Method for predicting cracks in cast product, method for manufacturing steel sheet, and device for predicting cracks in cast product

Abstract

Provided is a method for predicting cracks in a cast product, said method being capable of appropriately predicting the presence / absence of the formation of cracks by a rolling process on the basis of shape information for a surface of the cast product.　The method for predicting cracks in a cast product predicts whether cracks will form in a surface of a cast product cast by continuous casting when the cast product is subjected to a rolling process, and comprises: a shape information acquisition step for acquiring shape information regarding the surface of the cast product after casting; and a crack formation prediction step for predicting, on the basis of the shape information, whether cracks will form in the surface of the cast product during the implementation of the rolling process.

Description

Method for predicting slab cracking, method for manufacturing steel plate, and slab cracking prediction device

[0005]

[0001] The present invention relates to a method for predicting slab cracking, a method for manufacturing a steel plate, and a slab cracking prediction device that predict whether or not cracks will occur on the surface of a slab when the slab cast by continuous casting of molten steel is subjected to rolling treatment.

[0002] In continuous casting of molten steel, in order to prevent sticking between the mold and the solidified shell, the mold is vibrated periodically and continuously in the casting direction. This vibration is called "oscillation". Further, due to this vibration, concave grooves (oscillation marks) are formed on the surface of the slab. And when the depth of the groove is deep, it is known that it becomes the starting point of surface cracking during the subsequent rolling process due to the notch effect of the recess.

[0003] Therefore, when concave grooves (oscillation marks) are formed on the surface of a slab that has undergone continuous casting, generally, machining or grinding with a depth of about 1 to 6 mm is performed on the surface of the slab to remove the grooves, and then rolling treatment is performed.

[0004] Therefore, various technical developments have been carried out to grasp the presence or absence of grooves on the surface of a slab that has undergone continuous casting. For example, Patent Document 1 discloses a method for predicting the depth of oscillation marks formed on the surface of a slab based on casting conditions in continuous casting. Patent Document 2 first calculates the frictional force generated between the mold and the cast slab based on the time-discrete data of the load during mold vibration. Then, based on the comparison result between the frictional force variation coefficient of the calculated frictional force (= standard deviation of frictional force / average value × 100 [%]) and a preset threshold value, a method for predicting transverse cracking is disclosed.

[0005] JP-A-2010-120044 JP-A-2016-043372

[0006] However, the method disclosed in Patent Document 1 applies a model formula to predict the depth of oscillation marks, which makes it highly likely that errors will occur compared to the actual depth. Furthermore, the method disclosed in Patent Document 2 uses indirect information such as friction force to predict the occurrence of transverse cracks, which makes it highly likely that noise will be included in the calculations for prediction, resulting in errors compared to the actual crack occurrence situation.

[0007] The present invention has been made in view of the above circumstances, and its object is to provide a method for predicting cracks in a cast slab, a method for manufacturing a steel sheet, and a cast slab crack prediction device that can appropriately predict whether or not cracks will occur due to rolling based on shape information on the surface of the cast slab.

