A quantitative statistical and analytical method for precipitated phases in nickel-based superalloys

Abstract

The present invention belongs to the field of material microstructure research and analysis, and specifically relates to a quantitative statistics and analysis method for precipitated phases in nickel-based high-temperature alloys. In view of the shortcomings of traditional image analysis methods for precipitated phases in nickel-based high-temperature alloys in processing complex, multi-scale, and multi-phase coexisting microstructures, the present invention uses an energy dispersive spectrometer to detect nickel-based alloy samples, uses Image J, Image pro plus, and Origin software to analyze the detection spectrum, and judges the control ability of a certain set process parameter on the precipitated phase in the nickel-based alloy sample based on the analysis results. When p<0.05, it means that a certain set process parameter has a preliminary control ability on the precipitated phase, can promote the refinement and homogenization of the precipitated phase size to a limited extent, and the smaller the p value, the stronger the control ability; when p<0.01, it means that a certain set process parameter has a significant control effect on the precipitated phase, and the precipitated phase achieves statistical homogenization.

Description

Technical Field

[0001] The present invention belongs to the field of material microstructure research and analysis, and in particular relates to a quantitative statistical and analytical method for precipitation phases in nickel-based high-temperature alloys. Background Art

[0002] Nickel-based superalloys, due to their exceptional high-temperature strength, oxidation resistance, and corrosion resistance, exhibit excellent stability under extreme service conditions. They are widely used in key components of aircraft engines, pressure vessels in nuclear power systems, high-temperature heat exchangers and reactors in chemical plants, high-pressure pipelines transporting corrosive media, and thermal process equipment such as industrial high-temperature furnaces. One of the key factors influencing the service performance of these alloys is the type, quantity, and evolution of the various precipitates formed during heat treatment and service.

[0003] During long-term thermal exposure above 900°C, complex intermetallic compounds will precipitate in nickel-based high-temperature alloys, including MC carbides, M6C carbides, M 23 C6 carbides and Laves phases. Excessive carbide precipitation can cause local strain and reduce ductility. The grain boundary embrittlement of the Laves phase leads to reduced ductility and impact toughness of the alloy. Therefore, a deep understanding of the thermodynamic stability, formation mechanism, and evolution of these precipitated phases is crucial for optimizing alloy composition design and heat treatment processes, thereby improving service life.

[0004] Currently, the mainstream method for quantitative analysis of precipitates in nickel-based alloys is still based on microscopic image-based line-sectioning techniques. This method relies on the processing and statistical analysis of large amounts of image data, which has certain efficiency and accuracy limitations. Specifically, this technique uses multiple reference lines (i.e., line sections) drawn within the micrograph to perform statistical analysis of the microstructure. Typically, these line sections should cover the entire field of view as evenly as possible. Their orientation can be arbitrary or, depending on practical needs, set to perpendicular, parallel, or other specific directions. During the analysis, researchers record the number of intersections between each line section and precipitates or specific microstructural features (such as grain boundaries and phase boundaries). The statistical data from these intersections can be used to infer microstructural parameters such as the volume fraction, interface density, and particle size distribution of the precipitates in the material. The calculations of this method rely on classical geometric probability principles, and commonly used statistical formulas include volume fraction estimation formulas and interface density conversion formulas. However, this method is highly dependent on image representativeness and the number of line sections, which can introduce significant human errors and may also limit accuracy when dealing with heterogeneous structures or low-volume-fraction phases. Therefore, although this method has a certain universality, there is still room for improvement and replacement when high-resolution quantitative analysis is required.

[0005] Therefore, in view of the shortcomings of traditional image analysis methods of precipitate phases in nickel-based high-temperature alloys in dealing with complex, multi-scale, and multi-phase coexisting microstructures, it is necessary to develop a new method for quantitative analysis of precipitate phases that is more efficient and accurate. Summary of the Invention

[0006] In view of the technical issues discussed above, the present invention provides a method for quantitative statistical analysis of precipitates in nickel-based alloys, which is quick and convenient, and produces accurate results. By comparing this method with other currently available techniques in the same field, we believe that it offers advantages over other methods in quantitative statistical analysis of precipitates in alloys.

[0007] The present invention provides a quantitative statistical and analytical method for precipitated phases in nickel-based high-temperature alloys, comprising the following contents:

[0008] An energy dispersive spectrometer (EDS) is used to perform surface scanning on the nickel-based alloy sample to obtain an element surface scanning distribution detection map. The detection map is analyzed for elemental composition and phase using Image J, Image pro plus, and Origin software. Based on the analysis results, the ability of a certain set of process parameters to control the size of the precipitated phase in the nickel-based alloy sample and the uniformity of the distribution of the precipitated phase in the sample is determined. The certain set of process parameters refers to the main parameters in the preparation process of the nickel-based alloy, including but not limited to alloy composition, element content in the alloy, cooling rate, cooling time, etc.

