Shape distortion prediction model and method for constructing same, control command device for steel plate production line using shape distortion prediction model and method for operating same, and steel plate production method

Abstract

To effectively suppress shape distortion of a steel plate in a cooling process of a steel plate manufacturing.SOLUTION: A manufacturing technique of a steel plate using a shape distortion prediction model conformed to a steel plate manufacturing line to manufacture a steel plate is provided. The shape distortion prediction model performs machine learning using a prescribed machine learning algorithm so as to make a prediction value of the shape distortion amount of the steel plate approach to an actual value of the shape distortion amount of the steel plate in which components of the steel plate, size of the steel plate, heating temperature to the steel plate, and a prescribed cooling condition including top and bottom water volume ratio of cooling water supplied to the steel plate are explanatory variables, and shape distortion amount of the steel plate is an objective variable. In the shape distortion prediction model, the prediction value is outputted based on the explanatory variables at the time of operation of the steel plate manufacturing line, and the top and bottom water volume ratio is determined based on the prediction value by an inverse analysis method.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a shape distortion prediction model and a method for constructing the same, a control command device for a steel sheet production line using the shape distortion prediction model and an operation method thereof, and a steel sheet production method. [Background technology]

[0002] Conventionally, in a steel plate production line, a hot steel plate that has been hot-rolled is cooled under a predetermined controlled condition. In this cooling process, the hot steel plate is rapidly cooled by continuously spraying cooling water on both the top and bottom surfaces of the conveyed hot steel plate. In this manner, the strength and toughness of the steel plate are ensured by rapidly cooling the hot steel plate.

[0003] On the other hand, if the cooling process is not performed under appropriate control conditions, the steel plate may suffer from shape distortion. Such shape distortion of the steel plate caused by the cooling process is sometimes called shape distortion. Shape distortion includes a distortion called "C warp" in which the height of both ends in the width direction of the steel plate is higher than the height of the center, and a distortion called "reverse C warp" in which the height of the center is higher than the height of both ends, depending on the cooling process conditions. Such C warp and reverse C warp are caused by a thermal stress difference caused by the difference between the upper and lower surface temperatures of the steel plate. Therefore, suppressing the shape distortion of the steel plate during the cooling process is extremely important in ensuring the quality of the product. For this reason, in the cooling process, the occurrence of shape distortion is suppressed by adjusting the ratio of the amount of cooling water sprayed on the upper surface of the steel plate to the amount of cooling water sprayed on the lower surface (upper and lower water volume ratio).

[0004] For example, the following Patent Document 1 discloses a controlled cooling technique for reducing distortion of a small plate as a product plate cut from a large plate after controlled cooling. More specifically, in the method according to this technique, the upper surface side cooling water flow rate, the lower surface side cooling water flow rate, and the upper and lower flow rate ratio of the large plate to be processed are set based on the dimensions, components, and cooling temperature conditions of the large plate to be processed. Next, from a performance database in which the past results of the dimensions, components, cooling temperature conditions, and upper and lower flow rate ratios of the large plate to be processed and the past results of the shape information of the small plate are associated with each other, the past results of the upper and lower flow rate ratios of the large plate to be processed and the past results of the shape information of the small plate are extracted, and the set upper and lower flow rate ratio is corrected based on the correlation between the extracted upper and lower flow rate ratios and the past results of the shape information. Then, the set upper and lower flow rate of the cooling water flow rate and the lower surface side cooling water flow rate are controlled so that the corrected upper and lower flow rate ratio is satisfied, and cooling water is sprayed on both the upper and lower sides of the large plate to be processed, thereby performing controlled cooling of the large plate to be processed.

[0005] Also, the following Patent Document 2 discloses a technique for preventing C-warp of a thick steel plate by accurately adjusting the ratio of water and sewage flow rates during cooling of the thick steel plate. More specifically, the method according to the technique includes a water and sewage flow rate ratio determination step for determining a water and sewage flow rate ratio for a thick steel plate to be cooled such that at least one of the C-warp amount and curvature of the thick steel plate to be cooled falls within a target allowable range based on past operational conditions of the thick steel plate, a past water and sewage flow rate ratio for the thick steel plate when cooling under the past operational conditions of the thick steel plate was performed, and at least one of the past C-warp amount and curvature of the thick steel plate measured by a shape measuring device arranged on the outlet side of the cooling zone when cooling under the past operational conditions of the thick steel plate was performed, and a cooling water amount adjustment step for adjusting the amount of cooling water sprayed on the thick steel plate so as to achieve the water and sewage flow rate ratio determined in the water and sewage flow rate ratio determination step. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] JP 2015-199112 A [Patent Document 2] International Publication 2020 / 162004 Summary of the Invention [Problem to be solved by the invention]

[0007] In a steel plate production line, workers visually check the shape distortion of each steel plate that has been cooled, and adjust the ratio of the amount of water and sewage in cooling water according to the state of the shape distortion. More specifically, workers calculate the distortion correction rate according to the state of the shape distortion of the steel plate, and adjust the ratio of water and sewage by selecting the optimal ratio of water and sewage from a table that specifies the operating parameters.

[0008] Thus, the adjustment of the ratio of the amount of cooling water to the amount of wastewater depends largely on the experience of the worker, and there is a problem that there is a large variation. Also, since it takes a very long time to adjust the ratio of the amount of cooling water to the amount of wastewater, there is a problem that production efficiency does not increase.

[0009] On the other hand, the techniques disclosed in Patent Documents 1 and 2 aim to suppress shape distortion of the steel plate during the cooling treatment while eliminating the need for worker intervention in adjusting the water / sewage flow ratio as much as possible. However, in order to further improve product quality, further suppression of shape distortion is desired.

[0010] That is, as described above, the occurrence of shape distortion of a steel plate due to cooling treatment is largely due to the difference in thermal stress resulting from the temperature difference between the top and bottom surfaces of the steel plate. In particular, the leading end of a steel plate being transported is cooled differently by cooling water from the portion following the leading end (hereinafter referred to as the "steady portion"), and therefore the state of shape distortion may differ between the leading end and the steady portion. However, the current adjustment of the ratio of water and drainage does not take into account the respective cooling conditions of the leading end and the steady portion of the steel plate, and therefore it is not possible to effectively suppress the different shape distortions occurring at the leading end and the steady portion.

[0011] Therefore, an object of the present invention is to propose a technique for more effectively suppressing shape distortion of a steel sheet due to a cooling treatment in the production of the steel sheet.

[0012] More specifically, one of the objects of the present invention is to propose a technique for effectively suppressing shape distortion in both the leading edge and stationary portion of a steel plate caused by cooling treatment in the manufacture of the steel plate.

[0013] Another object of the present invention is to construct a shape distortion prediction model that can accurately predict shape distortion of a steel plate caused by a cooling treatment in the production of the steel plate.

