Steel slab scarfing apparatus, steel slab scarfing method, and steel slab manufacturing method

Abstract

To provide a steel slab scarfing apparatus and the like capable of determining an appropriate preheating state of a steel slab.SOLUTION: A steel slab scarfing apparatus comprises: a preheating unit that preheats a steel slab; a preheating determination unit that determines the preheating state of the steel slab preheated by the preheating unit; and a scarfing unit that scarfs the steel slab preheated by the preheating unit. The preheating determination unit includes: an image acquisition part that acquires an image of the preheated steel slab; a detection part that detects a morphology of a molten material of the steel slab in the image; and a preheating state information generation part that generates preheating state information indicating the preheating state of the steel slab on the basis of the morphology of the molten material of the steel slab. The scarfing unit scarfs the steel slab on the basis of the preheating state information.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a billet scalding apparatus for scalding a preheated billet, a billet scalding method, and a billet manufacturing method. [Background technology]

[0002] The surface of a billet obtained by blooming or continuous casting contains surface defects such as surface scratches and inclusions that are harmful when rolled into a product. These surface defects are removed, for example, by using a machine scarf to remove the surface of the billet to a desired depth.

[0003] In machine scarf cutting, the surface of the billet entering the machine is melted and preheated by ejecting a preheating flame, which is made by burning fuel gas and a gas containing oxygen, toward the billet.

[0004] The completion of preheating of the slab is determined by the operator. Specifically, the operator determines the preheating status based on the formation of a basin and the surface brightness in an image of the slab being preheated.

[0005] Image judgment by the operator depends on the individual's skill, and surface defects may not be completely removed. If machine scarfing is performed without completing preheating, uneven preheating occurs, and excessive scarification can cause gouges and deep cuts in the billet, or surface defects can remain unscathed.

[0006] Therefore, uneven preheating of a steel slab is detected. For example, Patent Document 1 discloses a method for determining the preheating state based on the extent of a molten metal puddle formed on the surface of a bloom of the steel slab.

[0007] Furthermore, Patent Document 2 discloses that the preheating state is determined based on the proportion of the area occupied by the basin in an image of the surface of the bloom of a steel slab.

[0008] Furthermore, Patent Document 3 discloses that the preheating state of a steel slab is determined based on the time when the surface brightness in a specified area of ​​the steel slab exceeds a threshold value and the preheating end time, which is the time when an instruction to end preheating is issued to the machine scarf device. [Prior art documents] [Patent documents]

[0009] [Patent Document 1] Japanese Patent Application Publication No. 5-269577 [Patent Document 2] Japanese Patent Application Publication No. 7-164143 [Patent Document 3] Japanese Patent Application Laid-Open No. 2015-167977 Summary of the Invention [Problem to be solved by the invention]

[0010] In Patent Documents 1 and 2, the preheating state is determined based on the area of ​​the molten metal pool in a captured image of the billet, but there is still a problem of uneven preheating in the width direction of the steel material.

[0011] In Patent Documents 1 to 3, the operator sets an area of ​​the billet where a puddle is expected to occur and then images the billet. Therefore, depending on the skill of the operator, there is a risk that the operator will not be able to set an appropriate area and will not be able to properly determine the preheating state.

[0012] The present invention has been made in consideration of the above problems, and has an object to provide a steel billet laser cutting device and the like that can appropriately determine the preheating state of the steel billet. [Means for solving the problem]

[0013] In order to solve the above problems, the present invention has the following features.

