Width prediction method for rough rolled material, width control method for rough rolled material, method for manufacturing hot-rolled steel sheet, and method for generating width prediction model for rough rolled material

Abstract

This method for predicting the width of a rough rolled material is a method for predicting the width of a rough rolled material in a hot rolling line, the hot rolling line comprising: a heating furnace for heating a slab; a roughing mill for roughing the heated slab to produce a roughed material; and a finishing mill for manufacturing a finish-rolled material by finishing-rolling the rough-rolled material, the method comprising: a prediction step for predicting statistical information on the width of the rough-rolled material using a width prediction model learned by a Gaussian process regression method, and predicting the statistical information on the width of the rough-rolled material; the width prediction model includes, as input data, one or more operating parameters selected from among the operating parameters of the roughing mill, and statistical information on the width of the roughing stock is set as output data.

Description

Technical Field

[0001] The present invention relates to a method for predicting the width of a rough-rolled piece in a hot rolling line, a method for controlling the width of a rough-rolled piece, a method for manufacturing a hot-rolled steel sheet, and a method for generating a width prediction model of a rough-rolled piece. Background Art

[0002] In a hot rolling line, first, a slab, which is a steel sheet material, is heated in a heating furnace, and the width of the slab is adjusted by a width reduction pressing device (sizing press). Then, the slab is roughly rolled by one or more than two rough rolling mills to manufacture a semi-finished steel sheet called a rough-rolled slab with a thickness of about 30 mm to 50 mm (hereinafter referred to as a rough-rolled piece). Next, after cutting the front and rear ends of the rough-rolled piece by a crop shear, the rough-rolled piece is finish-rolled by a finishing mill composed of 5 to 7 rolling mill stands capable of continuous rolling to manufacture a steel sheet with a thickness of about 1.0 mm to 25.0 mm (hereinafter referred to as a finish-rolled piece). And finally, the finish-rolled piece in a high-temperature state is cooled by a cooling device on an exit roller table and then wound by a coiler to become a hot-rolled steel sheet. In the hot rolling line, since plastic deformation is applied to the steel sheet in the thickness direction and the width direction in the width reduction pressing device, the rough rolling mill, and the finishing mill, the width of the steel sheet varies complexly in the manufacturing process of the hot-rolled steel sheet. On the other hand, the width accuracy of the hot-rolled steel sheet directly affects the product yield. Therefore, in the hot rolling line, the width of the rough-rolled piece at the stage before the rough rolling is completed and loaded into the finishing mill is controlled (rough rolling width control), and the width of the steel sheet is controlled during the process of passing through the finishing mill (finish rolling width control).

[0003] In the hot rolling line, since the width of the steel sheet varies for various reasons, various techniques for improving the width accuracy of the hot-rolled steel sheet have been proposed. For example, Patent Document 1 discloses the following method: a prediction model composed of measured values of the width and temperature of a slab rolled by a rough rolling mill is used to predict the width change amount of the slab before and after rolling, and based on the predicted width change amount, the opening of an edger is set. In addition, Patent Document 1 discloses that for the opening setting of the edger, the opening setting based on a motor and the opening adjustment based on oil pressure are combined to improve the setting accuracy of width reduction, and the parameters of the prediction model are adaptively corrected online based on the actual value of the width of the slab.

[0004] In addition, Patent Document 2 discloses the following method: the width of the rough-rolled slab is controlled by an edge rolling mill, and the width of the finished rolled piece is controlled to a target value, and the width of the finished rolled piece is estimated using a width prediction model, which shows the relationship between the performance data representing the width of the slab and the rolling process and the width of the finished rolled piece. In addition, Patent Document 2 describes the following method: for a rolled steel material (rolled material) that has just been rolled, the deviation between the estimated value and the measured value of the width of the rolled steel material after the finish rolling is stored, and the target value of the width of the slab after the finish rolling of the next rolling is modified. According to Patent Document 2, the width of the rolled steel material can be controlled with high precision according to the deviation of the width of the slab after continuous casting.

[0005] In addition, Patent Document 3 describes the following method: With a rough rolling mill of a hot rolling line as the object, it is provided with: a learning unit that learns the information of the rolled material on the input side of the edger, the roller diameter information of the edger and the horizontal rolling mill, and the relationship between the actual values of the width after rolling by the edger and the thickness after horizontal rolling and the actual value of the width after horizontal rolling; and a prediction unit that calculates the predicted value of the width of the rolled material after horizontal rolling according to the knowledge obtained by the learning unit, thereby setting the opening of the edger so that the difference between the predicted value of the width after horizontal rolling predicted by the prediction unit and the target value is zero. In addition, Patent Document 3 describes that the learning unit can be composed of a neural network.

[0006] Prior art literature

[0007] Patent Literature

[0008] Patent Document 1: Japanese Patent Application Laid-Open No. 7-303909

[0009] Patent Document 2: Japanese Patent Application Publication No. 2010-64103

[0010] Patent Document 3: Japanese Patent Application Laid-Open No. 9-225513

[0011] Non-patent literature

[0012] Non-patent literature 1: “Gaussian Processes and Machine Learning”, Daichi Mochihashi and Shigemasa Ohba, published on March 7, 2019, ISBN 978-4-06-152926-7, Kodansha Summary of the invention

[0013] Problems to be solved by the invention

[0014] However, the method described in Patent Document 1 uses a physical model as a prediction model, which predicts the width expansion behavior of slabs produced by an edger mill and a horizontal rolling mill. Therefore, a representative value of the slab width is calculated as the predicted value of the slab width, and the deviation of the slab width cannot be predicted. In addition, although Patent Document 1 describes that the width of the slab will deviate depending on the difference between the set opening and the actual value of the edger mill, the deviation of the slab width caused by factors other than the estimation accuracy of the set opening of the edger mill is not considered. Therefore, it is impossible to avoid the situation where the width of the rough-rolled piece is too small or too large due to deviations such as the temperature of the rough-rolled piece.

[0015] On the other hand, Patent Document 2 describes using a physical model representing the width contraction amount of the edger mill and the width widening amount of the horizontal rolling mill to estimate the width change of the slab in the roughing mill. However, the physical model is used to calculate the representative value of the slab width, and the deviation of the slab width cannot be predicted. In addition, Patent Document 2 describes a method of modifying the target value of the width after rolling of the slab for the next rolling based on the deviation between the estimated value and the measured value of the width of the rolled steel after finish rolling. However, the factors causing the width deviation of the rough-rolled piece are not only the width deviation of the slab after continuous casting, but also other factors. Therefore, the width deviation of the rough-rolled piece cannot be eliminated, and it is impossible to avoid the situation where the width of the rough-rolled piece is too small or too large.

[0016] In addition, Patent Document 3 describes predicting the width of the rough-rolled piece based on the operation performance data of the edger mill and the horizontal rolling mill through a learning unit such as a neural network. However, the prediction unit for the width of the rough-rolled piece is used to calculate the representative value of the width of the rough-rolled piece, and the deviation of the width of the rough-rolled piece cannot be predicted. Therefore, it is impossible to avoid the situation where the width of the rough-rolled piece is too small or too large due to deviations such as the temperature of the rough-rolled piece.

[0017] As described above, in the existing roughing width control, a method of improving the width accuracy of the rough-rolled piece by suppressing the deviation has been adopted by focusing on a specific phenomenon that causes the width deviation of the rough-rolled piece. However, for example, deviations such as the temperature of the slab and the deformation resistance also cause the width deviation of the rough-rolled piece. Therefore, it is difficult to completely eliminate the width deviation of the rough-rolled piece. As a result, there are situations where the width of the rough-rolled piece is too small or too large, and it is impossible to avoid the situation of poor width accuracy of the rough-rolled piece and a decrease in the product yield.

[0018] The present invention aims to solve the above problems, and an object thereof is to provide a method for predicting the width of a rough rolled piece, which can predict statistical information including the width deviation of the rough rolled piece. Another object of the present invention is to provide a method for controlling the width of a rough rolled piece, which takes into account the width deviation of the rough rolled piece and can accurately control the width of the rough rolled piece in the length direction. Still another object of the present invention is to provide a method for manufacturing a hot-rolled steel sheet, which can improve the yield rate of the hot-rolled steel sheet product. Yet another object of the present invention is to provide a method for generating a width prediction model of a rough rolled piece, which can generate a width prediction model for predicting statistical information including the deviation of the width of the rough rolled piece.

[0019] Means for solving the problems

[0020] The method for predicting the width of a rough rolled piece according to the present invention predicts the width of the rough rolled piece in a hot rolling line, the hot rolling line including: a heating furnace for heating a slab; a rough rolling mill for rough rolling the heated slab to manufacture the rough rolled piece; and a finish rolling mill for finish rolling the rough rolled piece to manufacture a finish rolled piece, and includes: a prediction step of predicting statistical information of the width of the rough rolled piece using a width prediction model learned by a Gaussian process regression method, the width prediction model including one or more operation parameters selected from the operation parameters of the rough rolling mill as input data and setting the statistical information of the width of the rough rolled piece as output data.

[0021] Alternatively, the hot rolling line includes a width adjusting stamping device disposed upstream of the rough rolling mill, the width adjusting stamping device intermittently adjusting the width of the slab heated by the heating furnace, and the width prediction model includes one or more operation parameters selected from the operation parameters of the width adjusting stamping device as input data.

[0022] Alternatively, the width prediction model includes one or more operation parameters selected from the operation parameters of the heating furnace as input data.

[0023] Alternatively, the width prediction model includes one or more parameters selected from the attribute information of the slab as input data

[0024] The method for controlling the width of a rough rolled piece according to the present invention includes: a resetting step of predicting statistical information of the width of the rough rolled piece using the method for predicting the width of the rough rolled piece according to the present invention and resetting one or more operation parameters selected from the operation parameters of the rough rolling mill in such a manner that the probability that the width of the rough rolled piece is lower than the target width of the rough rolled piece becomes smaller based on the predicted statistical information.