[0014] Another object of the present invention is to provide a control command device that can adjust operational parameters to control the cooling process using a constructed shape distortion prediction model in steel plate manufacturing so as to effectively suppress shape distortion of the steel plate due to the cooling process, as well as a steel plate manufacturing system and a steel plate manufacturing method using the same. [Means for solving the problem]

[0015] The present invention, which aims to solve the above problems, includes the following invention-specific matters and technical features.

[0016] The present invention according to one aspect is a geometric distortion prediction model adapted to a steel plate production line for producing a steel plate. The geometric distortion prediction model is constructed by performing machine learning using a predetermined machine learning algorithm, with predetermined cooling conditions including at least the components of the steel plate, the size of the steel plate, the heating temperature for the steel plate, and the ratio of the amount of cooling water to the amount of water and the amount of cooling water supplied to the steel plate as explanatory variables, and with the amount of geometric distortion of the steel plate as a response variable, so that a predicted value of the amount of geometric distortion of the steel plate approaches an actual value of the amount of geometric distortion of the steel plate.

[0017] Here, the predetermined cooling conditions further include at least one of a cooling start temperature and a cooling end temperature of the steel plate, a temperature of the cooling water, and a cooling zone used for cooling the steel plate.

[0018] The amount of shape distortion is the amount of C-warp or reverse C-warp based on the difference in height between the center and edge portions in the width direction of the steel plate.

[0019] In addition, the shape distortion prediction model is subjected to machine learning using a predetermined machine learning algorithm so that the predicted values ​​of the shape distortion amounts of the tip portion and the steady portion of the steel plate approach the actual values ​​of the shape distortion amounts of the tip portion and the steady portion.

[0020] Furthermore, the predetermined machine learning algorithm is executed under a monotonicity constraint between the upstream / downstream flow rate ratio and the amount of geometric distortion.

[0021] According to another aspect, the present invention provides a method for constructing a geometric distortion prediction model adapted to a steel plate production line for manufacturing a steel plate, the method including performing machine learning using a predetermined machine learning algorithm, with predetermined cooling conditions including at least the components of the steel plate, the size of the steel plate, the heating temperature for the steel plate, and the ratio of the amount of cooling water to the amount of water and the amount of cooling water supplied to the steel plate being explanatory variables, and with the amount of geometric distortion of the steel plate being an objective variable, such that a predicted value of the amount of geometric distortion of the steel plate approaches an actual value of the amount of geometric distortion of the steel plate.

[0022] According to yet another aspect, the present invention provides a control command device for controlling a steel plate production line for producing a steel plate, the control command device including: a shape distortion prediction model in which machine learning has been performed using a predetermined machine learning algorithm, with predetermined cooling conditions including at least a component of the steel plate, a size of the steel plate, a heating temperature for the steel plate, and a water / sewage flow ratio of cooling water supplied to the steel plate as explanatory variables, and a shape distortion amount of the steel plate as a target variable, so that a predicted value of the shape distortion amount of the steel plate approaches an actual value of the shape distortion amount of the steel plate; and a water / sewage flow control unit that determines the water / sewage flow ratio under the predetermined cooling conditions based on the predicted value output from the shape distortion prediction model.

[0023] The water / sewage volume control unit controls the supply of the cooling water based on the determined water / sewage volume ratio when a deviation between the determined water / sewage volume ratio and a predetermined reference value does not exceed a predetermined threshold value.

[0024] The water / sewage flow control unit determines the water / sewage flow ratio by an inverse analysis method according to the predicted value of the amount of geometric distortion. The inverse analysis method is a grid search method.

[0025] According to another aspect, the present invention provides a method for operating a control command device in a steel plate production line for producing steel plates. The method for operating the control command device includes: using a shape distortion prediction model that has been machine-learned using a predetermined machine learning algorithm so that a predicted value of the shape distortion amount of the steel plate approaches an actual value of the shape distortion amount of the steel plate, with predetermined cooling conditions including at least the components of the steel plate, the size of the steel plate, the heating temperature for the steel plate, and a water / sewage flow rate ratio of cooling water supplied to the steel plate as explanatory variables and a shape distortion amount of the steel plate as a target variable, outputting a predicted value of the shape distortion amount based on the explanatory variables during operation of the steel plate production line, using a shape distortion prediction model that has been machine-learned using a predetermined machine learning algorithm so that a predicted value of the shape distortion amount of the steel plate approaches an actual value of the shape distortion amount of the steel plate; determining the water / sewage flow rate ratio under the predetermined cooling conditions using an inverse analysis method based on the predicted value output from the shape distortion prediction model; and controlling cooling equipment to adjust the flow rate of the water / sewage flow rate ratio under the predetermined cooling conditions based on the determined water / sewage flow rate ratio.

[0026] According to another aspect, the present invention is a steel plate production line including a control command device, a cooling facility for performing a cooling process under a predetermined cooling condition on a heated and rolled steel plate under the control of the control command device, and a cutting facility for performing a cutting process for cutting the steel plate after the cooling process into small plates under the control of the control command device. The cutting facility includes a shape distortion measuring device for measuring a shape distortion amount of the small plate. The control command device includes a shape distortion prediction model that is machine-learned using a predetermined machine learning algorithm so that a predicted value of the shape distortion amount of the small plate approaches an actual value of the shape distortion amount of the small plate, with predetermined cooling conditions including at least the components of the steel plate, the size of the steel plate, the heating temperature for the steel plate, and the water / sewage flow ratio of cooling water supplied to the steel plate as explanatory variables, and the shape distortion amount of the small plate as an objective variable, and a water / sewage flow control unit that determines the water / sewage flow ratio based on the predicted value output by the shape distortion prediction model that is machine-learned. The water / sewage flow control unit controls the supply of the cooling water according to the determined water / sewage flow ratio.

[0027] According to another aspect, the present invention is a steel plate manufacturing method for determining and controlling a water / sewage flow ratio of the cooling water to be supplied to the steel plate using the shape distortion prediction model. The steel plate manufacturing method includes: the shape distortion prediction model subjected to the machine learning outputs the predicted value of the amount of shape distortion of the steel plate based on actual parameters collected during operation of the steel plate manufacturing line and consisting of predetermined cooling conditions including at least the components of the steel plate, the size of the steel plate, the heating temperature for the steel plate, and the water / sewage flow ratio of the cooling water supplied to the steel plate; a water / sewage flow control unit determines the water / sewage flow ratio using an inverse analysis method based on the output predicted value; and a water / sewage flow control unit controls the supply of the cooling water according to the determined water / sewage flow ratio.

[0028] Furthermore, the present invention may be a computer program for causing a computer to realize the steel plate manufacturing method under the control of a processor, and a computer-readable recording medium having the computer program non-transitoryly recorded thereon.

