Method for identifying surface cracking of a steel billet during hot working

Abstract

The present application belongs to the technical field of metal plastic working, and relates to a method for identifying surface cracking of a blank during hot working of steel. The specific steps are as follows: obtaining an infrared image of the surface of the blank; identifying a contour region of the surface of the blank; calculating the median value of the temperature values or infrared radiation intensity values of all pixel points in the contour region of the surface of the blank, and performing noise reduction processing on the infrared image by using the median value; identifying and calibrating the bare metal surface and the oxide scale covered surface in the contour region of the surface of the blank; calculating the temperature or infrared radiation intensity gradient of each pixel point in the contour region of the surface of the blank; for each identified crack, performing fractal calculation to determine the morphology of the crack, and classifying the possible causes of the crack. The present application solves the problem of online automatic identification of surface cracking of high-temperature metal during hot working, avoids false positives and false negatives caused by manual monitoring, and can output crack morphology data to provide data for detailed technical analysis.

Description

Technical Field

[0001] This invention belongs to the field of metal plastic processing technology, and specifically addresses an online identification and monitoring method for cracks in steel ingots or billets during high-temperature hot processing. Background Technology

[0002] Cracking of steel ingots or billets during hot working is a common defect. If cracks in ingots or billets are not detected promptly and hot working continues, the cracks will extend and deepen, leading to the scrapping of the billet and severely impacting product yield. Therefore, timely detection of cracks and prevention of further expansion is crucial. Early detection allows for timely salvage treatment, preventing greater losses.

[0003] In hot working processes, the commonly used method is to visually identify cracks with the naked eye. At high temperatures, the surface of a billet typically cools to lower temperatures than its interior due to heat dissipation to the environment. However, once a crack appears on the surface, the temperature in the cracked area is higher than the surface temperature. During processing, operators judge whether cracks have appeared on the billet surface based on the presence and shape of high-temperature areas. However, due to the complex surface conditions of billets, and the possibility of oxide scale covering carbon steel or alloy steel, it is difficult to accurately and promptly detect surface cracks due to limitations in personal experience.

[0004] Using keywords such as "crack," "identification," "infrared," and "continuous casting," the search results included the following patents: "Method and Apparatus for Online Non-destructive Testing of Hard Alloy Top Hammers" (CN201510888105.X), which discloses a method and apparatus for non-destructive testing of hard alloy top hammers. This invention cleverly utilizes the different near-infrared light scattering characteristics of dust and particles compared to top hammer cracks, achieving rapid removal of impurities by setting a threshold, thus improving detection accuracy. The patent "A Method for Measuring Fatigue Crack Propagation Rate Based on Infrared Thermographic Technology" (CN201810921058.8) discloses a method for measuring fatigue crack propagation rate based on infrared thermographic technology. The patent "An Infrared Thermographic Imaging System for Detecting Surface Defects on Online Slabs" (CN202120876014.5) discloses an infrared thermographic imaging system for detecting surface defects on online slabs. This system uses an infrared thermographic imaging device to collect the surface temperature of the slab, displaying crack defects and enabling real-time online detection. It overcomes the influence of iron oxide scale covering the slab surface, making defect identification more accurate and intuitive. This patent only discloses the image acquisition and signal transmission system for infrared thermal imaging, without explaining how to identify crack information in the image. Patent CN201410440283.1, titled "A Method and Apparatus for Detecting Surface Cracks in Continuously Cast Billets," discloses a method and apparatus for detecting surface cracks in continuously cast billets conveyed in the continuous casting machine's conveyor rollers. It involves irradiating the surface of the billet with a laser to form a laser band, acquiring an image of the laser band on the billet surface, extracting the shape of the laser band from the image, and finally identifying the surface cracks based on the shape of the laser band. The detection method and apparatus provided by this invention can automatically identify surface cracks in continuously cast billets and will not misidentify iron oxide scale as surface cracks, thereby improving the accuracy of the detection results.

[0005] Analysis of relevant patent search results revealed that there was no information involving online analysis of infrared thermography images, especially patents that use pixel temperature or infrared radiation intensity distribution characteristics of infrared images to identify cracks generated during metal heat treatment and to accurately identify them. Summary of the Invention

[0006] The purpose of this invention is to solve the problem of online automatic identification of surface cracks in high-temperature metals during hot working, avoid false alarms and missed alarms that occur with manual monitoring, and output crack morphology data to provide data for detailed technical analysis. This invention provides a method for identifying surface cracks in steel billets during hot working.