[0025] Alternatively, the statistical information of the width of the rough rolled piece includes the average value W of the width of the rough rolled piece mand the standard deviation W σ , the resetting step includes: using the target width W of the rough-rolled piece t to set one or more operation parameters selected from the operation parameters of the roughing mill in such a way that the following relationship shown in Equation (1) is satisfied.

[0026] [Equation 1]

[0027]

[0028] The method for manufacturing a hot-rolled steel sheet according to the present invention includes: a step of manufacturing a hot-rolled steel sheet using the method for controlling the width of a rough-rolled piece according to the present invention.

[0029] The method for generating a width prediction model of a rough-rolled piece according to the present invention is a method for generating a width prediction model of a rough-rolled piece for predicting the width of a rough-rolled piece in a hot-rolling line. The hot-rolling line includes: a heating furnace for heating a slab; a roughing mill for rough-rolling the heated slab to manufacture the rough-rolled piece; and a finishing mill for finishing the rough-rolled piece to manufacture a finished rolled piece. The method for generating a width prediction model of a rough-rolled piece includes: a learning data acquisition step of acquiring a plurality of learning data, the plurality of learning data including: one or more operation performance data selected from the operation performance data of the roughing mill, and the performance data of the width of the rough-rolled piece; and a step of generating the width prediction model using the plurality of learning data acquired in the learning data acquisition step by using a Gaussian process regression method. The width prediction model includes one or more operation performance data selected from the operation performance data of the roughing mill as input performance data, and sets the statistical information of the width of the rough-rolled piece as output data.

[0030] Advantages of the Invention

[0031] According to the method for predicting the width of a rough-rolled piece according to the present invention, it is possible to predict the statistical information including the width deviation of the rough-rolled piece. In addition, according to the method for controlling the width of a rough-rolled piece according to the present invention, considering the width deviation of the rough-rolled piece, it is possible to accurately control the width of the rough-rolled piece in the length direction. In addition, according to the method for manufacturing a hot-rolled steel sheet according to the present invention, it is possible to improve the product yield of the hot-rolled steel sheet. In addition, according to the method for generating a width prediction model of a rough-rolled piece according to the present invention, it is possible to generate a width prediction model for predicting the statistical information including the width deviation of the rough-rolled piece. Description of the Drawings

[0032] Figure 1 is a schematic diagram showing a structural example of a hot-rolling line to which the present invention is applied.

[0033] Figure 2 is shown as Figure 1Schematic diagram of the structural example of the heating furnace shown.

[0034] Figure 3 It is a schematic diagram showing the interior of the heating furnace as viewed from above. Figure 1 Schematic diagram of the interior of the heating furnace shown.

[0035] Figure 4 It is a schematic diagram showing Figure 1 Schematic diagram of the structural example of the width-adjusting stamping device shown.

[0036] Figure 5 It is a diagram showing the structure of the rolling mill housing that constitutes Figure 1 the rough rolling mill shown.

[0037] Figure 6 It is a schematic diagram for explaining the optical type width measurement method.

[0038] Figure 7 It is a flowchart showing the process of the model generation step according to an embodiment of the present invention.

[0039] Figure 8 It is a flowchart showing the process of the prediction step according to an embodiment of the present invention.

[0040] Figure 9 It is a block diagram showing the structure of the width prediction model generation unit according to an embodiment of the present invention.

[0041] Figure 10 It is a diagram for explaining the method of setting the control target width in the conventional width control method.

[0042] Figure 11 It is a block diagram showing the structure of the width prediction unit according to an embodiment of the present invention.

[0043] Figure 12 It is a diagram showing the relationship between the probability density distribution of the width on the rough rolling output side predicted by the width prediction model and the target width.

[0044] Figure 13 It is a diagram for explaining the method of setting the control target value of the rough rolling width control in the present embodiment.

[0045] Figure 14 It is a diagram showing the relationship between the probability density distribution of the width on the rough rolling output side predicted by the width prediction model and the target width.

[0046] Figure 15 It is a diagram showing the relationship between the deviation of the width on the rough rolling output side from the target width and the cut amount of the steel plate. Detailed implementation mode

[0047] Next, with reference to the accompanying drawings, a method for predicting the width of a rough-rolled product, a method for controlling the width of a rough-rolled product, a method for manufacturing a hot-rolled steel sheet, and a method for generating a width prediction model for a rough-rolled product according to an embodiment of the present invention will be described in detail.

[0048] 〔Hot rolling line〕

[0049] First, with reference to Figures 1 to 6 , the structure of a hot rolling line to which the present invention is applied will be described.

[0050] Figure 1 is a schematic diagram showing an example of the structure of a hot rolling line to which the present invention is applied. As Figure 1 shown, the hot rolling line 1 to which the present invention is applied includes a heating furnace 2, a descaling device 3, a width-adjusting stamping device 4, a rough rolling mill 5, a finish rolling mill 6, a cooling device 7, and a coiler (winding machine) 8. A cast slab (not shown) is heated to a predetermined set temperature after being charged into the heating furnace 2 and is taken out from the heating furnace 2 as a hot slab. The hot slab taken out from the heating furnace 2 removes the primary scale formed on the surface through the descaling device 3, and then the width is reduced to a predetermined set width through the width-adjusting stamping device 4. Then, the slab with the reduced width is rolled to a predetermined thickness in the rough rolling mill 5 to become a rough-rolled product and is conveyed to the finish rolling mill 6. In the finish rolling mill 6, the rough-rolled product is rolled to the product thickness by a continuous rolling mill composed of 5 to 7 rolling mill stands, thereby becoming a finish-rolled product. On the downstream side of the finish rolling mill 6, a cooling device 7 is provided in a device called an exit roller table, and the finish-rolled product is wound into a coil shape by the coiler 8 after being cooled to a predetermined temperature. In addition, in the middle of the transfer process of the hot rolling line 1, a plurality of width gauges are provided as width measurement units. In the example shown in Figure 1 , a rough-rolled output side width gauge 11 is provided on the output side of the rough rolling mill 5, and a finish-rolled output side width gauge 12 is provided on the output side of the finish rolling mill 6. And, a coiler front width gauge (coiler input side width gauge) 13 for measuring the width of the steel sheet before winding is provided on the output side of the cooling device 7. Hereinafter, the width of the rough-rolled product measured by the rough-rolled output side width gauge 11 may be referred to as the rough-rolled output side width, the width of the finish-rolled product measured by the finish-rolled output side width gauge 12 may be referred to as the finish-rolled output side width, and the width of the steel sheet measured by the coiler front width gauge 13 may be referred to as the coiler front width.

[0051] The hot rolling line 1 is equipped with a control controller (PLC) 90 for controlling each device constituting the hot rolling line 1, a control computer (process computer) 91 for applying control instructions to the control controller 90, and a host computer 92 for giving manufacturing instructions to the hot rolling line 1. The width control of the steel plate in the hot rolling line 1 is performed as follows: The host computer 92 or the control computer 91 sets the control target values of the rough rolling output side width (rough rolling control target width), the finish rolling output side width (finish rolling control target width), and the width in front of the coiler (coiler front control target width) based on the manufacturing instructions from the host computer 92, and sets the operating conditions of the roughing mill 5 and the finishing mill 6. Specifically, the host computer 92 or the control computer 91 sets the target width of the finish rolled piece (finish rolling target width) based on the coiler front target width (coiler front target width) determined according to the product specifications of the hot rolled steel plate, considering the width change amount of the steel plate generated between the output side of the finishing mill 6 and the width meter 13 in front of the coiler. Moreover, the host computer 92 or the control computer 91 sets the target width of the rough rolled piece, that is, the rough rolling target width (hereinafter, sometimes only referred to as the target width), considering the width change amount of the steel plate in the finishing mill 6 based on the set finish rolling target width. In this case, the target values of the width control (rough rolling control target width, finish rolling control target width, coiler front control target width) are sometimes set by presetting a width margin (allowance) relative to the rough rolling target width, the finish rolling target width, and the coiler front target width. And the host computer 92 or the control computer 91 sets the rolling conditions in each pass of the rough rolling so that the rough rolling output side width is consistent with the rough rolling control target width. And the host computer 92 or the control computer 91 sets the rolling conditions in each rolling mill stand of the finishing mill 6 so that the finish rolling output side width is consistent with the finish rolling control target width. Moreover, the host computer 92 or the control computer 91 sometimes sets the tension between the finishing mill 6 and the coiler 8 and the cooling conditions of the cooling device 7 so that the width in front of the coiler is consistent with the coiler front control target width. In this case, in the finishing mill 6, dynamic width control is sometimes performed while referring to the measured values of the rough rolling output side width and the finish rolling output side width. The control controller 90 has a function of collecting information obtained from various sensors (thickness gauge, thermometer, etc.) in addition to collecting information obtained from the width meters provided on the hot rolling line 1 at a specified sampling period and outputting them to the control computer 91 at the specified sampling period

[0052] A method for predicting the width of a rough rolled piece according to an embodiment of the present invention is a method for measuring the rough rolling output side width. In addition, a method for controlling the width of a rough rolled piece according to an embodiment of the present invention is a method for controlling the width of a rough rolled piece so that the rough rolling output side width satisfies a specified relationship with respect to the rough rolling target width

[0053] 〔Heating Furnace〕

[0054] Figure 2 is a schematic diagram showing the structural example of the heating furnace 2 as Figure 1 shown. As Figure 2 shown, in this embodiment, the slab SA is charged into the heating furnace 2 from Figure 2 the left side. Regarding the temperature of the slab SA charged into the heating furnace 2, sometimes the slab SA is cooled to near room temperature in the post-casting slab yard after casting, and sometimes it is maintained at a temperature of about 600°C during the cooling process. In addition, sometimes the slab SA is charged into the heating furnace 2 at a temperature of about 600°C to 800°C without passing through the slab yard after casting. The inside of the heating furnace 2 is divided into multiple zones. Generally, 2 to 8 heating zones and 1 to 3 soaking zones divided into zones are provided on the upstream side. In Figure 2 the example shown, 5 heating zones and 1 soaking zone are provided. Here, the two are collectively referred to as the "heating furnace zones". Different atmosphere temperatures are set for each heating furnace zone so that the average temperature of the slab SA charged into the heating furnace 2 gradually rises to reach the specified target heating temperature (the target value of the average temperature when the slab SA is drawn out from the heating furnace 2). In addition, thermometers 21 for measuring the atmosphere temperature in the heating furnace zone are provided above all the heating furnace zones.

