Shape prediction method for unequal angle steel, manufacturing method for unequal angle steel, rolling facility and generation method for shape prediction model

Abstract

To provide a shape prediction method for unequal angle steel that can accurately predict a bending amount of unequal angle steel.SOLUTION: In a shape prediction method for unequal angle steel, which predicts a bending amount of unequal angle steel that is manufactured by a rolling facility which has a plurality of rolling machines that hot-rolls heated steel billets to unequal angle steel, a cooling device that cools the unequal angle steel and a shape measurer that measures a bending amount of the cooled unequal angle steel, input data, which includes rolling-operation parameters of the rolling machines, a cooling-operation parameter of the cooling device and an attribute parameter of the unequal angle steel, is inputted to a shape prediction model and the bending amount is outputted, so as to predict the bending amount of the unequal angle steel.SELECTED DRAWING: Figure 5

Description

[Technical field]

[0001] The present invention relates to a shape prediction method for predicting the amount of bending of unequal leg angle steel, a manufacturing method for unequal leg angle steel, rolling equipment, and a method for generating a shape prediction model. [Background technology]

[0002] The background art will be described using an example of unequal thickness angle steel (hereinafter also referred to as "NAB"), which is a type of unequal angle steel and is used as a reinforcing steel material for structures such as ship hulls, bridges, and tanks. NAB is manufactured by hot rolling a steel billet as a material. FIG. 1 is a schematic diagram showing the cross-sectional shape of NAB. As shown in FIG. 1, NAB 10 is composed of long sides 12 that are relatively thin and wide, and short sides 14 that are relatively thick and narrow. There are about a dozen combinations of dimensions such as thickness and width of the long sides 12 and short sides 14. In addition, the mechanical properties that NAB should satisfy are determined by the application and the classification society's standard classification, and the chemical composition of NAB is appropriately adjusted accordingly. When NAB is used as a steel material for a structure, the dimensions, thickness, and chemical composition of the long sides 12 and short sides 14 are selected according to the required performance of the structure.

[0003] In the manufacture of such NABs, a steel slab, which is a material that has been preheated in a heating furnace, is generally subjected to rough rolling, intermediate rolling, and finish rolling using multiple rolling mills to form the slab into a cross-sectional shape with the desired dimensions. For example, in the rough rolling and intermediate rolling, the steel slab is rolled multiple times using a rough rolling mill and an intermediate rolling mill that have a pair of upper and lower rolls with multiple grooves called calibers, and is roughly formed to approximate the desired dimensions. Finally, one pass of finish rolling is performed using a finishing rolling mill that has a pair of upper and lower rolls with one caliber, forming the slab into a NAB with the desired cross-sectional shape.

[0004] The Japanese Industrial Standard JISG3192:2021 (Shape, Dimensions, Mass and Tolerances of Hot Rolled Section Steel) and corresponding foreign and international standards prescribe the tolerances for the cross-sectional dimensions of section steel. For example, the NAB prescribes the representative values ​​and their tolerances for each dimension such as long side width A, short side width B, long side thickness t1, and short side thickness t2, and requires that all of these dimensions be controlled so that they fall within the range of the above representative values ​​and the tolerances. Therefore, the rolling conditions such as the roll gap for each rolling mill and each pass are adjusted and modified so that each NAB dimension falls within the tolerance.

[0005] As shown in Fig. 1, NAB rolls a steel sheet in a V-shape with a vertically asymmetric cross-sectional shape, so that the reduction ratios of the long and short sides in one rolling pass are different. That is, the amount of elongation in the rolling direction differs between the long and short sides 12 and 14 in one rolling pass, so bending may occur along the longitudinal direction. In addition, when the rolling conditions, such as the distance between the upper and lower rolls in each pass, are changed for the purpose of dimensional adjustment, the amount of bending in the longitudinal direction due to rolling differs each time.

[0006] In addition, since the thickness of the short sides of the NAB is thicker than that of the long sides, the short sides are more difficult to cool than the long sides during the cooling process during rolling, and the temperature becomes higher. Thermal stress caused by the temperature difference between the long and short sides and stress caused by phase transformation may also cause bending along the longitudinal direction during or after cooling. Therefore, for example, a cooling device is provided before the finishing rolling mill to cool the long or short sides of the NAB during the hot rolling process, thereby eliminating the temperature difference between the long and short sides and reducing the amount of bending. Furthermore, when manufacturing a higher strength NAB, an accelerated cooling device is sometimes provided at the exit side of the finishing rolling mill to water-cool the long or short sides of the NAB in order to control the structure of the steel material.

[0007] As described above, in the manufacture of NAB, bending occurs due to the complex overlap of the effects of rolling and cooling, making it difficult to control the amount of NAB bending, and it has been difficult to keep the amount of NAB bending within the allowable range. Furthermore, if the amount of NAB bending is large at the product stage, additional processes such as pressing are required, which increases manufacturing costs and reduces productivity.

[0008] As a technique for reducing the amount of bending of the NAB, Patent Document 1 discloses a cooling method for cooling the long and short sides so that the temperature of the long sides is equal to or higher than the temperature of the short sides and the temperature difference is within 50° C. According to Patent Document 1, by controlling the cooling of the long and short sides of the NAB so as to achieve the above temperature difference, bending of the NAB can be suppressed.

[0009] Patent Document 2 discloses a control method in which predicted values ​​are calculated using a rolling deformation prediction model based on various setting data and measurement data related to rolling conditions, the amount of NAB bending is obtained from these predicted values, and the rolling conditions are set based on this amount of bending. According to Patent Document 2, it is possible to reduce warping and bending of the rolled material by setting and adjusting the optimal positions of the roll gap, entry guide, and exit guide based on the amount of NAB bending. [Prior art documents] [Patent documents]

[0010] [Patent Document 1] Japanese Patent Application Publication No. 188726 / 1983 [Patent Document 2] JP 2003-39107 A Summary of the Invention [Problem to be solved by the invention]

[0011] The method disclosed in Patent Document 1 is not particularly limited with respect to the dimensions or size of the NAB, the steel type, or the rolling conditions, and claims that bending can be suppressed as long as the temperature of the long side is equal to or higher than the temperature of the short side and the temperature difference is within 50° C. However, there was a problem that, depending on the dimensions or sizes of some of the dozens of NABs, the amount of NAB bending was not sufficiently reduced even when cooling was performed under the above cooling conditions.

[0012] In addition, the bending of the NAB is also affected by stress caused by phase transformation during cooling. There are several types of steel for NAB, and different types of steel cause different stresses caused by phase transformation even if the thermal history and temperature are the same. For this reason, there was a problem that the bending of the NAB could not be sufficiently reduced even when cooling was performed by the cooling method described in Patent Document 1, depending on the type of steel.

[0013] Furthermore, the amount of NAB bending may be affected not only by temperature but also by rolling conditions such as the roll gap of each rolling mill performed for dimensional adjustment, that is, by plastic deformation due to rolling. The method described in Patent Document 1 does not take into account the effects of rolling conditions, and therefore has a problem in that the amount of NAB bending cannot be sufficiently reduced depending on the rolling conditions.

[0014] Patent Document 2 discloses, as an example of a shape prediction model, a formula for predicting the NAB bending using a linear sum of each parameter of the rolling conditions. However, since the amount of NAB bending is affected not only by the rolling conditions but also by temperature, it is difficult to accurately predict the amount of bending using such a formula. In addition, in order to accurately predict the amount of NAB bending taking into account not only the temperature but also the rolling effects, two-dimensional or three-dimensional structural analysis is required, so that online prediction requires a very high calculation load, which is difficult to achieve with the current capabilities of computers.