[0007] The technical solution of this invention is as follows:

[0008] A method for identifying surface cracks in steel billets during hot working includes the following steps:

[0009] (1) Obtain an infrared image of the surface of the billet, wherein the infrared image contains the temperature value or infrared radiation intensity value of each pixel;

[0010] (2) Identify the surface contour region of the billet in the infrared image according to the preset boundary temperature threshold;

[0011] (3) Calculate the median value of the temperature value or infrared radiation intensity value of all pixels in the outline area of ​​the blank surface, and use the median value to perform noise reduction processing on the infrared image;

[0012] (4) Based on the denoised infrared image, identify and mark the exposed metal surface and oxide scale covered surface within the outline area of ​​the billet surface;

[0013] (5) Calculate the temperature or infrared radiation intensity gradient of each pixel point in the contour area of ​​the blank surface, and identify the continuous pixel area with gradient value exceeding the preset gradient threshold as crack boundary.

[0014] (6) For each identified crack, fractal calculation is performed using a support vector machine to determine the crack shape, and the possible causes of the crack are classified in conjunction with a pre-set crack shape database.

[0015] Furthermore, step (4) specifically includes:

[0016] Based on the statistical distribution curves of temperature or infrared radiation intensity of each pixel on the surface of the billet, the high-temperature peak region corresponding to the exposed metal and the low-temperature region corresponding to the oxide scale are distinguished, and the low-temperature region is temperature compensated so that its temperature value is close to or equal to the median temperature value of the high-temperature peak region.

[0017] The purpose of temperature compensation for the oxide-covered area (which may contain crack information) in this invention is to prevent the system from identifying the transition boundary between this area and the non-oxide-covered area as a crack boundary. Specifically, the compensation method is to calculate the average signal (radiation intensity or temperature) D of the pixel to be corrected and the surrounding 8 pixels (i.e., a 3×3 pixel area), a total of 9 pixels. a The corrected signal value (radiation intensity or temperature) of the pixel to be corrected is the original signal value of that point plus the median value and then minus the average value D mentioned above. a .

[0018] The purpose of temperature compensation is to reduce or eliminate the boundary information between the oxidized and non-oxidized zones, so as to avoid identifying the boundary as a crack boundary during crack identification.

[0019] Furthermore, step (5) is calculated using the Sobel operator, and the preset gradient threshold is 5℃ / pixel spacing. (The hot processing temperature is generally above 800℃, and the gradient threshold is set to a fixed value.)

[0020] Furthermore, in step (6), "using a support vector machine for fractal calculation," the feature vector input to the support vector machine includes at least the crack length, width, fractal dimension, and temperature difference between the crack center and edge. In this invention, the image information after crack boundary processing only contains the cracked region and the non-cracked region. This image no longer contains temperature information and can be understood as a binary representation of the image. Crack morphology is then analyzed based on this image. Therefore, the input information only contains the geometric morphology information of the crack.

[0021] Furthermore, step (6) "Crack Morphology Database" stores different crack morphologies and their corresponding cause classification labels. The cause classification includes at least thermal stress cracks, mechanical overload cracks, and material original defect cracks. The crack morphology database is established based on a combination of historical data and expert experience, using known crack morphology databases and the possible causes corresponding to each crack morphology. It can be adjusted and updated during use.

[0022] Furthermore, in step (1) "acquiring an infrared image of the billet surface", an infrared imager with a pixel resolution of not less than 640×480 is used for taking pictures.

[0023] Furthermore, the "preset boundary temperature threshold" is set according to the hot working temperature of the billet, and its value is lower than the lowest surface temperature of the billet and higher than the ambient temperature.

[0024] Furthermore, when the hot working temperature of the billet is 1200℃, the preset boundary temperature threshold is 450℃.

[0025] Furthermore, the method for identifying surface cracks in steel billets during hot working also includes an output step: outputting the location, shape, size of the cracks, and the possible causes of the cracks obtained from the classification.

[0026] This invention is primarily applicable to materials with low oxidation levels or uniform, dense oxide layers under hot working temperatures (typically above 800°C), such as stainless steel, high-temperature alloys, and titanium alloys. However, for ordinary carbon steel or alloy steel, the uneven oxide layer thickness and uneven shedding during deformation can lead to significant deviations in the identification results. Therefore, this method is not suitable for ordinary carbon steel, alloy steel, or tool steel.

[0027] This invention provides a method for identifying surface cracks in steel billets during hot working, the detailed steps of which are as follows:

[0028] Infrared images (temperature or infrared radiation intensity data) of the surface to be monitored are captured using a high-resolution infrared imager, and the infrared radiation intensity or temperature parameters of each pixel are obtained.