[0008] [1] A method for predicting whether or not cracks will occur on the surface of a cast slab when a rolling process is applied to a cast slab cast by continuous casting, comprising: a shape information acquisition step of acquiring shape information on the surface of the cast slab after casting; and a crack occurrence prediction step of predicting whether or not cracks will occur on the surface of the cast slab when the rolling process is applied, based on the shape information. [2] The method for predicting cast slab cracks according to [1], wherein the shape information includes information on the number of grooves and groove depth on the surface of the cast slab, the number of grooves is the average value of the number of grooves in the cast slab for each specific length along the casting direction in the continuous casting, and the groove depth is the average value of the groove depth in the cast slab. [3] The method for predicting cracks in a cast slab when the rolling process is performed, according to [2], wherein in the crack occurrence prediction step, the number of grooves is less than or equal to the average value of the number of grooves for each specific length in cast slabs in which no cracks occurred in past operational data × 0.95, and the groove depth is greater than or equal to the average value of the groove depth in cast slabs in which no cracks occurred in past operational data × 1.10. [4] The method for predicting cracks in a cast slab when the rolling process is performed, according to [1], wherein the shape information includes information on the number of grooves on the surface of the cast slab, and in the crack occurrence prediction step, a fast Fourier transform is performed on the shape information for each specific length along the casting direction in the continuous casting, and based on the amplitude value obtained by the fast Fourier transform, it is predicted whether or not cracks will occur on the surface of the cast slab when the rolling process is performed. [5] The method for predicting cracks in a cast slab when the rolling process is performed, according to [4], wherein if the maximum value of the amplitude value is 100 μm or more, it is predicted that cracks will occur on the surface of the cast slab when the rolling process is performed. [6] A method for predicting slab cracking according to any one of [2] to [5], wherein the specific length along the casting direction is 500 mm or less. [7] A method for manufacturing a steel plate, wherein a rolling process is applied to the slab that is predicted not to develop cracks on its surface according to the method for predicting slab cracking according to any one of [1] to [6] to manufacture a steel plate.[8] A cast slab cracking prediction device for predicting whether or not cracks will occur on the surface of a cast slab when a rolling process is applied to a cast slab cast by continuous casting, comprising: a shape information acquisition unit for acquiring shape information on the surface of the cast slab after casting; and a crack occurrence prediction unit for predicting whether or not cracks will occur on the surface of the cast slab when the rolling process is applied, based on the shape information. [9] The cast slab cracking prediction device according to [8], wherein the shape information includes information on the number of grooves and groove depth on the surface of the cast slab, the number of grooves is the average value of the number of grooves in the cast slab for each specific length along the casting direction in the continuous casting, and the groove depth is the average value of the groove depth in the cast slab.

[10] The slab cracking prediction device according to [9], wherein the number of grooves is less than or equal to the average value of the number of grooves for each specific length in slabs in which no cracks occurred in past operational data, and the groove depth is greater than or equal to the average value of the groove depth in slabs in which no cracks occurred in past operational data, multiplied by 1.10, and the cracking prediction unit predicts that cracks will occur on the surface of the slab when the rolling process is performed.

[11] The slab cracking prediction device according to [8], wherein the shape information includes information on the number of grooves on the surface of the slab, and the cracking prediction unit performs a fast Fourier transform on the shape information for each specific length along the casting direction in the continuous casting, and predicts whether or not cracks will occur on the surface of the slab when the rolling process is performed based on the amplitude value obtained by the fast Fourier transform.

[12] The slab cracking prediction device according to

[11] , wherein the cracking prediction unit predicts that cracks will occur on the surface of the slab when the rolling process is performed if the maximum value of the amplitude value is 100 μm or more.

[13] The slab cracking prediction device according to any one of [9] to

[12] , wherein the specific length along the casting direction is 500 mm or less.

[0009] According to the present invention, the presence or absence of cracks caused by the rolling process can be appropriately predicted based on the shape information on the surface of the cast slab.

[0010] Figure 1 is a schematic diagram showing the surface shape of a cast slab after casting. Figure 2 is a diagram showing a schematic configuration of an example of the cast slab cracking prediction device of the present invention.

[0011] The embodiments of the present invention will be described below. The following embodiments are illustrative examples of devices and methods for realizing the technical idea of ​​the present invention, and do not specify a particular configuration. That is, the technical idea of ​​the present invention can be modified in various ways within the technical scope described in the claims. Also, the configurations shown in the drawings are schematic and may differ from actual configurations.

[0012] In the following explanation, "grooves" refers to oscillation marks such as indentations formed on the surface of the cast slab during continuous casting. "Cracks" refers to fissures that originate from the "grooves" on the surface of the cast slab during the rolling process.

[0013] The inventors considered that cracks occurring during the rolling process are greatly influenced by the shape of grooves (oscillation marks) on the surface of the cast slab after casting, taking into account the mechanism of crack formation. Therefore, they considered it important to focus on the shape of the surface of the cast slab after casting in order to appropriately predict whether or not cracks will occur during the rolling process. They then considered that by analyzing information regarding this shape, it would be possible to appropriately predict whether or not cracks will occur during the rolling process.

[0014] The present inventors have discovered a method for predicting whether or not cracks will occur on the surface of a cast slab when it is subjected to a rolling process. This method involves acquiring information about the shape of the surface of the cast slab and predicting whether or not cracks will occur based on the acquired shape.

[0015] Specifically, the present invention's method for predicting slab cracking first includes a shape information acquisition step, which involves acquiring shape information on the surface of the slab after casting.