[0009] The main parameters of the energy dispersive spectrometer are set as follows:

[0010] (1) Setting the acceleration voltage to 5kV~10kV can excite most medium and heavy elements; the beam intensity is set to 13A~15A;

[0011] (2) Sample positioning and focusing: Switch to the secondary electron imaging mode (Secondary Electron) or backscattered electron imaging mode (Backscattered Electron), observe the sample morphology, select the target area, and perform automatic focus and aberration correction; define the surface scanning area in EDS to ensure that the surface scanning area contains the precipitate phase to be characterized, including NbC phase, Laves phase, M6C phase, γ′ phase and other types of precipitates;

[0012] (3) Acquisition parameter setting: Set the image resolution to ≥ 256 × 256 (or 512 × 512 for high resolution); the dwell time per pixel is usually set to 50 µs to 500 µs;

[0013] (4) Start the mapping acquisition program to generate elemental maps to intuitively display the coexistence relationship of elements.

[0014] The obtained element surface scan distribution map is processed by software. The main processing steps are as follows:

[0015] (5) Open the elemental surface scan distribution map in Image J, use the Trainable Weka Segmentation plug-in to separate the precipitated phase area from the non-precipitated phase area, and generate the segmented image;

[0016] (6) Convert the segmented image to 16-bit format in Image J. After the format conversion, perform binarization on the segmented image to obtain binary processed images of the precipitated phase area and the non-precipitated phase area, and save them in Tif format.

[0017] (7) Use Image J software to perform noise reduction on the binary image in Tif format in step (6). The noise radius is usually set to 2-3 pixels to obtain the binary image after noise reduction. Then use Image pro plus software to perform automatic bright object statistics to measure and count the precipitated phase area to obtain statistical data.

[0018] (8) The obtained statistical data of the precipitated phase area were analyzed for significance. The statistical data were subjected to Tukey's honestly significant difference test (Tukey HSD significance test) in Origin. The differences within or between statistical data groups were judged by the p-value, reflecting the ability of a certain set process parameter to regulate the size of the precipitated phase in the nickel-based alloy and the uniformity of the distribution of the precipitated phase in the sample;

[0019] ,

[0020] in: q is the Studentized range statistic; , are the means of the two groups to be compared; MSE is the mean square error of the one-way analysis of variance (from the ANOVA results); if the sample sizes of the two sets of statistical data are equal, then n is the number of samples in each set of statistical data; if the number of samples in the two sets of statistical data is not equal, then n is the equivalent sample size of the statistical data; the p-value is q The corresponding right-tail probability in the Studentized range distribution table;

[0021] When p < 0.05, it indicates that a certain set process parameter has a preliminary control ability on the precipitate phase, which can promote the refinement and homogenization of the precipitate size to a limited extent and promote the uniform distribution of the precipitate phase in the sample. The smaller the p value, the stronger the control ability. Furthermore, when p < 0.01, it indicates that a certain set process parameter has a significant control effect on the precipitate phase, and the precipitate phase achieves statistical homogenization. Furthermore, when p < 0.001, it indicates that a certain set process parameter has an extremely significant effect on the control of the precipitate phase in the alloy. However, when p > 0.05, it indicates that the control effect of a certain set process parameter on the precipitate phase has reached its upper limit, and the precipitate distribution has entered a stable stage.

[0022] In step (8) of the present invention, the following image analysis steps are specifically performed:

[0023] 1) Relative position of the median and mean: When the mean is greater than the median (right skew), it indicates that the data as a whole tends to be small-sized precipitates, but there are a small number of large-sized precipitates; when the mean is less than the median (left skew), it indicates that the precipitates as a whole tend to be large-sized, but small-sized precipitates occupy a high frequency; when the mean is close to the median, it indicates that the particle size of the precipitates presents an approximately normal distribution and good uniformity.

[0024] 2) Interquartile range (IQR) and whisker length: A large IQR and a significantly longer upper whisker indicate that the particle size distribution of the precipitate phase fluctuates greatly and has many maximum values. A smaller IQR and closer upper and lower whiskers indicate that the data concentration is increasing and the particle size of the precipitate phase is tending to be consistent. An extremely small IQR and nearly symmetrical upper and lower whiskers indicate that the size distribution of the precipitate phase is highly concentrated, reflecting good uniformity.

[0025] 3) Number and distribution of extreme values: If there are multiple upper outliers in the box plot (far away from the upper whisker), it indicates the presence of local coarse precipitates, which may cause microstructure heterogeneity. If the number of outliers gradually decreases and tends to disappear, it indicates that the precipitate control tends to be stable and abnormally sized precipitates are suppressed.

[0026] 4) Comprehensive judgment and conclusion of distribution trend: Based on the box plot characteristics and significance test results in the above steps, the following quantitative judgment can be made on the state of the precipitate phase under different treatment conditions: If the IQR gradually narrows, the mean / median coincides, the outliers disappear, and the difference between groups reaches the significance level (p < 0.05), it can be determined that the treatment condition can effectively promote the refinement and homogenization of the precipitate phase size. If the above trends tend to stabilize or disappear, and the significance level is no longer met (p > 0.05), it means that the control effect has reached the upper limit and the precipitate phase particle size distribution has been saturated. It is not recommended to continue to increase the treatment intensity (such as alloying element content or cooling rate).

