JP2021109186A - Control method for rolling device, control device for rolling device, and manufacturing method for steel plate - Google Patents

Abstract

To suppress a camber of a rolled material on a final rolling path exit side of a rolling device even when a camber occurs caused by a rolled material steel making condition and a heating condition.SOLUTION: A control method for a rolling device when a rolled material obtained by refining and casting in a steel making step and heating in a heating furnace is rolled using a rolling device including a plurality of rolling machines includes: a prediction step (step S101) of predicting a camber amount of the rolled material on an inlet side of the rolling device using an expression constructed by machine learning using at last one operator parameter out of an operation parameter group concerning the rolled material steel making step and an operation parameter group concerning the heating step; a calculation step (step S102) of calculating a reduction leveling amount of each of the rolling machines based on the predicated camber amount of the rolled material; and a control step (step S103) of controlling each of the rolling machines based on the calculated reduction leveling amount.SELECTED DRAWING: Figure 4

Description

The present invention relates to a method for controlling a rolling apparatus, a control apparatus for a rolling apparatus, and a method for manufacturing a steel sheet.

Generally, in a hot rolling step, a material to be rolled heated in a heating furnace is rolled down to a predetermined thickness by a group of rolling mills after being rolled down by a width reduction device. It is known that cambers and wedges are generated in the width reduction device due to the temperature deviation in the plate width direction of the material to be rolled and the amount of off-center on the entry side when the width is reduced. These left-right asymmetry factors, such as camber, wedge, and temperature deviation in the plate width direction of the material to be rolled after rolling reduction, cause camber in the material to be rolled in the rolling mill, which causes equipment damage such as roll scratches and side guides. Not only that, but also the product yield, quality deterioration and line operation rate decrease. Further, even in the hot rolling process that does not include the width reduction process, the temperature deviation in the plate width direction of the material to be rolled is cited as a factor that causes camber in the material to be rolled on the exit side of the rolling mill group in the rolling process. Be done. This temperature deviation in the plate width direction is caused by cooling from the extraction side by the inflowing outside air when the extraction door is opened when the material to be rolled heated in the heating furnace is extracted. Therefore, in the rolling process, various techniques for suppressing camber of the material to be rolled have been conventionally proposed.

Patent Document 1 describes the amount of camber measured on the inlet / output side of the rolling mill, at least the information on the cast strands of the material to be rolled, and the furnaces of the materials to be rolled that are adjacent to each other inside the heating furnace that heats the material to be rolled. Rolling that accumulates data on the distance between the inner materials and sets the leveling value of the rolling mill, assuming that the cast strand information and the material to be rolled, which can be regarded as having the same distance between the inner materials in the furnace, have similar temperature deviations in the width direction. The leveling setting method of the machine is disclosed.

Patent Document 2 describes the leveling value of the rolling mill based on the staying time of the material to be rolled in the heating furnace in the tropics, the distance between the adjacent materials to be rolled in the heating furnace, and the width of the material to be rolled. The leveling setting method of the rolling mill to be set is disclosed.

Patent Document 3 describes a method for setting the leveling of a rolling mill, which sets a leveling value of the rolling mill according to the temperature deviation measured on the inlet side of the measured width reduction device, the plate width, and the width reduction amount in the width reduction device. Is disclosed.

Japanese Unexamined Patent Publication No. 2017-225989 Japanese Patent No. 6269536 Japanese Patent No. 6311627

However, in the method described in Patent Document 1, the leveling of the rolling mill is only set in consideration of the difference in deformation resistance in the width direction due to the temperature deviation in the width direction of the material to be rolled before the start of rolling. , Not applicable when the material to be rolled before the start of rolling has a camber.

Further, the method described in Patent Document 2 only sets the leveling of the rolling mill in consideration of the heating conditions, and cannot be applied when the operating conditions of the steelmaking process and the width reduction process change.

In the method described in Patent Document 3, it is necessary to measure the plate thickness deviation in the plate width direction of the material to be rolled, but it is difficult to measure the plate thickness deviation of a steel sheet having a high temperature and an oxidation scale stably and accurately. be. Furthermore, it cannot be applied when camber occurs after the width reduction.

Further, as described above, even in the hot rolling process that does not include the width reduction process, the material to be rolled on the exit side of the rolling mill group in the rolling process due to the temperature deviation in the plate width direction generated in the material to be rolled in the heating process. May cause camber. Therefore, it is desired that it can be applied to a hot rolling process that does not include a width reduction process.

The present invention has been made in view of the above circumstances, and even if camber occurs due to the steelmaking conditions and heating conditions of the material to be rolled, the subject is covered on the final rolling pass exit side of the rolling mill. It is an object of the present invention to provide a control method for a rolling apparatus capable of suppressing camber of a rolled material, a control device for a rolling apparatus, and a method for manufacturing a steel plate.

The method for controlling a rolling apparatus according to the present invention is a method for rolling an object to be rolled, which is refined and cast in a steelmaking process and heated in a heating furnace, by using a rolling apparatus composed of a plurality of rolling mills. An equation constructed by machine learning using at least one operation parameter group related to the steelmaking process of the material to be rolled and an operation parameter group related to the heating process as explanatory variables of the recirculator, which is a control method. A prediction step for predicting the camber amount of the material to be rolled on the entry side of the rolling apparatus, and a calculation step for calculating the rolling leveling amount of each rolling mill based on the predicted camber amount of the material to be rolled. It is characterized by including a control step for controlling each rolling mill based on the calculated rolling reduction amount.

In the method for controlling the rolling mill according to the present invention, in the above invention, the operating parameters used in the prediction step are silicon equivalent as the operating parameters related to the steelmaking process and the soaking tropical furnace time as the operating parameters related to the heating process. It is characterized by including.