[0016] Here, a method for obtaining shape information on the surface of a cast slab will be explained using Figure 1. Figure 1 is a schematic diagram showing the surface shape of a cast slab after casting. In the figure, arrow A indicates the width direction of the cast slab S, and arrow B indicates the casting direction in continuous casting.

[0017] Figure 1 represents a processed image obtained by imaging the surface of the cast slab S with a surface shape measurement sensor and then applying image processing to the image data. More specifically, Figure 1 represents a processed image obtained by converting each pixel of the image data into a grayscale (white to black gradient) with 256 levels (0 to 255) based on the individual brightness information contained in each pixel of the image data. In Figure 1, the darker the "dark (black)" area, the more it indicates that there is no object (an object is located farther from the surface shape measurement sensor). On the other hand, the brighter the "bright (white)" area in Figure 1, the more it indicates that there is an object (an object is located close to the surface shape measurement sensor).

[0018] Therefore, as shown in Figure 1, the area around the cast slab S where no objects are located nearby is shown as a "dark (black)" region, and the flat surface of the cast slab S is shown as a "bright (white)" region. Furthermore, on the surface of the cast slab S, the "gray" region (G in the figure) that shows the grayscale between the surrounding "dark (black)" region and the "bright (white)" region corresponds to the grooves (oscillation marks) formed on the surface of the cast slab S. Then, the 256-level grayscale (0 to 255) is used as a reference to set a specific grayscale value for the flat surface of the cast slab S, and the grayscale values ​​of other areas in the processed image may be determined according to this reference.

[0019] In this invention, information regarding the number of grooves is obtained in relation to the region corresponding to the grooves. Specifically, information regarding the number of grooves is obtained by counting the number of "gray" regions. Furthermore, information regarding the groove depth is obtained based on whether the "gray" region corresponding to the groove is a grayscale close to a "dark (black)" state or a grayscale close to a "light (white)" state. In this case, a grayscale close to a "dark (black)" state indicates that the groove is "deeper," and a grayscale close to a "light (white)" state indicates that the groove is "shallower." Finally, shape information on the surface of the cast slab after casting is obtained, including the information regarding the number of grooves and the information regarding the groove depth.

[0020] Here, the index for converting grayscale information into groove depth information (specific numerical values ​​of groove depth) may be set in advance based on past operational performance data. Alternatively, image data may be created using a surface shape measurement sensor, and groove depth information (specific numerical values ​​of groove depth) may be obtained from the created image data using triangulation or other methods.

[0021] Next, the slab cracking prediction method of the present invention includes a crack occurrence prediction step that predicts whether or not cracks will occur on the surface of the slab when it is subjected to rolling, based on the acquired shape information.

[0022] As mentioned earlier, the inventors considered that cracks generated during the rolling process are greatly influenced by the shape of the grooves (oscillation marks) on the surface of the cast slab after casting, taking into account the mechanism of crack generation. Therefore, according to the present invention, it is possible to appropriately predict whether or not cracks will occur due to the rolling process based on the shape information on the surface of the cast slab.

[0023] As mentioned earlier, in continuous casting, the mold is periodically and continuously vibrated (oscillated) in the casting direction to prevent seizing between the mold and the solidified shell. For this reason, the inventors considered that grooves formed on the surface of the cast slab tend to be formed with a certain periodicity along the casting direction in continuous casting.

[0024] Therefore, it is preferable that the shape information acquired in the shape information acquisition step is such that the number of grooves is the average value of the number of grooves in the cast slab for each specific length along the casting direction in continuous casting, and the groove depth is the average value of the groove depth in the cast slab.

[0025] Regarding the prediction of cracks that occur during the rolling process, by using the average number of grooves at specific lengths along the casting direction, accurate predictions that take into account mold vibrations can be made. Furthermore, by using the average depth of the grooves in the cast slab being predicted, crack predictions in the cast slab can be made even more accurate.

[0026] Furthermore, based on past operational data regarding "cast slabs that did not crack," the inventors discovered that cracks occur during the rolling process when the number and depth of grooves on the surface meet predetermined conditions.

[0027] Specifically, in the crack occurrence prediction step, it is more preferable to predict that cracks will occur on the surface of the cast slab when the rolling process is performed if the number of grooves is less than or equal to the average number of grooves for a specific length in cast slabs in which no cracks occurred in past operational data multiplied by 0.95, and the groove depth is greater than or equal to the average groove depth in cast slabs in which no cracks occurred in past operational data multiplied by 1.10.

