the coefficient of variation; the closest distance between two inclusions. The calculation formula is: .

[0019] As a preferred embodiment of the detection and rating method for non-metallic inclusions of the present invention, the inclusion shape distribution In the formula, The maximum diameter of the inclusions is greater than the equivalent sphere diameter; S represents the number of spheres among the inclusions included in the calculation; P represents the number of strips among the inclusions included in the calculation. The influence coefficient of spherical inclusions; The influence coefficient for elongated inclusions; Reference dimensions for calculating spherical inclusions; Reference dimensions for calculating the shape of elongated inclusions. The diameter of the inclusion is the equivalent sphere diameter. The maximum diameter of the inclusion; when the aspect ratio of the inclusion is... At that time, the inclusion was considered to be spherical; when At that time, it was determined to be a long strip shape.

[0020] As a preferred embodiment of the detection and rating method for non-metallic inclusions of the present invention, step S1 further includes the following steps: the X-ray transmittance is 25-35%, the scanning parameters are set as follows: scanning voltage 110-160kV, scanning power 10W, single-image projection acquisition time 1-4s, distance between X-ray source and sample <8mm, and scanning angle 360°.

[0021] As a preferred embodiment of the detection and rating method for non-metallic inclusions of the present invention, the focal spot size of the X-ray tomography system is ≤10μm and the side length of the three-dimensional image voxel is ≤3μm.

[0022] As a preferred embodiment of the non-metallic inclusion detection and rating method of the present invention, step S1 further includes a sampling step, specifically, selecting a representative part from the part of the billet where inclusions are easily enriched, preferably at the position of 1 / 4 of the thickness direction on the inner arc side of the billet.

[0023] As a preferred embodiment of the non-metallic inclusion detection and rating method of the present invention, step S1 further includes a pretreatment step, including cleaning the sample with alcohol or acetone to remove oil and impurities from the processing, and fixing the sample so that the sample is exposed above the fixture by a size ≥ 5 mm, the perpendicularity deviation is ≤ 3°, and the coaxiality is ≤ 0.1 mm.

[0024] As a preferred embodiment of the non-metallic inclusion detection and rating method of the present invention, in step S1, the feature extraction specifically includes performing median filtering and Black Top Hat filtering on the image, and using a threshold segmentation tool to extract features from the filtered sample based on grayscale.

[0025] This invention creatively proposes a method based on the difference in X-ray absorption between a steel matrix and non-metallic inclusions. It utilizes a three-dimensional computed tomography (CT) scanner to perform 360° projection acquisition and reconstruction of the sample, obtaining three-dimensional grayscale data. The steel matrix appears as a high grayscale value (bright area) in the image, while the non-metallic inclusions appear as a low grayscale value (dark area), forming a "matrix-inclusion" boundary. This allows for precise location and contour extraction of the inclusions. By employing mean filtering and top-hat filtering techniques to process the image, the distinguishability between the "matrix-inclusion" boundary is further amplified, thereby improving the accuracy of inclusion shape and contour recognition. Specifically, this is reflected in three aspects: First, the equipment parameter requirements: This invention clearly defines the core parameters of the X-ray CT system, with a focal spot size ≤10μm and a three-dimensional image voxel side length ≤3μm. The extremely small focal spot and voxel size enable "high spatial resolution" imaging, avoiding distortion of inclusion contours due to pixel blurring (especially for small inclusions larger than 1μm, clearly restoring their edge details). Second, the scanning parameters are standardized (110~160kV, scanning power 10W, single-image projection acquisition time 1~4s, X-ray source and sample distance <8mm). This parameter combination balances the X-ray penetration depth and signal sensitivity, ensuring X-ray penetration of the steel sample (avoiding excessively dark images due to strong absorption) while accurately acquiring the weak absorption signals of inclusions, avoiding signal loss or noise interference. Third, for the three-dimensional grayscale data acquired by X-ray scanning, this invention further eliminates interference and enhances the "matrix-inclusion" image processing through a dual morphological specific operation image processing algorithm using "median filtering" and "Black Top Hat". The boundary comparison method performs pixel-level segmentation between dark areas (inclusions) below a threshold and bright areas (matrix) above a threshold, achieving "complete extraction" of inclusions. For spherical inclusions, it can accurately restore their spherical contours, avoiding misjudgment of "flattened spheres" due to blurred boundaries; for elongated inclusions, it can completely extract their length, diameter, and end morphology, avoiding omission of small branches or misjudgment as short particles; for irregular inclusions (such as clustered alumina), it can accurately present their spatial distribution morphology.

