A Method for Establishing a High-Temperature Alloy Hot Corrosion Model Based on Image Recognition

Abstract

This invention discloses a method for establishing a high-temperature alloy hot corrosion model based on image recognition, comprising: preparing a sample and conducting a hot corrosion test; photographing and recording the surface morphology of the sample after hot corrosion during the test, and preprocessing the images; performing grayscale analysis on the preprocessed results to calculate the mean grayscale G and grayscale distribution of the surface images after hot corrosion; calculating the fractal dimension D of the corrosion morphology of different sample surfaces using the box-ring method; and establishing a high-temperature alloy hot corrosion model based on the mean grayscale G and the fractal dimension D of the corrosion morphology. This invention can infer the hot corrosion temperature and time based on the surface morphology characteristics of the high-temperature alloy after hot corrosion, providing a technical basis for rapidly determining the hot corrosion temperature and time in engineering practice.

Description

Technical Field

[0001] This invention relates to the field of high-temperature alloy hot corrosion technology, specifically a method for establishing a high-temperature alloy hot corrosion model based on image recognition. Background Technology

[0002] Nickel-based alloys, for example, are widely used in key hot-end components of aero engines due to their excellent high-temperature mechanical properties. In the long-term high-humidity, high-salt-spray marine atmospheric environment, corrosive molten salts accumulate on the surface of turbine blades in marine gas turbines and carrier-based aircraft engines. These deposited salts can damage the protective oxide layer on the alloy surface under high-temperature conditions, leading to hot corrosion.

[0003] Traditional methods for describing the degree of hot corrosion of nickel-based superalloys include thermogravimetric analysis (TGA), which characterizes the hot corrosion process by measuring mass changes (see invention patent "Test Method for Hot Corrosion Performance of Nickel-Based Single Crystal Superalloys", application publication number CN110132826A); and metallographic characterization, which involves preparing metallographic samples through cutting and grinding processes and observing the thickness and morphology of the corrosion layer in the cross-section to characterize the degree of hot corrosion (see invention patents "An Analytical Method for Hot Corrosion Performance of Nickel-Based Single Crystal Superalloys", application publication number CN111896458A; invention patent "A Method for Damage Assessment of Thermal Barrier Coatings Based on Corrosion Kinetics", application publication number CN110132826A). These methods can establish models of sample mass, corrosion layer thickness, and hot corrosion temperature and time, but they are complex to operate, require sample destruction, and are difficult to apply in practice. Therefore, there is an urgent need for a more convenient and accurate hot corrosion temperature and time model to more accurately determine the degree of hot corrosion damage to the sample, the temperature of the hot corrosion environment, and the duration of hot corrosion. Summary of the Invention

[0004] To address the aforementioned issues, this invention proposes a method for establishing a high-temperature alloy hot corrosion model based on image recognition. This method characterizes the hot corrosion process by extracting surface morphology features of the sample after hot corrosion, and can represent the relationship between the mean gray level and fractal dimension of the hot-corroded surface and the hot corrosion temperature and time.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solution:

[0006] This invention is a method for establishing a high-temperature alloy hot corrosion model based on image recognition, comprising the following operations:

[0007] Prepare samples and conduct hot corrosion tests;

[0008] The surface morphology of the specimen after hot corrosion was photographed and recorded during the hot corrosion test, and the images were preprocessed.

[0009] Gray-scale analysis was performed on the preprocessing results to calculate the mean gray-scale G and gray-scale distribution of the surface image of the sample after hot corrosion.

[0010] The fractal dimension D of corrosion morphology on different sample surfaces was calculated using the box method.

[0011] A hot corrosion model for high-temperature alloys was established based on the mean gray level G and the fractal dimension D of the corrosion morphology.

[0012] A further improvement of the present invention is that: preparing a sample and conducting a hot corrosion test specifically includes:

[0013] Several circular pieces with a diameter of 20 mm and a thickness of 2 mm were made from GH4169 alloy. They were then polished, cleaned, and dried in sequence to obtain multiple spare samples.

[0014] Multiple samples were divided into several groups and embedded in a mixed salt of Na2SO4 and NaCl with a mass ratio of 3:1. They were then placed in high-temperature furnaces at different temperatures for thermal corrosion.

[0015] After the hot corrosion test begins, several samples from each group are taken at set intervals and cooled in air. They are then boiled in deionized water, ultrasonically cleaned, and dried in sequence until the hot corrosion test ends.

[0016] A further improvement of the present invention is that: the surface morphology of the sample after hot corrosion in the hot corrosion test is captured and recorded, and the image is preprocessed, specifically including:

[0017] The surface morphology of the sample after hot corrosion was captured by a camera under a fixed light source.

[0018] The sample surface is identified and excess background is removed. The color image is converted into a grayscale image, and a grayscale pixel matrix A is established.