[0027] The Tukey HSD (Honestly Significant Difference) significance test involved in step (8) of the present invention is a post hoc significance analysis method for comparing the means of multiple groups. It is mainly based on the standardized range distribution to control the first type error rate under multiple comparisons. This method calculates a minimum significant difference threshold (HSD) by setting a significance level (such as α=0.05), and compares the mean difference between groups with the threshold to determine whether there is a statistically significant difference. In other words, HSD is a critical value obtained by looking up the standardized range distribution table under the premise of a given significance level α, which is used to determine whether the mean difference between two groups reaches significance. The p-value is a probability value calculated from the actual observed data, which indicates the probability of observing the current mean difference or even more extreme differences under the condition that the original hypothesis (the means of each group are equal) is valid. Therefore, HSD is a "discrimination threshold" and the p-value is a "quantified probability of difference significance". The two are judgment criteria and evidence for each other, but cannot be used interchangeably.

[0028] when When the p-value is greater than HSD, or p < α (e.g., 0.05), it indicates a statistically significant difference between the two groups. In this case, the difference level should be clearly reported (e.g., p < 0.05, p < 0.01, p < 0.001), and the regulatory effect or mechanism of the variable should be explored in the context of the experiment.

[0029] when When ≤ HSD, or p ≥ α, the difference is not significant. This may mean that the variable has a limited impact on the system, or that factors such as sample size and data variability limit the expression of significance. Overinterpretation should be avoided, and further optimization of the experimental design or expansion of the sample size may be considered.

[0030] It should be noted that the Tukey HSD significance test involved in step (8) of the present invention is itself a statistical method for multiple comparisons after one-way analysis of variance (ANOVA), which is mainly used to determine whether there are significant differences in the means between multiple groups. The application premise of the Tukey HSD significance test is to meet the normal distribution of data and homogeneity of variance. However, under large sample conditions (such as the number of samples in each group in the present invention exceeds 30 or the overall sample size is large), according to the Central Limit Theorem, the sample mean distribution will be close to the normal distribution. At this time, even if the original data deviates slightly from the normality assumption, the Tukey HSD test still has strong robustness and can usually be used directly without performing a normality test. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 1 is a comparison diagram of the significance of precipitation phases of Inconel 625 nickel-based alloy after adding 0.009wt.% magnesium in Example 1; wherein, (a) is the intra-group comparison; (b) is the inter-group comparison.

[0032] Figure 2 This is a comparison diagram of the significance of precipitation phases in Inconel 625 nickel-based alloy after adding 0.012wt.% magnesium in Example 2; among them, (a) is the comparison within the group; (b) is the comparison between the groups.

[0033] Figure 3 This is a comparison diagram of the significance of precipitation phases in Inconel 625 nickel-based alloy after adding 0.029wt.% magnesium in Example 3; among them, (a) is the comparison within the group; (b) is the comparison between the groups.

[0034] Figure 4 The S is obtained after the Inconel 625 nickel-based alloy in Example 4 was subjected to the 20°C / min in-situ solidification experiment. 0-4 、S 1-4 、S 2-4 、S 3-4 Comparison of the significance of precipitated phases of the samples.

[0035] Figure 5 The S is obtained after the Inconel 625 nickel-based alloy in Example 5 is subjected to the 40°C / min in-situ solidification experiment. 0-5 、S 1-5 、S 2-5 、S 3-5 Comparison of the significance of precipitated phases of the samples. DETAILED DESCRIPTION

[0036] Below in conjunction with embodiment and accompanying drawing, the technical scheme of the present invention is clearly and completely described, and the operation process and application effect of the inventive method in specific experimental analysis are described in more detail.It should be pointed out that the embodiment described in the present invention is only used for further explanation and illustration, rather than limiting its scope of application.Based on the present invention, all other embodiments obtained by those skilled in the art without making creative work premise all belong to protection scope of the present invention.

[0037] The preparation method of Inconel 625 nickel-based alloy is as follows:

[0038] In a 25kg vacuum induction furnace (maximum operating temperature 1750°C, ultimate vacuum degree 5×10 -1Four Inconel 625 alloy ingots, labeled S0, S1, S2, and S3 (S1, S2, and S3 were subjected to varying degrees of magnesium microalloying), were prepared in a smelting process (1000 psi, 1000 psi, and 2000 psi, respectively). The specific alloy compositions are shown in Table 1. During the smelting process, the metals Ni, Cr, Mo, and Nb were first placed in a crucible. The vacuum was reduced to below 5 Pa, and nitrogen was introduced into the furnace to a pressure of 50 kPa. Heating began and the temperature was raised to 1500°C. Si, Mn, and C were then added for alloying, and 50% Al particles were added for deoxidation. Sponge titanium was added for alloying, and the remaining 50% of aluminum was used for deep deoxidation. Samples S0, S1, S2, and S3 were cooled by air.