In the method for controlling a rolling apparatus according to the present invention, in the above invention, the operation parameters used in the prediction step are the carbon equivalent of the material to be rolled, the silicon equivalent of the material to be rolled, and the material to be rolled as the operation parameters related to the steelmaking process. Includes the number of the smelting and cast casting machine, and is characterized by including the soaking time in the tropical furnace and the extraction temperature of the heating furnace as operating parameters related to the heating process.

The method for controlling a rolling apparatus according to the present invention is that, in the above invention, the rolling apparatus rolls a material to be rolled that has been subjected to width reduction using a width reduction device in a width reduction step after a heating step. The prediction step uses an equation constructed by machine learning using at least one operating parameter from each of the operating parameter group for the steelmaking process, the operating parameter group for the heating process, and the operating parameter group for the width rolling process. It is characterized by including a step of predicting the camber amount of the material to be rolled on the entry side of the rolling apparatus.

In the method for controlling the rolling mill according to the present invention, in the above invention, the operating parameters used in the prediction step are silicon equivalent as the operating parameters related to the steelmaking process, and the soaking tropical furnace time as the operating parameters related to the heating process. The operation parameter related to the width rolling down step is characterized by including the width rolling down amount.

In the method for controlling a rolling apparatus according to the present invention, in the above invention, the operating parameters used in the prediction step are the carbon equivalent of the material to be rolled, the silicon equivalent of the material to be rolled, and the material to be rolled as the operating parameters related to the steelmaking process. Includes the number of the smelting and casting casting machine, includes the soaking temperature in the tropical furnace and the extraction temperature of the heating furnace as the operating parameters related to the heating process, and the operating parameters related to the width rolling process include the width rolling amount and the mold usage amount. It is characterized by including the plate width of the material to be rolled before the width reduction and the plate thickness of the material to be rolled before the width reduction.

In the above invention, the control method of the rolling apparatus according to the present invention further includes a setting step of setting the leveling amount of the rolling mill according to the rolling leveling amount calculated by the calculation step, and the calculation step is predicted. On the exit side of the rolling mill, the camber amount of the material to be rolled on the inlet side of the rolling mill and the influence coefficient representing the relationship between the change in the camber amount on the exit side of the rolling mill with respect to the operation of the rolling reduction amount in the rolling mill are used. The control step includes a step of calculating a reduction leveling amount that can control the camber amount of the material to be rolled to a target value, and the control step is a step of controlling each rolling mill based on the reduction leveling amount set by the setting step. It is characterized by including.

The control device for the rolling apparatus according to the present invention is a control device for a rolling apparatus that rolls a material to be rolled, which is refined and cast in a steelmaking process and heated in a heating furnace, by using a rolling apparatus composed of a plurality of rolling mills. As an explanatory variable of the recirculator, there is a recirculator constructed by machine learning using at least one of the operation parameters related to the steelmaking process of the material to be rolled and the operation parameters related to the heating process. Then, based on the prediction unit that predicts the camber amount of the material to be rolled on the entrance side of the rolling apparatus using the recirculator, and the predicted camber amount of the material to be rolled, the rolling mill's rolling leveling amount. It is characterized in that it is provided with a calculation unit for calculating the above and a control unit for controlling each rolling mill based on the calculated rolling down leveling amount of each rolling mill.

In the above invention, the control device of the rolling mill according to the present invention is the regressor using the silicon equivalent as the operation parameter for the steelmaking process and the soaking tropical furnace time as the operation parameter for the heating process. It is characterized by constructing.

In the above-mentioned invention, the control device of the rolling apparatus according to the present invention describes the carbon equivalent of the material to be rolled, the silicon equivalent of the material to be rolled, and the refining and casting of the material to be rolled as operating parameters related to the steelmaking process. It is characterized in that the recirculator is constructed by using the caster number, the soaking time in the tropical furnace, and the extraction temperature of the heating furnace as operating parameters related to the heating process.

The control device for the rolling apparatus according to the present invention is the above-mentioned invention, wherein the rolling apparatus rolls a material to be rolled, which has been subjected to width reduction using a width reduction device in a width reduction step after a heating step. The prediction unit uses an equation constructed by machine learning using at least one operation parameter from each of the operation parameter group related to the steelmaking process, the operation parameter group related to the heating process, and the operation parameter group related to the width rolling process. It is characterized in that the camber amount of the material to be rolled on the entry side of the rolling apparatus is predicted.

In the above invention, the control device of the rolling mill according to the present invention relates to silicon equivalent as operating parameters related to the steelmaking process, soaking time in a tropical furnace, and the width reduction process as operating parameters related to the heating process. It is characterized in that the regressionr is constructed using the width rolling reduction amount as an operation parameter.

In the above-mentioned invention, the control device of the rolling apparatus according to the present invention describes the carbon equivalent of the material to be rolled, the silicon equivalent of the material to be rolled, and the refining and casting of the material to be rolled as operating parameters related to the steelmaking process. Casting machine number, as operating parameters related to the heating process, soaking time in tropical furnace, and heating furnace extraction temperature, as operating parameters related to the width reduction process, width reduction amount, mold usage amount, rolling before width reduction It is characterized in that the reversing device is constructed using the plate width of the material and the plate thickness of the material to be rolled before the width reduction.

In the above invention, the control device for the rolling apparatus according to the present invention further includes a setting unit for setting the leveling amount of the rolling mill according to the rolling down leveling amount calculated by the calculation unit, and the calculation unit further includes the prediction unit. The rolling apparatus using the camber amount of the material to be rolled predicted by the section and the influence coefficient representing the relationship between the change in the camber amount on the exit side of each rolling mill with respect to the operation of the rolling leveling amount in each rolling mill. The amount of rolling down leveling that can control the camber amount of the material to be rolled on the exit side of the final rolling pass is calculated, and the control unit calculates the rolling milling amount based on the amount of rolling down leveling set by the setting unit. It is characterized by controlling.