[0029] In this specification, the term "means" does not simply mean physical means, but also includes cases where the functions of the means are realized by software. The functions of one means may be realized by two or more physical means, or the functions of two or more means may be realized by one physical means. Furthermore, the term "system" refers to a logical collection of multiple devices (or functional modules that realize specific functions), and it does not matter whether each device or functional module is contained within a single housing. Effect of the Invention

[0030] According to the present invention, it is possible to more effectively suppress the shape distortion of a steel plate caused by the cooling treatment in the manufacture of the steel plate. In particular, according to the present invention, it is possible to construct a shape distortion prediction model capable of appropriately predicting the shape distortion of the tip portion and the steady portion of the shape distortion of the steel plate caused by the cooling treatment in the manufacture of the steel plate. Therefore, by using such a shape distortion prediction model, the cooling treatment is controlled by adjusting the operation parameters so as to effectively suppress the shape distortion of the steel plate caused by the cooling process, and therefore it is possible to suppress the shape distortion of the tip portion and the steady portion and manufacture a high quality steel plate.

[0031] Other technical features, objects, and operational effects or advantages of the present invention will become apparent from the following embodiments described with reference to the accompanying drawings. The effects described in this specification are merely examples and are not limited thereto, and other effects may also be present. [Brief description of the drawings]

[0032] [Figure 1] FIG. 1 is a diagram showing an example of a schematic configuration of a steel sheet production line according to an embodiment of the present invention. [Diagram 2] FIG. 2 is a diagram showing an example of a schematic configuration of a cooling facility in a steel plate production line according to an embodiment of the present invention. [Diagram 3]FIG. 3 is a diagram showing an example of a schematic configuration of a control command device in a steel sheet production line according to an embodiment of the present invention. [Figure 4] FIG. 4 is a diagram for explaining shape distortion of a steel plate in a steel plate production line according to one embodiment of the present invention. [Diagram 5] FIG. 5 is a diagram for explaining the state of the cooling treatment in the steel plate production line according to one embodiment of the present invention. [Figure 6] FIG. 6 is a graph showing the relationship between the ratio of the amount of cooling water and the amount of cooling water in the steady part of the steel plate during the cooling treatment in a steel plate production line according to an embodiment of the present invention and the shape distortion of the steel plate. [Figure 7] FIG. 7 is a graph showing the relationship between the ratio of the amount of water and wastewater to the shape distortion of the steel plate during the cooling treatment in the steel plate production line according to one embodiment of the present invention. [Figure 8] FIG. 8 is a flowchart for explaining an example of a process for constructing the geometric distortion prediction model 121 according to an embodiment of the present invention. [Figure 9] FIG. 9 is a flowchart for explaining an example of a process for setting the water / sewage flow ratio based on a predicted value by the geometric distortion prediction model 121 according to an embodiment of the present invention. [Figure 10] FIG. 10 is a flowchart for explaining an example of an inverse analysis method using the geometric distortion prediction model 121 according to an embodiment of the present invention. [Figure 11] FIG. 11 is a conceptual diagram for explaining an example of the inverse analysis technique. [Figure 12] FIG. 12 is a graph comparing the predicted results of steady-state distortion using a geometric distortion prediction model according to an embodiment of the present invention with actual values. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0033] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. However, the embodiment described below is merely an example, and there is no intention to exclude various modifications and application of techniques not specified below. The present invention can be implemented by various modifications (for example, combining the respective embodiments) within the scope of the gist of the present invention. In addition, in the following description of the drawings, the same or similar parts are represented by the same or similar reference numerals. The drawings are schematic and do not necessarily correspond to actual dimensions, ratios, etc. The drawings may also include parts with different dimensional relationships and ratios.

[0034] Fig. 1 is a diagram showing an example of a schematic configuration of a steel plate production line to which a geometric distortion prediction model according to an embodiment of the present invention is applied. In the present disclosure, an example in which a model for predicting geometric distortion of a steel plate is applied to a steel plate production line is described, but is not limited to this. In addition, the steel plate as a product produced in this steel plate production line may be, for example, a thick steel plate (steel plate for shipbuilding, steel plate for line pipe, steel plate for civil engineering construction, etc.).

[0035] As shown in the figure, the steel plate production line 1 includes, for example, a control command device 10, a heating facility 20, a rolling facility 30, a cooling facility 40, and a cutting facility 50.

[0036] The control command device 10 is a device that issues various commands to comprehensively control operations in the steel sheet production line 1. The control command device 10 controls each processing facility and / or equipment in the steel sheet production line 1, for example, in accordance with a predetermined control sequence program. The control command device 10 is configured to include, for example, a process computer 110 that controls each processing facility, a shape distortion control device 120, and a database that stores various performance parameters, operation parameters, etc. (see FIG. 4).

[0037] As described later, the control command device 10 includes a geometric distortion prediction model 121 constructed by performing machine learning to predict the amount or state of geometric distortion of a steel plate to be manufactured. Machine learning for the geometric distortion prediction model 121 may be performed by the geometric distortion control device 120 of the control command device 10, or may be performed by another off-site computer (not shown). In the present disclosure, the geometric distortion prediction model 121 on which such machine learning has been performed may be explicitly referred to as a trained geometric distortion prediction model 121. During operation of the steel plate production line 1, the control command device 10 uses the trained geometric distortion prediction model 121 to predict the value of the geometric distortion amount of the steel plate, determines the ratio of the amount of cooling water to the amount of cooling water on the basis of the predicted value (hereinafter referred to as the "predicted value"), and adjusts the operation parameters, thereby enabling the production of a steel plate in which geometric distortion is effectively suppressed. In the present disclosure, the ratio of the amount of cooling water to the amount of cooling water on the bottom surface of the steel plate relative to the amount of cooling water on the top surface of the steel plate (amount of cooling water sprayed on the top surface / amount of cooling water sprayed on the bottom surface).

[0038] The heating equipment 20 is, for example, equipment including a furnace for performing heat treatment on slabs (plate-shaped steel material). During operation, the heating equipment 20 heats the inside of the furnace by, for example, a heating device (not shown) to keep it at a high temperature, heats the slabs carried in from an entrance to a predetermined temperature, and carries them out from an exit. The slabs heated to a high temperature are transported to the rolling equipment 30 along a transport path.

[0039] The rolling equipment 30 is an equipment for performing rolling processing on a slab heated to a predetermined temperature. The rolling processing is classified into, for example, hot rolling processing, cold rolling processing, and warm rolling processing, etc., depending on the processing temperature, but is not limited thereto. The rolling equipment 30 is configured to include, for example, a plurality of rolling mills, that is, a roughing mill and a finishing mill, not shown. Each of the roughing mill and the finishing mill typically includes a work roll for rolling the slab and a backup roll for receiving the rolling load and suppressing the deflection of the work roll, but is not limited thereto. The slab is rolled by, for example, moving back and forth in the conveying direction in each of the roughing mill and the finishing mill, and becomes a steel plate. This steel plate is sometimes called a large plate before being cut. The steel plate that has been subjected to the rolling processing is transported to the cooling equipment 40 along the conveying path.