[0014] [1] An apparatus for laser-cutting a slab, comprising: a preheating section that preheats a slab; a preheating determination section that determines a preheat state of the slab preheated in the preheating section; and a laser-cutting section that laser-cuts the slab preheated in the preheating section, the preheating determination unit includes an image acquisition unit that acquires an image of the preheated steel billet; a detection unit that detects the shape of the molten steel billet in the image; a preheating state information generating unit that generates preheating state information indicating a preheating state of the slab based on a form of the molten slab, The slag cutting unit is a slag cutting device that slag cuts the slag based on the preheating state information. [2] The apparatus for laser cutting a billet of steel according to [1], wherein the preheating state information generation unit generates preheating state information indicating the preheating state of the billet of steel, using a puddle of the billet of steel and a streak-like form of the billet of steel as the form of the molten material of the billet of steel. [3] the detection unit has a first machine learning model that uses the image as input data and outputs the morphology of the molten steel billet, The preheating state information generation unit has a second machine learning model that uses the shape of the molten steel piece as input data and outputs preheating state information of the steel piece. [4] the detection unit has a first machine learning model that uses the image as input data and outputs the shape of the pouring pool and the streaks of the billet; The preheating state information generation unit has a second machine learning model that uses the basin of the billet and the streak shape as input data and outputs preheating state information of the billet. [5] a reference data acquisition unit that acquires reference data that serves as a reference for the preheating state of the slab, The apparatus for laser-cutting a steel slab according to any one of [1] to [4], wherein the preheating state information generating unit generates preheating state information of the steel slab based on the form of the molten material of the steel slab and the reference data. [6] A method for laser cutting a billet, comprising: a preheating step of preheating a billet; a preheating determination step of determining a preheated state of the billet preheated in the preheating step; and a laser cutting step of laser cutting the billet preheated in the preheating step, The preheating determination step an image capturing step of capturing an image of the preheated billet; a detecting step of detecting the morphology of the melt of the steel billet in the image; and a preheating state information generating step of generating preheating state information indicating a preheating state of the slab based on the form of the molten slab, In the laser cutting step, the steel billet is laser cut based on the preheating state information. [7] The method for hot-cutting a steel piece according to [6], wherein the preheating state information generating step generates preheating state information indicating the preheating state of the steel piece using a puddle and streak shape of the steel piece as the shape of the molten steel piece. [8] the detecting step is carried out using a first machine learning model that uses the image as input data and outputs the morphology of the molten steel billet; The method for hot-cutting a steel slab according to [6], wherein the preheating state information generation step is performed using a second machine learning model that uses the shape of the molten steel slab as input data and outputs preheating state information of the steel slab. [9] the detection step is carried out using a first machine learning model that uses the image as input data and outputs the shape of the pouring pool and the streaks of the steel billet; The method for hot-cutting a slab of steel described in [7], wherein the preheating state information generation step is performed using a second machine learning model that uses a puddle of the slab of steel and the streak form as input data as the form of the molten material of the slab of steel and outputs preheating state information of the slab of steel.

[10] a reference data acquisition step of acquiring reference data that serves as a reference for the preheating state of the steel slab, The method for laser cutting a steel slab according to any one of [6] to [9], wherein in the preheating state information generating step, preheating state information of the steel slab is generated based on the form of the molten material of the steel slab and the reference data.

[11] A method for producing a steel billet, comprising: A method for producing a billet, wherein the billet is produced using the method for spall cutting of a billet according to any one of [6] to

[10] . [Effects of the Invention]

[0015] The billet laser cutting device of the present invention has a preheating state information generating unit that generates billet preheating state information based on the morphology of the molten billet. This makes it possible to determine the preheating state including the morphology information of multiple billets, making it possible to determine an appropriate billet preheating state. [Brief explanation of the drawings]

[0016] [Figure 1] FIG. 2 is a block diagram showing the configuration of a steel billet laser cutting device. [Figure 2] 1 is a process flow showing a method for producing a steel slab. [Figure 3] This is a subroutine for determining preheating in FIG. [Figure 4] FIG. 2 is an explanatory diagram showing a preheating state of a steel slab. [Figure 5] FIG. 2 is an explanatory diagram showing a preheating state of a steel slab. [Figure 6] FIG. 2 is an explanatory diagram showing a preheating state of a steel slab. [Figure 7] FIG. 10 is a block diagram showing the configuration of a billet laser cutting device according to a second embodiment. [Figure 8] 10 is a subroutine for determining preheating according to the second embodiment. [Figure 9] 1 is a graph showing the relationship between the temperature of a steel slab and the preheating judgment time. DETAILED DESCRIPTION OF THE INVENTION

[0017] (First embodiment) An embodiment of the present invention will now be described with reference to the drawings. Fig. 1 shows the configuration of an apparatus for laser-cutting a billet. As shown in Fig. 1, the apparatus 100 for laser-cutting a billet includes a preheating section 10 that preheats the billet, a laser-cutting section 20 that laser-cuts the billet preheated in the preheating section 10, and a preheating determination section 30 that determines the preheating state of the billet preheated in the preheating section 10.

[0018] The preheating section 10 is provided with nozzles (not shown) that spray combustible gas and oxygen toward the billet. The preheating section 10 transports the billet to the laser cutting section 20 in accordance with the preheating state determined by the preheating determination section 30.