[0028] In continuous casting, the load applied to the cast slab by the periodic and continuous oscillation of the mold is always kept constant. Therefore, if the number of grooves formed on the surface is small, the depth of each groove becomes deeper (larger), and when deformation occurs in the cast slab, the load on the deep grooves becomes greater, leading to cracking.

[0029] Furthermore, the inventors, after conducting an investigation using past operational data, found specific threshold values ​​for the number of grooves and groove depth on the surface as detailed conditions for cracking to occur during the rolling process.

[0030] Regarding the number of grooves, we found that it should be "less than or equal to the average number of grooves for a specific length in slabs that did not crack in past operational data × 0.95". In other words, the condition is that there are fewer grooves (oscillation marks) formed after casting compared to slabs that did not crack in past operational data (normal slabs).

[0031] Regarding groove depth, we found that it should be "1.10 or more times the average groove depth of slabs in past operational data where no cracks occurred." In other words, the condition is that the groove depth formed after casting is deeper (larger) than that of slabs in past operational data where no cracks occurred (normal slabs).

[0032] Furthermore, it is preferable to predict that cracks will occur on the surface of the cast slab during the rolling process when both of these conditions regarding the number of grooves and groove depth are met. By setting specific thresholds for the number of grooves and groove depth formed on the surface of the cast slab after casting, it is possible to predict more accurately whether or not cracks will occur.

[0033] Here, regarding the counting of the number of grooves formed on the surface of the cast slab, the "specific length" in the casting direction is not limited to any particular length, as long as it is possible to predict the occurrence of cracks, taking into account the vibration of the mold during continuous casting. Furthermore, considering the resolution of the processed image showing the shape of the surface of the cast slab (see Figure 1), for example, the "specific length" may be set to 258 mm (1032 pixels (0.25 mm per pixel)). It is preferable to set the "specific length" in the casting direction to a length of 500 mm or less. This is because the reproducibility rate of the results obtained by implementing the cast slab crack prediction method of this embodiment can be made 70% or more.

[0034] Furthermore, it is preferable to use past operational data of slabs of the same steel type as the slab being targeted for crack prediction, as "slabs in past operational data where no cracks occurred." This is because using operational data for the same steel type enables more accurate crack prediction.

[0035] Furthermore, the inventors considered the situation in continuous casting where the mold is controlled to vibrate periodically and continuously in the casting direction, and focused on the periodicity of the groove shape (number of grooves) formed on the surface of the cast slab. They also considered that by applying the Fast Fourier Transform, which is effective for frequency analysis, to the groove shape information (number of grooves), it would be possible to predict the occurrence of cracks due to the rolling process.

[0036] Therefore, in the crack occurrence prediction step, it is preferable to perform a fast Fourier transform on the shape information (number of grooves) for each specific length along the casting direction in continuous casting, and predict whether or not cracks will occur on the surface of the cast slab when the rolling process is performed based on the obtained amplitude values.

[0037] By performing a Fast Fourier Transform on the surface shape information (number of grooves) of the cast slab at specific length intervals along the casting direction in continuous casting, it is possible to detect the presence or absence of abnormal frequencies in the groove formation morphology, and predict the occurrence of cracks due to the rolling process with high probability.

[0038] Furthermore, in the crack occurrence prediction step, it is even more preferable to predict that cracks will occur on the surface of the cast slab when the rolling process is performed if the maximum amplitude value exceeds 100 μm. By performing a Fast Fourier Transform on the shape information of the surface of the cast slab, it is possible to detect the presence or absence of abnormal frequencies and to detect shapes with large amplitude values, thereby predicting the occurrence of cracks due to the rolling process with a higher probability.

[0039] Here, when actually implementing the slab cracking prediction method according to the present invention, it is preferable to first determine whether the following conditions are met: "the number of grooves is less than or equal to the average number of grooves for a specific length in slabs that did not crack in past operational data × 0.95, and the groove depth is greater than or equal to the average groove depth in slabs that did not crack in past operational data × 1.10."