The method for producing a steel sheet according to the present invention is characterized by including a step of producing a steel sheet by using the method for controlling a rolling apparatus according to the above invention.

According to the present invention, in a rolling process of manufacturing a steel plate by rolling a material to be rolled, which has been refined and cast in a steelmaking process and heated in a heating furnace, by a rolling apparatus in a plurality of passes, the final rolling pass of the rolling apparatus. It is possible to suppress the camber of the material to be rolled on the output side.

FIG. 1 is a diagram showing a configuration example of a rolling apparatus according to an embodiment. FIG. 2 is a diagram for explaining the rolling reduction amount of the rolling mill. FIG. 3 is a diagram for explaining the camber amount of the material to be rolled. FIG. 4 is a flowchart showing an example of a control method of the rolling apparatus according to the embodiment.

Hereinafter, a method for controlling a rolling apparatus, a control apparatus for a rolling apparatus, and a method for manufacturing a steel sheet according to an embodiment of the present invention will be described with reference to the drawings. In the present embodiment, a rough rolling apparatus for a hot rolling line is exemplified as an example of a rolling apparatus to which the present invention is applied, but the present invention is not limited to the present embodiment. In addition, it should be noted that the drawings are schematic, and the dimensional relationship of each element, the ratio of each element, etc. may differ from the actual ones. Even between drawings, there may be parts with different dimensional relationships and ratios. Further, in each drawing, the same components are designated by the same reference numerals.

[1. Hot rolling line]
FIG. 1 is a diagram showing a configuration example of a rolling apparatus according to an embodiment. The control device 1 controls the reduction leveling amount of each rolling pass of the rough rolling apparatus 10 that sequentially performs rolling of a plurality of passes on a plurality of materials to be rolled on the hot rolling line 100. The hot rolling line 100 is arranged on the width reduction device 30, the heating furnace 40, the continuous casting machine 50, and the downstream side of the rough rolling apparatus 10 arranged in the transport path on the upstream side of the rough rolling apparatus 10. Equipment such as finish rolling equipment (not shown) is included. Continuous casting using the continuous casting machine 50 is included in the steelmaking process. The steelmaking process is a processing process performed in the order of hot metal pretreatment, converter, secondary refining, and continuous casting. For example, the material to be rolled that has been refined and cast by the continuous casting machine 50 is heated in the heating furnace 40, then sequentially conveyed along the transfer path, passed through the width reduction device 30, and then rolled by the rough rolling device 10. Is done. The material to be rolled after multiple-pass rolling (rough rolling) in the hot rolling process passes through various facilities of the hot rolling line 100 such as a finish rolling apparatus, and is then coiled by a coiler (not shown). The transport path is for sequentially transporting a plurality of materials to be rolled in the hot rolling line 100, and is composed of a plurality of transport rolls (not shown).

In the hot rolling line 100, the material 20 to be rolled is bent in the horizontal direction due to various factors such as a temperature deviation in the width direction of the material 20 to be rolled, a plate thickness deviation, and uneven opening in the width direction of the rolling roll. So-called camber occurs. If the camber amount of the material 20 to be rolled is large, equipment such as rolling rolls and side guides may be damaged. Further, when the camber amount generated by rough rolling is large, the material to be rolled 20 collides with the side guide and the edge portion is formed when the tail end portion of the material to be rolled 20 is pulled out in the finishing rolling mill (not shown). Rolling in a folded state, so-called "narrowing down", may occur.

In the width reduction device 30, the main factors that cause camber are temperature deviation in the plate width direction and off-center to the press die. The temperature deviation in the plate width direction occurs when the material 20 to be rolled heated in the heating furnace 40 is extracted and cooled from the extraction side by the inflowing outside air when the extraction door is opened. Further, in the heating furnace 40, the plate width is different because the heating state at the tip and tail end portion is different from that at the longitudinal central portion depending on whether the material 20 to be rolled having different lengths is adjacent or the charging position in the longitudinal direction of the material 20 to be rolled. A directional temperature deviation occurs. In addition, when the material 20 to be rolled is conveyed in an off-center state with respect to the press die and is subjected to width reduction, uneven reduction on the left and right occurs, and camber and wedges are generated.

On the other hand, in order to suppress camber, it is necessary to comprehensively consider the temperature difference in the width direction of the material 20 to be rolled and the operation parameters in the steelmaking process, heating process, width reduction process, and rolling process of the material 20 to be rolled. be. However, the parameters that can be directly used as a physical model are very limited in the operation parameter group. Conventionally, the camber amount on the exit side of the rolling mill has been predicted by using the temperature deviation in the plate width direction of the material 20 to be rolled, the camber amount that changes depending on the leveling operation in the rolling mill, and the like. However, the prediction is performed using a parameter group of only a limited part using a physical model, and the prediction accuracy of the camber amount is insufficient.

Therefore, the present inventors have analyzed in detail the relationship between the occurrence of camber in rough rolling in the hot rolling process and the operating conditions during rolling, and diligently studied a method for preventing camber. It has been found that it is possible to predict the camber on the rolling mill entry side with high accuracy by using at least one of the operation parameter groups related to the steelmaking process, the heating process, and the width rolling reduction process. Furthermore, it was found that the camber amount on the exit side of the rolling mill can be suppressed by changing the leveling control amount in the rough rolling mill by using the camber prediction method after width rolling by the regression machine. That is, in the present invention, an unprecedented improvement in accuracy has been confirmed by predicting the camber amount and setting the leveling by using a regression device that can comprehensively consider each operation parameter of the material 20 to be rolled. In addition, when it is simply described as "rolling machine entry side", it means the first rolling pass entry side of the rough rolling apparatus 10. Similarly, when simply described as "rolling machine exit side", it means the final rolling path exit side of the rough rolling apparatus 10.