[0019] The scalding section 20 is a so-called machine scarf that is provided with nozzles (not shown) that spray flammable gas and oxygen toward the steel billet. The scalding section 20 performs scalding on the preheated steel billet, thereby performing surface treatment on the steel billet.

[0020] The pre-heating determination unit 30 is a computer including a CPU. The pre-heating determination unit 30 has an input unit 31 that receives data from an external terminal and an output unit 32 that outputs data to the external terminal. The pre-heating determination unit 30 also has a memory unit 33 that stores various data.

[0021] The input unit 31 is connected to a camera 41 that captures images of the steel piece being preheated in the preheating unit 10, and a PLC (Programmable Logic Controller) 42 that controls the operation of the preheating unit 10 and the cutting unit 20, so that data can be transmitted.

[0022] The camera 41 is not particularly limited, but may be, for example, a CCD camera. The camera 41, for example, captures still images at predetermined time intervals and transmits the image data to the preheating assessment unit 30. However, the camera 41 is not limited to this mode, and may transmit video data to the preheating assessment unit 30 as image data.

[0023] The PLC 42 stores preheating environment information, including preheating time, for each billet to be subjected to laser cutting. The preheating environment information, for example, records the preheating temperature and preheating time for each identification number of the billet to be subjected to laser cutting. The PLC 42 transmits the preheating time for each identification number of the billet to the preheating determination unit 30.

[0024] The output unit 32 is connected to the thermal cutting unit 20 and the notification unit 43 so as to be able to communicate data with them. The preheating determination unit 30 transmits data relating to the determination of the preheating state of the steel piece to the thermal cutting unit 20 and the notification unit 43 via the output unit 32.

[0025] The notification unit 43 is not particularly limited, but may be a speaker that outputs sound and a display that can display characters and images. The notification unit 43 may have a speaker and a display. The notification unit 43 notifies the user of the data transmitted from the preheating determination unit 30 by sound, text, etc.

[0026] The storage unit 33 is a non-volatile memory such as an HDD (Hard Disk Drive) or an SSD (Solid State Drive). The storage unit 33 stores image data captured by the camera 41 in chronological order. The storage unit 33 also stores the results of the preheating state of the billet determined by the preheating determination unit 30.

[0027] The preheating assessment unit 30 has an image acquisition unit 35 that acquires an image of the billet preheated in the preheating unit 10, and a detection unit 36 ​​that detects the morphology of the molten material in the billet in the image. The preheating assessment unit 30 also has a preheating condition information generation unit 37 that generates preheating condition information that indicates the preheating condition of the billet based on the morphology of the molten material in the billet.

[0028] The input unit 31, output unit 32, memory unit 33, image acquisition unit 35, detection unit 36, and preheating state information generation unit 37 are communicably connected to one another via a bus 34. The image acquisition unit 35, detection unit 36, and preheating state information generation unit 37 realize their functions by executing programs stored in the memory unit 33.

[0029] The image acquisition unit 35 reads and acquires the image captured by the camera 41 from the storage unit 33 .

[0030] The detection unit 36 ​​has a first machine learning model 36a. The first machine learning model 36a receives the image acquired by the image acquisition unit 35 as input data, and outputs the morphology of the molten steel billet as an image. The detection unit 36 ​​detects the morphology of the molten steel billet using the first machine learning model 36a.

[0031] Here, examples of the shape of the molten steel billet include a puddle shape, a streak shape, and a spot shape. The first machine learning model 36a is a machine learning model trained using training data in which, for example, an image having a substantially rectangular region extending in the width direction of the billet is used as an example of a puddle. Such training data uses, for example, image data selected by an operator. In the operator's selection, objects that do not have an area larger than a predetermined value, for example, point-like objects, are excluded from the training data.

[0032] Furthermore, a pouring pool is formed when the billet is heated during preheating and melts. For this reason, a pouring pool tends to form near the nozzles from which combustible gas and oxygen are ejected. The shape of the pouring pool is a region that extends in the width direction of the billet. Therefore, it is preferable to use, as training data for the pouring pool, an image of the shape of the region that extends in the width direction of the billet and is formed near the nozzles from which combustible gas and oxygen are ejected. By using such training data, it is possible to improve the noise removal efficiency and the pouring pool detection accuracy.

[0033] The first machine learning model 36a is a machine learning model trained using training data in which, for example, images having linear shapes extending in the axial direction of the nozzle of the preheating section 10 are used as examples of streak shapes. The streak shapes are formed linearly along the axial direction of the nozzle. As the preheating state of the steel billet progresses, the number of streak shapes tends to increase and they tend to be more widely distributed in the width direction of the steel billet. Such training data is, for example, image data selected by an operator. In the operator's selection, images that do not have a line length longer than a predetermined value, for example, point-like images, are excluded from the examples of streak shapes during training.