[0040] Obtaining shape information regarding the number and depth of grooves formed on the surface of a cast slab, and comparing this information with threshold values ​​for the number and depth of grooves, allows for rapid calculation from a data processing perspective, enabling accurate and immediate prediction of cracks on the surface of the cast slab.

[0041] And when the "number of grooves is less than or equal to the average value of the number of grooves for each specific length in the slab where no cracks occurred in the past operation data × 0.95, but the groove depth is not greater than the average value of the groove depth in the slab where no cracks occurred in the past operation data × 1.10" or "the number of grooves is not less than the average value of the number of grooves for each specific length in the slab where no cracks occurred in the past operation data × 0.95, and the groove depth is greater than or equal to the average value of the groove depth in the slab where no cracks occurred in the past operation data × 1.10", it is preferable to perform a fast Fourier transform on the groove shape information (number of grooves) and predict that cracks will occur during the rolling process when the maximum value of the amplitude value reaches 100 μm or more. The situation where the maximum value of the amplitude value reaches 100 μm or more suggests that the periodicity of the groove shape information (number of grooves) is strong, that is, the number of grooves formed on the surface of the slab is extremely large.

[0042] The application of the fast Fourier transform to the groove shape information (number of grooves) requires some time for data processing, but it can predict the occurrence of cracks during the rolling process with a high probability.

[0043] Here, in order to efficiently perform image processing on the image data of the surface of the slab obtained using the surface shape measurement sensor, it is preferable to apply a high-pass filter to remove the gentle slope of the surface. Furthermore, in order to improve the quality of the image processing, a median filter may be applied to the image data to perform interpolation processing on the error portions in the image data. By performing these filter processes on the image data, it is possible to improve the prediction accuracy of the presence or absence of cracks in the slab.

[0044] Also, the method for predicting slab cracks according to the present invention may be applied to implement a method for manufacturing a steel plate. Specifically, it may be possible to perform a rolling process on a slab predicted not to have cracks on the surface by the above-described method for predicting slab cracks to manufacture a steel plate.

[0045] Next, a slab crack prediction device that enables the execution of the slab crack prediction method of the present invention will be described using the drawings. FIG. 2 is a diagram showing a schematic configuration as an example of the slab crack prediction device 10 of the present invention.

[0046] As shown in FIG. 2, the slab crack prediction device 10 includes a surface shape measurement sensor 1, a processing unit 2, an output unit 3, and a storage unit 4. The processing unit 2 includes an image processing unit 2a, a shape information acquisition unit 2b, and a crack occurrence prediction unit 2c.

[0047] The processing unit 2 is, for example, a CPU or the like, and functions as the image processing unit 2a, the shape information acquisition unit 2b, and the crack occurrence prediction unit 2c by executing various programs stored in the storage unit 4. Further, as the output unit 3, for example, an LCD or a CRT display or the like may be applied. As the storage unit 4, for example, an information recording medium such as a rewritable flash memory, a hard disk built-in or connected by a data communication terminal, a memory card, and its reading and writing device may be applied. The storage unit 4 stores programs, arithmetic expressions, etc. used in the processing unit 2. The storage unit 4 stores information regarding "an index for converting grayscale information into groove depth information" and information regarding "slabs without cracks" in past operation data.

[0048] The surface shape measurement sensor 1 captures an image of the surface of the slab S to create image data, and transmits the created image data to the image processing unit 2a of the processing unit 2. The image processing unit 2a performs image processing on the received image data. The shape information acquisition unit 2b acquires shape information on the surface of the slab S based on the image data subjected to image processing by the image processing unit 2a. At this time, the shape information acquisition unit 2b may acquire the shape information on the surface of the slab S based on the information regarding "an index for converting grayscale information into groove depth information" stored in advance in the storage unit 4. Then, the crack occurrence prediction unit 2c predicts whether cracks will occur on the surface of the slab when rolling processing is performed based on the shape information acquired by the shape information acquisition unit 2b. The crack occurrence prediction unit 2c displays the predicted result on the output unit 3.

[0049] Further, the shape information acquisition unit 2b executes a shape information acquisition step of acquiring shape information on the surface of the slab after casting. The crack occurrence prediction unit 2c executes a crack occurrence prediction step of predicting whether cracks will occur on the surface of the slab when rolling processing is performed based on the acquired shape information.