[0039] Table 1 Chemical composition of experimental alloys (wt.%):

[0040] element C Si Mn S Al Ti Cr Mo Nb Ni Mg <![CDATA[S0]]> 0.039 0.29 0.077 0.004 0.27 0.12 22.77 8.81 3.57 64.00 - <![CDATA[S1]]> 0.042 0.07 0.05 0.007 0.19 0.14 22.32 8.71 3.89 64.43 0.009 <![CDATA[S2]]> 0.041 0.05 0.05 0.006 0.20 0.14 22.55 8.66 3.83 64.33 0.012 <![CDATA[S3]]> 0.031 0.27 0.05 0.003 0.21 0.13 22.53 8.56 3.83 64.52 0.029

[0041] The analysis method of each embodiment is carried out according to the following content:

[0042] The main equipment used was a Thermo Scientific energy dispersive spectrometer (EDS). The main analysis software used in this invention was Image J, Image pro plus, and Origin 2024.

[0043] The main settings of the energy dispersive spectrometer are as follows:

[0044] (1) Setting the acceleration voltage to 5kV~10kV can excite most medium and heavy elements; the beam intensity is set to 13A~15A;

[0045] (2) Sample positioning and focusing: Switch to secondary electron imaging mode or backscattered electron imaging mode, observe the sample morphology, select the target area, perform automatic focus and aberration correction; define the surface scanning area in EDS to ensure that the surface scanning area contains the precipitated phase to be characterized;

[0046] (3) Acquisition parameter settings: Set the image resolution to 512×512; the dwell time per pixel is usually set to 500µs;

[0047] (4) Start the mapping acquisition program to generate elemental maps to intuitively display the coexistence relationship of elements.

[0048] (5) Open the elemental surface scan distribution map in Image J, use the Trainable Weka Segmentation plug-in to separate the precipitated phase area from the non-precipitated phase area, and generate the segmented image;

[0049] The specific steps are as follows: Open Image J, click File → Open, and select the main element surface scan distribution map of the precipitate phase to be counted. Enter WEKA in the Image J input box, select the Trainable Weka Segmentation image processing plug-in, click Run, select the precipitate phase to be counted in Add to class 1, and select the matrix and other precipitates to be stripped in Add to class 2. Since the main elements in different precipitates and matrices vary significantly, the area to be counted can be clearly marked in the element distribution map. After making your selection, click Train classifier to complete the area marking. Finally, click Create result.

[0050] (6) Convert the segmented image to 16-bit format in Image J, perform binarization on the segmented image after format conversion, and save it in Tif format;

[0051] The specific steps are as follows: Click Image → Type → 16 bit, Process → Binary → Make Binary, Process → Noise → Remove outliers → Radius, set the noise radius to 2, and click OK. File → Save as → Tif to any path. Click Toggle overlay to switch between the original and the rendered image. Achieving the best processing results often requires multiple adjustments. Add and delete unnecessary parts repeatedly in Add to class 1 and Add to class 2. Perform noise reduction on the binarized image, setting the noise radius to 2-3 pixels. This will yield a denoised binarized image, which can then be processed using Image Pro Plus software.

[0052] (7) Open the binary processed image in Image Pro Plus software, perform automatic bright objects statistics, measure and count the precipitated phase area, and obtain statistical data;

[0053] The specific steps are as follows: Open Image Pro Plus, open the Tif file you just saved, select the ruler, click Count and measure objects, click Automatic bright objects, click Measure → Select measurements → per area → OK. Click Count → File → Date to clipboard to copy the calculated data, and paste the statistical data into the Origin data sheet.

[0054] (8) Perform Tukey HSD significance analysis on the statistical data in Origin to obtain the significance level p. The p-value is used to judge the difference within or between statistical data groups, reflecting the ability of a certain set process parameter to regulate the precipitation phase in nickel-based alloys.

[0055] To do this, install and click the Paired comparison plot plugin in Origin. Add a data column to DateColumn, add a category column to Group columns, select the graphic type in Plot type, select the significance marker style in Significance mark, select Tukey in Mean comparison methods, and click OK. Plot types available for selection include Bar, Date points, and Scatter with errorbar.

[0056] Example 1

[0057] The initial Inconel 625 nickel-based alloy S0 was micro-alloyed with 0.009 wt.% magnesium to obtain sample S1 for processing. Samples S0 and S1 were tested and analyzed to investigate the ability of Mg content to regulate the harmful segregation phases NbC phase and Laves phase.

[0058] In step (2), the main precipitated phases NbC and Laves phase of the sample S1 to be tested are scanned.

[0059] In step (5), the distribution map of elements such as Nb and Mo is opened in Image J. Enter WEKA in the Image J software input box, select the Trainable Weka Segmentation image processing plug-in, click Run, select the main distribution area of ​​elements such as Nb and Mo in Add to class 1, and select the matrix and other precipitated phases to be stripped in Add to class 2. After the selection is completed, click Train classifier to complete the area marking. Finally, click Create result.

[0060] In step (7), click Count→File→Date to clipboard to copy the calculated data and paste the statistical data into the Origin data sheet. Select five representative NbC and Laves face scans and repeat steps (2) to (7) to improve the accuracy of the statistical results.