[0034] The molten material of a slab is formed as follows: When the slab is heated during preheating, it melts and a puddle forms. In the early stages of preheating, point-like spots form in various parts of the slab. Puddles are characterized by a larger area than point-like spots. As mentioned above, puddles tend to be concentrated near the nozzle from which oxygen is ejected. As the number of puddles increases, part of the puddle is blown away by the combustible gas and oxygen ejected from the nozzle, causing the molten material to take on a streak-like form. Therefore, it is advisable to use data extending from the puddle as training data for the streak form. Using such training data can improve noise removal efficiency and enhance the accuracy of streak form detection.

[0035] The preheating state information generation unit 37 has a second machine learning model 37a. The second machine learning model 37a receives input data such as a puddle and a streak as the shape of the molten steel piece detected by the detection unit 36, and outputs preheating state information for the slab. The preheating state information generation unit 37 generates preheating state information using the second machine learning model 37a. The generated preheating state information is stored in the memory unit 33 as the result of the preheating state of the slab determined by the preheating determination unit 30. Note that the preheating state information may be text data indicating a preheating state such as "preheating completed" or "preheating not completed," or may be a signal or the like that can identify these preheating states.

[0036] The second machine learning model 37a is a machine learning model trained using training data including, for example, the occurrence of puddles and the occurrence of streaks. The second machine learning model 37a is a machine learning model trained using training data that defines, for example, a situation in which puddles occur across the width direction of a slab as an example of completed preheating of a puddle. The second machine learning model 37a is also a machine learning model trained using training data that defines, for example, a situation in which puddles are formed across the width direction of a slab as an example of completed preheating of a streaky form. These training data are, for example, data selected by an operator.

[0037] The operator uses these images as training data and, for example, uses general-purpose machine learning model creation software to learn the feature values ​​of the pool and streak shapes from the training data.

[0038] Fig. 2 shows a method for manufacturing a slab using a method for laser cutting the slab. As shown in Fig. 2, first, a preheating step is performed in which the slab is preheated in a preheating section 10 (step S101). Specifically, when the slab to be preheated is stopped at a predetermined position, preheating in the preheating section 10 is started by an operator.

[0039] Next, the preheating determination unit 30 executes a preheating determination step as a subroutine to determine the preheating state of the steel slab preheated in the preheating step of step S101 (step S102).

[0040] Next, the laser cutting unit 20 performs a laser cutting step of laser cutting the steel billet preheated in the preheating step of step S101 (step S103).

[0041] After the completion of the cutting step in step S103, the steel billet is produced by appropriately checking the removal status of surface defects, checking the status of the marking numbers on the billet, and using a grinder to repair any defective parts that were not removed in the cutting section 20.

[0042] Fig. 3 shows a subroutine of the preheating determination step of step S102 in Fig. 2. When preheating by the preheating unit 10 starts, the camera 41 transmits image data to the preheating determination unit 30 at predetermined time intervals. The preheating determination unit 30 stores the received image data in the storage unit 33 sequentially.

[0043] As shown in FIG. 3, first, the image acquisition unit 35 reads out the captured image data from the storage unit 33, acquires an image to be used for preheating determination, and executes an image acquisition step (step S201).

[0044] Next, the detection unit 36 ​​executes a detection step (step S202) of detecting the morphology of the molten steel billet in the image data acquired by the image acquisition unit 35. Hereinafter, an example will be described in which the morphology of a puddle of the billet and the streaky morphology are used as the morphology of the molten steel billet.

[0045] In this embodiment, the detection unit 36 ​​detects the morphology of the molten steel billet using a first machine learning model 36a. That is, the first machine learning model 36a receives an image of the imaging data as input data and outputs the morphology of the molten steel billet.

[0046] Specifically, when the first machine learning model 36a detects a puddle of molten steel or a streaky form as the form of the molten steel, it specifies an area for the detected object using coordinates in the imaging data, etc., and outputs an image.

[0047] The preheating state information generating unit 37 generates preheating state information of the slab based on the form of the molten metal puddle and streaks detected in the detection step of step S202, and executes the preheating state information generating step (step S203).