[0029] Billet contour recognition processing. The received 2D image is used to identify the billet contour region based on the temperature or infrared radiation intensity values ​​of each pixel. A boundary temperature threshold (which can be pre-set according to the hot working temperature) is used as the recognition standard; areas below this threshold are considered environmental areas, and areas above or equal to this threshold are considered billet surface areas. A billet contour database is established using pixel coordinates as parameters.

[0030] Based on the pixel coordinates of the infrared image and the boundary contour database, the median value of the temperature distribution within the contour range of the billet surface is calculated.

[0031] The median value is used to denoise the image, removing noise caused by factors such as electromagnetic radiation, image transmission, and environmental conditions.

[0032] The radiation statistical distribution curve of each pixel on the billet surface is used to identify and determine the radiation intensity (temperature) of the exposed metal surface and the oxide scale-covered surface, and the exposed metal surface and the oxide scale-covered surface in the billet surface area are calibrated.

[0033] The Sober operator is used to calculate the temperature or infrared radiation intensity gradient of the graph, and the region where the infrared radiation intensity or temperature gradient exceeds the threshold is taken as the crack boundary, and the crack location is marked.

[0034] For each identified crack, fractal calculations were performed using the support vector machine method to determine the crack morphology. By comparing with the crack morphology database, the possible causes of each crack were preliminarily classified, and relevant information was output.

[0035] This invention employs infrared thermal imaging technology to record the temperature (or infrared radiation intensity) of each pixel position on the surface of the billet according to the image of an infrared imaging thermometer, forming a database of pixel coordinates and temperature (radiation intensity). A crack boundary algorithm is used to mark areas with abnormally high temperatures on the image and determine the boundaries of cracks. Then, a crack morphology algorithm is used to determine the shape, size, and location of cracks and output alarm information, providing basic information for timely judgment by operators and subsequent analysis by technicians.

[0036] This invention uses the temperature or infrared radiation intensity of each pixel in the image as a calculation parameter, which can effectively avoid intensity deviation caused by visible light illumination on the surface. It can accurately determine the location and shape of the actual cracks online, especially for cracks that are missed under oxide scale coverage. It has more reliable accuracy than methods that use visible light or laser irradiation to obtain images.

[0037] This invention enables real-time online monitoring and alarming of cracks appearing in billets during the forging process, and allows for preliminary classification of the causes of cracks, providing immediate information for operators and technicians.

[0038] The beneficial technical effects achieved by this invention are as follows:

[0039] Using the infrared radiation intensity of the billet surface or temperature parameters based on this data for surface crack identification can effectively avoid misjudgment caused by images generated when using visible light or laser irradiation.

[0040] Using the infrared radiation intensity of the billet surface as the basis for crack identification, it can effectively identify cracks that cannot be detected by visible light and laser light sources covered under oxide scale; it can realize real-time monitoring and provide monitoring information;

[0041] By combining a pre-learned crack database, the causes of cracks are initially classified, providing operators with reliable operational references. Detailed Implementation

[0042] To better understand the present invention, the following embodiments further illustrate the content of the present invention, but the content of the present invention is not limited to the following embodiments.

[0043] The technical solution of the present invention will be further described below with reference to specific embodiments. It should be understood that these embodiments are only used to describe the specific implementation of the present invention and are not intended to constitute any limitation on the scope of protection of the present invention.

[0044] Example: Crack monitoring process during the forging of large billets at a forging temperature of 1200℃

[0045] The first step is to use a high-resolution infrared imager with a resolution of 640×480 pixels to capture an infrared image of the surface to be monitored, obtain the temperature parameters of each pixel, and transmit the image containing the pixel temperature information to the computer.

[0046] The second step is billet contour recognition processing. The temperature threshold for billet edge recognition is set to 450℃, meaning areas in the image with a temperature higher than 450℃ are considered the billet surface area, and the remaining areas are considered the background. A billet contour database is established using pixel coordinates as parameters.

[0047] The third step is to calculate the median value of the temperature distribution within the surface contour range of the billet based on the pixel coordinates of the infrared image and the boundary contour database.

[0048] The fourth step is to use the median value to perform noise reduction on the image, removing noise caused by factors such as electromagnetic radiation, image transmission, and environmental conditions.

[0049] The fifth step involves identifying the temperature of the exposed metal surface and the oxide scale-covered surface by using the radiation statistical distribution curve of each pixel on the billet surface. The exposed metal surface and the oxide scale-covered surface in the billet surface area are calibrated, and temperature compensation is performed on the oxide scale-covered area where the temperature is lower than the median value.