[0061] After step (8), we get Figure 1 (a) and Figure 1 The significant comparison results in (b) are analyzed as follows:

[0062] (8-1) Comparison between median and mean. Figure 1 The box plots of the two types of precipitated phases (NbC phase and Laves phase) on the left and right show that their medians are close to the center of the box and their means are slightly higher (according to the "★" mark), but the difference is not large, indicating that the data distribution is approximately symmetrical; the right skewness of the distribution is weak, and the number of large-particle precipitates is small.

[0063] (8-2) Interquartile range (IQR) and whisker judgment. In this case, the IQR is short, indicating that the distribution of the precipitate phase particle size is highly concentrated; the lengths of the upper and lower whiskers are relatively symmetrical, indicating that there is no obvious deviation in the data.

[0064] (8-3) Outlier Analysis. A small number of outliers (marked with "●") are present in both the NbC phase and the Laves phase, indicating that the particle size of some analytes is slightly larger than the typical range. However, these outliers are relatively few in number and are considered occasional abnormalities that do not dominate the overall distribution.

[0065] (8-4) Significance Level (p-value) Analysis. The "☆" in the figure indicates statistically significant differences between groups, with the number of ☆s representing the degree of significance between samples. After adding 0.009 wt.% Mg, the precipitation areas of the NbC and Laves phases were significantly reduced compared to the initial group S0. This indicates that Mg, at the concentration specified in Example 1, effectively promotes particle size refinement and uniform distribution of the precipitated NbC and Laves phases.

[0066] (8-5) According to the above Tukey HSD analysis results, when 0.009wt.% Mg is added, the area distribution of the precipitated phases (NbC phase and Laves phase) shows a high degree of concentration; the distribution is approximately symmetrical, and the median is close to the mean; the number of outliers is small, and the abnormal precipitation behavior is suppressed; the significance analysis shows p<0.05, proving that the regulatory effect of Mg is statistically significant; this indicates that at this content, Mg has a significant effect on the size refinement and uniform distribution of the NbC phase and Laves phase in Inconel 625 nickel-based alloy.

[0067] Example 2

[0068] 0.012wt.% magnesium was added to the initial Inconel 625 nickel-based alloy S0 for microalloying treatment to obtain the processed sample S2. The ability of the Mg content in sample S2 to regulate the harmful segregation phases NbC phase and Laves phase was detected and analyzed.

[0069] In step (2), the main precipitated phases NbC and Laves phase of the sample S2 to be tested are scanned.

[0070] In step (5), the distribution map of elements such as Nb and Mo is opened in Image J. Enter WEKA in the Image J software input box, select the Trainable Weka Segmentation image processing plug-in, click Run, select the main distribution area of ​​elements such as Nb and Mo in Add to class 1, and select the matrix and other precipitated phases to be stripped in Add to class 2. After the selection is completed, click Train classifier to complete the area marking. Finally, click Create result.

[0071] In step (7), click Count→File→Date to clipboard to copy the calculated data and paste the statistical data into the Origin data sheet. Select five representative NbC and Laves face scans and repeat steps (2) to (7) to improve the accuracy of the statistical results.

[0072] After step (8), we get Figure 2 (a) and Figure 2 The significant comparison results in (b) are analyzed as follows:

[0073] (8-1) Comparison between median and mean. Figure 2It can be seen that the medians of NbC and Laves phase are very close to the mean (marked with "★") and almost overlap, indicating that the data skewness is basically eliminated and presents an approximately normal distribution. Compared with the case of S1 with 0.009wt.% Mg added, the distribution is more symmetrical and concentrated.

[0074] (8-2) Interquartile range (IQR) and whisker interpretation. The IQR is significantly reduced: the box height decreases, indicating that the sizes of most precipitates are within a very narrow range; the upper and lower whiskers are almost the same length, indicating a very balanced distribution with low volatility; this indicates that the variability of the precipitate particle size has significantly decreased and is becoming more consistent.

[0075] (8-3) Outlier analysis. The number of outliers has been further reduced (marked with "●"), with only a few data points remaining. The outliers are located near the ends of the whiskers, indicating that extremely coarse precipitates no longer appear. The regulatory effect of Mg has become more stable, almost eliminating the phenomenon of abnormal localized precipitate aggregation.

[0076] (8-4) Significance Level (p-value) Analysis. The "☆" in the figure indicates statistically significant differences between groups, with the number of ☆s representing the degree of significance between samples. After adding 0.012 wt.% Mg, the precipitation area under the current treatment conditions showed statistically significant differences compared to the other groups (0.009 wt.% Mg) (p < 0.05 or p < 0.01). This indicates that the Mg refinement and homogenization effects have been significantly optimized.

[0077] (8-5) According to the above Tukey HSD analysis results, when adding 0.012wt.% Mg, the particle size distribution of the precipitate phase is highly concentrated and the IQR is significantly narrowed; the median and the mean almost coincide, indicating an approximate normal distribution with excellent distribution symmetry; there are very few outliers, and there are no longer extremely coarse precipitates; the significance level results (p<0.05) further verify that Mg has a regulatory effect on enhancing the uniformity of precipitate distribution and particle size refinement; compared with 0.009wt.% Mg, 0.012wt.% Mg shows a more stable control effect, and the effect tends to optimize the critical point.