[0015] The present invention has been made in consideration of the above-mentioned problems, and its object is to provide a shape prediction method for unequal leg angle steel that can accurately predict the amount of bending of unequal leg angle steel, a manufacturing method for unequal leg angle steel using the shape prediction method for unequal leg angle steel, rolling equipment, and a method for generating a shape prediction model. [Means for solving the problem]

[0016] The means for solving the above problems are as follows. [1] A shape prediction method for unequal angle steel, which predicts the amount of bending of unequal angle steel manufactured in a rolling facility having a plurality of rolling mills that hot roll heated steel pieces into unequal angle steel, a cooling device that cools the unequal angle steel, and a shape meter that measures the amount of bending of the unequal angle steel after cooling, wherein input data including rolling operation parameters of the rolling mills, cooling operation parameters of the cooling device, and attribute parameters of the unequal angle steel are input into a shape prediction model, and the amount of bending is output to predict the amount of bending. [2] Identifying the cooling operation parameters such that the amount of bending predicted using the shape prediction method for unequal leg angle steel described in [1] falls within a predetermined allowable range; A method for manufacturing unequal angle steel, which manufactures unequal angle steel under manufacturing conditions including specified cooling operation parameters. [3] A rolling facility having a plurality of rolling mills for hot rolling heated steel billets into unequal angle steel, a cooling device for cooling the unequal angle steel, a shape meter for measuring the amount of bending of the unequal angle steel after cooling, and a shape prediction device for the unequal angle steel for predicting the amount of bending, The shape prediction device is a rolling facility having a shape prediction unit that inputs input data including rolling operation parameters of the rolling mill, cooling operation parameters of the cooling device, and attribute parameters of the unequal leg angle steel into a shape prediction model, and outputs the amount of bending. [4] A method for generating a shape prediction model that predicts the amount of bending of unequal angle steel manufactured in a rolling facility having a plurality of rolling mills that hot roll heated steel pieces into unequal angle steel, a cooling device that cools the unequal angle steel, and a shape meter that measures the amount of bending of the unequal angle steel after cooling, the method comprising: training a machine learning model using a plurality of data sets, each set being a set of actual values ​​of rolling operation parameters of the rolling mills, actual values ​​of cooling operation parameters of the cooling device, actual values ​​of attribute parameters of the unequal angle steel, and actual values ​​of the amount of bending, of unequal angle steel manufactured in the past, as training data; and generating a shape prediction model that uses the rolling operation parameters, the cooling operation parameters, and the attribute parameters as inputs and outputs the amount of bending. [5] The method for generating a shape prediction model described in [4], wherein the machine learning model is one of a neural network, decision tree learning, random forest, and support vector regression. Effect of the Invention

[0017] According to the present invention, the amount of bending of unequal leg angle steel is predicted using a shape prediction model of unequal leg angle steel that takes into account the effects of not only cooling but also rolling, so that the amount of bending of unequal leg angle steel can be predicted with high accuracy. Furthermore, by using the predicted amount of bending to control the cooling conditions in the manufacture of unequal leg angle steel, it becomes possible to stably manufacture unequal leg angle steel with the amount of bending within the allowable range. [Brief description of the drawings]

[0018] [Figure 1] FIG. 1 is a schematic diagram showing the cross-sectional shape of the NAB. [Diagram 2] FIG. 2 is a schematic diagram showing the progress of the target cross-sectional shape in each manufacturing process of the NAB. [Diagram 3] FIG. 3 is a schematic diagram showing the cross-sectional shapes of the ABS and the BP. [Figure 4] FIG. 4 is a perspective view for explaining the amount of bending of the NAB. [Diagram 5] FIG. 5 is a schematic diagram showing a schematic configuration of a rolling facility including an unequal leg angle steel shape prediction device capable of implementing the unequal leg angle steel shape prediction method according to this embodiment. [Figure 6] FIG. 6 is a schematic diagram showing a state in which the long and short sides of the NAB are being cooled by a cooling device. [Figure 7] FIG. 7 is a schematic diagram showing the warping of the NAB in the vertical direction. [Figure 8] FIG. 8 is a functional block diagram of a shape prediction device for an unequal leg angle steel. [Figure 9] FIG. 9 is a flow diagram showing a flow of a cooling operation parameter specification process performed by the cooling operation parameter specification unit. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0019] An embodiment of the present invention will be described below with reference to the drawings. In the following embodiment, the present invention will be described using an example in which the present invention is applied to a NAB, which is a type of unequal leg angle iron. Note that the following embodiment is merely an example of an apparatus and method for embodying the technical concept of the present invention, and the technical concept of the present invention is not limited to the following embodiment.

[0020] Figure 2 is a schematic diagram showing the progression of the target cross-sectional shape at each manufacturing process of NAB 10. NAB 10 is manufactured by heating a steel slab, which is the raw material, in a heating furnace, and then rolling the slab with each rolling mill in the rough rolling process, intermediate rolling process, and finish rolling process.

[0021] FIG. 3 is a schematic diagram showing the cross-sectional shapes of an unequal-leg equal-thickness angle steel 16 and a spherical flat steel 20. FIG. 3(a) shows the cross-sectional shape of an unequal-leg equal-thickness angle steel (hereinafter also referred to as "ABS") 16 in which the thickness of the short side 17 and the long side 18 are equal. FIG. 3(b) shows the cross-sectional shape of a spherical flat steel (hereinafter also referred to as "BP (valve plate)") 20 in which the short side 21 is tapered. ABS16 and BP20 are each a type of unequal-leg angle steel. Since ABS16 and BP20 can also be manufactured by the same manufacturing method as NAB10, the shape prediction method of unequal-leg angle steel, the manufacturing method of unequal-leg angle steel, and the rolling equipment according to this embodiment can also be applied to predicting the amount of bending of ABS and BP and manufacturing ABS and BP. Therefore, the unequal-leg angle steel targeted in this embodiment includes NAB, ABS, and BP.

[0022] 4 is a perspective view illustrating the amount of bending of the NAB 10. In this embodiment, the amount of bending d of the NAB 10 is indicated by a sign "+" when the NAB 10 bends so that the long side 12 side is convex with respect to a line connecting the bottom end of the short side 14 on the leading end side and the bottom end of the short side 14 on the trailing end side, and indicated by a sign "-" when the NAB 10 bends so that the long side 12 side is convex with respect to a line connecting the bottom end of the short side 14 on the trailing end side.

[0023] 5 is a schematic diagram showing a schematic configuration of a rolling facility 100 including a shape prediction device 58 for unequal leg angle steel in which the shape prediction method for unequal leg angle steel according to the present embodiment can be implemented. As shown in FIG. 5, the rolling facility 100 includes a heating furnace 30 for heating a steel slab to a predetermined temperature, a roughing mill 32 for rolling the steel slab heated in the heating furnace 30, intermediate rolling mills 34 and 36, a finishing rolling mill 38, a rolling control device 40 for controlling the operation of these multiple rolling mills, cooling devices 42 and 44 for cooling the unequal leg angle steel during rolling, thermometers 46, 48, and 50 for measuring the temperature of the unequal leg angle steel during rolling at the inlet and outlet sides of the cooling devices 42 and 44, a cooling control device 52 for controlling the operation of the cooling devices 42 and 44, a shape meter 54 for measuring the amount of bending of the unequal leg angle steel after cooling, a process computer 56, and a shape prediction device 58 for unequal leg angle steel.

[0024] A steel slab is charged into the heating furnace 30 and heated to a temperature above the austenite temperature range (for example, 1100 to 1300°C). The steel slab extracted from the heating furnace 30 is transported to the rolling mills by a plurality of table rollers installed on the outlet side of the heating furnace 30, and is hot rolled into the NAB 10 by the roughing mill 32, intermediate rolling mills 34 and 36, and finishing rolling mill 38. Furthermore, before and after the rolling mills, the NAB 10 is cooled by cooling devices 42 and 44 while being transported by the table rollers.

[0025] In rough rolling, the raw material, a rectangular cross-section steel billet, is rolled multiple times in a rough rolling mill 32 to adjust the size of the billet and form it into a crude billet. In intermediate rolling, the raw billet is rolled multiple times in intermediate rolling mills 34 and 36 to form it into a shape with an upward convex shape near the joint between the long and short sides. In finish rolling, the rolled material is shaped into the product shape by a finish rolling mill 38 having a caliber for the final product shape. The number of calibers and the number of passes through which the rolled material passes vary depending on the size of the product.