[0050] The sixth step is to use the Sober operator to calculate the temperature gradient of the graphic, and to define the crack boundary as the area where the absolute value of the temperature gradient exceeds the threshold (5℃ / pixel spacing), and to mark the location of the crack.

[0051] Step 7: Individual Crack Region Identification and Processing. Using the highest temperature point within the crack region as the center, a support vector machine method is employed for fractal calculations to determine the crack morphology. By comparing with a crack morphology database, the possible causes of each crack are preliminarily classified, and relevant information is output.

[0052] The method of the present invention is mainly applicable to the workpiece contour boundary point at the temperature change point between the workpiece area and the environment area under hot working temperature conditions (usually the working temperature is greater than 800℃). The workpiece temperature will not be lower than 450℃. Therefore, in this embodiment, 450℃ is used as the contour threshold for identifying the workpiece (blank).

[0053] The infrared imager in this invention is used to acquire high-resolution infrared images. It does not have special requirements for temperature measurement accuracy and frame rate, but it has high requirements for thermal sensitivity. The appropriate thermal sensitivity should be selected based on the specific application scenario. Generally, the thermal sensitivity required to meet the implementation requirements of this invention is less than 0.02℃.

[0054] Of course, those skilled in the art should recognize that the above embodiments are only used to illustrate the present invention and are not intended to limit the present invention. Any modifications or variations of the above embodiments that are within the spirit and essence of the present invention will fall within the scope of the claims of the present invention.

Claims

1. A method for identifying surface cracks in steel billets during hot working, characterized in that, Includes the following steps: (1) Obtain an infrared image of the surface of the billet, wherein the infrared image contains the temperature value or infrared radiation intensity value of each pixel; (2) Identify the surface contour region of the billet in the infrared image according to the preset boundary temperature threshold; (3) Calculate the median value of the temperature value or infrared radiation intensity value of all pixels in the outline area of ​​the blank surface, and use the median value to perform noise reduction processing on the infrared image; (4) Based on the denoised infrared image, identify and mark the exposed metal surface and oxide scale covered surface within the outline area of ​​the billet surface; (5) Calculate the temperature or infrared radiation intensity gradient of each pixel point in the contour area of ​​the blank surface, and identify the continuous pixel area with gradient value exceeding the preset gradient threshold as crack boundary. (6) For each identified crack, fractal calculation is performed using a support vector machine to determine the crack shape, and the possible causes of the crack are classified in conjunction with a pre-set crack shape database.

2. The method for identifying surface cracks in steel billets during hot working according to claim 1, characterized in that, The steps in step (4) specifically include: Based on the statistical distribution curves of temperature or infrared radiation intensity of each pixel on the surface of the billet, the high-temperature peak region corresponding to the exposed metal and the low-temperature region corresponding to the oxide scale are distinguished, and the low-temperature region is temperature compensated so that its temperature value is close to or equal to the median temperature value of the high-temperature peak region.

3. The method for identifying surface cracks in steel billets during hot working according to claim 1, characterized in that, The step (5) is calculated using the Sobel operator, and the preset gradient threshold is 5°C / pixel spacing.

4. The method for identifying surface cracks in steel billets during hot working according to claim 1, characterized in that, In step (6) "using support vector machine for fractal calculation", the feature vector input to the support vector machine includes at least the crack length, width, fractal dimension, and temperature difference between the crack center and edge.

5. The method for identifying surface cracks in steel billets during hot working according to claim 1, characterized in that, The "Crack Morphology Database" in step (6) stores different crack morphologies and their corresponding cause classification labels. The cause classification includes at least thermal stress cracks, mechanical overload cracks, and material original defect cracks.

6. The method for identifying surface cracks in steel billets during hot working according to claim 1, characterized in that, In step (1) "acquiring an infrared image of the billet surface", an infrared imager with a pixel resolution of not less than 640×480 is used for taking pictures.

7. The method for identifying surface cracks in steel billets during hot working according to claim 1, characterized in that, The "preset boundary temperature threshold" is set according to the hot working temperature of the billet, and its value is lower than the lowest surface temperature of the billet and higher than the ambient temperature.

8. The method for identifying surface cracks in steel billets during hot working according to claim 7, characterized in that, When the hot working temperature of the billet is 1200℃, the preset boundary temperature threshold is 450℃.

9. The method for identifying surface cracks in steel billets during hot working according to claim 1, characterized in that, It also includes an output step: outputting the location, shape, size of the crack, and the possible causes of the crack obtained from the classification.