[1-1. Rolling equipment to be controlled]
The rough rolling apparatus 10 performs a plurality of passes rolling (rough rolling of a plurality of rolling passes) with a total number of rolling passes N on the material 20 to be rolled, and the plurality of rolling machines share the multiple pass rolling. It is configured. As shown in FIG. 1, the rough rolling apparatus 10 is a group of rolling mills composed of all four stands of rolling mills, and the first rolling mills 11 and 2 are arranged in order along the transport direction of the material 20 to be rolled. A rolling mill 12, a third rolling mill 13, and a fourth rolling mill 14 are provided. The first to fourth rolling mills 11 to 14 are four-stage rolling mills each having a pair of rolling rolls and a pair of backup rolls. The pair of rolling rolls are arranged at positions facing the thickness direction D1 of the material 20 to be rolled with a transport roll (not shown) interposed therebetween. The transport direction of the material 20 to be rolled is the same as the longitudinal direction D2 of the material 20 to be rolled.

Further, the first to fourth rolling mills 11 to 14 each include first to fourth rolling mills 11a to 14a. Each rolling device 11a, 12a, 13a, 14a adjusts the rolling mills 11, 12, 13, 14 corresponding to each of the rolling mills 11a, 12a, 13a, 14a. As shown in FIG. 2, reduction leveling amount Lv _i is defined as the difference between the amount of reduction in between the ends of the rolling roll 11b, 11c roll axial direction of rolling a material being rolled 20 (pressure level). The definition of this reduction leveling amount Lv _i is the same for the first to fourth rolling mills 11 to 14. For example, when adjusting the reduction leveling amount of the first rolling mill 11, the first reduction device 11a is controlled, and when adjusting the reduction leveling amount of the second rolling mill 12, the second reduction device 12a is controlled. That is, the control device 1 can individually control the rolling leveling amount of each rolling mill 11-14.

[1-2. Control device]
The control device 1 includes a camber amount measuring unit 2, a storage unit 3, an arithmetic processing unit 4, and a control unit 5.

The camber amount measuring unit 2 measures the camber amount on the rolling path exit side by a plurality of rolling devices to be controlled, and optically detects and stores the camber generation direction and the generation amount by an imaging device or the like. It is transmitted to the unit 3 and the arithmetic processing unit 4. The amount of camber is defined as the curvature. For example, the bending state of the camber of the material 20 to be rolled can be distinguished such that the bending of the rolling roll of the rolling mill toward the working side is positive and the bending toward the driving side is negative. The details of the camber amount will be described later with reference to FIG.

The storage unit 3 stores various information necessary for controlling the reduction leveling amount of each rolling path. As shown in FIG. 1, the storage unit 3 stores the influence coefficient table 3a. Influence coefficient table 3a is a data table that contains the influence coefficient K _i used for the calculation of reduction leveling control. Storage unit 3, for the reduction leveling control of the rolling paths for rough rolling device 10, holds the influence coefficient table 3a, to manage, the influence coefficient K _i required processing in response to a request from the arithmetic processing section 4 It is provided to the arithmetic processing unit 4.

Influence coefficient K _i is a coefficient indicating the degree of reduction leveling operation for each rolling pass is affected by changes in the camber of each rolling pass the delivery side. Influence coefficient K _i is the camber amount of change in the i th i th affected reduction leveling operation rolling pass rolling pass exit side (∂Cam _i), reduction leveling by reduction leveling operation of the i-th rolling pass the ratio of the amount of change _{(∂Lv i) (∂Cam i /} ∂Lv i), that is represented by the following formula (1).

In the formula (1), reduction leveling amount Lv _i is the reduction leveling amount before reduction leveling operation for the i-th rolling pass. The reduction leveling operation amount dLv is the reduction leveling change amount due to the reduction leveling operation of the i-th rolling pass. The camber amount Cam _i (Lv _i ) is the camber amount on the exit side of the i-th rolling pass of the material 20 to be rolled when the rolling leveling amount of the i-th rolling pass is Lvi _i. The camber amount Cam _i (Lv _i + dLv) is the camber amount on the exit side of the i-th rolling pass of the material 20 to be rolled when the reduction leveling amount of the i-th rolling pass is (Lv _{i + dLv).}

Influence coefficient K _i can be obtained as follows. For example, when the rough rolling apparatus 10 is performing a plurality of passes rolling with a total number of rolling passes N on the material 20 to be rolled, one rolling mill is actually operated for reduction leveling, and before and after this reduction leveling operation, Each camber amount on the rolling pass exit side of the material to be rolled after rolling is measured. Each camber amount before and after the reduction leveling operation obtained in this way, and the reduction leveling amount and the reduction leveling operation amount before the reduction leveling operation at that time are substituted into the equation (1). Thus, the influence coefficient _{K i} can be identified from equation (1) (calculated).

Such multiple influence coefficient K _i obtained by the rolling mill for each of the rough rolling device 10, each width direction temperature deviation before rolling of the rolled material 20 is set for each rolling conditions of the rolled material 20. The plurality of influence coefficients _Ki are stored in the influence coefficient table 3a of the storage unit 3. As a result, the influence coefficient table 3a shows a plurality of influences associated with the rolling mills 11 to 14 of the rough rolling apparatus 10, the temperature deviation in the width direction of the material 20 to be rolled before rolling, and the rolling conditions of the material 20 to be rolled. The coefficient _Ki is included. In this embodiment, rolling conditions of the rolled material 20 to be associated with the influence coefficient K _i, for example, plate thickness and plate width targeted after rolling to be set to the rolled material 20, the rolling mill 11 to 14 The rolling reduction ratio set in, the material strength of the material 20 to be rolled, and the like.