[0048] In this embodiment, the preheating state information generator 37 generates the preheating state information of the billet using the second machine learning model 37a. That is, the second machine learning model 37a receives as input data the basin and streak morphology of the billet detected in the detection step of step S202, and outputs the preheating state information of the billet.

[0049] The above preheating determination subroutine is performed until preheating status information indicating that preheating is complete is generated in the preheating status information generation step of step S203. That is, when preheating status information indicating that preheating is incomplete is generated in step S203, the image acquisition step of step S201 is performed using imaging data captured a predetermined time after the imaging data used to generate the preheating status information. This process is repeated until preheating status information indicating that preheating is complete is generated, and the preheating determination subroutine ends.

[0050] When preheating state information indicating that the preheating of the slab is incomplete is generated, the preheating assessment unit 30 causes the notification unit 43 to display, for example, the words "Preheating incomplete." When preheating state information indicating that the preheating of the slab is completed is generated, the preheating assessment unit 30 causes the notification unit 43 to display, for example, the words "Preheating completed."

[0051] 4 to 6 show the preheating state of a slab. FIG. 4 is an example of image data taken immediately after the start of preheating in the preheating section 10. As shown in FIG. 4, immediately after the start of preheating, the temperature of the surface of the slab is in the process of rising. Therefore, no puddles or streaks are formed on the surface of the slab. Note that point-like spots SP are formed in various locations on the slab.

[0052] When the image of the imaging data shown in Fig. 4 is input, the detection unit 36 ​​outputs the image as if a pool of hot water and streaks were not detected. The preheating status information generation unit 37 generates and outputs preheating status information indicating that preheating is incomplete. The preheating determination unit 30 transmits the preheating status information to the notification unit 43, causing the notification unit 43 to notify the preheating status information.

[0053] FIG. 5 is an example of image data taken after the start of preheating in the preheating section 10. The surface temperature of the billet in FIG. 5 is higher than that in FIG. 4, but preheating is not yet complete. In FIG. 5, the thick linear region extending from the upper left arrow to the lower right arrow is the pouring pool MS. The pouring pool MS can also be visually recognized as a thin, approximately rectangular region. The thin linear region extending from the pouring pool MS to the left of the image is the streak-like structure SF. The direction connecting the upper left arrow and the lower right arrow corresponds to the width direction of the billet.

[0054] In Figure 5, because preheating was not sufficient, molten pools MS are formed in a patchy pattern in the width direction of the billet. In addition, the streak-like pattern SF is formed in a mixture of narrow and wide intervals in the width direction of the billet. That is, in Figure 5, the streak-like pattern SF is formed unevenly in the width direction of the billet. In other words, the streak-like pattern SF is distributed unevenly in the width direction of the billet.

[0055] When the image of the imaging data shown in Fig. 5 is input, the detection unit 36 ​​determines that the hot water reservoir MS and the streak formation SF have been detected, and outputs an image by specifying each area using the coordinates, etc. The preheating status information generation unit 37 inputs the image, designating each area as the hot water reservoir MS and the streak formation SF, and generates and outputs preheating status information indicating that preheating is incomplete. The preheating determination unit 30 transmits the preheating status information to the notification unit 43, causing the notification unit 43 to notify the preheating status information.

[0056] FIG. 6 is an example of image data obtained when preheating in the preheating section 10 is completed. In FIG. 6, the thick linear region extending from the arrow in the upper left to the arrow in the lower right is the pouring pool MS. The pouring pool MS of the billet in FIG. 6 is wider than that in FIG. 5 and is formed in a substantially rectangular shape. Furthermore, the streak pattern SF has a more uniform spacing in the width direction of the billet than in the embodiment shown in FIG. 5. That is, in FIG. 6, the streak pattern SF is uniformly formed so that the spacing is equal to or less than a predetermined spacing across the width direction of the billet. In other words, the streak pattern SF is uniformly distributed in the width direction of the billet. The spacing of the streak pattern SF tends to correspond to the spacing between the burners that heat the billet.

[0057] When the image of the imaging data shown in FIG. 6 is input, the detection unit 36 ​​detects the pouring basin MS and streak morphology SF and outputs the respective regions using their coordinates, etc. The preheating state information generation unit 37 inputs the respective regions as the pouring basin MS and streak morphology SF, generates preheating state information indicating that preheating is complete, and outputs this information. The preheating assessment unit 30 transmits the preheating state information to the notification unit 43, causing the notification unit 43 to notify the preheating state information. The preheating assessment unit 30 also transmits the preheating state information to the preheating unit 10 and the laser-cutting unit 20. The slab is then transported to the laser-cutting unit 20, where laser-cutting begins.