[0050] Next, we will describe an example in which the slab cracking prediction method of the present invention is applied to a slab cast by continuous casting, and whether or not cracks will occur on the surface of the slab when it is subjected to rolling.

[0051] First, a slab was cast using a vertical bending type continuous casting machine, and the slab was cut using a fixed cutter installed downstream in the casting direction. The surface of the cut slab was then imaged using a surface shape measuring sensor. The slab was imaged by placing multiple surface shape measuring sensors around the slab to be imaged. More specifically, six surface shape measuring sensors were placed on each of the multiple long sides of the slab, perpendicular to the casting direction, and two sensors were placed on each of the multiple short sides of the slab to capture images and generate image data. In this example, an Atlas Copco sensor (SlabMaster) was used as the surface shape measuring sensor to image the slab.

[0052] For each surface shape measurement sensor, image data of the cast slab was acquired with a size of 1992 x 1032 pixels (250 μm per pixel). The obtained image data was then converted into a numerical file and subjected to image processing. The image processing was performed on TIFF format image data so that the values ​​were based on a gray value (white to black gradient).

[0053] Furthermore, to understand the grooves formed on the surface of the cast slab, the calculation area on the surface of the cast slab was identified from the image data. To identify the calculation area, first, the edge of the cast slab in the width direction was detected based on the grayscale of the image data that had been processed. Then, the calculation area was defined as the region on the surface of the cast slab from 10 to 90 pixels away from the detected edge (80 pixels (20 mm)). By identifying the calculation area, it becomes possible to speed up the subsequent acquisition of shape information related to the number and depth of grooves.

[0054] Subsequently, within the identified computational domain, the presence or absence of grooves was determined, and shape information regarding the number of grooves and groove depth was also acquired. Specifically, groove depth information (numerical values ​​of groove depth) was obtained from image data acquired by surface shape measurement sensors based on triangulation.

[0055] Here, shape information regarding the number of grooves and groove depth in the processed image exists at the pixel level within the image. Therefore, after acquiring groove depth information, locations (in pixels) that are not recognized as the surface of the cast slab (without grooves) but are understood to have groove depth are recognized as "groove regions" that are at least included in the grooves. Furthermore, within the area recognized as a "groove region," the location (in pixels) that can be recognized as the deepest location was defined as the "peak position."

[0056] Furthermore, if the groove depth at the peak position of the groove region is 50 μm or more, and the distance between the peak position of the groove region and the peak position of another nearby groove region is 2.5 mm or more, the groove region is recognized as a groove. In this case, the groove depth of the groove region recognized as a groove will be 50 μm or more.

[0057] The reason for setting the groove depth threshold to 50 μm or more is that, based on past operational experience, the depth of the oscillation marks formed on the surface of the cast slab is, on average, 50 μm. Furthermore, the reason for setting the distance threshold between the peak position of another nearby groove region and the peak position of the groove region in question to 2.5 mm or more is as follows: This takes into account the normal oscillation period in the mold (300 cpm, Vc = 1.0 m / min), and the situation in which the cast slab moves 3 mm per period (0.2 sec) in the casting direction.

[0058] Based on these considerations, in each embodiment, the average number of grooves in the cast slab for each specific length along the casting direction was calculated as the number of grooves, and the average depth of the grooves in the cast slab was calculated as the groove depth. Based on these calculations, it was predicted that cracks would occur on the surface of the cast slab during rolling if the number of grooves was less than or equal to the average number of grooves for each specific length in cast slabs where no cracks occurred (based on past operational data), and the groove depth was greater than or equal to the average groove depth in cast slabs where no cracks occurred (based on past operational data), multiplied by 1.10. The total length of the cast slab after cutting with a fixed cutter was approximately 3-4 m. Furthermore, the "specific length" was set to 258 mm (1032 pixels (0.25 mm per pixel)).

[0059] Furthermore, in each embodiment, a Fast Fourier Transform was performed on the shape information (number of grooves) of the cast slab at specific lengths along the casting direction. Based on the amplitude values ​​obtained by the Fast Fourier Transform, it was predicted that cracks would occur on the surface of the cast slab during the rolling process if the maximum amplitude value exceeded 100 μm.