[0026] This invention creatively proposes four parameters for rating inclusions: volume fraction, size distribution concentration, spatial distribution uniformity, and shape distribution. This quantifies the spatial distribution of inclusions, which cannot be quantitatively characterized by other methods, including size and shape. Specifically, volume fraction, as a core indicator, quantifies the "total scale" of inclusions in steel materials, directly reflecting the basic level of steel cleanliness and being a core factor affecting material strength and toughness. Size distribution concentration is a risk-focusing indicator. By using the median equivalent sphere diameter and its dispersion, it focuses on the risk of "large-sized inclusions." The closer the size is to the reference value (median size of qualified batches) and the more concentrated the distribution, the lower the proportion of large-particle inclusions and the lower the risk of material failure caused by inclusions. Conversely, a higher proportion of large-particle inclusions makes them more prone to stress concentration. This can compensate for the limitations of "only looking at the total amount." For example, some materials may have a low volume fraction of inclusions, but a high concentration of large-sized inclusions poses a far greater risk than "small-particle-dense" inclusions. This parameter accurately reflects this difference. Spatial distribution uniformity is a stress risk indicator that quantifies the "dispersion" of inclusions within steel. It uses the coefficient of variation of nearest neighbor distances to determine the presence of "local enrichment." A lower index indicates a more uniform spatial distribution of inclusions, resulting in more balanced stress transmission. Conversely, a higher index indicates local enrichment, which can easily lead to stress concentration in enriched areas, significantly reducing the steel's fatigue and fracture resistance. Spatial distribution uniformity addresses the limitation of traditional methods in evaluating "spatial distribution," serving as a quantitative value for the impact of local inclusion concentration on steel performance. Shape distribution is a hazard type indicator, distinguishing the "morphological risk" of inclusions. Using three-dimensional aspect ratio, inclusions are categorized into spherical and elongated shapes. Elongated inclusions are more prone to stress concentration and pose a significantly greater hazard than spherical inclusions. This parameter is weighted to enhance the risk weight of elongated inclusions.

[0027] This invention creatively proposes to fully cover the core influencing factors of inclusions from four dimensions: "total scale - size risk - spatial distribution - morphological hazards". Then, through weighting (adapting to the performance requirements of different steel grades), it is finally integrated into the IEI value, realizing the standardized transformation of "multi-dimensional differences → unified quantitative value". This solves the problems of inconsistent evaluation standards for inclusions of different steel grades and the inability to intuitively represent the comprehensive hazards.

[0028] In this invention, the correlation coefficient factors for common typical steel grades can adopt the recommended values ​​as shown in Table 1.

[0029] Table 1 Recommended values ​​of correlation coefficient factors for typical steel grades