[0055] Figure 3 is a schematic diagram of observing the inside of the heating furnace 2 as Figure 1 shown from above. As Figure 3 shown, the slab SA charged into the heating furnace 2 passes through each heating furnace zone in sequence through a conveying device called a walking beam 22 inside the heating furnace 2. In addition, multiple slabs SA are charged into the heating furnace 2 at the same time and are drawn out from the output side outlet of the heating furnace 2 in the order of being charged into the heating furnace 2 for hot rolling. The slab SA is water-cooled inside the walking beam 22, and through components in direct contact with the slab SA called skids, there will be parts that locally hinder the temperature rise of the slab SA. The part of the slab SA in contact with the skids is called a skid mark, and its temperature is lower than the part not in contact with the skids. The skid mark is one of the reasons for the deviation in the width on the rough rolling output side.

[0056] 〔Width Adjusting and Stamping Device〕

[0057] Figure 4 is a schematic diagram showing as Figure 1Schematic diagram of the structural example of the width-adjusting stamping device 4 shown. After the slab SA heated by the heating furnace 2 removes the primary scale formed on its surface through the descaling device 3, it is width-adjusted to a specified set width by the width-adjusting stamping device 4. As Figure 4 shown, the width-adjusting stamping device 4 includes a pair of width-adjusting dies 41. The width-adjusting dies 41 press down on the slab SA in the width direction. The width-adjusting stamping device 4 drives the width-adjusting dies 41 through the driving device 42 while conveying the slab SA, and intermittently width-adjusts the slab SA from both sides in the width direction of the slab SA.

[0058] In the width-adjusting stamping device 4, the slab SA is conveyed using a pinch roller 43 or the like. The width-adjusting stamping device 4 can change the feed pitch of the slab SA between width-adjusting passes by changing the driving amount of the pinch roller 43. The feed pitch refers to the conveying distance of the slab SA per width-adjusting pass in the width-adjusting stamping device 4. The driving amount of the pinch roller 43 is controlled by the control controller 90 of the width-adjusting stamping device 4. On the surface of the width-adjusting die 41 in contact with the slab SA, a parallel portion 41a parallel to the conveying direction of the slab SA and an inclined portion 41b extending in the width direction toward the direction opposite to the conveying direction of the slab SA are sequentially formed from the front end side in the conveying direction of the slab SA. In the width-adjusting die 41, in order to suppress the occurrence of slip relative to the slab SA, one or more parallel portions 41a are sometimes provided between the inclined portions 41b. The shape of the width-adjusting die 41 changes the deformation state of the slab SA and affects the width on the rough rolling output side.

[0059] 〔Roughing mill〕

[0060] Return to Figure 1 . The roughing mill 5 includes a reversible rolling mill 5a capable of reverse rolling and an irreversible rolling mill 5b capable of rolling only in the downstream conveying direction. As Figure 1The arrow (solid line) shown below the roughing mill 5 indicates the reduction pass (the rolling pass for thinning the thickness). In the reversing mill 5a, typically, about 5 to 11 reduction passes are performed in the reversing direction (from the upstream side to the downstream side or from the downstream side to the upstream side). In the final reduction pass, since rolling and transfer to the next mill are carried out simultaneously, the number of rolling passes of the reversing mill must be odd, and the steel plate is transferred to the mill on the downstream side while rolling. At this time, the time from when the rear end of the steel plate disengages from the mill in the current reduction pass until the rolling direction is reversed and the steel plate is bitten into the mill in the next reduction pass is called the inter-pass time of the reversing pass. In addition, the time from when the rear end of the steel plate disengages from the reversing mill 5a in the final rolling pass of the reversing pass until it is bitten into the non-reversing mill 5b is called the inter-pass time of the continuous pass. Further, the inter-pass time of the reversing pass and the inter-pass time of the continuous pass are combined and called the inter-pass air cooling time. The inter-pass air cooling time represents the time when the slab is air-cooled during transfer and affects the temperature change of the steel plate.

[0061] Figure 5 is shown as Figure 1 a schematic diagram showing the structure of the rolling mill housing constituting the roughing mill 5. As Figure 5 shown, the roughing mill 5 includes a horizontal rolling mill 51 for thinning the thickness of the steel plate SB and an edging mill (vertical rolling mill) 52 for reducing the width of the steel plate SB. The edging mill 52 is a rolling mill in which a pair of rolls are arranged longitudinally and is disposed adjacent to the horizontal rolling mill 51. In each rolling pass of roughing, usually, width rolling (width reduction) of the steel plate SB using the edging mill 52 is performed before horizontal rolling. Therefore, when one edging mill 52 is arranged for the reversing mill 5a, the width reduction of the edging mill 52 is only performed in the forward direction, i.e., odd passes, and not in the even passes, i.e., reverse passes. However, when edging mills 52 are arranged on both sides of the reversing mill 5a, width reduction can be performed in all rolling passes. The roll opening (roll gap) of the horizontal rolling mill 51 and the opening of the edging mill 52 in each pass of roughing are set by the control computer 91. In the roughing mill 5, a descaling header for spraying descaling water toward the steel plate SB is provided, and descaling is performed on the input side of the horizontal rolling mill 51. However, descaling water is not limited to being sprayed in all rolling passes of roughing, and a specified descaling mode can be set according to the material of the steel plate SB, etc., for example, spraying only in the first pass or only in odd passes. The steel plate SB in the state after all the rolling passes preset by the roughing mill 5 are completed is called a roughing bar or a thin slab, and is called a roughing piece in this specification.

[0062] 〔Width gauge〕

[0063] Return to Figure 1The width of the steel plate in the hot rolling line 1 is measured by a rough rolling output side width meter 11, a finish rolling output side width meter 12, and a width meter 13 before the coiler. Most of these width meters use an optical width measurement method. The optical width measurement method is to arrange a light source below the pass line for transporting the steel plate and an image sensor above it, and measure the width of the steel plate based on the shadow length in the width direction of the light emitted from the light source during the passage of the steel plate. Also, the width meter can measure the width by determining the positions of the end portions in the width direction of the steel plate with a camera. Figure 6 Figs. (a) and (b) are schematic views showing a structural example of a rough rolling output side width meter using a camera. In Figure 6 the example shown in Fig. (a), a set of cameras 15a, 15b provided in the rough rolling output side width meter capture an image of the steel plate SB including the end portions in the width direction of the steel plate SB. The cameras 15a, 15b use CMOS or CCD sensors. Also, the image processing unit provided in the rough rolling output side width meter determines the positions of the end portions in the width direction of the steel plate SB based on the images captured by the cameras 15a, 15b, and calculates the width W of the steel plate S based on the setting interval between the cameras 15a, 15b. The reference numeral 14 in the figure indicates the pass line. However, since this width meter captures the end portions in the width direction of the steel plate SB obliquely, it is likely to cause a width measurement error when the steel plate SB floats from the pass line 14. Therefore, most width meters have a function of correcting the width measurement error corresponding to the floating of the steel plate SB from the pass line. Specifically, as Figure 6 shown in Fig. (b), when the steel plate SB is transported at a height of the floating amount H from the pass line 14, the width measurement error is corrected as follows. That is, first, a set of cameras 15a, 15b arranged in the width direction of the steel plate SB determine the two end portions in the width direction of the steel plate SB, thereby determining the measured value W1 of the width. Next, another set of cameras 15c, 15d arranged in the width direction of the steel plate SB also determine the two end portions in the width direction of the steel plate SB, thereby determining the measured value W2 of the width. Then, based on the positional relationship between the two sets of cameras (in Figure 6 the example shown in Fig. (b), the interval D in the width direction between the cameras 15a, 15b and the interval L in the width direction between the camera 15a (15b) and the camera 15c (15d)), the actual width W of the steel plate SB is calculated by the following formula (2) to obtain the measured value of the width of the steel plate SB.

[0064] [Formula 2]

[0065]

[0066] However, as other methods for measuring the width, the following method is also used: a laser is emitted in the width direction of the steel plate, the reflected light from the end face of the steel plate is received, and the width of the steel plate is measured based on the distances to both end faces of the steel plate. The width meter also has a thermal expansion correction function for converting to the width of the cooled steel plate based on the temperature of the steel plate. The width of the steel plate is measured using the width meter during the conveyance of the steel plate. Therefore, the width measurement value of the steel plate obtained by the width meter is time-series numerical information corresponding to the sampling interval of the width meter. Also, using the information on the conveyance speed of the steel plate at the position where the steel plate passes through the width meter, it is converted into the relationship between the position in the length direction of the steel plate and the actual value of the width of the steel plate. And in the control computer 91, a representative value of the width of the steel plate is calculated based on the measured value of the width of the steel plate obtained. The representative value of the width of the steel plate can use the average value of the width of the steel plate in the length direction of the steel plate (average width), the measured value of the width of the steel plate in the stable part of the steel plate except for the front and rear ends (stable width), the measured value of the width of the steel plate in the front end part of the steel plate (front width), the measured value of the width of the steel plate in the tail end part of the steel plate (tail end width), etc. And sometimes the minimum value (minimum width) and the maximum value (maximum width) of the width of the steel plate in the length direction of the steel plate are calculated, etc. The width of the steel plate measured by the width meter is sometimes expressed using the deviation from a preset target width.

[0067] 〔Gaussian process regression〕

[0068] The width prediction method for the rough rolled product according to an embodiment of the present invention is used to predict the width on the rough rolling output side on the above-mentioned hot rolling line 1. The width prediction method for the rough rolled product according to an embodiment of the present invention uses a width prediction model learned by a method of Gaussian process regression. The method of Gaussian process regression includes, as input data, one or more operation parameters selected from the operation parameters of the rough rolling mill 5, and uses the statistical information of the rough rolling output side width as output data. Hereinafter, the method of Gaussian process regression applied in the width prediction method for the rough rolled product according to an embodiment of the present invention will be described.