The arithmetic processing unit 4 outputs the predicted camber amount and the predicted camber amount by the relational expression between the predicted value of the camber amount on the rolling mill entry side constructed by using the regression machine and each operation parameter. the reduction leveling amount by each path rolling at controllable rough rolling device 10 to the control target value of the side is calculated using the influence coefficient K _i. That is, the reduction leveling amount is calculated (set) for each rolling pass by the arithmetic processing unit 4.

The control unit 5 executes reduction leveling control of each rolling pass of the rough rolling apparatus 10 that performs a plurality of rolling passes on the material 20 to be rolled. The control unit 5 controls the rolling devices 11a, 12a, 13a, 14a of the rolling mills 11, 12, 13, 14 based on the rolling mill leveling amount calculated by the arithmetic processing unit 4. The control unit 5 controls the rolling pass leveling amount at the time of the plurality of rolling passes by the rough rolling apparatus 10 through the control of the plurality of rolling gears 11a to 14a.

[2. Camber amount of material to be rolled]
Here, the camber amount will be described with reference to FIG. Figure 3 is a diagram for explaining a camber amount Cam _i of the rolled material 20. The longitudinal direction D2 of the material 20 to be rolled is the same as the transport direction of the material 20 to be rolled, the tip side (forward direction) of the material 20 to be rolled is positive, and the tail end side (reverse direction) is negative. .. The width direction D3 of the material 20 to be rolled is the same as the roll axis direction of the transport roll and the roll axis direction of each rolling roll of each rolling mill 11-14. Further, in the width direction D3, the left side (working side) is positive and the right side (drive side) is negative when facing the positive side in the longitudinal direction D2. Further, the thickness direction (plate thickness direction) D1, the longitudinal direction D2, and the width direction D3 are directions perpendicular to each other.

As shown in FIG. 3, the camber amount Cam _i of the rolled material 20, a bending amount of the positive or negative side in the width direction D3 with respect to the longitudinal direction D2 of the material to be rolled 20 at each rolling pass the delivery side of the multiple passes rolling Defined. Specifically, the camber amount Cam _i is defined as the maximum value of the distance between the center position S1 in the width direction of the material 20 to be rolled and the reference position S2 of the material 20 to be rolled on the exit side of the i-th rolling pass.

The sign of the positive and negative camber amount Cam _i (occurrence direction of the camber) is determined by the relationship between the positional deviation of the direction and the width direction D3 of the widthwise center position S1 to the reference position S2 of the material to be rolled 20. In the rolled material 20 shown in FIG. 3, since the widthwise center position S1, is located offset to the positive side in the width direction D3 (working side) relative to the reference position S2, the camber amount Cam _i is a positive value become. That is, although not shown, if the widthwise center position S1 is located offset to the negative side in the width direction D3 with respect to the reference position S2, the camber amount Cam _i is a negative value. The reference position S2 is represented by a straight line (reference line) passing through the width direction center position Wa at the tip portion 20a of the material to be rolled 20 and the width direction center position Wb at the tail end portion 20b. Camber amount Cam _i will distance between the widthwise center position S1 and reference position S2 at the center position in the longitudinal direction D2 of the material to be rolled 20.

[3. Rolling equipment control method]
Next, a control method of the rolling mill will be described with reference to FIG. FIG. 4 is a flowchart showing an example of the control method in the embodiment. The control method shown in FIG. 4 is executed by the control device 1. The control device 1 sequentially executes steps S101 to S103 shown in FIG.

As shown in FIG. 4, the arithmetic processing unit 4 predicts the camber amount of the material to be rolled 20 on the entry side of the rough rolling apparatus 10 (step S101). The arithmetic processing unit 4 predicts the camber amount on the entrance side of the rolling mill using a model constructed in advance.

In step S101, at least the camber amount of the material 20 to be rolled and the amount of camber to be rolled on the side of the material 20 to be rolled are the characteristic amounts used when calculating (predicting) the camber amount of the material 20 to be rolled on the entry side of the rough rolling apparatus 10. A relational expression for obtaining the camber amount on the rolling mill entrance side of the material 20 to be rolled is created by using the operation parameter group related to each specifications, the steelmaking process, the heating process, and the width reduction process obtained at the time of extracting the material 20 from the heating furnace. That is, this relational expression is an expression constructed by machine learning using at least one operation parameter from each of the operation parameter related to the steelmaking process, the operation parameter related to the heating process, and the operation parameter related to the width reduction process as the explanatory variables of the regression device. Is. All the features may be used for constructing the relational expression, or the features may be selected by using a method such as principal component analysis or contribution analysis. Further, as the feature quantity, not only quantitative variables such as temperature and dimensions but also qualitative variables such as extraction furnace number, cast strand number and steel grade can be used.

For example, the following features are selected from the importance of variables by machine learning using XGBost, which is one of the machine learning methods. Specifically, the operating parameters related to the steelmaking process include the carbon equivalent (C equivalent) of the material 20 to be rolled, the silicon equivalent (Si equivalent) of the material 20 to be rolled, and the number of the casting machine in which the material 20 to be rolled is refined and cast. As the operating parameters related to the heating process, the soaking time in the tropical furnace, the extraction temperature of the heating furnace, and the operating parameters related to the width rolling process, the width rolling down amount, the mold usage amount, and the plate width and width of the material 20 to be rolled before the width rolling down. The plate thickness of the material 20 to be rolled before rolling is selected.