[0030] Steel grade GCr15 bearing steel 0.0015 5.00 0.70 0.30 0.60 0.50 1.00 0.50 20 30 2.0 1.3 0.5 0.2 High-purity ESR bearing steel GCr15ESR 0.0015 6.00 0.70 0.30 0.60 0.50 0.95 0.50 22 32 1.9 1.4 0.5 0.2 300M high-strength steel for aviation 0.0020 7.00 0.65 0.35 0.60 0.50 0.90 0.50 25 35 1.8 1.3 0.6 0.3 18Ni martensitic aging steel 0.0025 8.00 0.65 0.35 0.60 0.50 0.95 0.55 25 35 1.6 1.3 0.8 0.4 Nuclear power pressure vessel steel SA508-3 0.0025 8.00 0.65 0.35 0.60 0.50 0.95 0.55 25 35 1.6 1.3 0.8 0.4 Turbine disk alloy steel FGH95 0.0030 10.00 0.60 0.40 0.60 0.50 0.90 0.50 28 38 1.5 1.2 0.9 0.4 H13 hot work die steel 0.0028 9.00 0.65 0.35 0.60 0.50 0.90 0.55 26 36 1.5 1.3 0.8 0.4 Cold work die steel Cr12MoV 0.0028 9.00 0.65 0.35 0.60 0.50 0.90 0.60 26 36 1.5 1.2 0.9 0.5 P20 plastic mold steel 0.0032 12.00 0.60 0.40 0.60 0.50 0.85 0.50 30 40 1.4 1.2 1.0 0.4 Gear steel 20CrNiMoA 0.0030 11.00 0.60 0.40 0.60 0.50 0.90 0.60 28 38 1.4 1.1 1.1 0.5 Carburized gear steel 18Cr2Ni4WA 0.0040 15.00 0.55 0.45 0.60 0.50 0.80 0.45 35 45 1.3 1.0 1.3 0.4 Pipeline steel APIX80 0.0032 12.00 0.60 0.40 0.60 0.50 0.85 0.50 30 40 1.4 1.2 1.1 0.4 Pressure vessel steel 12Cr2Mo1R 0.0035 12.00 0.60 0.40 0.60 0.50 0.85 0.50 30 40 1.3 1.2 1.1 0.4 High-strength bolt steel 35CrMo 0.0060 18.00 0.55 0.45 0.60 0.50 0.80 0.40 40 50 1.2 1.0 1.4 0.4 Automotive Duplex Steel Q&P980 0.0015 5.00 0.70 0.30 0.60 0.50 1.00 0.50 20 30 2.0 1.3 0.5 0.2 Spring steel 60Si2MnA 0.0015 6.00 0.70 0.30 0.60 0.50 0.95 0.50 22 32 1.9 1.4 0.5 0.2 Welded structural steel Q345B 0.002 7.00 0.65 0.35 0.60 0.50 0.90 0.50 25 35 1.8 1.3 0.6 0.3 Axle steel 42CrMo 0.0025 8.00 0.65 0.35 0.60 0.50 0.95 0.55 25 35 1.6 1.3 0.8 0.4 High-strength sheet Q690 0.0025 8.00 0.65 0.35 0.60 0.50 0.95 0.55 25 35 1.6 1.3 0.8 0.4 Plain carbon steel Q235B 0.003 10.00 0.60 0.40 0.60 0.50 0.90 0.50 28 38 1.5 1.2 0.9 0.4 Q355B structural steel for buildings 0.0028 9.00 0.65 0.35 0.60 0.50 0.90 0.55 26 36 1.5 1.3 0.8 0.4 HRB400 steel reinforcement 0.0028 9.00 0.65 0.35 0.60 0.50 0.90 0.60 26 36 1.5 1.2 0.9 0.5

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] 1. This invention achieves three-dimensional non-destructive imaging based on X-ray CT technology, overcoming the limitations of "two-dimensional cross-sectional observation" methods such as metallography and electron microscopy. It can completely capture the true morphology and spatial location of inclusions larger than 1μm, avoiding the omission of small-sized inclusions and misjudgment of the distribution of large-particle inclusions. Through equipment parameter optimization (focal point ≤10μm, voxel ≤3μm) + dual filtering processing, the distinction between the matrix and inclusion boundaries is enhanced, accurately identifying spherical, elongated, and irregularly shaped inclusions, solving the problems of contour distortion and morphological misjudgment in traditional methods.

[0033] 2. This invention quantifies four dimensions of indicators: volume fraction, size distribution concentration, spatial distribution uniformity, and shape distribution. This overcomes the shortcomings of traditional methods that use a single parameter for characterization or qualitative description, and comprehensively covers the core hazard factors of total inclusions, size, space, and morphology, effectively improving the objectivity of inclusion rating.

[0034] 3. This invention addresses the pain point of inconsistent inclusion evaluation standards across different steel grades by constructing an IEI weighted evaluation model and using influencing factors adapted to the performance requirements of different steel grades. It achieves a standardized transformation from "multi-dimensional differences to unified quantitative values." The directly output IEI rating results provide intuitive evidence for steel quality control and performance prediction, supporting industrial applications. Detailed Implementation

[0035] The technical solutions described below in conjunction with the embodiments will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0036] Example 1

[0037] This embodiment takes bearing steel GCr15 produced by a certain factory as an example, and the method for detecting and rating non-metallic inclusions provided includes the following steps:

[0038] S0, Sampling and Pretreatment

[0039] Samples were taken from representative locations in the billet where inclusions are prone to accumulate. In this embodiment, samples were taken at 1 / 4 of the thickness of the inner arc side of the billet. The samples were cleaned with alcohol or acetone to remove oil and impurities from the processing.