[0019] A further improvement of this invention lies in: performing grayscale analysis on the preprocessing results to calculate the mean grayscale G and grayscale distribution of the surface image of the sample after hot corrosion. Specific operations include:

[0020] The grayscale is divided into 256 levels. The element range of the grayscale pixel matrix A is [0, 255], which represents the grayscale value of the corresponding pixel. Black is 0 and white is 255. The mean grayscale G of each sample is calculated to characterize the overall color level of the sample surface after hot corrosion. The mean grayscale of the original sample surface before hot corrosion is denoted as G0.

[0021] A further improvement of this invention lies in: calculating the fractal dimension D of the corrosion morphology on different sample surfaces using the box method, specifically including:

[0022] A gray-level pixel matrix A is covered with squares of a certain scale. The number of boxes containing the image is counted based on the comparison between the average gray level of the image within the box and the average gray level of the entire image.

[0023] By changing the size of the boxes, plot the number of boxes and the box size on a double logarithmic coordinate system. The slope of the linear region represents the fractal dimension D of the corrosion morphology. The fractal dimension of the corrosion morphology of the original sample surface image before hot corrosion is denoted as D0.

[0024] A further improvement of this invention is that the high-temperature alloy hot corrosion model includes a grayscale model based on the mean grayscale G of the surface image after hot corrosion of the sample and a high-temperature hot corrosion fractal dimension model based on the fractal dimension D of the corrosion morphology of the sample surface. The relationship between the mean grayscale G and the test temperature in the grayscale model is expressed as follows:

[0025] S=K G T B

[0026]

[0027] Where S is the average grayscale attenuation rate of the surface, T is the hot corrosion temperature, and K is the kJ / m² temperature. G Here, B is the material condition coefficient, G0 is the grayscale variation index, and G0 is the mean grayscale of the original sample surface before hot corrosion.

[0028] The relationship between the fractal dimension D of the corrosion morphology on the sample surface and the hot corrosion time in the high-temperature hot corrosion fractal model is expressed as follows:

[0029] D=K D ·t+C

[0030] Where t is the thermal corrosion time, K D Here, C is the material condition coefficient, and C is a constant.

[0031] The beneficial effects of the present invention are: 1. The present invention provides a new approach for estimating hot corrosion temperature and time. The method of extracting hot corrosion morphology feature parameters by image recognition can accurately and effectively characterize the degree of hot corrosion of the sample.

[0032] 2. This invention has a certain degree of universality. The exponential relationship between the mean gray level and the hot corrosion temperature in the gray-scale model, and the linear relationship between the fractal dimension of corrosion morphology and corrosion time in the high-temperature hot corrosion fractal dimension model, are commonly found in the hot corrosion of high-temperature alloys. The material condition coefficient K can be revised according to different sample materials and hot corrosion conditions. G K D .

[0033] 3. This invention can be applied in engineering practice to quickly determine the hot corrosion damage status of high-temperature alloy materials, and further determine the hot corrosion temperature and time range. The method of extracting hot corrosion feature parameters by image recognition also greatly avoids the deviation caused by the traditional manual visual method to determine the degree of hot corrosion. Attached Figure Description

[0034] Figure 1 This is a flowchart of the method in an embodiment of the present invention;

[0035] Figure 2 This is a schematic diagram of the surface morphology image processing of the sample after hot corrosion in an embodiment of the present invention;

[0036] Figure 3 This is a schematic diagram of grayscale model data fitting in an embodiment of the present invention;

[0037] Figure 4 This is a schematic diagram of data fitting for the high-temperature thermal corrosion fractal model in an embodiment of the present invention. Detailed Implementation

[0038] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0039] like Figure 1 As shown in this embodiment, a method for establishing a high-temperature alloy hot corrosion model based on image recognition includes the following operations:

[0040] Step 1, prepare the sample and conduct a high-temperature alloy hot corrosion test T, specifically including:

[0041] Step 1.1: Cut the high-temperature alloy GH4169 into circular pieces with a diameter of 20 mm and a thickness of 2 mm. Polish the sample with SiC sandpaper of increasing grit until smooth. Soak it in ethanol solution and clean it with an ultrasonic cleaner for 5 minutes to remove abrasive residue and oil stains from the surface. Dry it in a drying oven at 150°C for 30 minutes.

[0042] Step 1.2: Place the sample in a quartz crucible. Weigh Na₂SO₄ and NaCl at a mass ratio of 75% and 25% respectively, mix thoroughly, pour into the quartz crucible, cover the sample, and place in a tube furnace. Conduct four sets of tests, with temperatures set at constant 650℃, 750℃, and 900℃, and variable temperature from 650℃ to 900℃. According to the test plan, every 6 or 12 hours, remove two samples from each group and allow them to cool naturally in air. Then, sequentially boil them in deionized water for 30 minutes, ultrasonically clean them in ethanol for 10 minutes to remove residual salts from the surface, and finally dry them in a drying oven at 150℃ for 30 minutes. Take high-resolution images of the front and back of the sample under a fixed light source using a digital camera until the 72-hour test is completed.