Among the above-mentioned operating parameter groups, the characteristic quantities that contribute highly in each process are the silicon equivalent of the material to be rolled 20 as the operating parameters related to the steelmaking process, and the soaking time in the tropical furnace and the width reduction as the operating parameters related to the heating process. The width rolling reduction amount is selected as the operation parameter related to the process.

Next, all of the above-mentioned feature amounts and the camber amount on the rolling mill entry side were input to the regressionr to create a relational expression between each feature amount and the camber amount on the rolling mill entry side of the material 20 to be rolled. The input data was normalized by the maximum value and the minimum value, learned with 7,000 coils in the operation data of 10,000 coils, and the model prediction accuracy was verified with the remaining 3000 coils. The amount of camber on the exit side of the rolling mill of 10,000 coils had an average error of 15.1 mm and a standard deviation of 28.6 mm. First, a neural network was used as a regressor, and here, the intermediate layer was set to 3 layers, and the number of nodes was set to 100 each. The sigmoid function was used as the activation function. The model prediction accuracy was an error average of 2.2 mm and a standard deviation of 19.1 mm.

Similarly, it was created using XGBost as a regression device. Here, the learning rate is 0.2, the minimum value of loss reduction is 0.1, the maximum value of the tree depth is 12, the minimum weight of the child node is 1, and the number of boosting is 500. The model prediction accuracy was an error average of 0.2 mm and a standard deviation of 15.3 mm. From this result, as a regressor for predicting the camber amount after rolling using each feature amount, the prediction using XGBost is more advantageous.

However, the regressionr using machine learning is not particularly limited to that method. For example, decision trees, random forests, support vector regression, Gaussian processes, neural networks, XGBost, etc. can be used. Furthermore, ensemble machine learning that combines at least two or more regressionrs can also be used.

After executing step S101, the control device 1 calculates the rolling-down leveling amount of each rolling pass at the time of multiple-pass rolling by the rough rolling apparatus 10 (step S102).

The settable range of the reduction leveling amount of the rough rolling apparatus 10 is set to the settable range of each reduction leveling amount of the first to fourth rolling mills 11 to 14 (hereinafter, abbreviated as "the settable range of the reduction leveling amount"). Equivalent to. That is, each rolling mill has an upper limit value and a lower limit value that can be set based on the structure (equipment specifications) of each rolling mill, and the upper limit of the rolling milling amount that can be set for each rolling mill. A finite range below the value and above the lower limit is the settable range for the rolling down leveling amount. Therefore, the arithmetic processing unit 4 calculates the reduction leveling operation amount that can be executed within the settable range of the reduction leveling amount.

After executing step S102, the control device 1 controls the reduction leveling amount of each rolling pass when the material to be rolled 20 is rolled in a plurality of passes by the rough rolling apparatus 10 (step S103). In step S103, the control unit 5 controls each rolling device 11a, 12a, 13a, 14a based on the rolling pass reduction leveling amount calculated in step S102 described above, and the rough rolling apparatus 10 performs a plurality of rolling devices. The amount of reduction leveling of each rolling pass during pass rolling is controlled. Then, after executing step S103, the control device 1 ends this process. Further, the present invention can also be implemented as a method for manufacturing a steel sheet by, for example, rolling the material 20 to be rolled in the hot rolling process by the control method of the rolling apparatus in the above-described embodiment. That is, the method for manufacturing the steel sheet in the embodiment includes the step of manufacturing the steel sheet by using the method for controlling the rolling apparatus described above.

The hot rolling line 100 does not have to be provided with the width reduction device 30. In this case, it is possible to carry out the above-mentioned calculation process without using the operation parameters related to the width reduction process. For example, it is a hot rolling line in which the width reduction device 30 is removed from the hot rolling line 100 shown in FIG. 1 described above. In the hot rolling line 100 not provided with the width reduction device 30, the width reduction process is not performed. Therefore, in order to suppress camber, the temperature difference in the width direction of the material 20 to be rolled, the steelmaking process of the material 20 to be rolled, and the heating process are performed. , The operation parameters in the rolling process may be comprehensively taken into consideration when deciding. That is, the operation parameter group related to the width rolling process is omitted from the feature amount when calculating the camber amount of the material 20 to be rolled on the outlet side of the rough rolling apparatus 10 and setting the reduction leveling amount of each rolling mill 11 to 14. .. In short, using the operation parameter group related to the steelmaking process, heating process, and rolling process of the material 20 to be rolled, the rolling down leveling amount in each rolling mill that can control the camber amount on the exit side of the rolling mill to the control target value is calculated. A machine-learned returner may be constructed, and each rolling mill may be controlled based on the rolling reduction amount calculated using this returner.

In the hot rolling line 100 not provided with the width reduction device 30, one of the factors causing camber on the exit side of the rolling mill in the rough rolling device 10 is a temperature deviation in the plate width direction. The temperature deviation in the plate width direction is cooled from the extraction side by the inflowing outside air when the material to be rolled 20 heated in the heating furnace 40 is extracted, even if the width reduction device 30 is not provided. It is caused by being rolled. Further, in the heating furnace 40, the plate width is different because the heating state at the tip and tail end portion is different from that at the longitudinal central portion depending on whether the material 20 to be rolled having different lengths is adjacent or the charging position in the longitudinal direction of the material 20 to be rolled. A directional temperature deviation occurs.

Here, the effect of the present invention will be described with reference to verification examples (Examples 1 to 3) assuming a case where the above-described embodiment in the hot rolling process is applied.

(Example 1)
The present invention was verified on a hot rolling line 100 having a width rolling device 30 provided with a mold for rolling the material 20 to be rolled in the width direction and a four-stage rolling mill composed of a working roll and a reinforcing roll. In Example 1, the rolling target was a slab having a length of 6000 to 8000 mm, a thickness of 260 mm, and a width of 900 to 1500 mm. The thickness is 190-210 mm.