[0026] The rolling operation parameters in such a hot rolling process include the gap between the upper and lower rolls, the leveling amount, and the thrust amount. The gap between the upper and lower rolls is the gap between the upper and lower rolls provided in the roughing mill 32, the intermediate mills 34 and 36, and the finishing mill 38, and these devices are structured so that the gap between the upper and lower rolls can be adjusted as desired. By adjusting the gap between the rolls, the thickness of the material being rolled, i.e., the reduction rate, can be changed.

[0027] The leveling amount is the amount of inclination of the upper and lower rolls provided in the roughing mill 32, the intermediate mills 34 and 36, and the finishing mill 38, and although rolling is normally performed while maintaining a horizontal state, the structure allows for rolling by inclining the upper and lower rolls or one of them from a horizontal state. By adjusting the leveling amount, the reduction ratios of the long sides 12 and the short sides 14 of the NAB 10 can be changed, respectively.

[0028] The thrust amount is the difference in the relative axial positions of the upper and lower rolls provided in the roughing mill 32, the intermediate mills 34 and 36, and the finishing mill 38, and these devices are structured so that the axial positions of the upper and lower rolls can be adjusted as desired. By adjusting this thrust amount, each dimension of the NAB 10 can be changed.

[0029] The rolling conditions of the gap between the upper and lower rolls, the leveling amount, and the thrust amount are adjusted so that each dimension of the NAB 10 falls within the range of the allowable values. In this way, when the rolling conditions are changed to adjust each dimension, the reduction ratio of the long side 12 and the short side 14 of the NAB 10 changes. That is, the longitudinal elongation ratio of the long side 12 and the short side 14 changes, and the amount of bending of the NAB 10 may increase. As such, the amount of bending of the NAB 10 is affected not only by temperature but also by the rolling conditions, so that the effect of the rolling conditions must be taken into consideration in order to accurately predict the amount of bending of the NAB 10. In particular, the finishing mill 38 has a particularly strong correlation with the amount of bending of the NAB 10 because it performs rolling last.

[0030] When the widths of the long sides 12 and short sides 14 of the NAB 10 are the same, different manufacturing sizes can be produced by combining the distance between the upper and lower rolls and the thrust and leveling amounts. Also, when continuously rolling NAB 10 of the same cross-sectional dimensions, the number of rolling passes in each rolling process of rough rolling, intermediate rolling, and finish rolling is basically not changed.

[0031] The cooling devices 42, 44 cool the NAB 10 under predetermined cooling conditions. The cooling devices 42, 44 are divided into predetermined lengths (e.g., 10 m), and each division is called a bank. The cooling devices 42, 44 shown in FIG. 5 each have six banks, but this is not limiting, and the number of banks may be increased or decreased from six. Cooling operation parameters that are the cooling conditions for the cooling devices 42, 44 include the water flow density of the cooling water and the cooling time. The cooling time can be calculated by dividing the cooling length, which is calculated by multiplying the length per bank by the number of banks, by the conveying speed.

[0032] Increasing the water flow density of the cooling water can increase the cooling speed and temperature drop of the NAB 10. Also, slowing down the conveying speed and lengthening the cooling time can increase the temperature drop of the NAB 10. By adjusting the water flow density and cooling time, which are cooling operation parameters of the cooling devices 42 and 44, the surface temperatures of the long sides 12 and short sides 14 of the NAB 10 after cooling can be controlled.

[0033] The cooling devices 42, 44 are capable of cooling the long sides 12 and the short sides 14 of the NAB 10 individually, and the density of the water sprayed onto the long sides 12 and the short sides 14 may be changed. Due to changes over time such as deterioration of the equipment and changes in temperature, the temperature of the long sides 12 or the short sides 14 may vary and not always be the same even for NABs 10 of the same cross-sectional dimensions depending on the manufacturing conditions, and bending may not always be prevented by the same cooling conditions. In this way, by changing the cooling conditions for the long sides 12 and the short sides 14, the temperatures of the long sides 12 and the short sides 14 of the NAB 10 can be adjusted individually.

[0034] Also, the cooling operation parameters of each cooling device, such as the cooling devices 42 and 44, installed before and after the finishing mill 38, such as the water flow rate density and the cooling time, may be changed independently. This is because the amount of bending of the NAB 10 is correlated with the temperature difference between the long side 12 and the short side 14 after finish rolling, and the mechanical properties of the steel material are correlated with the cooling stop temperature. Also, the cooling device 44 may be an accelerated cooling device.

[0035] A typical flat spray nozzle, square spray nozzle, or multi-hole jet nozzle can be used for the cooling devices 42, 44. In this case, if it is difficult to cool the entire flange with a single flat spray or square spray nozzle, multiple nozzles may be arranged in the width direction of the long side 12 and short side 14 for cooling.

[0036] Fig. 6 is a schematic diagram showing the state in which the long sides 12 and short sides 14 of the NAB 10 are being cooled by a cooling device 44. Fig. 6 shows the state in which the long sides 12 and short sides 14 of the NAB 10 are being cooled by a flat spray 45, which is an example of the cooling device 44. As shown in Fig. 6, it is preferable to cool the long sides 12 and short sides 14 of the NAB 10 from both the top and bottom, respectively.

[0037] FIG. 7 is a schematic diagram showing the vertical warpage of the NAB 10. When either the top or bottom of the long side 12 and short side 14 of the NAB 10 is cooled, a temperature difference occurs in the plate thickness direction, and the vertical warpage y of the NAB 10 may occur as shown in FIG. 7. In addition, the cooling water sprayed from the bottom surface falls from the bottom surfaces of the long side 12 and short side 14 of the NAB 10 in a short time due to gravity after contacting the long side 12 and short side 14 of the NAB 10. Therefore, even if the top and bottom surfaces are cooled with the same amount of water, the cooling capacity of the bottom surface is lower than that of the top surface. For this reason, it is preferable that the amount of cooling water on the bottom surface is 1.1 times or more and 1.5 times or less than the amount of cooling water on the top surface. This allows the top and bottom surfaces of the long side 12 and short side 14 of the NAB 10 to be uniformly cooled, and the vertical warpage y of the NAB 10 can be reduced.

[0038] Referring again to Figure 5, thermometers 46, 48, 50 are provided at the inlet and outlet of cooling devices 42, 44 to measure the surface temperatures of long sides 12 and short sides 14 of NAB 10 before and after cooling by cooling devices 42, 44. Thermometers 46, 48, 50 are devices that measure the temperatures of long sides 12 and short sides 14 of NAB 10 using scanning thermometers that scan temperature measurement points in the width direction of long sides 12 and short sides 14, or a method in which multiple spot thermometers are arranged in the width direction of long sides 12 and short sides 14. When the latter method is adopted, it is preferable to measure the temperatures by arranging spot thermometers at at least five or more different positions in the width direction of long sides 12 and short sides 14 in order to grasp the temperature distribution in the width direction of long sides 12 and short sides 14.

[0039] The amount of bending of NAB10 correlates with the temperature difference between the long side 12 and the short side 14 after finish rolling. According to the inventors' studies, the bending of NAB10 is particularly correlated with the temperature at the width center positions (A / 2 and B / 2 positions in FIG. 1) of each of the long side 12 and the short side 14. For this reason, it is preferable that the thermometer 48 measures at least the temperature at the width center positions of each of the long side 12 and the short side 14. Furthermore, the mechanical properties of NAB10 depend on the structure of the steel material. When the cooling device 44 is an accelerated cooling device, the cooling device plays a role in controlling the structure of the steel material, i.e., controlling the mechanical properties, so the thermometer 50 measures the surface temperatures of the long side 12 and the short side 14 on the exit side of the cooling device 44.