[0043] Step 2 involves preprocessing the image showing the surface morphology of the sample after hot corrosion. For example... Figure 2 As shown, it specifically includes:

[0044] Step 2.1: Digitize the image data. Scan the image from left to right and top to bottom respectively, identify the φ20mm circular area with data fluctuations, define it as the sample, and define the surrounding area with uniform data as the background. Delete the background to obtain the hot corrosion surface morphology image containing only the sample.

[0045] Step 2.2: Convert the image to grayscale and establish a grayscale pixel matrix A based on the image pixels.

[0046] Step 3: Perform grayscale analysis on the preprocessing results to calculate the mean grayscale G and grayscale distribution of the surface image of the sample after hot corrosion.

[0047] The grayscale is divided into 256 levels. The grayscale value corresponding to pure white is normalized to 255. All element values ​​in the grayscale pixel matrix A are mapped to the range [0, 255] by this normalization factor. Pure black is 0 and pure white is 255. This is denoted as matrix A'.

[0048] (1) The mean value G of matrix A' is denoted as the mean gray level of the image, and the mean gray level of the image of the sample before hot corrosion is denoted as G0. Detailed data are shown in Table 1:

[0049] Table 1. Mean gray values ​​of samples before and after hot corrosion

[0050]

[0051]

[0052] Step 4: Calculate the fractal dimension D of the corrosion morphology on different sample surfaces using the box method. This includes:

[0053] Step 4.1: Based on the row and column size of matrix A', normalize it into a 256x256 matrix A", and use the mean value for each element.

[0054] Step 4.2, according to the side length 2 n The matrix A” is divided into squares (n is a number from 1 to 8), and the divided unit cells are processed using the box method.

[0055] Step 4.3: Plot the results on a double logarithmic coordinate system using 8 squares. Record the slope as the fractal dimension D of the corrosion morphology. Record the fractal dimension of the sample image before hot corrosion as D0. Detailed data for samples subjected to constant temperature at 900℃ and variable temperature at 650–900℃ are shown in Table 2.

[0056] Table 2 Fractal dimension of sample surface morphology before and after high-temperature thermal corrosion

[0057]

[0058] Step 5: Establish a high-temperature alloy hot corrosion model based on the mean gray level G and the fractal dimension D of the corrosion morphology. The high-temperature alloy hot corrosion model includes a gray level model based on the mean gray level G of the surface image after hot corrosion of the sample, and a high-temperature hot corrosion fractal dimension model based on the fractal dimension D of the corrosion morphology of the sample surface. The principle of the gray level model is as follows: In the hot corrosion of high-temperature alloys, as hot corrosion occurs, the surface changes from an initial metallic silver-white to being covered with a dark brown corrosion scale. Different temperatures result in different colors and coverage of the corrosion scale, which can be characterized by the mean gray level G of the image. Its value is mainly related to temperature and independent of corrosion time. The principle of the high-temperature hot corrosion fractal dimension model is as follows: For low-temperature hot corrosion samples (650℃), surface hot corrosion occurs relatively uniformly and slowly, and the change in the fractal dimension D of the corrosion morphology of the image with corrosion time is not obvious. However, for high-temperature hot corrosion samples (900℃), the surface corrosion layer exhibits numerous pits, plowshares, and other obvious corrosion characteristics, and the fractal dimension D of the corrosion morphology shows a decreasing trend with corrosion time. In addition, the complexity of the sample surface after corrosion is definitely greater than that before corrosion. As the hot corrosion deepens, when the fractal dimension of the sample surface image is close to the original fractal dimension D0 of the sample before corrosion, it can be approximately considered that the hot corrosion has reached the critical point, the growth and shedding of the corrosion layer reach equilibrium, and the thickness does not change. This moment is the critical corrosion time cct.

[0059] In this embodiment, the surface mean grayscale attenuation rate S is calculated according to Table 1, and the expression is:

[0060]

[0061] The calculation results are shown in Table 3 below:

[0062] Table 3 Surface Mean Gray-Scale Attenuation Rate

[0063]

[0064] The surface mean grayscale attenuation rate is related to the hot corrosion temperature as follows:

[0065] S=K G T B

[0066] Where T is the hot corrosion temperature, K G B is the material working condition coefficient, and B is the grayscale change index.

[0067] The least squares method was used to fit the data in Table 3, and the fitting results are as follows. Figure 3 Show.