[0040] The rolling equipment 30 may include a descaling equipment, which is not shown in the figure. The descaling equipment is equipment for removing (descaling) scale formed on the surface of a heated steel plate in accordance with the descaling information. The descaling equipment includes, for example, injection nozzles (not shown) provided on the entry side and exit side of the rolling mill so as to face the upper and lower surfaces of the steel plate, respectively, and injects a high-pressure water flow from the injection nozzle onto the surface of the steel plate during rolling processing by the rolling equipment, thereby removing the generated scale.

[0041] The cooling equipment 40 is equipment for supplying cooling water to the rolled steel plate in a high temperature state according to predetermined cooling conditions, and performing a cooling process. The cooling equipment 40 is typically divided into several cooling zones, as described below, and cools the steel plate by spraying or jetting cooling water onto the top and bottom surfaces of the steel plate. Such cooling process improves the material properties (e.g., strength, toughness, etc.) of the steel plate. In the present disclosure, the cooling equipment 40 sprays cooling water at a water / upper water volume ratio determined and set according to a predicted value of the amount of geometric distortion of the steel plate by the learned geometric distortion prediction model 121.

[0042] The cutting equipment 50 is equipment for cutting or shearing a cooled steel plate (large plate) to a predetermined length. Small plates cut out from the large plate by the cutting equipment 50 become steel plates as products. Typically, multiple small plates are cut out from one large plate. The steel plate that has been cooled and transported to the cutting equipment 50 has its leading edge (hereinafter referred to as the "leading edge") cut off. In the present disclosure, in particular, for the small plate from which the leading edge has been cut off, the shape distortion of the leading edge and the steady portion is taken into consideration, and for the other small plates, the shape distortion of the steady portion is taken into consideration.

[0043] In the present disclosure, the cutting equipment 50 is configured to include a geometric distortion measuring device 52. Alternatively, the geometric distortion measuring device 52 may be provided independently of the cutting equipment 50. The geometric distortion measuring device 52 is a device for measuring the amount or state of geometric distortion (e.g., C-warp or reverse C-warp) of the cut steel plate (small plate). The geometric distortion measuring device 52 transmits data obtained by measurement (measurement data) to the control command device 10.

[0044] FIG. 2 is a diagram showing an example of a schematic configuration of a cooling facility in a steel plate production line according to one embodiment of the present invention. More specifically, FIG. 2 shows an example of a schematic configuration of a cooling facility using a drain roll.

[0045] That is, the cooling equipment 40 shown in Fig. 2 has a plurality of cooling zones Z(1) to Z(n) partitioned by a pair of upper and lower rolls 41. The steel sheet passes through the cooling zones Z(1) to Z(n) in order to be cooled to a desired temperature.

[0046] The rolls 41 typically include a conveying roll 41a provided on the lower side and a water draining roll 41b provided on the upper side. In each cooling zone Z, a cooling header 42 is provided so as to face the upper and lower surfaces of the steel plate being conveyed, respectively. The cooling header 42 includes a plurality of injection nozzles 42a connected to a flow rate control valve 43. In order to cool the steel plate, the injection nozzles 42a inject cooling water, the injection flow rate of which is adjusted by the flow rate control valve 43, toward the steel plate under the control of the control command device 10. The control command device 10 can independently control the upper and lower flow rate control valves 43 to adjust the ratio of the amount of water and the amount of water. The control command device 10 also controls the flow rate control valve 43 to adjust the amount of water by the injection nozzles 42a corresponding to the leading end and stationary portion of the steel plate.

[0047] A radiation thermometer 44 is provided at the inlet and outlet of each cooling zone Z. The temperatures measured by the radiation thermometer 44 are notified to the control command device 10. The control command device 10 adjusts various operation parameters based on the notified temperatures, and controls the steel plate production line 1.

[0048] 3 is a diagram showing an example of a schematic configuration of a control command device in a steel sheet production line according to an embodiment of the present invention. As shown in the figure, the control command device 10 includes, for example, a process computer 110, a geometric distortion control device 120, and a database 130. The process computer 110, the geometric distortion control device 120, and the database 130 are connected to each other so as to be able to communicate with each other via a communication network such as an in-house LAN.

[0049] The process computer 110 is a computer for controlling the processing equipment. In this example, a heating equipment process computer 110a for controlling the heating equipment 20, a rolling equipment process computer 110b for controlling the rolling equipment 30, a cooling equipment process computer 110c for controlling the cooling equipment 40, and a cutting equipment process computer 110d for controlling the cutting equipment 50 are provided, but this is not limited thereto. For example, one process computer 110 may be responsible for controlling several processing equipment.

[0050] For example, the cooling equipment process computer 110c controls the flow rate regulating valve 43 to adjust the water / sewage flow ratio based on operation parameters related to cooling conditions that are set according to the predicted value of the geometric distortion of the steel plate. In addition, the cutting equipment process computer 110d controls the geometric distortion measuring device 52 to measure the shape of the cut steel plate and transmits the measurement data to the geometric distortion control device 120.

[0051] The geometric distortion control device 120 is a computer for controlling the geometric distortion of a steel plate manufactured in the steel plate manufacturing line 1. In this example, the geometric distortion control device 120 may include a geometric distortion prediction model 121 and a water / sewage flow control unit 122. The water / sewage flow control unit 122 issues a command to the process computer 110 based on a geometric distortion amount predicted according to the geometric distortion amount predicted by the geometric distortion prediction model 121. In this disclosure, the water / sewage flow control unit 122 outputs a control command to the cooling equipment process computer 110c to adjust an operation parameter related to the cooling condition in the cooling process (e.g., a ratio of water / sewage flow rates, etc.) according to the predicted geometric distortion amount.

[0052] The database 130 accumulates operation parameters of each process in the steel plate production line 1 and various performance parameters acquired during operation. The performance parameters include, for example, heating process data, rolling process data, cooling process data, cutting process data, and measurement data of shape distortion. The database 130 can also store the composition performance of the steel plate and material test data measured in advance by a predetermined measurement method.

[0053] 4 is a diagram for explaining shape distortion of a steel plate in a steel plate production line according to one embodiment of the present invention, showing a cross section of the steel plate in the width direction. A shape distortion measuring device 52 measures the surface shape of the steel plate, and thereby the state of shape distortion of the steel plate is determined.