[0040] S1. Non-metallic inclusion detection

[0041] S101 The sample is fixed so that the sample protrudes 5 mm from the fixture, the perpendicularity deviation is 3°, and the coaxiality is 0.1 mm. An X-ray tomography system is used to scan the sample to acquire three-dimensional grayscale data. The image is processed by median filtering and BlackTop Hat filtering. A threshold segmentation tool is used to extract features from the filtered sample based on grayscale to identify and mark non-metallic inclusions. The X-ray transmittance is set to 25%, the scanning parameters are set as follows: scanning voltage 150 kV, scanning power 10 W, single-image acquisition time 2 s, scanning angle 360°, the focal size of the X-ray tomography system is 70 μm, and the voxel side length of the three-dimensional image is 1 μm.

[0042] S102 Perform connected cell analysis on the non-metallic inclusions marked in step S101 to obtain the three-dimensional morphological parameters and spatial location information of the non-metallic inclusions;

[0043] S2, Statistics of Non-metallic Inclusions

[0044] The volume fraction of inclusions is obtained by statistically analyzing the three-dimensional morphological parameters and spatial locations obtained in step S102. Inclusion size distribution concentration Spatial distribution uniformity of inclusions Distribution of inclusion shapes Data, among which, ; ; ; The values ​​of the correlation coefficient factor are shown in Table 1.

[0045] S3, Non-metallic inclusion rating

[0046] S301 An inclusion evaluation index is calculated from the data obtained in step S2 using an evaluation model. The evaluation model is...

[0047] The values ​​of the correlation coefficient factors are shown in Table 1.

[0048] S302 Based on the inclusion evaluation index obtained in step S201, the non-metallic inclusions of the steel grade are rated.

[0049] The IEI value in this embodiment was measured to be 1.35.

[0050] Example 2

[0051] This embodiment takes 20CrNiMoA gear steel produced by a certain factory as an example. The method for detecting and rating non-metallic inclusions provided includes the following steps:

[0052] S0, Sampling and Pretreatment

[0053] Samples were taken from representative locations in the billet where inclusions are prone to accumulate. In this embodiment, samples were taken at 1 / 4 of the thickness of the inner arc side of the billet. The samples were cleaned with alcohol or acetone to remove oil and impurities from the processing.

[0054] S1. Non-metallic inclusion detection

[0055] S101 The sample is fixed so that the sample protrudes 5 mm from the fixture, the perpendicularity deviation is 3°, and the coaxiality is 0.1 mm. An X-ray tomography system is used to scan the sample to acquire three-dimensional grayscale data. The image is processed by median filtering and BlackTop Hat filtering. A threshold segmentation tool is used to extract features from the filtered sample based on grayscale to identify and mark non-metallic inclusions. The X-ray transmittance is set to 25%, the scanning parameters are set as follows: scanning voltage 150 kV, scanning power 10 W, single-image acquisition time 2 s, scanning angle 360°, focal spot size of the X-ray tomography system 100 μm, and voxel side length of the three-dimensional image ≤ 1 μm.

[0056] S102 Perform connected cell analysis on the non-metallic inclusions marked in step S101 to obtain the three-dimensional morphological parameters and spatial location information of the non-metallic inclusions;

[0057] S2, Statistics of Non-metallic Inclusions

[0058] The volume fraction of inclusions is obtained by statistically analyzing the three-dimensional morphological parameters and spatial locations obtained in step S102. Inclusion size distribution concentration Spatial distribution uniformity of inclusions Distribution of inclusion shapes Data, among which, ; ; ; The values ​​of the correlation coefficient factor are shown in Table 1.

[0059] S3, Non-metallic inclusion rating

[0060] S301 An inclusion evaluation index is calculated from the data obtained in step S2 using an evaluation model. The evaluation model is...

[0061] The values ​​of the correlation coefficient factor are shown in Table 1.

[0062] S302 Based on the inclusion evaluation index obtained in step S201, the non-metallic inclusions of the steel grade are rated.

[0063] The IEI value in this embodiment was measured to be 1.12.

[0064] It should be noted that, based on the above embodiments of the present invention, those skilled in the art can fully realize the scope of the independent claims and dependent claims of the present invention, and the implementation process and methods are the same as those in the above embodiments; and the parts of the present invention not described in detail belong to the well-known technology in the art. However, the protection scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the protection scope of the present invention.