[0078] Example 3

[0079] The processed sample S3 was obtained by adding 0.029wt.% magnesium to the initial Inconel 625 nickel-based alloy S0 for microalloying treatment. The ability of the Mg content in the S3 sample to regulate the harmful segregation phases NbC phase and Laves phase was detected and analyzed.

[0080] In step (2), the main precipitated phases NbC and Laves phase of the sample S3 to be tested are scanned.

[0081] In step (5), the distribution map of elements such as Nb and Mo is opened in Image J. Enter WEKA in the Image J software input box, select the Trainable Weka Segmentation image processing plug-in, click Run, select the main distribution area of ​​elements such as Nb and Mo in Add to class 1, and select the matrix and other precipitated phases to be stripped in Add to class 2. After the selection is completed, click Train classifier to complete the area marking. Finally, click Create result.

[0082] In step (7), click Count→File→Date to clipboard to copy the calculated data and paste the statistical data into the Origin data sheet. Select 5 representative NbC and Laves face scans and repeat steps (2) to (7) to improve the accuracy of the statistical results.

[0083] After step (8), we get Figure 3 (a) and Figure 3 The significant comparison results in (b) are analyzed as follows:

[0084] (8-1) Comparison between median and mean. Figure 3 It can be seen that the median and mean of NbC and Laves phase almost completely coincide with each other (marked with "★"), indicating that the particle size data of the precipitated phase are symmetrically distributed, close to normal, and almost unbiased. Compared with the distribution under the conditions of no Mg addition or low Mg concentration, the symmetry of the particle size distribution is significantly enhanced.

[0085] (8-2) Interquartile range (IQR) and whisker analysis. The IQRs of the two precipitates are significantly reduced, and the boxes are extremely short. The upper and lower whiskers are essentially symmetrical and of similar length. This indicates that the concentration of the precipitate size distribution has reached an optimal state, and the data fluctuation range has been minimized.

[0086] (8-3) Outlier analysis. Figure 3 There are very few outliers (marked with "●"), and only a few outliers are seen, which are close to the tail of the whisker; there are no extreme values ​​far away from the main body, indicating that the abnormal coarse precipitates have been almost completely eliminated; this shows that when Mg is 0.029 wt.%, the ultimate regulation effect is achieved, and the precipitation behavior is stable and controllable.

[0087] (8-4) Significance level (p-value) analysis. The presence of a significant mark ("☆") in the figure indicates that the differences in the areas of NbC and Laves phases compared to the control group (e.g., S0 without Mg) or the low-Mg content group are significant (p < 0.05 or p < 0.01). This suggests that the 0.029 wt.% Mg treatment effectively inhibits the excessive precipitation and agglomeration growth of NbC and Laves phases.

[0088] (8-5) According to the above Tukey HSD analysis results, under the condition of adding 0.029wt.% Mg, the Tukey significance analysis results show that: the particle size distribution is the most concentrated, almost showing an ideal normal distribution; the IQR shrinks significantly, indicating that the NbC and Laves phase sizes are highly consistent and the volatility is minimal; there are very few outliers, and no large coarse precipitates exist; the significance test results show that this treatment method can statistically significantly improve the uniformity of the structure; indicating that the amount of Mg added may have reached the saturation point of the precipitate phase refinement control, and further improvement will tend to stabilize.

[0089] Example 4

[0090] In order to study the solidification process of Inconel 625 alloy, an in-situ observation experiment was conducted using a laser confocal microscope at a cooling rate of 20°C / min. After the experiment, four completely melted samples S under magnesium gradients (0wt.%, 0.009wt.%, 0.012wt.%, and 0.029wt.%) were obtained. 0-4 、S 1-4 、S 2-4 、S 3-4 , detect and analyze S 0-4 、S 1-4 、S 2-4 、S 3-4 The ability of Mg to regulate the harmful segregation phases NbC phase and Laves phase in the alloy at a set cooling rate.

[0091] In step (2), the sample S to be tested 0-4 、S 1-4 、S 2-4 、S 3-4 Surface scanning was performed on the main precipitated phases NbC and Laves phase.

[0092] In step (5), the distribution map of elements such as Nb and Mo is opened in Image J. Enter WEKA in the Image J software input box, select the Trainable Weka Segmentation image processing plug-in, click Run, select the main distribution area of ​​elements such as Nb and Mo in Add to class 1, and select the matrix and other precipitated phases to be stripped in Add to class 2. After the selection is completed, click Train classifier to complete the area marking. Finally, click Create result.

[0093] In step (7), click Count→File→Date to clipboard to copy the calculated data and paste the statistical data into the Origin data sheet. Select 5 representative NbC and Laves face scans and repeat steps (2) to (7) to improve the accuracy of the statistical results.