[0054] For example, the geometric distortion measuring device 52 may be configured to include a plurality of distance sensors (not shown) arranged in the gap between the conveying rolls at a first interval in a direction (i.e., width direction of the steel plate) perpendicular to the conveying direction (i.e., longitudinal direction of the steel plate). The geometric distortion measuring device 52 measures the distance from the imaginary surface on the conveying roll to the lower surface of the steel plate conveyed in the conveying direction by the conveying roll (not shown) using the plurality of distance sensors. In this way, the geometric distortion measuring device 52 can obtain measurement data related to the distance from the imaginary surface to the lower surface of the steel plate in the longitudinal direction and width direction of the steel plate. The geometric distortion measuring device 52 transmits the obtained measurement data to the control command device 10. The control command device 10 judges the state of geometric distortion of the steel plate (C-warp or reverse C-warp) based on the measurement data transmitted from the geometric distortion measuring device 52. In the present disclosure, as described later, the state of geometric distortion of the steel plate is judged using a representative value calculated and / or selected from the measurement data.

[0055] More specifically, the geometric distortion measuring device 52 measures the distance from the conveying table at several sampling points (e.g., three points at the center of the leading edge and both edges) in the width direction at the leading edge and steady edge of the manufactured steel plate (i.e., small plate), and transmits measurement data based on the obtained distance values ​​to the control command device 10. The geometric distortion control device 120 of the control command device 10 calculates the difference between the distance value at the center and the distance value at the edge of the transmitted measurement data, and if the difference value is negative, it determines that there is C-warp, and if the difference value is positive, it determines that there is reverse C-warp. Note that the geometric distortion control device 120 determines that there is no warp if the difference is within a predetermined tolerance.

[0056] That is, when the amount of water on the upper side is greater than the amount of water on the lower side during the cooling process, and the cooling rate of the upper side of the steel plate is greater than that of the lower side, thermal stress is generated in the direction in which the upper side of the steel plate contracts, resulting in C-warp (Figure 1(a)). Conversely, when the amount of water on the upper side is less than the amount of water on the lower side, and the cooling rate of the lower side of the steel plate is greater than that of the upper side, thermal stress is generated in the direction in which the lower side of the steel plate contracts, resulting in reverse C-warp (Figure 1(b)). Thus, in the present disclosure, the amount of shape distortion indicating the above-mentioned C-warp / reverse C-warp is calculated for each of the tip and steady parts of the steel plate.

[0057] Next, the mechanism by which shape distortion occurs at the tip and stationary portions of a steel plate due to cooling treatment will be described. Fig. 5 is a diagram for explaining the state of cooling treatment in a steel plate production line according to one embodiment of the present invention. Fig. 6 is a graph showing the relationship between the ratio of the amount of cooling water to the amount of water and the shape distortion of the steel plate in the stationary portion of the steel plate during cooling treatment in a steel plate production line according to one embodiment of the present invention.

[0058] For example, as shown in FIG. 2, the cooling equipment 40 has each cooling zone Z separated by rolls 41, and is configured to be able to control, under the control of the control command device 10, the conveying speed of the steel plate, the amount of cooling water sprayed from the spray nozzle 42a, and the like in each cooling zone Z.

[0059] Here, as shown in Fig. 5, immediately after the steel material (large plate) is transported and enters the cooling zone Z, its leading edge is open, so that no pool of cooling water is formed on the upper surface of the steel plate by the cooling water sprayed from the spray nozzle 42a, and a sufficient cooling effect cannot be ensured compared to the steady portion. Therefore, as shown in Fig. 6(a), the control command device 10 adjusts the operation parameters so that a larger amount of cooling water is supplied to the leading edge of the steel plate than to the steady portion (dotted line in the figure), and in response to this, the cooling equipment 40 controls the flow rate adjustment valve 43. Meanwhile, the actual value of the amount of cooling water supplied to the steel plate changes continuously from the leading edge to the steady portion, as shown by the solid line.

[0060] Under the control of the control command device 10, the cooling equipment 40 adjusts the upper and lower water volume ratio so that the upper water volume is greater than the lower water volume, and controls the upper water volume so that the amount of cooling water on the upper surface decreases when the leading edge of the steel plate reaches the discharge roll 41. In this way, the cooling process for the leading edge of the steel plate is controlled differently from that for the steady portion. Therefore, as shown in Figures 6(b) and (c), the state of shape distortion may differ between the leading edge and the steady portion of the steel plate.

[0061] As shown in Fig. 1B, the leading edge and the steady state of the steel plate are not clearly distinguished, and in fact, the shape distortion at the boundary between the two changes gradually or continuously. Since the leading edge is cut off in the cutting equipment 50, in this disclosure, the leading edge of the steel plate refers to the leading edge of the small plate from which the leading edge is cut off, and the steady state refers to the portion following the leading edge. When multiple small plates are cut out from a large plate, the entirety of the other small plates other than the small plate from which the leading edge is cut off can be regarded as the steady state.

[0062] Next, the geometric distortion prediction model 121 will be described. In the present disclosure, in consideration of the above-mentioned mechanism of geometric distortion occurrence, various operational parameters including the water / sewage flow ratio are used as explanatory variables, and the geometric distortion amount of the steel plate is used as a target variable, and machine learning is performed on the geometric distortion prediction model 121 using a predetermined machine learning algorithm so that the predicted value of the geometric distortion amount of the steel plate approaches the actual value. The geometric distortion prediction model 121 may be configured to include, for example, two prediction models (a first prediction model and a second prediction model) internally so as to be able to predict each of the geometric distortions of the leading end portion and the steady portion of the steel plate.

[0063] The input (explanatory variables) to the shape distortion prediction model 121 takes into consideration the influence on the output (target variable) and includes at least some of, for example, the components of the steel plate, the size of the steel plate (e.g., plate thickness, plate width, plate length, etc.), the heating temperature in the heating equipment, the ratio of water and water to each of the tip portion and the steady portion, as well as various cooling conditions (e.g., cooling start temperature, cooling end temperature, cooling zone used), and the cooling water temperature.

[0064] As described above, the leading edge of the steel plate (large plate) cooled in the cooling equipment 40 is cut in the cutting equipment 50 to become a steel plate (small plate) as a product. In the present disclosure, the longitudinal length of the leading edge cut from the large plate is referred to as the cut length. The shape distortion amount of each cut small plate is measured by the shape distortion measuring device 52. The manufactured steel plate is required to have no shape distortion or to have a shape distortion within a substantially negligible range (for example, within ±1 mm). Therefore, the shape distortion prediction model 121 is machine-learned so that the shape distortion amount (predicted value) of the steel plate predicted based on various operation parameters including the water and sewage flow ratio approaches the value (actual value) of the shape distortion amount of the actual steel plate. In the present disclosure, the leading edge and steady portion of the steel plate whose shape distortion amount is to be predicted are defined as follows.