[0040] The data measured by the thermometers 46, 48, 50 are output to the process computer 56 as temperature information. The temperature information is, for example, temperature measurements at representative points at the widthwise center positions of the long sides 12 and short sides 14 (five points at equal intervals in the width direction from positions A / 2 and B / 2 to the leading and trailing ends). The data measured by the thermometers 46, 48, 50 may also be approximated by a polynomial of degree two or higher, and parameters that can specify the function of the approximated polynomial may be output as temperature information. The temperature information is used to confirm that the temperature of the NAB 10 is within a predetermined range as a result of controlling the cooling devices 42, 44.

[0041] The shape meter 54 measures the amount of bending of the NAB 10 after it has been cooled by the cooling device 44. The shape meter 54 may be any device capable of measuring the amount of bending of the NAB 10, and may, for example, use a two-dimensional laser displacement meter to measure the bending profile along the longitudinal direction of the NAB 10. In this case, if the shape meter 54 is installed below the NAB 10, it becomes difficult to perform stable measurements for a long period of time due to factors such as the scale falling off. For this reason, it is preferable to install the shape meter 54 above the NAB 10. When performing measurements using a two-dimensional laser displacement meter, it is preferable to set the measurement pitch in the longitudinal direction to a pitch of 1 to 50 mm in order to increase the measurement accuracy of the shape.

[0042] Alternatively, a plurality of one-dimensional laser displacement meters may be provided in the longitudinal direction of the NAB 10, and the amount of bending of the NAB 10 may be measured using these laser displacement meters. In this case, the shape of the NAB 10 at different positions in the longitudinal direction can be measured without scanning with the laser displacement meters as described above. In this case as well, it is preferable to provide the laser displacement meters at a pitch of 20 to 50 mm in the longitudinal direction of the short side 14 to improve measurement accuracy.

[0043] The shape data measured by the shape meter 54 may be used as the actual value of the amount of bending of the NAB 10 as it is, or may be approximated by a polynomial of degree 2 or higher, and the distance between a straight line connecting the lower ends of the short side 14 (or long side 12) at the leading and trailing ends of the NAB 10 and the lower end of the curved short side 14 (or long side 12) may be used as the actual value of the amount of bending of the NAB 10. Alternatively, a parameter that can specify a polynomial function that approximates the bending shape of the NAB 10 may be used as the actual value of the amount of bending of the NAB 10. In addition, the shape meter 54 does not necessarily have to be disposed on an extension line of the conveying direction of the NAB 10, and may be disposed above a cooling bed or the like so that the shape of the NAB 10 can be measured on the cooling bed after passing through the cooling device 44. The actual value of the amount of bending of the NAB 10 measured by the shape meter 54 is output to the process computer 56.

[0044] The process computer 56 may be a general-purpose computer such as a workstation or a personal computer. The process computer 56 is connected to the rolling control device 40, the cooling control device 52, the thermometers 46, 48, 50, and the shape meter 54 by wire or wirelessly, and controls the manufacturing process of the NAB 10. The process computer 56 also acquires attribute parameters of the NAB 10 from a higher-level computer. The attribute parameters include target dimensional values ​​of the NAB 10, information on the chemical composition (contents of C, Si, Mn, Cr, Mo, V, etc.), target values ​​of the mechanical properties of the NAB 10 (yield stress, tensile strength, elongation, toughness, hardness, etc.), and information on the steel type classification.

[0045] In addition to the attribute parameters, the process computer 56 stores information regarding the heating temperature of the NAB 10 and the cooling stop temperature at the outlet of the cooling devices 42, 44 required to obtain the desired material. The process computer 56 performs heat transfer calculations based on the internal model, and specifies the water flow rate and cooling time, which are cooling operation parameters of the cooling device, for achieving the cooling stop temperature.

[0046] The cooling operation parameters thus identified are sent to the cooling control device 52. Based on the cooling operation parameters, the cooling control device 52 controls the operating pressure and the number of operating cooling water pumps in the cooling devices 42 and 44, the opening degree of the flow rate valve of the cooling nozzle, and the rotation speed of the motor that drives the table roller.

[0047] In addition, the process computer 56 sets rolling operation parameters such as the number of passes and roll intervals of each rolling mill according to the attribute parameters of the NAB10. The set values ​​of normal rolling operation parameters are set as table values ​​associated with the attribute parameters of the NAB10 based on past rolling results. However, when the rolling rolls are rearranged, the same rolling results may not be obtained even if the preceding and succeeding rolling materials are rolled with the same parameters. For example, even if the same set values ​​are used before and after rearrangement of the rolling rolls, the reference state (e.g., the set state of the zero point) that is the basis of the set values ​​actually changes, so that the same cross-sectional dimensions may not be obtained if the rolling conditions are used as is. Even if the rolling rolls are not rearranged, the mill stretch behavior of the rolling mill and the friction characteristics between the housing and the roll chocks change due to the change in the spacing of the rolling rolls for the NAB10 with different cross-sectional dimensions, and nonlinear characteristics may occur in the relationship between the spacing of the rolling rolls and the rolling load. In such a case, the cross-sectional dimensions of the NAB10 may vary even if the spacing of the rolling rolls is the same. Therefore, in such a case, the operator may reset and change the values ​​in the table each time.

[0048] The process computer 56 sets the gap between the upper and lower rolls, the leveling amount, and the thrust amount for each pass in the roughing mill 32, the intermediate mills 34, 36, and the finishing mill 38 as rolling operation parameters for the rolling process. The process computer 56 outputs the set rolling operation parameters to the rolling control device 40. The rolling control device 40 controls the gap between the upper and lower rolls, the leveling amount, and the thrust amount for the roughing mill 32, the intermediate mills 34, 36, and the finishing mill 38 based on the rolling operation parameters.

[0049] The process computer 56 also collects and stores temperature information and bending information of the NAB 10 from the thermometers 46, 48, 50 and the shape meter 54. The rolling operation parameters, cooling operation parameters, temperature information, actual values ​​of the bending amount of the NAB 10, and actual values ​​of the attribute parameters are stored in a database of the process computer 56 in association with the identification number of the NAB 10 being manufactured.

[0050] The unequal leg angle steel shape prediction device 58 obtains the rolling operation parameters, cooling operation parameters, and attribute parameters of the NAB10 manufactured by the rolling facility 100 from the process computer 56, inputs input data including these into a shape prediction model, and predicts the amount of bending of the NAB10 by outputting the amount of bending of the NAB10. The unequal leg angle steel shape prediction device 58 also specifies cooling operation parameters that bring the predicted amount of bending of the NAB10 within the allowable range, and outputs the specified cooling operation parameters to the process computer 56 to set them as cooling operation parameters in the manufacture of the NAB10.

[0051] Next, a shape prediction device 58 for unequal-leg angle steel that predicts the amount of bending of the NAB10 will be described. FIG. 8 is a functional block diagram of the shape prediction device 58 for unequal-leg angle steel. The shape prediction device 58 for unequal-leg angle steel can be a general-purpose computer such as a workstation or a personal computer. The shape prediction device 58 for unequal-leg angle steel has a control unit 60, an input unit 62, an output unit 64, and a storage unit 66. The control unit 60 is, for example, a CPU, and functions as a data acquisition unit 68, a shape prediction unit 70, a cooling operation parameter identification unit 72, and a shape prediction model generation unit 74 by executing a program stored in the storage unit 66.

[0052] The input unit 62 is, for example, a keyboard, a touch panel provided integrally with a display, or the like. The output unit 64 is, for example, an LCD or CRT display, or the like. The storage unit 66 is, for example, an information recording medium such as an updatable flash memory, a hard disk built-in or connected via a data communication terminal, a memory card, or the like, and a read / write device for the information recording medium. The storage unit 66 stores programs and data for realizing each function of the shape prediction device 58 for unequal leg angle steel. The storage unit 66 further stores a database 76 and a shape prediction model 78. The database 76 stores 500 or more, more preferably 2000 or more, data sets each including rolling operation parameters, cooling operation parameters, attribute parameters, and actual values ​​of flange warpage of NAB10 manufactured in the past by the same rolling equipment 100 (the roll sets of the roughing mill 32, intermediate rolling mills 34, 36, and finishing rolling mill 38 are the same).