[0094] After step (8), we get Figure 4 The specific analysis is as follows:

[0095] (8-1) Comparison between median and mean. 0-4 The median of the group is slightly lower than the mean, and the data are right-skewed, indicating that some large-sized precipitates exist under the condition of no magnesium; S 1-4 With S 2-4 The group median gradually approaches the mean, and the distribution symmetry improves; S 3-4 The group median and mean almost completely coincide with each other (marked with “★”), indicating that the particle size distribution of the precipitated phase tends to be normal and the control effect is ideal.

[0096] (8-2) Interquartile range (IQR) and whisker judgment. As the magnesium content increases, the S 0-4 to S 3-4 In the S group, the IQR was significantly reduced and the box height was reduced, indicating that the particle size fluctuation range was narrowed and the distribution concentration was enhanced. 3-4 The group box is extremely short, and the upper and lower whiskers are symmetrical, indicating that the size distribution of the precipitated phase is stable and uniform.

[0097] (8-3) Outlier analysis. 0-4 There are many outliers (marked with "●") on the upper side of the whiskers in the S group, indicating that there are many abnormal coarse precipitates; 1-4 With S 2-4 Outliers decrease and tend to be closer to the tail; S 3-4 There are only a few edge outliers in the group, indicating that the abnormal size precipitation behavior is basically suppressed and the alloy structure tends to be uniform.

[0098] (8-4) Analysis of significance level (p value). The “☆” mark in the figure indicates S 3-4 With S 0-4 、S 1-4 、S 2-4 There are statistically significant differences between the two groups (p<0.05 or p<0.01). 3-4 The precipitation area of ​​the group was significantly reduced compared with that of the other groups. The regulation of Mg on the precipitation behavior at this content has reached a highly effective level.

[0099] (8-5) According to the above Tukey HSD analysis results, at a cooling rate of 20℃ / min, as the Mg content gradually increases from 0 to 0.029wt.%, the precipitation area of ​​NbC and Laves phase in Inconel 625 alloy shows a trend of gradually decreasing, concentrated distribution, and significant enhancement. 3-4 The group exhibited the best particle size uniformity and the fewest outliers, and the significance analysis showed that the control effect was statistically valid. Therefore, it can be inferred that under this cooling rate condition, 0.029wt.% Mg is the optimal content range for precipitate control, and further increasing the concentration may lead to a control saturation effect.

[0100] Example 5

[0101] In order to study the solidification process of Inconel 625 alloy, an in-situ observation experiment was conducted at a cooling rate of 40°C / min using a laser confocal microscope. After the experiment, four completely melted samples S under magnesium gradients (0wt.%, 0.009wt.%, 0.012wt.%, and 0.029wt.%) were obtained. 0-5 、S 1-5 、S 2-5 、S 3-5 , detect and analyze S 0-5 、S 1-5 、S 2-5 、S 3-5 The ability of Mg to regulate the harmful segregation phases NbC phase and Laves phase in the alloy at a set cooling rate.

[0102] In step (2), the sample S to be tested 0-5 、S 1-5 、S 2-5 、S 3-5 Surface scanning was performed on the main precipitated phases NbC and Laves phase.

[0103] In step (5), the distribution map of elements such as Nb and Mo is opened in Image J. Enter WEKA in the Image J software input box, select the Trainable Weka Segmentation image processing plug-in, click Run, select the main distribution area of ​​elements such as Nb and Mo in Add to class 1, and select the matrix and other precipitated phases to be stripped in Add to class 2. After the selection is completed, click Train classifier to complete the area marking. Finally, click Create result.

[0104] In step (7), click Count→File→Date to clipboard to copy the calculated data and paste the statistical data into the Origin data sheet. Select 5 representative NbC and Laves face scans and repeat steps (2) to (7) to improve the accuracy of the statistical results.

[0105] After step (8), we get Figure 5 The specific analysis is as follows:

[0106] (8-1) Comparison between median and mean. Figure 5 As can be seen, S 0-5 The median of the group deviates significantly from the mean, and the distribution is slightly right-skewed, indicating that some coarse precipitates exist under the condition of no Mg addition; and with the increase of Mg content, S 1-5 and S 2-5 The group median gradually approaches the mean; S 3-5 The group median and mean almost completely coincide with each other (marked with “★”), indicating that the particle size distribution is well symmetrical and tends to be an ideal normal distribution.

[0107] (8-2) Interquartile range (IQR) and whisker judgment. 0-5 The group IQR is large, the box is elongated, and the size fluctuation is significant; S 1-5 With S 2-5 The IQR of the group gradually narrowed, S 3-5 The height of the group box is the lowest, showing the high concentration of precipitation area. 3-5 The lengths in the group are close and the shapes are symmetrical, indicating that the particle size distribution is uniform and the fluctuation is very low.

[0108] (8-3) Outlier analysis. 0-5 The group shows multiple outliers (marked with "●") far away from the whiskers on the box, indicating the presence of a large number of abnormal coarse precipitates; while S 1-5 and S 2-5 The number of outliers in the group decreased significantly; S 3-5 There are almost no distal outliers in the group, and only individual data points close to the whiskers exist, indicating that Mg at 0.029wt.% basically suppresses the heterogeneous precipitation behavior.