[0065] That is, as described above, in the cooling process, since no pool is formed at the leading edge of the steel plate immediately after it is carried into the cooling zone Z, the amount of water for the leading edge is controlled to be greater than that for the steady portion. Meanwhile, the cooling conditions depend on the transport speed of the steel plate and the response speed (e.g., about 3 to 4 seconds) of the flow control valve 43 that adjusts the amount of water for the leading edge and steady portion accordingly. Therefore, the leading edge and steady portion of the steel plate are determined by taking into consideration the cut length and operational parameters related to the cooling conditions (the transport speed at the start of cooling, the response speed of the flow control valve 43, etc.).

[0066] The shape distortion measuring device 52 measures the shape distortion of the tip of the steel material, which is determined according to the cutting length, and the subsequent steady portion. In this case, for example, the maximum distortion amount at the tip is defined as the tip distortion amount, and the average distortion amount of the steady portion is defined as the steady portion distortion amount, and these are used as representative values ​​of the objective variables.

[0067] The amount of shape distortion of the steel plate is strongly correlated with the ratio of water flow rate to downstream water flow rate, and it is considered that there is a monotonically increasing phenomenon between the ratio and the amount of shape distortion. In other words, when the amount of water flow rate on the upper side increases relative to the amount of water flow rate on the lower side, the steel plate is C Warping is likely to occur, and when the amount of water on the upper side decreases relative to the amount of water on the lower side, reverseC Warping becomes more likely to occur.

[0068] That is, as shown in Fig. 7, the steady part of the manufactured steel plate reverse If it is determined that warping has occurred, adjust the upper and lower water volume ratio so that the upper water volume increases (the lower water volume decreases). reverse C-warp C The amount of shape distortion can be suppressed so that the warping cancels out (arrow A in the figure). C If warping is determined to be occurring, adjust the ratio of water volume above and below to decrease the amount of water above (increase the amount of water below). 、C Warping reverse The amount of shape distortion can be suppressed so that the warping cancels out (arrow B in the figure). Note that while the figure shows the relationship between shape distortion in the steady section and the ratio of water and drainage volume, the same explanation can be applied to the tip section.

[0069] Therefore, in the present disclosure, machine learning is performed on the shape distortion prediction model 121 using a predetermined machine learning algorithm (e.g., a gradient boosting decision tree) executed under a monotonicity constraint between the water / sewage volume ratio, which is one of the explanatory variables, and the shape distortion amount, which is the objective variable, and the water / sewage volume ratio is determined using an inverse analysis technique based on the shape distortion amount predicted by the learned shape distortion prediction model 121 and the target distortion amount.

[0070] In addition, in the present disclosure, a gradient boosting decision tree is used in a machine learning algorithm to consider a monotonicity constraint between a specific explanatory variable and a response variable, thereby constructing a geometric distortion prediction model that can satisfactorily explain the correlation between the water / sewage volume ratio, which is one of the explanatory variables, and the geometric distortion amount, which is the response variable. Furthermore, since such a monotonicity constraint can be considered, it becomes easier to apply the inverse analysis method. However, the machine learning algorithm is not limited to this, and any algorithm that can consider the monotonicity constraint may be used.

[0071] Next, construction of the geometric distortion prediction model 121 will be described. The geometric distortion prediction model 121 is constructed by performing machine learning in advance based on various performance parameters acquired during the operation of the steel sheet production line 1. Fig. 8 is a flowchart for explaining an example of a process for constructing the geometric distortion prediction model 121 according to one embodiment of the present invention.

[0072] As shown in the figure, in a steel plate production line 1, a steel plate is manufactured through each process (S801). That is, a slab (steel material) is turned into a steel plate as a product through each process of heating, rolling, cooling, and cutting. Performance parameters acquired in each process are stored in a database 130 via a process computer 110.

[0073] Furthermore, in the cutting process, the geometric distortion measuring device 52 measures the geometric distortion of the cut steel plate (small plate) (S802). Typically, a plurality of small plates are cut out from one large plate, and for the small plate whose leading edge is cut off, the geometric distortion is measured at a plurality of points on the leading edge and the steady portion, and for the other small plates, the geometric distortion is measured at a plurality of points on the steady portion. The geometric distortion measuring device 52 transmits the acquired measurement data to the geometric distortion control device 120.

[0074] On the other hand, the shape distortion control device 120 controls the operating parameters related to the cutting length and the cooling conditions. To Based on this, the tip portion and the steady portion of the steel plate are determined (S803).

[0075] Next, the geometric distortion control device 120 calculates the tip portion distortion amount and the steady portion distortion amount for the determined tip portion and steady portion based on the measurement data transmitted from the geometric distortion measuring device 52 (S804). The geometric distortion control device 120 stores the calculated tip portion distortion amount and steady portion distortion amount in the database 130 as performance parameters (S805).

[0076] Furthermore, the geometric distortion control device 120 judges whether it is time to update the geometric distortion prediction model 121 (S806). For example, the geometric distortion control device 120 may judge whether it is time to update the geometric distortion prediction model 121 based on whether a predetermined time has passed since the previous update. Alternatively, the geometric distortion control device 120 may judge whether it is time to update the geometric distortion prediction model 121 by evaluating deterioration of the prediction accuracy of the geometric distortion prediction model 121. For example, the geometric distortion control device 120 obtains the prediction accuracy of the geometric distortion prediction model 121 by statistically evaluating the deviation (e.g., root mean square error (RMSE)) between the predicted value of the geometric distortion amount by the trained geometric distortion prediction model 121 and the actual value of the geometric distortion amount measured by the geometric distortion measurement device 52, and judges that the prediction accuracy has deteriorated when it exceeds a predetermined allowable value.

[0077] When the shape distortion control device 120 determines that it is time to update the shape distortion prediction model 121 (Yes in S807), it performs machine learning on the shape distortion prediction model 121 using operational parameters related to the current water / sewage flow ratio as explanatory variables so that the objective variable becomes a predetermined value (S808).

[0078] If it is determined that it is time to update the geometric distortion prediction model 121 (No in S807), the geometric distortion control device 120 returns to the process of S801 and repeats the above process until the next update timing.

[0079] Next, a description will be given of a process for setting the water / sewage flow ratio using the trained geometric distortion prediction model 121. Fig. 9 is a flowchart for explaining an example of a process for setting the water / sewage flow ratio based on a prediction result using the geometric distortion prediction model 121 according to an embodiment of the present invention. All or a part of this process can be executed by, for example, the water / sewage flow control unit 122 in the geometric distortion control device 120 of the control instruction device 10.

[0080] As shown in the figure, the shape distortion control device 120 collects operational parameters and performance parameters before the cooling process from the cooling equipment 40 via the cooling equipment process computer 110c (S901). The operational parameters are setting data including, for example, the cooling start temperature, the cooling end temperature, the cooling zone to be used, the temperature of the cooling water, etc., and the performance parameters are data actually measured by a sensor or the like.