Further, in Table 1, as a conventional technique, the leveling amount is set as the initial value so that the roll gap difference at no load is not different between the drive side and the work side before the start of rolling, and the leveling amount is constant during rolling. (Conventional example: constant) and a method of performing a leveling operation based on visual information by an operator (conventional example: hand) will be illustrated. In addition, as an example using machine learning, Table 1 shows a neural network using all of the features shown in the embodiment after extraction from the heating furnace and before the start of rolling for the target material whose camber should be controlled. An example of a method of calculating the camber amount after width rolling (Example: XGBost: total feature amount) using a camber prediction model constructed using the above and a camber amount prediction model constructed using XGBost. Further, in Table 1, among the selected feature quantities, the feature quantities having a high contribution in each step are limited to the silicon equivalent in the steelmaking step, the soaking time in the tropical furnace in the heating step, and the width reduction amount in the width reduction step. Then, it is calculated using the predicted amount of camber amount constructed by using XGBost, and the leveling operation amount that becomes the control target camber amount is calculated using the preset influence coefficient with respect to the calculated camber amount. A method of determining and setting the reduction leveling (Example: XGBost: total feature amount) is exemplified.

Then, in the example shown in Table 1, the control target value was set to 0 mm. As shown in Table 1, when a total of 500 coils were rolled by applying this technology, the leveling setting method using all the features by XGBost was the most dominant, and the camber amount was 0 mm on average and the standard deviation was 18 mm. .. As described above, according to Example 1, it was confirmed that the camber amount can be controlled according to the control target value by applying the hot rolling method according to the present invention.

(Example 2)
Similar to the first embodiment, the present invention is performed on a hot rolling line 100 having a width rolling device 30 provided with a mold for rolling the material 20 to be rolled in the width direction and a four-stage rolling mill composed of a working roll and a reinforcing roll. The invention was verified. In the second embodiment, the rolling target is a slab having a length of 6000 to 8000 mm and a slab thickness of 230 mm and a width of 900 to 1500 mm. The plate thickness is 160 to 180 mm.

Further, in Table 2, as a conventional technique, the leveling amount is set as the initial value so that the roll gap difference at no load is not different between the drive side and the work side before the start of rolling, and the leveling amount is constant during rolling. (Conventional example: constant) and a method of performing a leveling operation based on visual information by an operator (conventional example: hand) will be illustrated. Further, in Table 2, as an example using machine learning, rolling down leveling is performed using a prediction model of the rolling mill entry side camber constructed using all the features described in the embodiment using XGBost. An example of the set method (Example: XGBost: total feature amount).

Then, in the example shown in Table 2, the control target value was set to 0 mm. As shown in Table 2, when a total of 500 coils were rolled by applying this technology, the leveling setting method using all the features by XGBost was the most dominant, and the camber amount was 1 mm on average and the standard deviation was 16 mm. .. As described above, according to Example 2, it was confirmed that the camber amount can be controlled according to the control target value by applying the hot rolling method according to the present invention.

(Example 3)
Similar to the first embodiment, the present invention is performed on a hot rolling line 100 having a width rolling device 30 provided with a mold for rolling the material 20 to be rolled in the width direction and a four-stage rolling mill composed of a working roll and a reinforcing roll. The invention was verified. In the third embodiment, the rolling target is a slab having a length of 6000 to 8000 mm, a slab thickness of 260 mm, and a width of 900 to 1500 mm. The plate thickness of the apparatus 10 on the exit side of the final rolling pass was 190 to 210 mm. That is, the third embodiment is an embodiment targeting a manufacturing process in which the width reduction process is not essential, that is, a hot rolling line 100 not provided with the width reduction device 30.

Further, in Table 3, as a conventional technique, the leveling amount is set as the initial value so that the roll gap difference at no load is not different between the drive side and the work side before the start of rolling, and the leveling amount is constant during rolling. (Conventional example: constant) and a method of performing a leveling operation based on visual information by an operator (conventional example: hand) will be illustrated. Further, in Table 3, as an example using machine learning, a camber prediction model constructed using all of the features described in the embodiment using XGBost and a rolling mill leveling set value are used. , A method in which rolling down leveling is set (Example: XGBost: total feature amount) is illustrated.

Then, in the example shown in Table 3, the control target value was set to 0 mm. As shown in Table 3, when a total of 500 coils were rolled by applying this technology, the leveling setting method using all the features by XGBost was the most dominant, and the camber amount was 0 mm on average and the standard deviation was 16 mm. .. As described above, according to Example 3, it was confirmed that by applying the hot rolling method according to the present invention, the camber amount can be controlled according to the control target value even when the rolling conditions are different. As described above, according to the present invention, the camber amount can be controlled to a low level, and as a result, it is possible to reduce a decrease in the operating rate and equipment damage due to poor threading. By applying the hot rolling method according to the present invention, it was confirmed that the camber amount can be controlled according to the control target value.

As described above, according to the embodiment, the material 20 to be rolled, which has been refined and cast in the continuous casting machine 50 and heated in the heating furnace 40, is rolled in a plurality of passes by the rough rolling apparatus 10 to produce a steel sheet. In the rolling process, it is possible to suppress the camber of the material to be rolled 20 on the exit side of the final rolling pass of the rough rolling apparatus 10.

The present invention is not limited to the above-described embodiment, and can be appropriately modified without departing from the object of the present invention. For example, the rolling apparatus to be controlled may be a rolling apparatus other than the rough rolling apparatus such as a finishing rolling apparatus.