[0109] (8-4) Analysis of significance level (p value). The “☆” mark in the figure indicates that S 3-5 Group and S 0-5 、S 1-5 、S 2-5 There were statistically significant differences between the two groups (p<0.05 or p<0.01), especially in the Laves phase, where the area decreased significantly, confirming that high Mg concentration effectively controlled the excessive growth of precipitates.

[0110] (8-5) According to the above Tukey HSD analysis results, under the cooling rate of 40℃ / min, when the Mg content gradually increases from 0wt.% to 0.029wt.%, the precipitation area of ​​NbC and Laves phase is significantly reduced, the distribution symmetry is enhanced, the volatility is weakened, the abnormal particles are greatly reduced, and the significance between groups is enhanced. 3-5 The indicators of the group were the best, indicating that the Mg regulation effect tended to be stable at this time and was close to the saturation point of precipitation regulation ability.

Claims

1. A quantitative statistical and analytical method for precipitation phases in nickel-based high-temperature alloys, characterized in that: Includes the following: An energy dispersive spectrometer is used to perform surface scanning on nickel-based alloy samples to obtain a surface scanning distribution detection map. The surface scanning distribution detection map is processed using software, and the detection map is analyzed for elemental composition and phase. Based on the analysis results, the ability of a set process parameter to control the size of the precipitated phase in the nickel-based alloy sample and the uniformity of the distribution of the precipitated phase in the sample is determined; The steps for processing the surface scan distribution detection map using software are as follows: (1) Open the element surface scan distribution map, separate the precipitated phase area from the non-precipitated phase area, and generate a segmented image; (2) Binarization is performed on the segmented image to obtain binary images of the precipitation phase region and the non-precipitation phase region; (3) Automatic bright area statistics are performed on the binary processed image to measure and count the precipitation phase area to obtain the precipitation phase area statistics; (4) The statistical data of the precipitated phase area obtained were analyzed for significance, and the statistical data were subjected to the Tukey's honest significant difference test. The differences within or between statistical data groups were judged by the p-value, reflecting the ability of a certain set process parameter to regulate the size of the precipitated phase in the nickel-based alloy and the uniformity of the distribution of the precipitated phase in the sample; ， in: q is the standardized range statistic, , is the mean of the two sets of statistical data to be compared, MSE is the mean square error of the one-way ANOVA, n is the number of samples of the statistical data, and the p-value is q The corresponding right-tail probability in the standardized range distribution table; When p<0.05, it indicates that a certain set process parameter has the ability to regulate the precipitated phase, can promote the refinement and homogenization of the precipitated phase size, and promote the uniform distribution of the precipitated phase in the sample. The smaller the p value, the stronger the regulation ability.

2. The quantitative statistical and analytical method for precipitated phases in nickel-based high-temperature alloys according to claim 1, characterized in that: The energy dispersive spectrometer parameters are set as follows: (1) Set the accelerating voltage to 5kV~10kV and the beam current to 13A~15A; (2) Switch to secondary electron imaging mode or backscatter imaging mode, observe the sample morphology, and define the surface scanning area, which contains the precipitated phase to be characterized; (3) Set the image resolution to ≥256×256 and the dwell time per pixel to 50µs~500µs; (4) Start the surface scan acquisition program to generate an element surface scan distribution map to show the element coexistence relationship.

3. The quantitative statistical and analytical method for precipitated phases in nickel-based high-temperature alloys according to claim 1, characterized in that: The set process parameters include alloy composition, element content in the alloy, cooling rate, and cooling time.

4. A quantitative statistical and analytical method for precipitated phases in nickel-based high-temperature alloys according to claim 1 or 2, characterized in that: The precipitated phases include NbC phase, Laves phase, M6C phase and γ′ phase.

5. The quantitative statistical and analytical method for precipitated phases in nickel-based high-temperature alloys according to claim 1, characterized in that: In step (3), the binary processed image in step (2) is first subjected to denoising, and the noise radius is set to 2-3 pixels to obtain the denoised binary processed image, and then automatic bright area statistics are performed.

6. The quantitative statistical and analytical method for precipitated phases in nickel-based high-temperature alloys according to claim 1, characterized in that: In step (4), when p < 0.01, it indicates that a certain set process parameter has a significant effect on the regulation of the precipitated phase, and the precipitated phase is statistically homogenized.

7. The quantitative statistical and analytical method for precipitated phases in nickel-based high-temperature alloys according to claim 1, characterized in that: In step (4), when p>0.05, it means that the effect of a certain set process parameter on the regulation of the precipitated phase has reached the upper limit, and the distribution of the precipitated phase has entered a stable stage.

8. The quantitative statistics and analysis method for precipitated phases in nickel-based high-temperature alloys according to claim 1, characterized in that: In step (4), if the sample sizes of the two sets of statistical data are equal, then n is the number of samples in each set of statistical data; if the number of samples in the two sets of statistical data is not equal, then n is the equivalent sample number of the statistical data.

Images