[0081] Next, the shape distortion control device 120 determines the current adjustment mode (S902). In this example, the adjustment mode includes a first adjustment mode in which the water / sewage flow rate ratio is adjusted for both the tip portion and the steady portion, and a second adjustment mode in which the water / sewage flow rate ratio is adjusted for only the steady portion, but is not limited thereto. For example, the adjustment mode may further include a third adjustment mode in which the water / sewage flow rate ratio is adjusted for only the tip portion. The adjustment mode is, for example, selected and set in advance by an operator, but is not limited thereto.

[0082] When the geometric distortion control device 120 determines that the current adjustment mode is the first adjustment mode, it first predicts the tip distortion amount using the learned geometric distortion prediction model 121, and searches for and determines the optimal water / sewage flow ratio using an inverse analysis method based on the predicted value (S903). As an example of the inverse analysis method, there is a grid search method.

[0083] Next, the geometric distortion control device 120 similarly predicts the steady-state distortion amount using the learned geometric distortion prediction model 121, and based on the predicted value, searches for and determines the optimal water / sewage flow ratio using an inverse analysis method (S904).

[0084] If the geometric distortion control device 120 determines that the current adjustment mode is the second adjustment mode, it skips the process of S903 and searches for and determines an optimal water / sewage flow rate ratio based on the predicted value of the steady-state distortion amount (S904).

[0085] Next, the shape distortion control device 120 checks the rationality and equipment constraints of the determined water / sewage flow ratio (S905). The check of rationality and equipment constraints is performed by comparing with a table (operation parameter table) in the database 130 that specifies operation parameters related to the water / sewage flow ratio. For example, the shape distortion control device 120 calculates the deviation between the determined water / sewage flow ratio and the classification water / sewage flow ratio corresponding to the steel plate in the operation parameter table as a predetermined reference value, and if the deviation exceeds a predetermined threshold value, it considers that the predicted value is wrong and uses the water / sewage flow ratio in the operation parameter table as the operation parameter without using the water / sewage flow ratio determined based on the predicted value. Alternatively, the shape distortion control device 120 may use a value in the operation parameter table that is corrected as necessary so that the calculated deviation falls within a predetermined limit range. By imposing such a limit on the calculated deviation, it is possible to prevent the determined water / sewage flow ratio from deviating significantly from the predetermined reference value. The geometric distortion controller 120 updates the operational parameters based on the determined water / sewage ratio after checking for rationality and equipment constraints.

[0086] Then, the geometric distortion control device 120 controls the upper and lower flow rate adjustment valves 43 in the cooling equipment 40 to adjust them in accordance with the updated operation parameters (S906).

[0087] Next, a description will be given of an inverse analysis method using the geometric distortion prediction model 121. Fig. 10 is a flowchart for explaining an example of the inverse analysis method using the geometric distortion prediction model 121 according to an embodiment of the present invention. Fig. 11 is a conceptual diagram for explaining an example of the inverse analysis technique.

[0088] The inverse analysis method is a method of calculating the values ​​of explanatory variables so as to satisfy the value of the objective variable (target value). A grid search method is known as an example of the inverse analysis method. The grid search method is a method of comprehensively calculating candidates for a solution (i.e., the value of the explanatory variable), selecting a candidate solution close to the target value, and further dividing and narrowing it down. In the present disclosure, a water and sewage volume ratio that can bring the predicted value of the amount of geometric distortion predicted by the trained geometric distortion prediction model 121 close to the target value (i.e., amount of geometric distortion=0) is searched for and determined (see FIG. 7).

[0089] That is, as shown in the figure, first, the geometric distortion control device 120 sets a range of the water / sewage flow ratio (solution) as a solution to be obtained (S1001). Next, the geometric distortion control device 120 divides the set solution range into a predetermined number of divisions (e.g., 30) (S1002), and further predicts the amount of geometric distortion at the division points using the geometric distortion prediction model 121 (S1003 and FIG. 11(a)). The amount of geometric distortion is predicted multiple times.

[0090] Next, the geometric distortion control device 120 compares the average values ​​of adjacent predicted geometric distortion amounts with the target value, and selects the pair of adjacent average values ​​of predicted geometric distortion amounts that are closest to the target value as the optimal pair (S1004 and FIG. 11(b)).

[0091] Next, the geometric distortion control device 120 judges whether the above process has been repeated a predetermined number of times (S1005), and if it is judged that the process has not been repeated the predetermined number of times (No in S1005), it sets the solution range based on the optimal pair (S1006 and FIG. 11(c)), and returns to the process of S1002. The number of repetitions is, for example, 3.

[0092] On the other hand, when the geometric distortion control device 12 judges that the process has been repeated a predetermined number of times (Yes in S1005), it determines the ratio of water and upper and lower volumes for the current optimal pair (S1007). For example, referring to Fig. 8, if the distortion amount in the steady state is +0.2 mm, the geometric distortion control device 12 sets the ratio of water and upper volumes to 0.7 so that the upper water volume decreases. As a result, the geometric distortion control device 12 controls the flow rate adjustment valve 43 to be adjusted based on the determined ratio of water and upper volumes.

[0093] (Verification example) 12 is a graph comparing the predicted value and actual value of the steady-state part distortion by the geometric distortion prediction model according to one embodiment of the present invention. In the graph, the horizontal axis represents the predicted value of the steady-state distortion amount by the geometric distortion prediction model 121, and the vertical axis represents the actual value of the geometric distortion amount of the steel plate actually manufactured. In other words, each white plot in the graph indicates the predicted value and the actual value of each verified steel plate.

[0094] As shown in the figure, of the population of the verified steel plates, the group belonging to the first quadrant (predicted value > 0 and actual value > 0) was 17.5%, the group belonging to the third quadrant (predicted value < 0 and actual value < 0) was 2.3%, and the groups belonging to the second and fourth quadrants were 0.1%, respectively.

[0095] In this example, when the amount of shape distortion is ±1 mm or less, it is considered that there is no shape distortion or that the distortion can be ignored. In this case, 80% of the groups had a steady-state distortion amount of ±1 mm, while the actual value was ±1 mm.

[0096] From the above, it can be seen that the directionality indicated by the predicted value by the geometric distortion prediction model 121 is almost consistent with the directionality of the actual value. Specifically, the coefficient of determination R2 value between the predicted value and the actual value was 0.48. Therefore, the geometric distortion prediction model 121 of the present disclosure can effectively predict the geometric distortion amount of the steel plate, and an optimal water / sewage flow ratio can be determined by using an inverse analysis method based on the predicted value.

[0097] As described above, according to the present embodiment, it is possible to more effectively suppress the shape distortion of the steel plate caused by the cooling treatment in the manufacture of the steel plate. In particular, according to the present invention, it is possible to construct a shape distortion prediction model capable of appropriately predicting the shape distortion of the tip portion and the steady portion of the shape distortion of the steel plate caused by the cooling treatment in the manufacture of the steel plate. Therefore, by using such a shape distortion prediction model, the cooling treatment is controlled by adjusting the operation parameters so as to effectively suppress the shape distortion of the steel plate caused by the cooling process, so that the shape distortion of the tip portion and the steady portion can be suppressed and a high quality steel plate can be manufactured.