[0060] Furthermore, in this embodiment, the value "48.0" was used as the "average number of grooves for a specific length in slabs where no cracks occurred in past operational data" (hereinafter referred to as "Bcount"). The value "217.0 μm" was used as the "average groove depth in slabs where no cracks occurred in past operational data" (hereinafter referred to as "Bdep"). For the "slabs where no cracks occurred in past operational data," slabs of the same steel type as those used in each embodiment were adopted.

[0061] In each example, if the case fell under either "the number of grooves is Bcount × 0.95 or less, but the groove depth is not Bdep × 1.10 or more" or "the number of grooves is not Bcount × 0.95 or less, but the groove depth is Bdep × 1.10 or more", it was predicted that cracking would occur during the rolling process if the maximum amplitude value obtained by the Fast Fourier Transform was 100 μm or more. In addition, in each example, after predicting whether or not cracking would occur in the cast slab, the rolling process was actually performed and the presence or absence of cracking on the surface of the cast slab was confirmed. The results for each example are shown in Table 1.

[0062]

[0063] In Invention Example 1, the number of grooves in the cast slab was less than or equal to "Bcount × 0.95", and the groove depth was greater than or equal to "Bdep × 1.10", so it was predicted that cracks would occur on the surface of the cast slab when it was rolled. And indeed, when the rolling process was performed, cracks did occur on the surface. Therefore, it was possible to appropriately predict whether or not cracks would occur due to the rolling process based on the shape information on the surface of the cast slab.

[0064] In Invention Example 2, the number of grooves in the cast slab exceeded "Bcount × 0.95", and the groove depth was less than "Bdep × 1.10", so it was predicted that no cracks would occur on the surface of the cast slab when it was rolled. In fact, no cracks occurred on the surface when the rolling process was performed. Therefore, it was possible to appropriately predict whether or not cracks would occur due to the rolling process based on the shape information on the surface of the cast slab.

[0065] In Invention Example 3, although the number of grooves in the cast slab was less than or equal to "Bcount × 0.95", the groove depth was less than "Bdep × 1.10". Therefore, based on the result that the maximum amplitude value obtained by the Fast Fourier Transform was 100 μm or more, it was predicted that cracks would occur on the surface of the cast slab when the rolling process was performed. And indeed, when the rolling process was performed, "cracks" occurred on the surface. Thus, it was possible to appropriately predict whether or not cracks would occur due to the rolling process based on the shape information on the surface of the cast slab.

[0066] In Invention Example 4, although the number of grooves in the cast slab exceeded "Bcount × 0.95", the groove depth was greater than or equal to "Bdep × 1.10". Therefore, based on the result that the maximum amplitude value obtained by the Fast Fourier Transform was less than 100 μm, it was predicted that no cracks would occur on the surface of the cast slab when rolled. And indeed, no cracks occurred on the surface when the rolling process was performed. Thus, it was possible to appropriately predict whether or not cracks would occur due to the rolling process based on the shape information on the surface of the cast slab.

[0067] In Invention Example 5, although the number of grooves in the cast slab exceeded "Bcount × 0.95", the groove depth was greater than or equal to "Bdep × 1.10". Therefore, based on the result that the maximum amplitude value obtained by the Fast Fourier Transform was 100 μm or more, it was predicted that cracks would occur on the surface of the cast slab when the rolling process was performed. And indeed, when the rolling process was performed, "cracks" occurred on the surface. Thus, it was possible to appropriately predict whether or not cracks would occur due to the rolling process based on the shape information on the surface of the cast slab.

[0068] In Invention Example 6, although the number of grooves in the cast slab exceeded "Bcount × 0.95", the groove depth was greater than or equal to "Bdep × 1.10". Therefore, based on the result that the maximum amplitude value obtained by the Fast Fourier Transform was less than 100 μm, it was predicted that no cracks would occur on the surface of the cast slab when rolled. And indeed, no cracks occurred on the surface when the rolling process was performed. Thus, it was possible to appropriately predict whether or not cracks would occur due to the rolling process based on the shape information on the surface of the cast slab.

[0069] Furthermore, the inventors concluded that the reproducibility of the results obtained by implementing the slab cracking prediction method of the present invention depends on the "specific length along the casting direction" in each embodiment. Based on this, the inventors concluded that the reproducibility can be increased to 70% or more by setting the "specific length along the casting direction" to a length of 500 mm or less.