[0069] Gaussian process regression is also called Gaussian process regression and Gaussian process, etc., and is a kind of non-linear regression model for estimating the function from the input variable to the output variable. The output can be set as a probability distribution. The case of using a Gaussian distribution determined by two parameters, the mean value and the variance, is called a Gaussian process. The Bayesian estimation method is used to obtain the probability distribution, indicating the reliability and uncertainty of the estimation. For example, when m variables are selected as the input of the width prediction model, the input variable is represented by the input vector x. And the output variable corresponding to the input vector x as the learning data is set as y. Hereinafter, specifically, the use of n learning data x （1） ~x （n） 、y（1） ~y （n） and obtain the statistical information y of the width on the rough rolling output side corresponding to the new input vector x by means of Gaussian process regression ＊ The m variables constituting the input vector x respectively represent different physical quantities. Therefore, the m variables can also be standardized (normalized) in advance. Specifically, for each of the m variables, the mean value and the standard deviation can be calculated based on n learning data, and the calculated mean value and standard deviation are used to standardize each variable. This is because the learning efficiency of the hyperparameters described later is improved by standardizing the m variables in advance. In this case, in order to convert the m variables into physical quantities, it is only necessary to perform an inverse transformation using the calculated mean value and standard deviation ＊ method. The m variables constituting the input vector x respectively represent different physical quantities. Therefore, the m variables can also be standardized (normalized) in advance. Specifically, for each of the m variables, the mean value and the standard deviation can be calculated based on n learning data, and the calculated mean value and standard deviation are used to standardize each variable. This is because the learning efficiency of the hyperparameters described later is improved by standardizing the m variables in advance. In this case, in order to convert the m variables into physical quantities, it is only necessary to perform an inverse transformation using the calculated mean value and standard deviation

[0070] In Gaussian process regression, a probability model is used in which a function f(x) representing the output variable y is estimated based on the input variable x and Gaussian noise is added. For example, the probability model is expressed as shown in the following equation (3). In this case, it is assumed that the function f(x) follows a multivariate Gaussian distribution. And, it is assumed that the Gaussian noise ε （i） follows a Gaussian distribution with a mean of zero and a variance of σ e （i）2 However, the Gaussian noise ε （i） can also be set to depend on the values of the n learning data x （1） ~x （n） or can be constant noise that does not depend on the n learning data x （1） ~x （n）

[0071] [Equation 3]

[0072]

[0073] The Gaussian distribution refers to a distribution in which the probability density N is represented by the following equation (4). In equation (4), μ represents the mean value, and σ represents the standard deviation (σ 2 represents the variance). That is, the Gaussian distribution is a probability density determined by the mean value μ and the standard deviation σ or the variance σ 2 . Gaussian process regression uses a multivariate normal distribution that extends this Gaussian distribution to multiple dimensions

[0074] [Equation 4]

[0075]

[0076] ​In Gaussian process regression, the mean function (mean vector) representing the multivariate Gaussian distribution is set to a constant (e.g., zero), and the covariance matrix is represented by a kernel function. The kernel function is a function used to calculate the similarity of data. The kernel function uses the input vectors x （i） , x （j） as independent variables and is expressed as k(x （i） , x （j） ), and outputs the similarity between the input vector x （i） and the input vector x （j） . As the kernel function, well-known kernel functions such as the white kernel function, linear kernel function, polynomial kernel function, Gaussian kernel function, and Matern kernel function can be used. Illustrating several kernel functions, using the parameter θ, they are expressed as shown in the following equations (5) to (7). Equation (5) represents the linear kernel function, equation (6) represents the quadratic polynomial kernel function, and equation (7) represents the Gaussian kernel function.

[0077] [Equation 5]

[0078]

[0079] [Equation 6]

[0080]

[0081] [Equation 7]

[0082]

[0083] Based on the above assumptions, the function f(x) representing the output variable y in terms of the input variable x is expressed using the probability density N as shown in the following equation (8).

[0084] [Equation 8]

[0085]

[0086] In this case, when the Gaussian noise is a constant value σ e 2 independent of the learning data, when defining the covariance matrix K n determined by the kernel function and the covariance matrix Σ n including the Gaussian noise using the following equations (9) and (10), equation (8) is expressed as the following equation (11) or equation (12). Here, I represents the identity matrix.

[0087] [Equation 9]

[0088]

[0089] [Equation 10]

[0090]

[0091] [Formula 11]

[0092]

[0093] [Formula 12]

[0094]

[0095] The covariance matrices K n , Σ n contained in the right - hand sides of Formulas (11) and (12) take the learning data x （1） ~x （n） as inputs, and the learning data y （1） ~y （n） are contained in the left - hand sides of Formulas (11) and (12) as outputs. Therefore, as long as the hyperparameters (parameters θ and Gaussian noise σ e 2 ) contained in the kernel function are determined in such a way that the relationships shown in Formula (11) or Formula (12) hold. Any method selected from well - known methods can be used to determine the hyperparameters. For example, the likelihood function of the learning data can be calculated, and the hyperparameters can be calculated by maximizing the log - likelihood represented by the logarithm of the calculated likelihood function. In this case, as the calculation method for maximizing the log - likelihood, optimization methods such as the Monte Carlo method and the conjugate gradient method can be used. Also, methods such as cross - validation and maximizing the marginal likelihood can be used.

[0096] Next, a method for estimating the statistical information y ＊ of the rough - rolling output - side width corresponding to a new input vector x ＊ using the function f(x) for which the hyperparameters of the kernel function have been determined will be described. The estimated value corresponding to an unknown input vector x ＊ not included in the learning data can be applied with Bayesian estimation and is expressed as Formula (13) shown below. In this case, the vector of the newly determined kernel function k ＊ is defined as Formula (14) shown below.

[0097] [Formula 13]

[0098]

[0099] [Formula 14]

[0100]

[0101] Thus, Formula (13) can be expressed as Formula (15) shown below. Then, by Formulas (16) and (17) shown below, the mean is set to W m , and the variance is set to Wσ , which can calculate the statistical information y of the width on the rough rolling output side relative to the input vector x * . For details of the Gaussian process regression method, reference can be made to well-known documents (such as Non-Patent Document 1) etc. * .

[0102] [Equation 15]

[0103]

[0104] [Equation 16]

[0105]

[0106] [Equation 17]

[0107]

[0108] In the present embodiment, the step of determining hyperparameters for representing the relationship of the above formula (11) or formula (12) is referred to as the model generation step. Specifically, in the model generation step, as Figure 7 shown, first, n learning data x （1） ~x （n） , y （1） ~y （n） are obtained from the data set accumulated in a database or the like (step S1). Next, a kernel function used in machine learning is selected from, for example, formulas (5) to (7) (step S2). Next, as the Gaussian noise ε （i） , a Gaussian distribution with a mean of zero and a variance of σ e 2 is set (step S3). Next, the covariance matrices K n , Σ n determined by the kernel function are calculated using formulas (9) and (10) (step S4). Then, the parameters θ and the Gaussian noise σ e 2 as hyperparameters are determined using formulas (11) and (12) representing the relationship between the input and output of the learning data and by a learning method using a likelihood function (step S5). The kernel function representing the input-output relationship of the learning data is determined by the hyperparameters thus determined. The determined hyperparameters may also be stored in the storage device of the computer that executes the model generation step. On the other hand, the step of calculating the mean value W m , the standard deviation W σ as the statistical information y ＊ of the width on the rough rolling output side relative to the input vector x ＊ is referred to as the prediction step. Specifically, in the prediction step, as Figure 8 shown, first, a new input vector x is obtained＊ (Step S11). Next, obtain the hyperparameters stored in the storage device of the computer that executes the model generation step, and use Equation (14) to determine the kernel function k ＊ corresponding to the new input vector x ＊ (Step S12). Then, use the relationships shown in Equations (16) and (17) to calculate the statistical information y ＊ corresponding to the input vector x ＊ of the width on the rough rolling output side as the predicted value of the average W m and the standard deviation W σ .

[0109] [Method for generating width prediction model]

[0110] Next, as a method for generating a width prediction model for rough-rolled pieces according to an embodiment of the present invention, an embodiment applying the above Gaussian process regression method will be described.

[0111] Figure 9 is a block diagram showing the structure of a width prediction model generation unit according to an embodiment of the present invention. As Figure 9 shown, the width prediction model generation unit 100 of the present embodiment includes a database unit 101 and a machine learning unit 102. The database unit 101 stores one or more operation performance data selected from the operation performance data of the rough rolling mill 5 and the performance data of the width on the rough rolling output side. The database unit 101 may also store, as needed, one or more operation performance data selected from the operation performance data of the width adjustment stamping device 4, one or more operation performance data selected from the operation performance data of the heating furnace 2, and one or more performance data selected from the performance data of the attribute information of the slab SA. The specific performance data stored in the database unit 101 will be described later.

[0112] The actual data stored in the database unit 101 can be appropriately obtained from the control controller 90, the control computer 91, or the host computer 92. Also, a data acquisition unit 103 may be provided to collect such actual data. The actual data is temporarily stored in the data acquisition unit 103, and after generating a data set in which multiple actual data are correlated, it is stored in the database unit 101. Since the timings at which the data stored in the database unit 101 are obtained are sometimes different, it is easy to form a data set in which they are in a corresponding relationship by correlating multiple actual data in the data acquisition unit 103. Regarding the data set stored in the database unit 101, at least one actual data is obtained for one steel plate manufactured from one slab. For example, when using the average width of the steel plate as the actual data of the width on the rough rolling output side, the operation actual data of the rough rolling mill 5 may use the representative value as the actual data. In this case, regarding the operation actual data of the width adjustment stamping device 4, the operation actual data of the heating furnace 2, and the actual data of the attribute information of the slab SA, the representative value for one steel plate may be used as the actual data.