Claims

1. A method for detecting and rating non-metallic inclusions, characterized in that, Includes the following steps: S1. Non-metallic inclusion detection S101 uses an X-ray tomography system to scan the sample, acquire three-dimensional grayscale data, and extract features based on the three-dimensional grayscale data to identify and mark non-metallic inclusions. S102 Perform connected cell analysis on the non-metallic inclusions marked in step S101 to obtain the three-dimensional morphological parameters and spatial location information of the non-metallic inclusions; S2, Statistics of Non-metallic Inclusions The volume fraction of inclusions is obtained by statistically analyzing the three-dimensional morphological parameters and spatial locations obtained in step S102. Inclusion size distribution concentration Spatial distribution uniformity of inclusions Distribution of inclusion shapes data; S3. Quantitative Evaluation of Non-metallic Inclusions S301 An inclusion evaluation index is calculated from the data obtained in step S2 using an evaluation model. The evaluation model is... ， In the formula, IEI is the inclusion evaluation index. , , and These are the volume influence factor, size distribution concentration influence factor, spatial distribution uniformity influence factor, and shape distribution influence factor of inclusions, respectively. S302 Based on the inclusion evaluation index obtained in step S201, the non-metallic inclusions of the steel grade are rated. The smaller the IEI index, the smaller the impact of the inclusions on the performance of the steel product.

2. The method for detecting and rating non-metallic inclusions as described in claim 1, characterized in that, The volume fraction of the inclusions In the formula, For reference volume fraction, Let n be the volume of the i-th inclusion; n is the total number of inclusions. To analyze the total volume of the region.

3. The method for detecting and rating non-metallic inclusions as described in claim 1, characterized in that, The concentration of the size distribution of the inclusions In the formula, These are reference dimensions, typically the median size of a qualified batch. This is the dimensional influence coefficient; This is the influence coefficient of size distribution. The interquartile interval is the equivalent sphere diameter. The median of the equivalent sphere diameter. To standardize the threshold; the size of the inclusions mainly considers the equivalent sphere diameter. .

4. The method for detecting and rating non-metallic inclusions as described in claim 1, characterized in that, The spatial distribution uniformity of the inclusions In the formula, Set of nearest neighbor distances Standard deviation; Set of nearest neighbor distances The mean, The standardized threshold for the coefficient of variation; the closest distance between two inclusions. The calculation formula is: .

5. The method for detecting and rating non-metallic inclusions as described in claim 1, characterized in that, The shape distribution of the inclusions In the formula, The maximum diameter of the inclusions is greater than the equivalent sphere diameter; S represents the number of spheres among the inclusions included in the calculation; P represents the number of strips among the inclusions included in the calculation. The influence coefficient of spherical inclusions; The influence coefficient for elongated inclusions; Reference dimensions for calculating spherical inclusions; Reference dimensions for calculating the shape of elongated inclusions. The diameter of the inclusion is the equivalent sphere diameter. The maximum diameter of the inclusion; when the aspect ratio of the inclusion is... At that time, the inclusion was considered to be spherical; when At that time, it was determined to be a long strip shape.

6. The method for detecting and rating non-metallic inclusions as described in claim 1, characterized in that, Step S1 also includes steps such as having a radiation transmittance of 25-35%, setting scanning parameters as follows: scanning voltage of 110-160kV, scanning power of 10W, single-image projection acquisition time of 1-4s, distance between X-ray source and sample of <8mm, and scanning angle of 360°.

7. The method for detecting and rating non-metallic inclusions as described in claim 1, characterized in that, The focal spot size of the X-ray tomography system is ≤10μm and the side length of the three-dimensional image voxel is ≤3μm.

8. The method for detecting and rating non-metallic inclusions as described in claim 1, characterized in that, Step S1 also includes a sampling step, specifically, selecting representative parts from areas in the billet where inclusions are easily formed and enriched, with priority given to the location at 1 / 4 of the thickness direction on the inner arc side of the billet.

9. The method for detecting and rating non-metallic inclusions as described in claim 1, characterized in that, Step S1 also includes a pretreatment step, which includes cleaning the sample with alcohol or acetone to remove oil and impurities from the processing, and fixing the sample so that the sample is exposed above the fixture by a size ≥5mm, a perpendicularity deviation ≤3°, and a coaxiality ≤0.1mm.

10. The method for detecting and rating non-metallic inclusions as described in claim 1, characterized in that, In step S1, the feature extraction specifically includes performing median filtering and Black Top Hat filtering on the image, and using a threshold segmentation tool to extract features from the filtered samples based on grayscale.

Images