[0053] The shape prediction model 78 is a trained machine learning model that has been trained using the data set stored in the database 76 as training data. The shape prediction model 78 according to this embodiment is a trained machine learning model that receives input data including rolling operation parameters, cooling operation parameters, and attribute parameters as input, and outputs the amount of bending of the NAB 10.

[0054] Next, there will be described the processing executed by the data acquisition unit 68 and the shape prediction unit 70. The data acquisition unit 68 acquires rolling operation parameters, cooling operation parameters, and attribute parameters from the process computer 56 as input data.

[0055] The data acquisition unit 68 acquires, from the process computer 56, for example, the distance between the upper and lower rolls of the finishing mill 38 and the leveling amount as rolling operation parameters. The finishing mill 38 is the final rolling that finishes the product to the target dimensions, and the distance between the upper and lower rolls and the leveling amount of this rolling directly affect the cross-sectional shape of the product. For this reason, the distance between the upper and lower rolls of the finishing mill 38 and the leveling amount have a particularly strong correlation with the amount of bending of the NAB 10. Therefore, it is preferable to use the distance between the upper and lower rolls of the finishing mill 38 and the leveling amount as rolling operation parameters, which improves the prediction accuracy of the amount of bending of the NAB 10.

[0056] The data acquisition unit 68 may further acquire the thrust amounts of the upper and lower rolls of the finishing mill 38 from the process computer 56 as rolling operation parameters. As described above, the finishing mill 38 is the final rolling that finishes the product to the target dimensions, so the thrust amounts of the upper and lower rolls of this rolling also greatly affect the cross-sectional shape of the product. For this reason, the thrust amounts of the upper and lower rolls of the finishing mill 38 are also strongly correlated with the bending amount of the NAB 10. Therefore, it is preferable to use the thrust amounts of the upper and lower rolls of the finishing mill 38 as rolling operation parameters, which further improves the prediction accuracy of the bending amount of the NAB 10.

[0057] The data acquisition unit 68 may further acquire, as rolling operation parameters from the process computer 56, the spacing between the upper and lower rolls, the leveling amount, and the thrust amount in each pass of the roughing mill 32. The data acquisition unit 68 may further acquire, as rolling operation parameters from the process computer 56, the spacing between the upper and lower rolls, the leveling amount, and the thrust amount in each pass of the intermediate rolling mills 34, 36. Since these values ​​also affect the amount of bending of the NAB 10, the prediction accuracy of the amount of bending of the NAB 10 is further improved by including the spacing between the upper and lower rolls, the leveling amount, and the thrust amount in each pass of the roughing mill 32 and the intermediate rolling mills 34, 36 in the rolling operation parameters.

[0058] The data acquisition unit 68 acquires the water flow rate and cooling time of the cooling devices 42, 44 as cooling operation parameters from the process computer 56. In the cooling devices 42, 44, the higher the water flow rate and the longer the cooling time, the larger the temperature difference between the inner and outer surfaces of the long sides 12 and short sides 14 of the NAB 10 becomes, and the larger the amount of bending becomes. Therefore, by using the water flow rate and cooling time of the cooling devices 42, 44 as cooling operation parameters, the accuracy of predicting the amount of bending of the NAB 10 is further improved.

[0059] Furthermore, the data acquisition unit 68 acquires the target dimension value, chemical composition, and mechanical property target values ​​of NAB10 as attribute parameters from the process computer 56. By including the target dimension value, chemical composition, and mechanical property target values ​​of NAB10 in the input data of the shape prediction model in this way, the shape prediction model takes into account the target dimension value, chemical composition, and mechanical property, and the amount of bending can be predicted using the same shape prediction model for NAB10 with different target dimension values, chemical composition, and mechanical properties. The data acquisition unit 68 may acquire the steel type classification of NAB10 instead of the chemical composition of NAB10. The data acquisition unit 68 outputs the acquired input data to the shape prediction unit 70.

[0060] When the shape prediction unit 70 acquires input data from the data acquisition unit 68, it reads out the shape prediction model 78 from the storage unit 66, inputs the input data into the shape prediction model 78, and causes it to output the amount of bending of the NAB 10. In this manner, by causing the shape prediction model 78 to output the amount of bending of the NAB 10, the shape prediction unit 70 predicts the amount of bending of the NAB 10 to be manufactured. Note that the shape prediction unit 70 may output the output amount of bending to the output unit 64, and cause the output unit 64 to display the amount of bending of the NAB 10. This allows the operator to visually check the output unit 64 and confirm the predicted value of the amount of bending of the NAB 10.

[0061] Next, a description will be given of the processing of the cooling operation parameter specifying unit 72 and the shape prediction model generating unit 74. The cooling operation parameter specifying unit 72 specifies cooling operation parameters that bring the amount of bending of the NAB 10 predicted by the shape prediction unit 70 within the allowable range, and outputs the cooling operation parameters to the process computer 56 to set them as cooling conditions.

[0062] Fig. 9 is a flow diagram showing the flow of the cooling operation parameter specification process by the cooling operation parameter specification unit 72. The flow shown in Fig. 9 is started, for example, by receiving an input from an operator to start the process.

[0063] First, the data acquisition unit 68 acquires rolling operation parameters of the NAB 10 to be manufactured from the process computer 56 (step S101). The data acquisition unit 68 also acquires attribute parameters of the NAB 10 to be manufactured from the process computer 56 (step S102). The data acquisition unit 68 outputs the acquired rolling operation parameters and attribute parameters to the shape prediction unit 70.

[0064] The cooling operation parameter specifying unit 72 sets arbitrary cooling operation parameters (step S103). The cooling operation parameter specifying unit 72 outputs the set cooling operation parameters to the shape prediction unit 70. The shape prediction unit 70 reads out the shape prediction model 78 from the storage unit 66, and inputs the acquired rolling operation parameters, cooling operation parameters, and attribute parameters to the shape prediction model 78 to output the amount of bending of the NAB 10, thereby predicting the amount of bending of the NAB 10 to be manufactured (step S104). The shape prediction unit 70 outputs the predicted value of the amount of bending of the NAB 10 that has been output to the cooling operation parameter specifying unit 72.

[0065] The range of the allowable value of the amount of bending of the NAB10 may be determined in advance and stored in the storage unit 66, or may be input by the operator through the input unit 62. When the cooling operation parameter specifying unit 72 acquires the predicted value of the amount of bending of the NAB10 and the range of the allowable value of the amount of bending of the NAB10, it determines whether the predicted value of the amount of bending of the NAB10 is within the range of the allowable value of the amount of bending (step S105). When the cooling operation parameter specifying unit 72 determines that the predicted value of the amount of bending of the NAB10 is outside the range of the allowable value (step S105: No), it changes the cooling operation parameters according to a predetermined condition (step S106), returns the process to step S104, and repeats the processes of steps S104 to S106 until the predicted amount of bending of the NAB10 is within the range of the allowable value.

[0066] On the other hand, if it is determined in step S105 that the predicted value of the amount of bending of NAB 10 is within the allowable range (step S105: Yes), the cooling operation parameter specifying unit 72 specifies that the cooling operation parameters used to predict the amount of bending of NAB 10 are cooling operation parameters that can bring the amount of bending of NAB 10 within the allowable range (step S107), and the flow of the cooling operation parameter specifying process shown in Fig. 8 ends. The cooling operation parameter specifying unit 72 outputs the specified cooling operation parameters to the process computer 56, thereby setting the specified cooling operation parameters as the cooling operation parameters of the cooling devices 42, 44.