[0113] On the other hand, in the data acquisition unit 103, multiple data sets may be generated for one steel plate manufactured from one slab and stored in the database unit 101. For example, when obtaining actual data related to the widths at the front end, the stable part, and the tail end of the rough rolling piece as the actual data of the width on the rough rolling output side, regarding the operation actual data of the rough rolling mill 5, the operation actual data obtained at the front end, the stable part, and the tail end of the rough rolling piece may be correlated with the actual data of the width on the rough rolling output side at the corresponding positions. However, regarding the actual data such as the attribute information of the slab that is determined regardless of the position in the length direction of the rough rolling piece, the actual data of the same attribute information is correlated with the actual data of the width on the rough rolling output side at the front end, the stable part, and the tail end of the rough rolling piece.

[0114] Moreover, in the data acquisition unit 103, the actual data of the width on the rough rolling output side can also be acquired for each position divided in the length direction with respect to one rough rolled piece, and the operation actual data acquired for each position in the length direction of the rough rolled piece is associated with the actual data of the width on the rough rolling output side measured at each position, and stored in the database unit 101. That is, the number of divisions in the length direction of the rough rolled piece is set to, for example, around 20 to 200, and the actual data of the width on the rough rolling output side in each division interval is associated with the operation actual data corresponding to each position. In this case, although the length of the steel plate SB rough rolled by the rough rolling mill 5 varies according to each rough rolling pass, as long as the operation actual data at the position corresponding to the division in the length direction of the rough rolled piece is acquired, a data set corresponding to each division interval can be formed in the data acquisition unit 103. When data sets corresponding to a plurality of positions divided in the length direction of the rough rolled piece are stored in the database unit 101, a width prediction model that varies according to each position in the length direction of the rough rolled piece can also be generated in the machine learning unit 102.

[0115] The width prediction model generation unit 100 can be provided in the control computer 91 for controlling the manufacture of the steel plate based on the hot rolling line 1. Moreover, the width prediction model generation unit 100 can also be provided in the host computer 92 that gives the manufacturing instruction to the control computer 91, and can also be provided in an independent computer capable of communicating with other devices. Also, the machine learning unit 102 can be configured as a device different from the database unit 101 using a device capable of receiving the data sets stored in the database unit 101. 100 or more data sets are stored in the database unit 101. It is also possible to store preferably 10,000 or more, more preferably 100,000 or more data sets in the database unit 101. The data stored in the database unit 101 may be screened as needed.

[0116] The machine learning unit 102 uses the data sets stored in the database unit 101 and performs machine learning by a method based on Gaussian process regression to generate a width prediction model M. The learning data used by the machine learning unit 102 is a plurality of data sets stored in the database unit 101, including one or more operation performance data selected from the operation performance data of the rough rolling mill 5 and the performance data of the width on the rough rolling output side. The machine learning unit 102 uses those learning data to perform machine learning by the method of Gaussian process regression and generate a width prediction model M. In the method of Gaussian process regression, one or more operation performance data selected from the operation performance data of the rough rolling mill 5 are included as input performance data, and the statistical information of the width on the rough rolling output side is used as output data. In addition, the machine learning unit 102 may also use the data sets stored in the database unit 101, and use one or more operation performance data selected from the operation performance data of the width adjustment stamping device 4, one or more operation performance data selected from the operation performance data of the heating furnace 2, and one or more performance data selected from the performance data of the attribute information of the slab SA as input performance data, perform machine learning by the method of Gaussian process regression, and generate a width prediction model M.

[0117] The machine learning in this case refers to determining the hyperparameters applied in the Gaussian process regression through Figure 7 the model generation steps shown. And the width prediction model M refers to the hyperparameters thus determined. This is because by determining the hyperparameters, the input-output relationship of the learning data is determined, so that the statistical information of the width on the rough rolling output side corresponding to unknown inputs can be predicted. And in the present embodiment, the statistical information of the width on the rough rolling output side, which is the output of the width prediction model M learned by Gaussian process regression, includes the prediction result of the average value of the width of the rough rolled piece and an index indicating its fluctuation, namely the standard deviation or variance. Thus, it is possible to predict the average value of the width on the rough rolling output side and simultaneously predict the fluctuation of the width on the rough rolling output side.

[0118] On the other hand, as described in Patent Document 3, for example, the prior art predicts the average value or representative value of the width change amount of the rough rolled piece based on a database generated using the performance data related to the width change amount of the rough rolled piece. However, in the prior art, information related to the deviation of the width change amount of the rough rolled piece cannot be obtained. Therefore, it is necessary to add a pre-set extra width (margin) to set the target value of the width of the rough rolled piece. Specifically, the conventional width control method is as Figure 10 shown, with respect to the target width (target width) W t of the width on the rough rolling output side, an extra width (margin) W r is pre-set, and the sum of the target width W t and the extra width W r is set as the control target value W of the width on the rough rolling output sidec Set the extra width W r The reason is that there is a certain fluctuation in the width on the rough rolling output side, so it is necessary to prevent the width on the rough rolling output side from being lower than the target width W t If the width on the rough rolling output side is lower than the target width Wt, the width of the steel plate after finish rolling may sometimes be lower than the product target width, and the product cannot be picked up due to insufficient width. Therefore, the manufactured hot-rolled steel plate is used as waste or the delivery destination of the product is changed, etc., which sometimes leads to a decrease in the product yield and a delay in delivery. On the other hand, if the set extra width W r is too large, the width of the steel plate after finish rolling becomes too large compared to the product target width. Therefore, trimming (edging) of the steel plate is required to obtain the product, resulting in a decrease in the product yield. That is, as a hot-rolled steel plate, if the occurrence of insufficient width is to be prevented, the width on the rough rolling output side will become too large. If the decrease in the yield caused by trimming is to be suppressed, insufficient width is likely to occur. Therefore, in the prior art, actual data of the width on the rough rolling output side is collected according to the thickness and width classification of the hot-rolled steel plate, and the extra width is preset according to its fluctuation, and the factory supervisor regularly monitors whether the setting of the extra width is appropriate.

[0119] In contrast, according to the present embodiment, the average value and statistical fluctuation of the width on the finish rolling output side can be predicted based on the operating conditions of the hot rolling line as input. Therefore, an appropriate extra width can be set according to the operating conditions of the rough rolled piece, rather than according to the thickness and width classification of the hot-rolled steel plate as in the prior art. Thereby, it is possible to suppress a decrease in the product yield caused by insufficient width and width surplus of the hot-rolled steel plate due to fluctuations in the width on the rough rolling output side.

[0120] 〔Slab attribute information〕

[0121] The slab attribute information that can be used in the input of the width prediction model M refers to information related to the slab size and information related to the component composition of the slab that affect the width change of the slab in the width adjustment stamping device 4 and the rough rolling mill 5. The information related to the slab size is information related to the thickness, width, length, and weight of the slab. The information related to the component composition of the slab is information related to the content of the components contained in the slab, and examples include the C content, Si content, Mn content, P content, S content, Nb content, Ti content, Cu content, Ni content, Mo content, B content, etc. of the slab. The information related to the slab size affects the temperature change of the slab in the hot rolling line 1, and thus affects the fluctuation of the width on the rough rolling output side. And the information related to the component composition of the slab affects the deformation resistance of the slab and the composition and thickness of the oxide film formed on the surface of the slab. Thereby, it affects the frictional force at the interface between the roll and the slab, changes the deformation state of the slab, and thus affects the deviation of the width on the rough rolling output side.

[0122] 〔Operating parameters of the reheating furnace〕

[0123] The operating parameters of the reheating furnace that can be used as inputs to the width prediction model M refer to the parameters representing the operating conditions of the reheating furnace 2 when heating the slab in the reheating furnace 2, and the information that affects the width change of the slab in the width-adjusting stamping device 4 and the roughing mill 5. Among the operating parameters of the reheating furnace 2, the temperature of the slab when it is charged into the reheating furnace 2, the residence time of the slab in the specified furnace zone in the reheating furnace 2, the atmosphere temperature of the final furnace zone of the reheating furnace 2, and the temperature of the slab withdrawn from the reheating furnace 2 can be used. By these parameters, the temperature drop during rough rolling of the slab is affected, and thus these parameters affect the width on the roughing output side.

[0124] In addition, information such as the charging position of the slab in the reheating furnace 2 and the positional relationship between the slab in the reheating furnace 2 and other slabs can also be used. As Figure 3 shown, the charging position of the slab in the reheating furnace 2 can also use the information related to the in-furnace charging position P, which represents the distance between the walking beam 22 at one end of the reheating furnace 2 and the end of the slab SA in the length direction. According to the in-furnace charging position P, the position of the black mark on the slideway in the length direction of the slab SA changes, affecting the width fluctuation in the length direction of the slab SA, and thus affecting the deviation of the width on the roughing output side. In the charging position of the slab SA, the distance D1 between the end of the slab SA in the length direction and the furnace wall of the reheating furnace 2 and the parameter representing the interval between the walking beam (fixed slideway) 22 or the moving slideway 23 of the reheating furnace 2 into which the slab SA is charged can also be used. In addition, as the information on the positional relationship between the slab in the reheating furnace 2 and other slabs, the distance (charging interval) D2 between the slab and adjacent other slabs in the reheating furnace 2 can be used. When the charging interval D2 in the reheating furnace 2 changes, the temperature of the end face of the slab SA changes. Thus, it affects the temperature distribution of the slab SA in the hot rolling line 1 and affects the deviation of the width on the roughing output side. The information on the positional relationship between the slab in the reheating furnace 2 and other slabs can use the length or thickness of other slabs, the difference between the distance between the front end of the slab to be predicted and the furnace wall of the reheating furnace and the distance between the front ends of other slabs, etc.