[0098] The above-described embodiments are merely examples for explaining the present invention, and are not intended to limit the present invention to these embodiments. The present invention can be embodied in various forms without departing from the spirit of the present invention.

[0099] For example, in the methods disclosed herein, steps, operations, or functions may be performed in parallel or in a different order, provided that the results are consistent. The steps, operations, and functions described are provided merely as examples, and some of the steps, operations, and functions may be omitted or combined into one, or other steps, operations, or functions may be added, without departing from the spirit of the invention.

[0100] In addition, although various embodiments are disclosed in this specification, specific features (technical matters) in one embodiment can be added to or replaced with specific features in other embodiments, with appropriate modifications, and such forms are also included in the spirit of the present invention. [Explanation of symbols]

[0101] 1...Steel plate production line 10...Control instruction device 110...Process computer 110a...Process computer for heating equipment 110b...Process computer for rolling equipment 110c…Process computer for cooling equipment 110d...Process computer for cutting equipment 120...Shape distortion control device 121…Shape distortion prediction model 122...Water and sewage volume control unit 130…Database 20…Heating equipment 30...Rolling equipment 40…Cooling equipment 50…Cutting equipment 52...Shape distortion measuring device

Claims

1. A control command device used in a steel plate production line for producing steel plates, a shape distortion prediction model in which machine learning is performed using a predetermined machine learning algorithm, with predetermined cooling conditions including at least the components of the steel plate, the size of the steel plate, the heating temperature for the steel plate, and the water / sewage flow rate ratio of cooling water supplied to the steel plate as explanatory variables, and with the amount of shape distortion of the steel plate as a response variable, so that predicted values ​​of the amount of shape distortion of each of the tip portion and the steady portion of the steel plate approach actual values ​​of the shape distortion of each of the tip portion and the steady portion; and a water / sewage flow rate control unit that determines the water / sewage flow rate ratio of the cooling water to be supplied to each of the tip portion and the steady portion under the predetermined cooling condition based on the predicted values ​​of the amount of shape distortion of each of the tip portion and the steady portion output from the shape distortion prediction model. Control command device.

2. The predetermined cooling conditions further include at least one of a cooling start temperature and a cooling end temperature of the steel plate, a temperature of the cooling water, and a cooling zone used to cool the steel plate. The control command device according to claim 1 .

3. The amount of shape distortion is the amount of C-warp or reverse C-warp based on the difference in height between the center and edge portions in the width direction of the steel plate.

3. The control command device according to claim 1 or 2.

4. The predetermined machine learning algorithm is executed under a monotonicity constraint between the upper and lower flow rate ratio and the amount of shape distortion. The control command device according to claim 1 .

5. The shape distortion prediction model outputs the prediction value based on actual parameters collected during operation of the steel plate production line, the actual parameters being predetermined cooling conditions including at least components of the steel plate, a size of the steel plate, a heating temperature for the steel plate, and the water / sewage rate ratio of the cooling water supplied to the steel plate. The control command device according to any one of claims 1 to 4.

6. the water / sewage volume control unit determines the water / sewage volume ratio by an inverse analysis method in accordance with the predicted value of the amount of geometric distortion. The control command device according to claim 5.

7. For each of the tip portion and the steady portion, when it is determined that the C-shaped warp is occurring based on the predicted value of the amount of shape distortion, the water / sewage flow control unit determines the water / sewage flow ratio so that the upper water volume is less than the lower water volume, and when it is determined that the reverse C-shaped warp is occurring based on the predicted value of the amount of shape distortion, the water / sewage flow control unit determines the water / sewage flow ratio so that the upper water volume is more than the lower water volume. The control command device according to claim 3.

8. A method for operating a control command device in a steel plate production line for producing steel plates, comprising the steps of: performing machine learning on a geometric distortion prediction model using a predetermined machine learning algorithm, with predetermined cooling conditions including at least the components of the steel plate, the size of the steel plate, the heating temperature for the steel plate, and the water / sewage flow rate ratio of cooling water supplied to the steel plate as explanatory variables, and with the amount of geometric distortion of the steel plate as a response variable, so that predicted values ​​of the amounts of geometric distortion of each of the tip portion and the steady portion of the steel plate approach actual values ​​of the amounts of geometric distortion of each of the tip portion and the steady portion; Using the machine learning-based geometric distortion prediction model, a predicted value of the geometric distortion amount is output based on the explanatory variables during operation of the steel sheet production line; and determining the water / sewage flow ratios of the tip portion and the steady portion under the predetermined cooling conditions using an inverse analysis method based on the predicted values ​​of the tip portion and the steady portion output from the shape distortion prediction model; and controlling the cooling equipment so as to adjust the flow rates of the water / sewage flow ratios of the tip portion and the stationary portion under the predetermined cooling conditions based on the determined water / sewage flow ratio. A method for operating a control command device.

9. A method for constructing a shape distortion prediction model for outputting a predicted value of a shape distortion amount of a steel plate, the prediction value being used for determining a water / sewage ratio of cooling water to be supplied to a steel plate in which shape distortion has occurred during a cooling process in a steel plate production line for producing the steel plate, the method comprising: the method includes performing machine learning on a geometric distortion prediction model using a predetermined machine learning algorithm, with predetermined cooling conditions including at least the components of the steel plate, the size of the steel plate, the heating temperature for the steel plate, and the water / sewage flow rate ratio of cooling water supplied to the steel plate as explanatory variables, and with the amount of geometric distortion of the steel plate as an objective variable, so that predicted values ​​of the amounts of geometric distortion of each of the tip portion and the steady portion of the steel plate approach actual values ​​of the geometric distortion of each of the tip portion and the steady portion. How to build a shape distortion prediction model.

10. A steel plate manufacturing method, comprising: determining and controlling a ratio of a water flow rate and a wastewater flow rate of the cooling water to be supplied to the steel plate by using the shape distortion prediction model according to claim 9; The machine learning-based geometric distortion prediction model outputs the predicted values ​​of the geometric distortion amounts of the tip portion and the steady portion based on actual parameters collected during the operation of the steel plate production line, the actual parameters being predetermined cooling conditions including at least the components of the steel plate, the size of the steel plate, the heating temperature for the steel plate, and the water / sewage rate ratio of the cooling water supplied to each of the tip portion and the steady portion of the steel plate; a water / sewage flow control unit that determines the water / sewage flow ratios of the tip portion and the steady portion by using an inverse analysis method based on the predicted values ​​of the tip portion and the steady portion that are output; The water / sewage flow control unit controls the supply of the cooling water according to the determined water / sewage flow ratios of the tip portion and the steady portion. Steel plate manufacturing method.

Images