[0067] In this way, by specifying the cooling operation parameters, it is possible to specify the cooling operation parameters that can control the amount of bending of the NAB 10 to within the tolerance range. Then, by manufacturing the NAB 10 under manufacturing conditions that include the cooling operation parameters, it becomes possible to stably manufacture the NAB 10 whose amount of bending of the NAB 10 falls within the tolerance range, thereby suppressing an increase in the manufacturing cost of the NAB 10 and a decrease in productivity.

[0068] Next, a description will be given of a method for generating a shape prediction model used to predict the amount of bending of the NAB 10. The data acquisition unit 68 acquires from the process computer 56 the actual values ​​of the rolling operation parameters, the actual values ​​of the cooling operation parameters, the actual values ​​of the attribute parameters, and the actual values ​​of the amount of bending of the NAB 10 of previously manufactured NABs 10, and stores a data set consisting of these in the database 76 of the storage unit 66. The number of data sets stored in the database 76 is preferably at least 500, and more preferably 2000 or more.

[0069] The shape prediction model generating unit 74 reads out the pre-stored machine learning model from the storage unit 66, and uses the data set stored in the database 76 as training data to train the machine learning model, thereby generating a trained machine learning model. This trained machine learning model becomes the shape prediction model. The machine learning model used in the shape prediction method for unequal leg angle steel and the shape prediction device 58 for unequal leg angle steel according to the present embodiment may be any of commonly used neural networks, decision tree learning, random forests, and support vector regression. In addition, instead of the bending amount of the NAB10, the output of the shape prediction model may be a machine learning model that outputs data that binarizes the judgment result of whether the bending amount of the NAB10 is within the range of the allowable value into two values, pass / fail, and in this case, a classification model such as k-nearest neighbor method or logistic regression may be used.

[0070] The shape prediction model may be updated to a new shape prediction model by re-training the model through machine learning, for example, once a month or once a year. The data acquisition unit 68 acquires performance data every time the NAB 10 is manufactured, and stores the data in the database 76. Since performance data of newly manufactured NAB 10 is stored in the database 76, the amount of stored performance data increases. The more the amount of performance data increases, the more accurate the shape prediction becomes. For this reason, by periodically training the model through machine learning and updating the shape prediction model, the amount of bending of the NAB 10 can be predicted with even higher accuracy.

[0071] As described above, the shape prediction method of the unequal leg angle steel according to the present embodiment and the shape prediction model used in the shape prediction device 58 of the unequal leg angle steel include rolling operation parameters in the input data. As described above, the rolling operation parameters of the rolling process affect the bending amount of the unequal leg angle steel. Therefore, by including the rolling operation parameters in the input data of the shape prediction model predicting the bending amount of the unequal leg angle steel, the shape prediction model becomes a shape prediction model that takes the rolling process into consideration, and the prediction accuracy of the bending amount of the unequal leg angle steel increases. Furthermore, if the bending amount of the unequal leg angle steel can be predicted with high accuracy in this way, it is possible to prevent the production of unequal leg angle steels whose bending amount exceeds the allowable range, and it is possible to prevent an increase in the manufacturing cost of the unequal leg angle steels and a decrease in productivity.

[0072] Also, by including attribute parameters of unequal leg angle steel in the input data of the shape prediction model, it becomes possible to predict the amount of bending of unequal leg angle steel with different target cross-sectional dimensions, chemical components and / or mechanical properties using the same shape prediction model. Furthermore, by identifying cooling operation parameters that bring the amount of bending of unequal leg angle steel predicted by the shape prediction model within an allowable range, and manufacturing unequal leg angle steel under manufacturing conditions that include the cooling operation parameters, it becomes possible to stably manufacture unequal leg angle steel with an amount of bending within an allowable range.

[0073] In the rolling equipment 100 shown in FIG. 5, an example having the process computer 56 and the shape prediction device 58 of the unequal angle steel is shown, but the present invention is not limited thereto. For example, the process computer 56 may have the function of the shape prediction device 58 of the unequal angle steel, and these may be configured as one device. In addition, in the shape prediction device 58 of the unequal angle steel shown in FIG. 8, an example has been shown in which the control unit 60 has the cooling operation parameter specifying unit 72 and the shape prediction model generating unit 74, but the present invention is not limited thereto. If the bending amount of the NAB10 is predicted by the shape prediction device 58 of the unequal angle steel, the control unit 60 does not need to have the cooling operation parameter specifying unit 72. Furthermore, if the shape prediction model 78 is generated externally and the generated shape prediction model 78 is stored in the storage unit 66 via the data acquisition unit 68, the shape prediction device 58 of the unequal angle steel does not need to have the shape prediction model generating unit 74. EXAMPLES

[0074] [Example 1] In order to verify the effect of the present invention, an NAB having a material of SS400, a long side width of 200 mm, a long side thickness of 10 mm, a short side width of 90 mm, a short side thickness of 14 mm, and a length of 20 m was manufactured using the NAB manufacturing equipment 100 shown in Fig. 5. The steel slab as the raw material was rolled into an NAB with a desired cross section through five passes in rough rolling, a total of six passes in intermediate rolling, and one pass in finish rolling.

[0075] The cooling devices 42 and 44 before and after the finishing mill 38 were composed of flat spray nozzles as shown in Fig. 6, and cooled the long and short sides of the NAB from the top and bottom. The spray widths of the flat spray nozzles on the long and short sides were 170 mm and 70 mm, respectively, and the nozzles were arranged at a pitch of 200 mm in the conveying direction. The total length of the cooling devices 42 and 44 was 18 m, and the devices were divided into a total of 6 banks, each 3 m apart. The thermometer 46 was installed 5 m upstream from the inlet of the cooling device 42, the thermometer 48 was installed 5 m downstream from the finishing mill 38, the thermometer 50 was installed 20 m downstream from the outlet of the cooling device 44, and the shape meter 54 was installed 25 m downstream from the outlet of the cooling device 44. The temperature and bending amount of the NAB manufactured were measured using these thermometers 46, 48, and 50 and the shape meter 54. The temperatures measured by thermometers 46, 48, 50 were used to verify that the temperature of the NABs cooled by cooling devices 42, 44 was within a predetermined range.

[0076] <Example 1> In Example 1 of the invention, a data set of about 5,000 items was prepared as learning data, each set consisting of the most recent past NAB rolling operation parameters, cooling operation parameters, attribute parameters, and actual values ​​of the NAB bending amount, and a shape prediction model was used in which a neural network model was machine-trained using the data set.

[0077] The rolling operation parameters used were the gap between the upper and lower rolls and the leveling amount in each pass of the roughing mill 32, the intermediate mills 34 and 36, and the finishing mill 38. The cooling operation parameters used were the water flow density of the cooling water in each of the cooling devices 42 and 44 and the cooling time.

[0078] In Example 1 of the present invention, calculations were performed repeatedly using the shape prediction model according to the flowchart shown in Figure 9, and cooling operation parameters were identified that satisfied the tolerance range of the predicted value of the NAB bending amount of less than ±0.003L mm (less than 0.3% of the longitudinal length L of the NAB, 60 mm in this case).

[0079] The cooling operation parameter to be reset is the cooling time, and the cooling operation parameter finally specified is to set the water flow density of the cooling water on the long side of the cooling device 42 to 1000 L / m 2 When the cooling time was 1.8 s and the water flow density of the cooling water on the short side was 0, the cooling time was 1.8 s and the water flow density of the cooling water on the long side in the cooling device 44 was 1200 L / m 2 min, the cooling time is 4.4 s, and the water density of the cooling water on the short side is 1200 L / m 2 The cooling time was 2.9 s when the temperature of the long side measured by the thermometers 46, 48, and 50 was 901°C, 825°C, and 701°C, respectively, and the temperature of the short side measured by the thermometers 46, 48, and 50 was 932°C, 827°C, and 680°C, respectively.