[0125] 〔Operating parameters of the width-adjusting stamping device〕

[0126] The operating parameters of the width-adjusting stamping device 4 that can be used as the input for the width prediction model M refer to the operating conditions when adjusting the width of the heated slab. The operating parameters of the width-adjusting stamping device 4 can use the operating parameters related to the width adjustment amount of the slab. The operating parameters related to the width adjustment amount of the slab include the width adjustment amount at the representative position in the length direction of the slab and the feed pitch of the slab between the width adjustment passes. The operating parameters related to the width adjustment amount of the slab may also include the length of the parallel portion 41a of the width-adjusting die 41 in contact with the slab, i.e., the width adjustment start position, in the first width adjustment pass for the front end portion of the slab. The operating parameters related to the width adjustment amount of the slab affect the shape of the dog-bone (thickness distribution in the width direction) formed on the width-adjusted slab, and thus affect the deviation of the width on the rough rolling output side. In addition, even if the width adjustment amount of the slab is a fixed amount in the length direction of the slab, the dog-bone shapes are different on the stable portion, the front end portion, and the rear end portion of the slab. As a result, the average value of the width on the rough rolling output side fluctuates, affecting the deviation of the width on the rough rolling output side. In addition, even if the feed pitch of the slab between the width adjustment passes is a fixed pitch, the dog-bone shapes deviate on the stable portion and the front and rear end portions of the slab. As a result, it affects the deviation of the width on the rough rolling output side.

[0127] In addition, the operating parameters of the width-adjusting stamping device 4 can use the operating parameters related to the die shape applied to the width-adjusting stamping device 4. The operating parameters related to the die shape are representative values fixed with respect to the length direction of the slab. For example, the length of the parallel portion 41a of the width-adjusting die 41 and the angle of the inclined portion 41b as shown in Figure 4 can be used. Depending on the shape of the width-adjusting die 41, the deformation state of the slab is different, which affects the deviation of the width on the rough rolling output side. In addition, the operating parameters of the width-adjusting stamping device 4 can also use the total weight of the slab and the cumulative value of the slab length after the width-adjusting die 41 is ground offline and assembled into the width-adjusting stamping device 4 and the width adjustment is performed using the assembled width-adjusting die 41. This is because the increase in the total weight of the slab and the cumulative value of the slab length causes the wear and damage of the width-adjusting die 41 to intensify, resulting in an error between the set value and the actual value of the width adjustment amount, and causing a deviation in the width on the rough rolling output side.

[0128] 〔Operating parameters of the rough rolling mill〕

[0129] The operating parameters of the roughing mill for the input of the width prediction model M refer to the rolling operation conditions that affect the width of the steel plate in any rolling pass during rough rolling of the roughing mill 5. Preferably, the operating parameters of the roughing mill 5 include the rolling conditions of the horizontal rolling mill 51 and the edging mill 52 that make up the roughing mill 5. As the rolling conditions of the horizontal rolling mill 51, the roll opening, the work roll diameter, the input side plate thickness, the output side plate thickness, the reduction ratio, the roughing target width, the rolling load, and the steel plate temperature in any rolling pass can be used. This is because these will affect the width expansion behavior of the steel plate during horizontal rolling, thereby affecting the deviation of the width on the roughing output side. As the rolling conditions of the edging mill 52, the edging mill opening, the edging mill roll diameter, the plate thickness, the width adjustment ratio, and the width adjustment load in any rolling pass can be used. These rolling conditions will affect the width expansion behavior of the steel plate during horizontal rolling and affect the deviation of the width on the roughing output side. In addition, even if the operating conditions of width rolling are fixed, dog-bone formation behavior will occur at the stable part and the front and rear ends of the steel plate, so it will affect the width distribution in the length direction of the steel plate, thereby affecting the deviation of the width on the roughing output side.

[0130] In addition, for the operating parameters of the roughing mill 5, the cumulative values of the total weight and length of the steel plate after rough rolling can also be used, where the work rolls used in the horizontal rolling mill 51 and the edging mill rolls used in the edging mill 52 are assembled to the roughing mill 5 after off-line grinding. This is because as the cumulative values of the total weight and length of the steel plate increase, the wear and damage of the work rolls and the edging mill rolls are aggravated, and an error occurs between the set value and the actual value of the reduction ratio during horizontal rolling and the width adjustment amount of the edging mill, thereby causing a deviation in the width of the steel plate. Since the operating parameters of the roughing mill 5, that is, the rolling conditions, have a great influence on the width of the steel plate, it is preferably to use the operating parameters selected from the operating parameters of all rolling passes in the roughing process for the input of the width prediction model M. Specifically, for the operating parameters of the roughing mill 5, it is preferred that the operating parameters of the roughing mill 5 include the set value or the actual value of the output side plate thickness from the first pass to the final pass of roughing. That is, the so-called roughing pass schedule. If the pass schedule is different, the thickness of the steel plate will change when it is transferred between roughing passes, resulting in a change in the temperature distribution during air cooling. Thus, depending on the position in the length direction of the steel plate, the deformation behavior in the width direction of the steel plate is different, affecting the deviation of the width on the roughing output side. Furthermore, the operating parameters of the roughing mill 5 can also include information related to the air cooling time between roughing passes and the presence or absence of spray descaling water in any rolling pass. This is because the temperature fluctuation of the steel plate will affect the deviation of the width on the roughing output side.

[0131] 〔Width prediction method for rough-rolled pieces〕

[0132] The width prediction method for a rough-rolled product according to an embodiment of the present invention includes a prediction step of using the width prediction model M generated as described above to predict the statistical information of the width on the rough-rolled output side. The width prediction unit that executes the prediction step may be provided in the control computer 91 for controlling the hot rolling line 1. Also, the width prediction unit may be provided in the host computer 92 that gives manufacturing instructions to the control computer 91, or may be provided in an independent computer capable of communicating with other devices. Hereinafter, with reference to Figure 11 The operation of the width prediction unit according to an embodiment of the present invention will be described.

[0133] For the steel plate manufactured in the hot rolling line 1, as Figure 11 shown, the operation of the width prediction unit 110 is executed before the rough-rolled output side width meter 11 measures the width on the rough-rolled output side. The operation of the width prediction unit 110 can be executed, for example, at the stage when the steel plate to be predicted is loaded into the heating furnace 2 as a slab. At the stage when the slab is loaded into the heating furnace 2, the slab size and information related to the composition of the slab are determined as the slab attribute information in the host computer 92. In addition, this is because, at the stage when the slab is loaded into the heating furnace 2, the manufacturing specifications of the hot-rolled steel plate are set, and the corresponding standard operation parameters are determined. Therefore, the prediction step can be executed by inputting the actual data as the slab attribute information and the set values of other operation parameters that can be determined by the standard conditions corresponding to the manufacturing specifications of the hot-rolled steel plate. In addition, the prediction step can be executed at the stage after the steel plate to be predicted is taken out of the heating furnace 2 as a slab. In this case, the actual data of the operation parameters of the heating furnace 2 can be obtained for input to the width prediction model M. Furthermore, for example, the prediction step can be executed at the stage after the steel plate to be predicted is width-adjusted by the width-adjusting stamping device 4. For the width-adjusted slab, the actual data of the operation parameters of the width-adjusting stamping device 4 are obtained by the control computer 91 or the host computer 92. And, this is because the obtained actual data are used for input to the width prediction model M, and since the set values of the operation parameters of the roughing mill 5 are set by the control computer 91, the input of the width prediction model M can be determined.

[0134] The operation of the width prediction unit 110 can be executed, for example, during the rolling passes of the steel plate in the roughing mill 5. This is because the actual data of the operation parameters of the roughing mill 5 in the rolling passes before the current rolling pass is obtained by the control computer 91 or the host computer 92, so that the set values of the operation parameters of the roughing mill 5 in the rolling passes after the current rolling pass can be obtained by the control computer 91 or the host computer 92. In short, the width prediction unit 110 can predict the statistical information of the roughing output side width by inputting the actual values of the operation parameters in the upstream processes or passes of the current process and the set values of the operation parameters during the period until the end of roughing into the width prediction model M.

[0135] As Figure 11 shown, the input data acquisition unit 111 of the width prediction unit 110 acquires the actual values or set values of the operation parameters of the hot rolling line 1 held by the control computer 91 or the host computer 92 in the above manner. The width prediction unit 110 inputs the input data acquired by the input data acquisition unit 111 to the prediction unit 112. The prediction unit 112 acquires the hyperparameters determined in the width prediction model generation unit 100. Then, the prediction unit 112 calculates the kernel function for the input data (new input vector) through Figure 8 the prediction steps shown, and calculates the statistical information of the roughing output side width as the output data. As described above, regarding the operation of the width prediction unit 110, since it can be executed at each stage of each process until the output side of the roughing mill 5 of the steel plate passes through the hot rolling line 1, it can be executed multiple times during the process of manufacturing one steel plate. The statistical information of the roughing output side width output in the above manner can also be displayed on a monitor or the like connected to the width prediction unit 110. Based on the output display of the statistical information of the roughing output side width, at least one of the operation parameters of the roughing mill 5 and the finishing mill 6 can be reset to suppress the occurrence of width defects in the hot rolled steel plate.

[0136] 〔Width control method for rough rolled pieces〕

[0137] In the width control method for rough rolled pieces according to an embodiment of the present invention, based on the statistical information of the roughing output side width predicted in the above manner, one or more operation parameters selected from the operation parameters of the roughing mill 5 are reset so that the probability that the roughing output side width is less than the target width Wt becomes smaller. In the above width prediction method for rough rolled pieces, the statistical information of the roughing output side width, which is the output of the width prediction model M, is determined, for example, as the average width W m and the standard deviation W σ of the roughing output side width. In this case, the roughing output side width W is predicted according to the probability density distribution g (W) shown in the following formula (18).