[0058] It is also preferable that an operator also judge the preheating state of the imaging data used to generate the preheating state information of the preheating state information generation unit 37. If the operator's judgment result differs from the preheating state information of the preheating state information generation unit 37, the preheating state judgment should be corrected to the operator's judgment. This makes it possible to improve the judgment accuracy when the second machine learning model 37a starts operating.

[0059] As described above, the billet laser cutting device and the like of the present invention has a preheating state information generating unit 37 that generates billet preheating state information based on the morphology of the molten billet. This makes it possible to determine the preheating state including the morphology information of a plurality of billets, making it possible to determine an appropriate billet preheating state.

[0060] That is, the preheating state information generating unit 37 generates preheating state information using the state of the morphology of the molten steel billet. As a result, even if it is difficult to detect the occurrence of a molten metal pool due to, for example, dirt on the lens of the camera 41 or interference with peripheral equipment, preheating state information can be generated based on the morphology of the molten steel billet. This makes it possible to improve the accuracy of determining the preheating state.

[0061] In this way, the preheating state of the slab is determined without relying on the judgment of the operator, and the preheating state can be determined appropriately without being affected by the skill of the operator. This makes it possible to reduce uneven preheating of the slab and appropriately remove surface defects. It also reduces the monitoring burden on the operator.

[0062] Note that preheating state information was generated for the image data of Figures 4 to 6 using the steel billet laser-cutting apparatus 100 described in the above embodiment. In this case, preheating state information indicating that preheating was incomplete was generated for the image data of Figures 4 and 5. Preheating state information indicating that preheating was complete was generated for the image data of Figure 6. Therefore, it was found that the steel billet laser-cutting apparatus 100 has sufficient accuracy in determining the preheating state.

[0063] Although the example in which the detection unit detects the morphology of the molten steel slab as a puddle and a streak has been described, the present invention is not limited to this, and the detection unit may detect a spot-like morphology in addition to the puddle and streak. In this way, the preheating state of the slab can be determined more precisely. As a result, the accuracy of determining the preheating state can be further improved.

[0064] Furthermore, the processing time for preheating a billet of steel when billet laser-cutting apparatus 100 generated preheating state information indicating that preheating was complete was equivalent to the processing time for preheating a billet of steel when the operator determined that preheating was complete. Furthermore, when laser-cutting was performed on a billet of steel for which billet laser-cutting apparatus 100 generated preheating state information indicating that preheating was complete, no peeling residue due to insufficient preheating occurred.

[0065] Furthermore, in the above-described embodiment, an example has been described in which the detection unit 36 ​​detects a puddle and a streak form as the form of the molten steel slab, and the preheating state information generation unit 37 generates preheating state information based on the puddle and streak form. However, the generation of preheating state information by the preheating state information generation unit 37 is not limited to this mode, and may be performed based on an image acquired by the image acquisition unit 35. In this case, the preheating state information generation unit 37 may generate preheating state information based on, for example, the area of ​​a region in the image that corresponds to the puddle and streak form and is brighter than the surrounding area.

[0066] (Second embodiment) In the above-described embodiment, an example has been described in which the detection unit is configured using a first machine learning model and the preheating state information generation unit is configured using a second machine learning model. The detection unit and the preheating state information generation unit may be configured without using these machine learning models. Note that the same components as those in the first embodiment are denoted by the same reference numerals and will not be described again.

[0067] 7 shows the configuration of a billet melt-cutting apparatus 200 according to the second embodiment. The billet melt-cutting apparatus 200 differs from the first embodiment in the configuration of the preheating judgment unit 50. Other points are the same as those of the first embodiment, so a description thereof will be omitted.

[0068] 7, the preheating determination unit 50 includes an input unit 51, an output unit 52, a memory unit 53, an image acquisition unit 55, a detection unit 56, a reference data acquisition unit 57, and a preheating state information generation unit 58, which are connected to one another via a bus 54. The input unit 51, the output unit 52, and the image acquisition unit 55 are the same as the input unit 31, the output unit 32, and the image acquisition unit 35 in the first embodiment, and therefore a description thereof will be omitted.

[0069] In addition to the data stored in memory unit 33, memory unit 53 stores first reference data that serves as a standard for the shape of the molten steel billet, such as the puddle and streak shapes of the billet, and second reference data that serves as a standard for the preheating state of the billet.