[0080] When a NAB with material SS400, long side width 200 mm, long side thickness 10 mm, short side width 90 mm, short side thickness 14 mm, and length 20 m was manufactured under manufacturing conditions including the cooling operation parameters specified as above, the predicted amount of bending by machine learning was +12 mm, whereas the actual amount of bending measured by shape meter 54 was +17 mm. From these results, it was confirmed that the amount of bending of the NAB can be predicted with high accuracy by using a shape prediction model including rolling operation parameters, and that by manufacturing a NAB with cooling operation parameters specified using the predicted results, a NAB with an amount of bending of less than ±0.003L mm, which is within the allowable range, can be manufactured.

[0081] <Comparative Example 1-1> In Comparative Example 1-1, a NAB was manufactured with the same material as in Invention Example 1, namely, SS400, long side width 200 mm, long side thickness 10 mm, short side width 90 mm, short side thickness 14 mm, and length 20 m. The manufacturing equipment was the same as in Invention Example 1.

[0082] In Comparative Example 1-1, a data set of about 5000 items, each of which is a set of the most recent past NAB cooling operation parameters and the actual value of the NAB bending amount, was prepared as learning data, and a shape prediction model was used in which a neural network model was machine-learned using the data set. That is, in Comparative Example 1-1, a shape prediction model in which the input data does not include rolling operation parameters was used.

[0083] In Comparative Example 1-1, calculations were performed repeatedly using the shape prediction model according to the flowchart shown in Fig. 9, and cooling operation parameters were identified that satisfied the predicted value of the NAB bending amount of less than ±0.003 L mm. The cooling operation parameters finally identified were: water flow density of the cooling water on the long side in the cooling device 42 of 1000 L / m 2 When the cooling time was 1.5 s and the water flow density of the cooling water on the short side was 0, the water flow density of the cooling water on the long side in the cooling device 44 was 1200 L / m 2 min, the cooling time is 4.4 s, and the water density of the cooling water on the short side is 1200 L / m 2 The cooling time was 2.9 s when the temperature of the long side measured by thermometers 46, 48, and 50 was 902°C, 835°C, and 711°C, respectively, and the temperature of the short side measured by thermometers 46, 48, and 50 was 933°C, 865°C, and 701°C, respectively.

[0084] When a NAB with material SS400, long side width 200 mm, long side thickness 10 mm, short side width 90 mm, short side thickness 14 mm, and length 20 m was manufactured under manufacturing conditions including the cooling operation parameters identified as described above, the predicted amount of bending by machine learning was +45 mm, whereas the actual amount of bending measured by shape meter 54 was +62 mm. From these results, it was confirmed that when a shape prediction model that does not include rolling operation parameters is used, the amount of bending of the NAB cannot be predicted with high accuracy, and that even if a NAB is manufactured using cooling operation parameters identified using the prediction results, it is not possible to manufacture a NAB with an amount of bending of less than ±0.003L mm, which is within the allowable range.

[0085] <Comparative Example 1-2> In Comparative Example 1-2, a NAB was manufactured with the same material as in Invention Example 1, namely, SS400, long side width 200 mm, long side thickness 10 mm, short side width 90 mm, short side thickness 14 mm, and length 20 m. The manufacturing equipment was the same as in Invention Example 1.

[0086] In Comparative Example 1-2, the amount of NAB bending was predicted using an equation expressed as a linear sum of each parameter of the rolling conditions disclosed in Patent Document 2. Moreover, unlike Invention Example 1 and Comparative Example 1-1, the cooling conditions did not involve repeated prediction of the amount of bending, but instead used cooling operation parameters set based on the results of heat transfer calculations previously performed offline using the finite element method. The cooling operation parameters used in Comparative Example 1-2 were a water flow density of the cooling water on the long side in the cooling device 42 of 1000 L / m 2 When the cooling time was 1.5 s and the water flow density of the cooling water on the short side was 0, the water flow density of the cooling water on the long side in the cooling device 44 was 1200 L / m 2 min, the cooling time is 4.4 s, and the water density of the cooling water on the short side is 1200 L / m 2 The cooling time was 2.9 s when the temperature of the long side measured by thermometers 46, 48, and 50 was 901°C, 833°C, and 713°C, respectively, and the temperature of the short side measured by thermometers 46, 48, and 50 was 931°C, 867°C, and 699°C, respectively.

[0087] When an NAB with material SS400, long side width 200 mm, long side thickness 10 mm, short side width 90 mm, short side thickness 14 mm, and length 20 m was manufactured under manufacturing conditions including these cooling operation parameters, the predicted amount of bending was +50 mm, whereas the actual amount of bending measured by shape meter 54 was +65 mm. From this result, it was confirmed that the amount of bending of the NAB cannot be predicted with high accuracy even using the equation expressed as a linear sum of each parameter of the rolling conditions disclosed in Patent Document 2, and that it is not possible to manufacture an NAB with an amount of bending of less than ±0.003L mm, which is within the allowable range.

[0088] As described above, in Example 1 of the invention, by using a shape prediction model in which rolling operation parameters and cooling operation parameters are input data and the amount of NAB bending is output data to predict the amount of NAB bending after passing through a cooling device, it is possible to predict the amount of NAB bending with higher accuracy than in Comparative Examples 1-1 and 1-2.It was confirmed that, further, by manufacturing an NAB with cooling operation parameters specified using the prediction results, it is possible to manufacture an NAB with an amount of bending within the allowable range.

[0089] [Example 2] In Example 2, an NAB having a long side width of 200 mm, a long side thickness of 10 mm, a short side width of 90 mm, a short side thickness of 14 mm, and a length of 20 m was manufactured using the same rolling equipment as in Example 1. The differences between Example 1 and Example 2 are the dimensions and steel type of the NAB to be manufactured, and the spray widths of the flat spray nozzles for cooling the long sides and short sides in the cooling devices 42 and 44, which were 370 mm and 75 mm, respectively.

[0090] <Example 2> In Example 2 of the invention, a data set of about 5,000 items was prepared as learning data, each set consisting of the most recent past NAB rolling operation parameters, cooling operation parameters, attribute parameters, and actual values ​​of the NAB bending amount, and a shape prediction model was used in which a neural network model was machine-learned using the data set.

[0091] The rolling operation parameters used were the gap between the upper and lower rolls and the leveling amount in each pass of the roughing mill 32, the intermediate mills 34 and 36, and the finishing mill 38. The cooling operation parameters used were the water flow density of the cooling water in each of the cooling devices 42 and 44 and the cooling time.

[0092] In Example 2 of the present invention, calculations were performed repeatedly using the shape prediction model according to the flowchart shown in Figure 9, and cooling operation parameters were identified that satisfied the tolerance range of the predicted value of the NAB bending amount of less than ±0.003L mm (less than 0.3% of the longitudinal length L of the NAB, 45 mm in this case).

[0093] The cooling operation parameter to be reset is the cooling time, and the cooling operation parameter finally specified is to set the water flow density of the cooling water on the long side of the cooling device 42 to 1000 L / m 2 When the cooling time was 2.3 s and the water flow density of the cooling water on the short side was 0, the water flow density of the cooling water on the long side in the cooling device 44 was 1200 L / m 2 min, the cooling time is 9.0 s, and the water density of the cooling water on the short side is 1200 L / m 2 The cooling time was 6.0 s when the temperature of the long side measured by the thermometers 46, 48, and 50 was 920°C, 849°C, and 650°C, respectively, and the temperature of the short side measured by the thermometers 46, 48, and 50 was 944°C, 850°C, and 615°C, respectively.

[0094] When a NAB with material SM490A, long side width 400 mm, long side thickness 13 mm, short side width 100 mm, short side thickness 16 mm, and length 15 m was manufactured under manufacturing conditions including the cooling operation parameters specified as described above, the predicted amount of bending by machine learning was +9 mm, whereas the actual amount of bending measured by shape meter 54 was +12 mm. From these results, it was confirmed that the amount of bending of the NAB can be predicted with high accuracy by using a shape prediction model including rolling operation parameters, and that by manufacturing a NAB with cooling operation parameters specified using the predicted results, a NAB with an amount of bending of less than ±0.003L mm, which is within the allowable range, can be manufactured.