[0138] [Formula 18]

[0139]

[0140] Figure 12 is an example of schematically showing the probability density distribution g(W) of the width W on the rough rolling output side predicted by the width prediction model M together with the target width W t In Figure 12 the example shown, the average value W of the width W on the rough rolling output side predicted by the width prediction model M is predicted m to be less than the target width W t , and based on the deviation of the width W on the rough rolling output side, it is predicted that the width W on the rough rolling output side is likely to be less than the target width W t . That is, in this case, according to the operating conditions of the hot rolling line 1 set currently, it is predicted that the width W on the rough rolling output side is likely to be insufficient in width. Therefore, in this example, one or more operating parameters selected from the operating parameters of the rough rolling mill 5 are reset to reduce the probability that the predicted value W of the width W on the rough rolling output side is less than the target width W t . Specifically, the operating parameters of the rough rolling mill 5 are corrected so that the probability density distribution g(W) represented by the solid line in Figure 12 becomes the probability density distribution represented by the dashed line. In this case, as the operating parameters of the rough rolling mill 5 to be reset, the operating parameters in the rolling pass that is not currently being rolled are selected. Additionally, as the reset operating parameters, it is preferable to select from the operating parameters for the input to the width prediction model M. In Figure 12 the example shown, the operating parameters of the rough rolling mill 5 to be reset are reset as the input to the width prediction model M, and the statistical information of the width W on the rough rolling output side is output, and it is confirmed whether the probability that the output statistical information of the width W on the rough rolling output side is less than the target width W t is reduced. Thereby, it can be determined whether the appropriate operating conditions have been reset.

[0141] In addition, as shown in Figure 13 , in the case where it is predicted that the probability of being less than the target width W t is high compared to the probability density distribution g(W) of the width W on the rough rolling output side predicted by the width prediction model M, the control target value W c of the rough rolling width control can also be set to reduce the probability that the width W on the rough rolling output side is less than the target width W t . Thereby, by performing the rough rolling width control of the hot rolling line 1 using the set control target value W c , even when a deviation occurs in the width W on the rough rolling output side, the probability of width deficiency with respect to the target width W t can be reduced. Further, it is preferable to reset one or more operating parameters selected from the operating parameters of the rough rolling mill 5 so that the target width Wt Satisfies the formula (1) shown below.

[0142] [Formula 19]

[0143]

[0144] Figure 14 Is a schematic diagram showing the relationship between the probability density distribution g(W) of the rough rolling output side width W predicted by the width prediction model M and the target width W t Among them. In the above formula (1), it means that the target width W t Is within the range determined by the average value W m Of the rough rolling output side width W output from the width prediction model M and the standard deviation W σ And determine W m -2.5W σ And W m -1.5W σ In this case, the target width W t Can be made constant and one or more operating parameters selected from the operating parameters of the rough rolling mill 5 are reset in such a way that the probability density distribution g(W) of the rough rolling output side width W output from the width prediction model M satisfies the above formula (1). Figure 15 Is a graph showing the relationship between the deviation of the rough rolling output side width W from the target width W t And the cut-off amount of the steel plate. It can be seen from Figure 15 That regardless of whether the rough rolling output side width W is too large or too small relative to the target width Wt, the cut-off amount of the steel plate increases and the product yield of the hot-rolled steel plate decreases. Specifically, when the rough rolling output side width W is too large relative to the target width W t , trimming needs to be carried out in subsequent processes to make the hot-rolled steel plate meet the width specifications. On the other hand, when the rough rolling output side width W is too small relative to the target width W t , it cannot be formed into a steel plate product and a part needs to be scrapped. However, compared with the case where the rough rolling output side width W is too small relative to the target width W t , the case where the rough rolling output side width W is too large relative to the target width W t Suppresses the cut-off amount of the steel plate. Therefore, the above formula (1) is a formula for setting the operating parameters of the rough rolling mill 5 in such a way that even if there is a fluctuation in the rough rolling output side width W relative to the target width W t , the probability of the rough rolling output side width W being too small becomes lower. That is, the lower limit value of the formula (1) relative to the target width W t Is used to reduce the probability of the rough rolling output side width W being too small, and the upper limit value relative to the target width W t Is used to avoid the rough rolling output side width W being too large relative to the target width W tand becomes too large.

[0145] In the width control method of the rough-rolled piece in the present embodiment, not only corresponding to the classification of the steel type and size of the slab, the product size of the hot-rolled steel sheet, etc., but also corresponding to different operating conditions for each rough-rolled piece, statistical information on the width at the rough-rolled output side is output. Therefore, fluctuations in the width at the rough-rolled output side of each rough-rolled piece can be predicted. Thus, it is not necessary to preset the surplus width in advance according to the classification of the steel type and size of the steel sheet as in the prior art. An appropriate surplus width can be given to each rough-rolled piece manufactured on the hot-rolling line, improving the product yield of the hot-rolled steel sheet.

[0146] Example

[0147] 〔Example 1〕

[0148] As an example of the present invention, width prediction and width control of rough-rolled pieces were performed in the hot-rolling line 1. The hot-rolling line 1 includes a width-adjusting stamping device 4 disposed on the downstream side of the heating furnace 2, a roughing mill 5 composed of 4 reversible rolling mills 5a and 1 irreversible rolling mill 1b, and a finishing mill 6 composed of 7 rolling mill stands. In the present example, through the hot-rolling line 1, slabs with a slab thickness of 250 mm to 270 mm and a slab width of 600 mm to 1600 mm were heated by the heating furnace 2 to manufacture hot-rolled steel sheets with a plate thickness of 30 mm to 35 mm on the output side of the roughing mill 5 and a plate thickness of 2 mm to 3 mm on the output side of the finishing mill 6. In addition, the hot-rolling line 1 is equipped with a rough-rolled output-side width gauge 11, a finishing-rolled output-side width gauge 12, and a width gauge 13 before the coiler.

[0149] The control computer 91 and the host computer 92 of the hot-rolling line 1 collected the actual values of the operation parameters of the rough-rolled pieces manufactured in the hot-rolling line 1, and the data acquisition unit 103 acquired the actual data. The data acquisition unit 103 acquired the edger opening, the diameter of the edger roll, and the total rolling length after grinding of the edger roll in the entire rolling passes of rough rolling as the operation actual data of the roughing mill 5. In addition, as the operation actual data of the roughing mill 5, the work roll diameter of the horizontal rolling mill, the input-side plate thickness, the output-side plate thickness, and the total rolling length after grinding of the work roll in the entire rolling passes of rough rolling were acquired. Further, the data acquisition unit 103 acquired the data of the thickness and width of the slab as the actual data of the slab attribute information.

[0150] On the other hand, the data acquisition unit 103 calculates the average width of the stable part based on the actual value of the width measured by the rough rolling output side width meter 11 and sets it as the actual data of the rough rolling output side width. Regarding the actual data of the rough rolling output side width, it is associated with the above-mentioned operation actual data by the data acquisition unit 103, thereby constituting a data set for one rough rolled piece and storing it in the database unit 101. Then, at the stage when 30,000 data sets are stored in the database unit 101, these data sets are divided into 20,000 learning data and 10,000 test data, and the learning data is used to generate the width prediction model M by the machine learning unit 102. In this embodiment, as the kernel function of Gaussian process regression, the radial basis function (RBF) kernel is used. And, as Gaussian noise, noise with a constant variance σ that does not depend on the learning data is used. e 2 The kernel function used in this embodiment is represented by the following formula (19). Among them, ||x （i） -x （j） || represents the Euclidean distance between the output vectors.

[0151] [Equation 20]

[0152]

[0153] In this embodiment, the hyperparameters of the width prediction model M are determined by the method of Gaussian process regression using the learning data. Then, the operation actual data of the test data is input to the prediction unit 112 of the width prediction unit 110, and the average value W m and the standard deviation W σ of the rough rolling output side width, which are the outputs of the width prediction model M, are obtained. Then, according to the deviation between the actual data W a of the test data, that is, the rough rolling output side width W, and the average value W m , the RMSE (root mean square error) is calculated. In addition, the number of test data in which the actual data W a of the rough rolling output side width W is within the range of W m ±W σ and within the range of W m ±1.96W σ is respectively obtained, and the ratio of the number of test data falling into those ranges to all the test data is calculated.

[0154] As a result, the RMSE calculated based on the deviation between the actual data W a of the rough rolling output side width W and the average value W m is good, being 0.1 mm. Further, the rough rolling output side width W falls within W m ±W σThe probability within the range of is 67.3%, and it falls into W m ±1.96W σ The probability within the range of is 95.3%. This means that, assuming the deviation of the width W on the rough rolling output side follows a normal distribution, the probabilities are 68.3% and 95.0% respectively. Therefore, through the width prediction model M of this embodiment, the deviation of the width W on the rough rolling output side can be predicted with high accuracy. On the other hand, the width prediction model M generated as described above is stored in the prediction unit 112 of the width prediction unit 110, and width control is performed on the rough rolled pieces within the range of the slab width from 1000 mm to 1200 mm. In this case, for each steel plate, the statistical information of the width W on the rough rolling output side is calculated using the width prediction model M when the slab is drawn out from the heating furnace 2, and the operation parameters of the rough rolling mill 5 are reset so that the target width W of the rough rolled piece set for each steel plate t Satisfies the relationship shown in formula (1). As the operation parameter of the rough rolling mill 5 reset, the opening of the edger of the rough rolling mill 5 is selected. In this way, for the rough rolled pieces that have undergone rough rolling, finish rolling is continued to manufacture hot-rolled steel plates. The manufactured hot-rolled steel plates are 400 coils. As a result, compared with the conventional example where the surplus width was preset in advance, the proportion of coils with a width of the hot-rolled steel plate smaller than the target width before the coiler decreased by 35%. In addition, through this embodiment, the average trimming amount cut off due to the width being too large compared to the target width before the coiler decreased by 0.8 mm. Thus, it was confirmed that through this embodiment, the statistical information of the width on the rough rolling output side can be predicted, and by applying the predicted statistical information of the width on the rough rolling output side to the width control of the rough rolled pieces, the width defects of the hot-rolled steel plates can be reduced and the product yield can be improved.

[0155] 〔Embodiment 2〕

[0156] As an embodiment of the present invention, another example of the width prediction method for rough rolled pieces will be described. In the above Embodiment 1, when storing the actual performance data in the database unit 101, the data acquisition unit 103 acquired the operation actual performance data of the width adjustment stamping device 4 and the operation actual performance data of the heating furnace 2, and stored them corresponding to the operation actual performance data of the rough rolling mill 5 and the actual performance data of the slab attribute information. This embodiment uses the actual performance data stored in the database unit 101 in this way, changes the input data for the width prediction model M, and evaluates the width prediction accuracy of the rough rolled pieces.