[0070] The first reference data may be, for example, image data of the pouring pool and streak morphology when preheating of the slab is completed. Alternatively, the first reference data may be, for example, image data obtained by averaging the characteristics of various pouring pool and streak morphologies.

[0071] The second reference data may be, for example, image data captured when preheating of the slab is complete. Alternatively, the second reference data may be, for example, image data obtained by averaging various basin shapes and various streak shapes when preheating is complete.

[0072] The detection unit 56 may be configured to detect the puddle and streak forms of the billet by, for example, image processing that uses the first reference data to extract the puddle and streak forms of the billet as the form of the molten billet from the imaging data.

[0073] The reference data acquisition unit 57 reads and acquires the first reference data and the second reference data from the storage unit 53.

[0074] The preheating state information generating unit 58 generates preheating state information for the slab in accordance with the degree of match between the morphology of the pouring pool and streaks in the image data acquired by the image acquiring unit 55 and the second reference data. The degree of match between the morphology of the pouring pool and streaks and the second reference data can be calculated, for example, by a known method of calculating the similarity between images. One known method of calculating the similarity between images is to create a histogram of numerical values ​​indicating the feature amounts of the images, compare them, and calculate the similarity, i.e., the degree of match.

[0075] Fig. 8 shows a subroutine for preheating determination according to the second embodiment. As shown in Fig. 8, first, the image acquisition unit 55 reads out imaging data from the storage unit 53, acquires an image to be subjected to preheating determination, and executes an image acquisition step (step S301).

[0076] The reference data acquisition unit 57 reads the first reference data and the second reference data from the storage unit 53, and executes a reference data acquisition step (step S302).

[0077] Next, the detection unit 56 executes a detection step of detecting a puddle and streaky forms of the molten steel billet in the image data acquired by the image acquisition unit 55 (step S303).

[0078] Specifically, the detection unit 56 extracts the shapes of the billet and streaks in the imaging data that have a high degree of agreement with the first reference data. For example, when the detection unit 56 detects the shapes of the billet and streaks as the shapes of the molten billet, the detection unit 56 specifies and extracts an area of ​​the detected object using coordinates or the like in the imaging data.

[0079] The preheating state information generator 58 determines the degree of agreement between the morphology of the basin and streaks detected in the detection step of step S303 and the second reference data. The preheating state information generator 58 determines whether the degree of agreement between the morphology of the basin and streaks and the second reference data is equal to or greater than a predetermined threshold (step S304).

[0080] If the degree of coincidence is less than the threshold value in the determination of step S304 (step S304: NO), the preheating state information generating unit 58 generates preheating state information indicating that preheating is incomplete, and executes a preheating state information generating step (step S305).

[0081] If the degree of coincidence is equal to or greater than the threshold value in the determination of step S304 (step S304: YES), the preheating state information generating unit 58 generates preheating state information indicating that preheating is complete, and executes a preheating state information generating step (step S306).

[0082] The above preheating determination subroutine is performed until the preheating status information generating step of step S306 is executed, which determines that preheating is complete. That is, when the preheating status information generating step of step S305 is executed, which determines that preheating is incomplete, the image acquiring step of step S301 is executed using imaging data captured a predetermined time after the imaging data used to generate the preheating status information. This process is repeated until the preheating status information generating step of step S306 is executed, which determines that preheating is complete, and the preheating determination subroutine ends.

[0083] Even when the steel billet melt-cutting apparatus 200 is configured as described above, it is possible to determine the preheating state including the morphological information of multiple steel billets, just like the steel billet melt-cutting apparatus 100 described in the first embodiment, making it possible to determine the appropriate preheating state of the steel billets.

[0084] It is known that there is a correlation between the temperature of a slab and the preheating judgment time. Fig. 9 is a graph showing the relationship between the temperature of a slab and the preheating judgment time. Here, the preheating judgment time is the time from when preheating starts to when it is judged that preheating is completed.