[0070] Here, recall is one of the evaluation metrics mainly used in fields such as statistics, machine learning, and information retrieval. In particular, recall can be useful information as an indicator for evaluating how accurately and "without missing" the inferred information obtained using an inferential model that predicts a specific event.

[0071] 1 Surface shape measurement sensor 2 Processing unit 2a Image processing unit 2b Shape information acquisition unit 2c Crack occurrence prediction unit 3 Output unit 4 Storage unit A Width direction B Casting direction G Groove S Cast slab

Claims

1. A method for predicting whether or not cracks will occur on the surface of a cast slab when a rolling process is applied to a cast slab produced by continuous casting, comprising: a shape information acquisition step of acquiring shape information of the surface of the cast slab after casting; and a crack occurrence prediction step of predicting whether or not cracks will occur on the surface of the cast slab when the rolling process is applied, based on the shape information.

2. The method for predicting slab cracking according to claim 1, wherein the shape information includes information relating to the number of grooves and groove depth on the surface of the slab, the number of grooves being the average value of the number of grooves in the slab for each specific length along the casting direction in the continuous casting, and the groove depth being the average value of the groove depth in the slab.

3. The method for predicting cracks in a cast slab according to claim 2, wherein, in the crack occurrence prediction step, if the number of grooves is less than or equal to the average value of the number of grooves for each specific length in cast slabs in which no cracks occurred in past operational data × 0.95, and the groove depth is greater than or equal to the average value of the groove depth in cast slabs in which no cracks occurred in past operational data × 1.10, it is predicted that cracks will occur on the surface of the cast slab when the rolling process is performed.

4. The method for predicting cracks in a cast slab according to claim 1, wherein the shape information includes information regarding the number of grooves on the surface of the cast slab, and in the crack occurrence prediction step, a fast Fourier transform is performed on the shape information for each specific length along the casting direction in the continuous casting, and based on the amplitude values ​​obtained by the fast Fourier transform, it is predicted whether or not cracks will occur on the surface of the cast slab when the rolling process is performed.

5. The method for predicting cracking of a cast slab according to claim 4, wherein if the maximum value of the amplitude becomes 100 μm or more, it is predicted that cracks will occur on the surface of the cast slab when the rolling process is performed.

6. The method for predicting slab cracking according to any one of claims 2 to 5, wherein the specific length along the casting direction is 500 mm or less.

7. A method for manufacturing a steel sheet, comprising: applying a rolling treatment to a cast slab that is predicted not to develop cracks on its surface according to the cast slab cracking prediction method described in any one of claims 1 to 6; and manufacturing a steel sheet.

8. A cast slab cracking prediction device that predicts whether or not cracks will occur on the surface of a cast slab when a rolling process is applied to a cast slab cast by continuous casting, comprising: a shape information acquisition unit that acquires shape information of the surface of the cast slab after casting; and a crack occurrence prediction unit that predicts whether or not cracks will occur on the surface of the cast slab when the rolling process is applied, based on the shape information.

9. The slab cracking prediction device according to claim 8, wherein the shape information includes information relating to the number of grooves and groove depth on the surface of the slab, the number of grooves being the average value of the number of grooves in the slab for each specific length along the casting direction in the continuous casting, and the groove depth being the average value of the groove depth in the slab.

10. The slab cracking prediction device according to claim 9, wherein the number of grooves is less than or equal to the average number of grooves for each specific length in slabs in which no cracks occurred in past operational data × 0.95, and the groove depth is greater than or equal to the average groove depth in slabs in which no cracks occurred in past operational data × 1.10, and the crack occurrence prediction unit predicts that cracks will occur on the surface of the slab when the rolling process is performed.

11. The cast slab cracking prediction device according to claim 8, wherein the shape information includes information regarding the number of grooves on the surface of the cast slab, and the crack occurrence prediction unit performs a fast Fourier transform on the shape information for each specific length along the casting direction in the continuous casting, and predicts whether or not cracks will occur on the surface of the cast slab when the rolling process is performed based on the amplitude values ​​obtained by the fast Fourier transform.

12. The cast slab cracking prediction device according to claim 11, wherein when the maximum value of the amplitude becomes 100 μm or more, the crack occurrence prediction unit predicts that cracks will occur on the surface of the cast slab when the rolling process is performed.

13. The slab cracking prediction device according to any one of claims 9 to 12, wherein the specific length along the casting direction is 500 mm or less.

Images