[0157] The operation performance data of the width-adjusting stamping device 4 stored in the database unit 101 are the width adjustment amount SPW of the slab, the feed pitch SPP of the slab between width adjustment passes, and the width adjustment start position SPS. As the width adjustment amount SPW of the slab and the feed pitch SPP of the slab between width adjustment passes, the width adjustment amount and the feed pitch in the stable part of the slab are used. In addition, the operation performance data of the heating furnace 2 stored in the database unit 101 are the in-furnace time IFT from when the slab is charged into the heating furnace 2 until it is discharged from the heating furnace 2, the temperature of the slab discharged from the heating furnace 2 (discharge temperature) ET, and the operation parameter related to the charging position of the slab in the heating furnace 2, that is, the distance D1 between the end of the slab in the length direction and the furnace wall of the heating furnace 2.

[0158] In this embodiment, at the stage when 30,000 data sets are stored in the database unit 101, these data sets are divided into 20,000 learning data and 10,000 test data, and a width prediction model M is generated by the machine learning unit 102 using the learning data. In this case, the machine learning unit 102 changes the variables of the input data for the width prediction model M and performs machine learning to generate a width prediction model M corresponding to each condition. In any condition, a radial basis function (RBF) kernel is used as the kernel function of Gaussian process regression.

[0159] Table 1 shows the input data used in each width prediction model. In Table 1, the width prediction models of No.1 to No.3 all include the operation parameters of the rough rolling mill 5 as input data. As the operation parameters of the rough rolling mill 5, the edge mill opening EG in the stable part of the entire rolling pass of the edge rolling roughing and the diameter ED of the edge mill roll are used. Further, as the operation parameters of the rough rolling mill 5, the input side plate thickness HI, the output side plate thickness HO, and the work roll diameter HWD of the horizontal rolling mill in the stable part of the entire rolling pass of the rough rolling are used. Among them, in this embodiment, no parameters of the slab attribute information are used as the input data of the width prediction model M.

[0160] The width prediction model of No.1 recorded in Table 1 only uses the above parameters of the rough rolling mill 5 as input data. The width prediction model of No.2 uses the operation parameters of the width-adjusting stamping device as input data in addition to the parameters of the rough rolling mill 5. The operation parameters of the width-adjusting stamping device used are the width adjustment amount SPW, the feed pitch SPP, and the width adjustment start position SPS. The width prediction model of No.3 uses the operation parameters of the heating furnace 2 as input data in addition to the parameters of the rough rolling mill 5. The operation parameters of the heating furnace 2 used are the in-furnace time IFT, the discharge temperature ET, and the distance D1 from the furnace wall representing the charging position.

[0161] In this embodiment, the hyperparameters of the width prediction models of No.1 to No.3 were determined by using the Gaussian process regression method with learning data. Moreover, the operation performance data of the test data were input to the prediction unit 112 of the width prediction unit 110, and the average value Wm and the standard deviation Wσ of the width W on the rough rolling output side, which are the outputs of the width prediction model, were obtained. And the RMSE (root mean square error) was calculated based on the deviation between the actual value data Wa of the width W on the rough rolling output side as the test data and the average value Wm. In addition, the number of test data for which the actual value data Wa of the width W on the rough rolling output side is within the range of Wm±Wσ was obtained respectively, and the ratio of the number of test data falling into those ranges to all the test data was calculated.

[0162] Table 1 shows the results of the prediction accuracy. In No.1, the RMSE calculated based on the deviation between the actual value data Wa of the width W on the rough rolling output side and the average value Wm was 1.0 mm. In addition, the probability that the width W on the rough rolling output side falls within the range of Wm±Wσ was 70.0%, and the probability was close to 68.3% assuming that the deviation of the width W on the rough rolling output side follows a normal distribution. In No.2, the RMSE was 0.1 mm, and the prediction accuracy was improved compared to No.1. On the other hand, the probability that the width W on the rough rolling output side falls within the range of Wm±Wσ was 66.3%, and the probability was close to 68.3% assuming that the deviation of the width W on the rough rolling output side follows a normal distribution. In No.3, the RMSE was 0.0 mm, and a high prediction accuracy was obtained for the average value of the width W on the rough rolling output side. In addition, the probability that the width W on the rough rolling output side falls within the range of Wm±Wσ was 66.0%, and the probability was close to 68.3% assuming that the deviation of the width W on the rough rolling output side follows a normal distribution. Based on the above, it was confirmed that any of the width prediction models can accurately predict the statistical information of the width on the rough rolling output side, that is, the average value Wm and the standard deviation Wσ of the width on the rough rolling output side.

[0163] [Table 1]

[0164]

[0165] As described above, the embodiments to which the invention completed by the present inventor is applied have been described, but the present invention is not limited by the description and drawings that are part of the disclosure of the present invention based on this embodiment. That is, other embodiments, examples, and application techniques completed by those skilled in the art based on this embodiment are all included in the scope of the present invention.

[0166] Industrial Applicability

[0167] According to the present invention, a method for predicting the width of a rough rolling piece can be provided, which can predict statistical information including fluctuations in the width of the rough rolling piece. Further, according to the present invention, a method for controlling the width of a rough rolling piece can be provided, which can take into account fluctuations in the width of the rough rolling piece and accurately control the width in the length direction of the rough rolling piece. Further, according to the present invention, a method for manufacturing a hot-rolled steel sheet can be provided, which can improve the product yield of the hot-rolled steel sheet. Further, according to the present invention, a method for generating a width prediction model of a rough rolling piece can be provided, which can generate a width prediction model for predicting statistical information including fluctuations in the width of the rough rolling piece.

[0168] Description of reference numerals

[0169] 1 Hot rolling line; 2 Heating furnace; 3 Descaling device; 4 Width-adjusting stamping device; 5 Rough rolling mill; 5a Reversing mill; 5b Non-reversing mill; 6 Finishing mill; 7 Cooling device; 8 Coiler (winding machine); 11 Width gauge on the rough rolling output side; 12 Width gauge on the finishing rolling output side; 13 Width gauge before the coiler (width gauge on the coiler input side); 14 Pass line; 15a, 15b, 15c, 15d Cameras; 21 Thermometer; 22 Walking beam (fixed slideway); 23 Moving slideway; 41 Die for width adjustment; 41a Parallel part; 41b Tapered part; 42 Driving device; 43 Pinch roll; 51 Horizontal rolling mill; 52 Edging mill (vertical rolling mill); 90 Controller for control; 91 Control computer; 92 Host computer; 100 Width prediction model generation unit; 101 Database unit; 102 Machine learning unit; 103 Data acquisition unit; 110 Width prediction unit; 111 Input data acquisition unit; 112 Prediction unit; M Width prediction model; SA Slab; SB Steel plate.

Claims

1. A method for predicting the width of a rough-rolled piece, which predicts the width of a rough-rolled piece in a hot rolling line, and the hot rolling line includes: A heating furnace for heating a slab; a rough rolling mill for rough rolling the heated slab to produce the rough rolled product; and a finish rolling mill for finish rolling the rough rolled product to produce a finish rolled product, the method for predicting the width of the rough rolled product comprising: A prediction step of predicting statistical information on the width of the rough rolled product using a width prediction model learned by a Gaussian process regression method, the width prediction model including one or more operating parameters selected from the operating parameters of the rough rolling mill as input data and setting the statistical information on the width of the rough rolled product as output data.

2. The method for predicting the width of a rough rolled product according to claim 1, wherein the hot rolling line includes a width adjusting stamping device disposed upstream of the rough rolling mill, the width adjusting stamping device intermittently adjusting the width of the slab heated by the heating furnace, and the width prediction model includes one or more operating parameters selected from the operating parameters of the width adjusting stamping device as input data.

3. The method for predicting the width of a rough rolled product according to claim 1 or 2, wherein the width prediction model includes one or more operating parameters selected from the operating parameters of the heating furnace as input data.

4. The method for predicting the width of a rough rolled product according to any one of claims 1 to 3, wherein the width prediction model includes one or more parameters selected from the attribute information of the slab as input data.

5. A method for controlling the width of a rough rolled product, comprising: A resetting step of predicting statistical information on the width of the rough rolled product using the method for predicting the width of a rough rolled product according to any one of claims 1 to 4, and based on the predicted statistical information, resetting one or more operating parameters selected from the operating parameters of the rough rolling mill in such a way that the probability that the width of the rough rolled product is lower than the target width of the rough rolled product becomes smaller.

6. The method for controlling the width of a rough rolled product according to claim 5, wherein The statistical information on the width of the rough-rolled piece includes the average value W of the width of the rough-rolled piece m and the standard deviation W σ , and the resetting step includes: setting one or more operating parameters selected from the operating parameters of the roughing mill in such a way that the target width W of the rough-rolled piece t satisfies the relationship shown in the following formula (1), [Equation 1] 。 7. A method for manufacturing a hot rolled steel sheet, comprising: A step of manufacturing a hot rolled steel sheet using the method for controlling the width of a rough rolled product according to claim 5 or 6.

8. A method for generating a width prediction model of a rough rolling piece, which generates a width prediction model for predicting the width of a rough rolling piece in a hot rolling line, and the hot rolling line includes: A heating furnace for heating a slab; a rough rolling mill for rough rolling the heated slab to produce the rough rolled product; and a finish rolling mill for finish rolling the rough rolled product to produce a finish rolled product, the method for generating the width prediction model of the rough rolled product comprising: A step of obtaining learning data, obtaining a plurality of learning data, the plurality of learning data including: one or more operation performance data selected from the operation performance data of the rough rolling mill and the performance data of the width of the rough rolled product; and A step of using the plurality of learning data obtained in the step of obtaining learning data and generating the width prediction model using a Gaussian process regression method, the width prediction model including one or more operation performance data selected from the operation performance data of the rough rolling mill as input performance data and setting the statistical information on the width of the rough rolled product as output data.

Images