[0068] K G =2.47095E18, B=-5.97438

[0069] There is a relationship between the fractal dimension D of the corrosion morphology on the sample surface and the hot corrosion time t during high-temperature hot corrosion:

[0070] D=K D ·t+C

[0071] The material condition coefficient K was calculated based on the data in Table 2. D The constant C, and the fitting results are as follows: Figure 4 As shown. Simultaneously, the critical corrosion time cct is calculated when the fractal dimension D of the corrosion morphology is equal to the fractal dimension D0 of the original sample surface image before hot corrosion:

[0072]

[0073] In summary, this embodiment clearly demonstrates the relationship between the mean grayscale and fractal dimension of the hot-corroded surface and the hot-corrosion temperature and time, providing a technical basis for quickly determining the hot-corrosion temperature and time in engineering practice.

[0074] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless defined as herein.

[0075] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for establishing a high-temperature alloy hot corrosion model based on image recognition, characterized in that: This includes the following operations: Prepare samples and conduct hot corrosion tests; The surface morphology of the specimen after hot corrosion was photographed and recorded during the hot corrosion test, and the images were preprocessed. Gray-scale analysis was performed on the preprocessing results to calculate the mean gray-scale G and gray-scale distribution of the surface image of the sample after hot corrosion. The fractal dimension D of corrosion morphology on different sample surfaces was calculated using the box method. A hot corrosion model for high-temperature alloys was established based on the mean gray level G and the fractal dimension D of the corrosion morphology. The high-temperature alloy hot corrosion model includes a grayscale model based on the mean grayscale G of the surface image after hot corrosion of the sample and a high-temperature hot corrosion fractal dimension model based on the fractal dimension D of the corrosion morphology of the sample surface. The relationship between the mean grayscale G and the experimental temperature in the grayscale model is expressed as follows: S=K G T B Where S is the average grayscale attenuation rate of the surface, T is the hot corrosion temperature, and K is the kJ / m² temperature. G Here, B is the material condition coefficient, G0 is the grayscale variation index, and G0 is the mean grayscale of the original sample surface before hot corrosion. The relationship between the fractal dimension D of the corrosion morphology on the sample surface and the hot corrosion time in the high-temperature hot corrosion fractal model is expressed as follows: D=K D ·t+C Where t is the thermal corrosion time, K D Here, C is the material condition coefficient, and C is a constant.

2. The method for establishing a high-temperature alloy hot corrosion model based on image recognition according to claim 1, characterized in that: The preparation of the sample and the conduct of the hot corrosion test specifically include: Several circular pieces with a diameter of 20 mm and a thickness of 2 mm were made from GH4169 alloy. They were then polished, cleaned, and dried in sequence to obtain multiple spare samples. Multiple samples were divided into several groups and embedded in a mixed salt of Na2SO4 and NaCl with a mass ratio of 3:

1. They were then placed in high-temperature furnaces at different temperatures for thermal corrosion. After the hot corrosion test begins, several samples from each group are taken at set intervals and cooled in air. They are then boiled in deionized water, ultrasonically cleaned, and dried in sequence until the hot corrosion test ends.

3. The method for establishing a high-temperature alloy hot corrosion model based on image recognition according to claim 1, characterized in that: The process of capturing and recording the surface morphology of the sample after hot corrosion in the hot corrosion test, and preprocessing the images, specifically includes: The surface morphology of the sample after hot corrosion was captured by a camera under a fixed light source. The sample surface is identified and excess background is removed. The color image is converted into a grayscale image, and a grayscale pixel matrix A is established.

4. The method for establishing a high-temperature alloy hot corrosion model based on image recognition according to claim 3, characterized in that: The grayscale analysis of the preprocessing results, calculating the mean grayscale G and grayscale distribution of the surface image after hot corrosion of the sample, specifically includes the following operations: The grayscale is divided into 256 levels. The element range of the grayscale pixel matrix A is [0, 255], which represents the grayscale value of the corresponding pixel. Black is 0 and white is 255. The mean grayscale G of each sample is calculated to characterize the overall color level of the sample surface after hot corrosion. The mean grayscale of the original sample surface before hot corrosion is denoted as G0.

5. The method for establishing a high-temperature alloy hot corrosion model based on image recognition according to claim 3, characterized in that: The specific steps for calculating the fractal dimension D of corrosion morphology on different sample surfaces using the box method include: A gray-level pixel matrix A is covered with squares of a certain scale. The number of boxes containing the image is counted based on the comparison between the average gray level of the image within the box and the average gray level of the entire image. By changing the size of the boxes, plot the number of boxes and the box size on a double logarithmic coordinate system. The slope of the linear region represents the fractal dimension D of the corrosion morphology. The fractal dimension of the corrosion morphology of the original sample surface image before hot corrosion is denoted as D0.

Images