[0085] As shown in FIG. 9, there is a relationship in which the preheating judgment time shortens as the temperature of the slab increases. Therefore, in the preheating state information generation step of step S203 in FIG. 3, preheating state information may be generated using the correlation shown in FIG. 9. Furthermore, in the judgment step of step S304 in FIG. 8, the correlation shown in FIG. 9 may also be used to make a judgment. By performing the above steps in this manner, it becomes possible to make more accurate judgments that take the preheating time into consideration. Furthermore, it is possible to exclude judgments in abnormal states, such as when the preheating judgment time relative to the slab temperature is higher or lower than the actual result. Therefore, the hot cutting process can be stopped before problems such as poor hot cutting due to insufficient preheating or deep cutting of the slab or equipment damage due to excessive preheating occur. [Explanation of symbols]

[0086] 100 Steel billet cutting equipment 200 Steel billet slicing equipment 10 Preheating section 20 Welding part 30 Preheating judgment unit 35 Image acquisition unit 36 Detector 36a First machine learning model 37 Preheating state information generation unit 37a Second machine learning model 41 Camera

Claims

1. An apparatus for laser-cutting a slab, comprising: a preheating section that preheats a slab; a preheating determination section that determines a preheat state of the slab preheated in the preheating section; and a laser-cutting section that laser-cuts the slab preheated in the preheating section, the preheating determination unit includes an image acquisition unit that acquires an image of the preheated steel billet; a detection unit that detects the shape of the molten steel billet in the image; a preheating state information generating unit that generates preheating state information indicating a preheating state of the slab based on a form of the molten slab, The slag cutting unit is a slag cutting device that slag cuts the slag based on the preheating state information.

2. 2. The apparatus for laser-cutting a billet of steel according to claim 1, wherein the preheating state information generation unit generates preheating state information indicating a preheating state of the billet of steel, using a puddle of the billet of steel and a streak-like form of the billet of steel as the form of the molten material of the billet of steel.

3. the detection unit has a first machine learning model that uses the image as input data and outputs the morphology of the molten steel billet, The apparatus for laser-cutting a billet of steel according to claim 1 , wherein the preheating state information generation unit has a second machine learning model that receives input data representing the shape of the molten material of the billet of steel and outputs preheating state information of the billet of steel.

4. the detection unit has a first machine learning model that uses the image as input data and outputs the shape of the pouring pool and the streaks of the steel billet, 3. The apparatus for laser-cutting a billet of steel according to claim 2, wherein the preheating state information generation unit has a second machine learning model that uses the shape of the basin of the billet of steel and the streaks as input data and outputs preheating state information of the billet of steel.

5. a reference data acquisition unit that acquires reference data that serves as a reference for the preheating state of the slab, 5. The apparatus for laser-cutting a billet of steel according to claim 1, wherein the preheating state information generating unit generates preheating state information of the billet of steel based on the form of the molten material of the billet of steel and the reference data.

6. A method for laser cutting a billet, comprising: a preheating step of preheating a billet; a preheating determination step of determining a preheated state of the billet preheated in the preheating step; and a laser cutting step of laser cutting the billet preheated in the preheating step, The preheating determination step an image capturing step of capturing an image of the preheated billet; a detecting step of detecting the morphology of the melt of the steel billet in the image; and a preheating state information generating step of generating preheating state information indicating a preheating state of the slab based on a form of the molten slab, In the laser cutting step, the steel billet is laser cut based on the preheating state information.

7. The method for laser cutting a steel piece according to claim 6, wherein the preheating state information generating step generates preheating state information indicating the preheating state of the steel piece using a puddle and streak shape of the steel piece as the shape of the molten steel piece.

8. the detecting step is performed using a first machine learning model that uses the image as input data and outputs the morphology of the molten steel billet; 7. The method for hot-cutting a billet of steel according to claim 6, wherein the preheating state information generating step is performed using a second machine learning model that uses the shape of the molten material of the billet of steel as input data and outputs preheating state information of the billet of steel.

9. the detection step is carried out using a first machine learning model that uses the image as input data and outputs the shape of the pouring pool and the streaks of the billet; 8. The method for hot-cutting a billet of steel according to claim 7, wherein the preheating state information generating step is performed using a second machine learning model that uses a puddle of the billet of steel and the streak form as input data as the form of the molten material of the billet of steel, and outputs preheating state information of the billet of steel.

10. a reference data acquisition step of acquiring reference data that serves as a reference for the preheating state of the steel slab, The method for laser cutting a billet of steel according to any one of claims 6 to 9, wherein in the preheating state information generating step, preheating state information of the billet of steel is generated based on a form of the molten material of the billet of steel and the reference data.

11. A method for producing a steel billet, comprising: A method for producing a steel billet, wherein the steel billet is produced using the method for laser cutting a steel billet according to any one of claims 6 to 9.

12. A method for producing a steel billet, comprising: A method for producing a steel billet, wherein the steel billet is produced using the method for laser cutting the steel billet according to claim 10.

Images