[0095] <Comparative Example 2-1> In Comparative Example 2-1, a NAB was manufactured with the material SM490A, long side width 400 mm, long side thickness 13 mm, short side width 100 mm, short side thickness 16 mm, and length 15 m, as in Invention Example 2. The manufacturing equipment was the same as that of Invention Example 1.

[0096] In Comparative Example 2-1, a data set of about 5000 items, each of which is a set of the most recent past NAB cooling operation parameters and the actual value of the NAB bending amount, was prepared as learning data, and a shape prediction model was used in which a neural network model was machine-learned using the data set. That is, in Comparative Example 2-1, a shape prediction model in which the input data does not include rolling operation parameters was used.

[0097] In Comparative Example 2-1, calculations were performed repeatedly using the shape prediction model according to the flowchart shown in Fig. 9, and cooling operation parameters were identified that satisfied the predicted value of the NAB bending amount of less than ±0.003 L mm. The cooling operation parameters finally identified were: water flow density of the cooling water on the long side in the cooling device 42 of 1000 L / m 2 When the cooling time was 1.5 s and the water flow density of the cooling water on the short side was 0, the water flow density of the cooling water on the long side in the cooling device 44 was 1200 L / m 2 min, the cooling time is 9.0 s, and the water density of the cooling water on the short side is 1200 L / m 2 The cooling time was 6.0 s when the temperature of the long side measured by the thermometers 46, 48, and 50 was 921°C, 860°C, and 660°C, respectively, and the temperature of the short side measured by the thermometers 46, 48, and 50 was 931°C, 880°C, and 640°C, respectively.

[0098] When a NAB with material SM490A, long side width 400 mm, long side thickness 13 mm, short side width 100 mm, short side thickness 16 mm, and length 15 m was manufactured under manufacturing conditions including the cooling operation parameters identified as described above, the predicted amount of bending by machine learning was +36 mm, whereas the actual amount of bending measured by shape meter 54 was +47 mm. From these results, it was confirmed that the use of a shape prediction model that does not include rolling operation parameters makes it impossible to predict the amount of bending of the NAB with high accuracy, and that even if a NAB is manufactured using cooling operation parameters identified using the prediction results, it is not possible to manufacture a NAB with an amount of bending of less than ±0.003L mm, which is within the allowable range.

[0099] <Comparative Example 2-2> In Comparative Example 2-2, a NAB was manufactured with the material SM490A, long side width 400 mm, long side thickness 13 mm, short side width 100 mm, short side thickness 16 mm, and length 15 m, as in Invention Example 2. The manufacturing equipment was the same as that of Invention Example 1.

[0100] In Comparative Example 2-2, the amount of NAB bending was predicted using an equation expressed as a linear sum of each parameter of the rolling conditions disclosed in Patent Document 2. Moreover, unlike Invention Example 2 and Comparative Example 2-1, the cooling conditions did not involve repeated prediction of the amount of bending, but instead used cooling operation parameters set based on the results of heat transfer calculations performed in advance offline using the finite element method. The cooling operation parameters used in Comparative Example 2-2 were a water flow density of the cooling water on the long side in the cooling device 42 of 1000 L / m 2 When the cooling time was 1.5 s and the water flow density of the cooling water on the short side was 0, the water flow density of the cooling water on the long side in the cooling device 44 was 1200 L / m 2 min, the cooling time is 9.0 s, and the water density of the cooling water on the short side is 1200 L / m 2 The cooling time was 6.0 s when the temperature of the long side measured by thermometers 46, 48, and 50 was 920°C, 861°C, and 662°C, respectively, and the temperature of the short side measured by thermometers 46, 48, and 50 was 930°C, 883°C, and 635°C, respectively.

[0101] When an NAB was manufactured under manufacturing conditions including these cooling operation parameters, using material SM490A, with a long side width of 400 mm, long side thickness of 13 mm, short side width of 100 mm, short side thickness of 16 mm, and length of 15 m, the predicted amount of bending was +30 mm, whereas the actual amount of bending measured by shape meter 54 was +49 mm. From these results, it was confirmed that the amount of bending of the NAB cannot be predicted with high accuracy even using the equation expressed as a linear sum of each parameter of the rolling conditions disclosed in Patent Document 2, and that it is not possible to manufacture an NAB with an amount of bending of less than ±0.003L mm, which is within the allowable range.

[0102] As described above, in Example 2 of the invention, by using a shape prediction model in which rolling operation parameters and cooling operation parameters are input data and the amount of NAB bending is output data, the amount of NAB bending can be predicted with higher accuracy than in Comparative Examples 2-1 and 2-2.Furthermore, it was confirmed that by manufacturing an NAB with cooling operation parameters specified using the prediction results, an NAB with an amount of bending within the allowable range can be manufactured. [Explanation of symbols]

[0103] 10 NAB 12 Long side 14 Short Side 16 ABS 17 Short Side 18 Long Side 20 BP 21 Short Side 30 Furnace 32 Roughing mill 34 Intermediate rolling mill 36 Intermediate rolling mill 38 Finishing Rolling Mill 40 Rolling control device 42 Cooling device 44 Cooling device 45 Flat Spray 46 Thermometer 48 Thermometer 50 thermometer 52 Cooling control device 54 Shape meter 56 Process Computer 58 Shape prediction device for unequal leg angle steel 60 Control section 62 Input section 64 Output section 66 Storage area 68 Data Acquisition Section 70 Shape Prediction Section 72 Cooling operation parameter specification section 74 Shape prediction model generation unit 76 Databases 78 Shape Prediction Model 100 Rolling equipment

Claims

1. A method for predicting the amount of bending of unequal angle steel produced by a rolling facility having a plurality of rolling mills for hot rolling heated steel billets into unequal angle steel, a cooling device for cooling the unequal angle steel, and a shape meter for measuring the amount of bending of the unequal angle steel after cooling, comprising: A shape prediction method for unequal arm angle steel, comprising inputting input data including rolling operation parameters of the rolling mill, cooling operation parameters of the cooling device, and attribute parameters of the unequal arm angle steel into a shape prediction model, and outputting the amount of bending to predict the amount of bending.

2. Identifying the cooling operation parameters such that the amount of bending predicted using the method for predicting the shape of an unequal leg angle steel according to claim 1 falls within a range of a predetermined allowable value; A method for manufacturing unequal angle steel, which manufactures unequal angle steel under manufacturing conditions including specified cooling operation parameters.

3. A rolling facility comprising a plurality of rolling mills for hot rolling heated steel billets into unequal angle steel, a cooling device for cooling the unequal angle steel, a shape meter for measuring the amount of bending of the unequal angle steel after cooling, and a shape prediction device for the unequal angle steel for predicting the amount of bending, The shape prediction device is a rolling facility having a shape prediction unit that inputs input data including rolling operation parameters of the rolling mill, cooling operation parameters of the cooling device, and attribute parameters of the unequal leg angle steel into a shape prediction model, and outputs the amount of bending.

4. A method for generating a shape prediction model for predicting the amount of bending of unequal angle steel produced by a rolling facility having a plurality of rolling mills for hot rolling heated steel billets into unequal angle steel, a cooling device for cooling the unequal angle steel, and a shape meter for measuring the amount of bending of the unequal angle steel after cooling, comprising: A machine learning model is trained using a plurality of data sets, each set being a set of actual values ​​of rolling operation parameters of the rolling mill, actual values ​​of cooling operation parameters of the cooling device, actual values ​​of attribute parameters of the unequal leg angle steel, and actual values ​​of the amount of bending of the unequal leg angle steel, as training data; A method for generating a shape prediction model, the method comprising: inputting the rolling operation parameters, the cooling operation parameters, and the attribute parameters; and generating a shape prediction model having the amount of bending as an output.

5. The method for generating a shape prediction model according to claim 4 , wherein the machine learning model is any one of a neural network, a decision tree learning, a random forest, and a support vector regression.

Images