Further, the rolling mill to be controlled (rough rolling mill 10) is not limited to a configuration including a rolling mill having a total of 4 stands or a total of 5 stands. The number of rolling mills (number of stands) constituting the rolling apparatus to be controlled may be one or a plurality as long as the material 20 to be rolled can be rolled in a plurality of passes. That is, in the present invention, the number of stands of the rolling mill constituting the rolling mill to be controlled is not particularly limited. In addition, the number of roll stages of the rolling mill is not limited to four, and may be a desired number of stages. That is, the number of roll stages of the rolling mill and the total number of rolling passes for multi-pass rolling are not particularly limited.

Further, the rolling mill constituting the rolling apparatus is a rolling mill (irreversible rolling mill) that rolls the material to be rolled in the forward direction, or is rolled in the forward direction and in the reverse direction. It is not particularly limited whether it is a rolling mill (reversible rolling mill) performed on the material 20. That is, in the present invention, the rolling mill to be controlled may include a plurality of rolling mills all of which are irreversible or reversible, or one or more irreversible rolling mills and one or more reversible rolling mills. It may be equipped with a rolling mill of the same type. Not only when the first rolling mill is a reversible rolling mill as in the second embodiment described above, but also when the second and subsequent rolling mills are reversible, the reversible rolling mill is used. The rolling in the forward direction and the rolling in the reverse direction may be performed by any of the multiple-pass rolling with respect to the material to be rolled 20. In this way, regardless of whether the rolling mill is a reversible type or a lossy type, if the camber amount of the material to be rolled is measured on the first rolling pass entry side in the rolling apparatus, the camber amount due to the subsequent multi-pass rolling is performed. Can be predicted.

Further, the definition of the positive and negative of the camber amount (definition of the camber generation direction) may be negative on the working side and positive on the driving side, contrary to the above-described embodiment. That is, the bending state of the camber of the material 20 to be rolled may be distinguished such that the bending toward the driving side of the rolling roll of the rolling mill is positive and the bending toward the working side is negative.

Further, the present invention also includes a configuration in which the above-mentioned components are appropriately combined. In addition, other embodiments, examples, operational techniques, and the like made by those skilled in the art based on the above-described embodiments are all included in the scope of the present invention.

1 Control device 2 Camber amount measurement unit 3 Storage unit 4 Arithmetic processing unit 5 Control unit 10 Rough rolling equipment 11, 12, 13, 14 1st to 4th rolling mills 11a, 12a, 13a, 14a 1st to 4th rolling mills 11b, 11c Rolling roll 20 Material to be rolled 30 Width reduction device 40 Heating furnace 50 Continuous casting machine 100 Hot rolling line

Claims (15)

A method for controlling a rolling apparatus when a material to be rolled, which is refined and cast in a steelmaking process and heated in a heating furnace, is rolled by using a rolling apparatus consisting of a plurality of rolling mills.
As an explanatory variable of the regressor, the rolling apparatus is used by using an equation constructed by machine learning using at least one operation parameter group of the operation parameter group related to the steelmaking process of the material to be rolled and the operation parameter group related to the heating process. A prediction step for predicting the camber amount of the material to be rolled on the entry side of
A calculation step for calculating the rolling leveling amount of each rolling mill based on the predicted camber amount of the material to be rolled, and
A method for controlling a rolling apparatus, which comprises a control step for controlling each rolling mill based on a calculated reduction leveling amount. The operating parameters used in the prediction step are
As an operation parameter related to the steelmaking process, silicon equivalent,
The method for controlling a rolling apparatus according to claim 1, wherein the operating parameters related to the heating step include a soaking time in a tropical furnace. The operating parameters used in the prediction step are
The operating parameters related to the steelmaking process include the carbon equivalent of the material to be rolled, the silicon equivalent of the material to be rolled, and the number of the casting machine in which the material to be rolled is refined and cast.
The method for controlling a rolling apparatus according to claim 1 or 2, wherein the operating parameters related to the heating step include a soaking time in a tropical furnace and an extraction temperature in a heating furnace. The rolling apparatus rolls the material to be rolled, which has been subjected to width reduction using the width reduction device in the width reduction process after the heating step.
The prediction step uses an equation constructed by machine learning using at least one operation parameter from each of the operation parameter group related to the steelmaking process, the operation parameter group related to the heating process, and the operation parameter group related to the width rolling process. The method for controlling a rolling apparatus according to claim 1, further comprising a step of predicting a camber amount of the material to be rolled on the entry side of the rolling apparatus. The operating parameters used in the prediction step are
As an operation parameter related to the steelmaking process, silicon equivalent,
As an operation parameter related to the heating process, the soaking time in the tropical furnace,
The control method for a rolling apparatus according to claim 4, wherein the operation parameter related to the width reduction step includes a width reduction amount. The operating parameters used in the prediction step are
The operating parameters related to the steelmaking process include the carbon equivalent of the material to be rolled, the silicon equivalent of the material to be rolled, and the number of the casting machine in which the material to be rolled is refined and cast.
The operating parameters related to the heating process include the soaking time in the tropical furnace and the extraction temperature of the heating furnace.
4. 5. The method for controlling a rolling apparatus according to 5. Further including a setting step of setting the leveling amount of the rolling mill according to the rolling down leveling amount calculated by the calculation step.
The calculation step uses the predicted camber amount of the material to be rolled on the rolling mill entry side and the influence coefficient representing the relationship between the change in the camber amount on the rolling mill exit side with respect to the operation of the rolling down leveling amount in the rolling mill. Including the step of calculating the rolling down leveling amount in which the camber amount of the material to be rolled on the exit side of the rolling mill can be controlled to the target value.
The rolling apparatus according to any one of claims 1 to 6, wherein the control step includes a step of controlling each rolling mill based on the rolling reduction amount set by the setting step. Control method. A control device for a rolling apparatus that rolls a material to be rolled that has been refined and cast in a steelmaking process and heated in a heating furnace using a rolling apparatus consisting of a plurality of rolling mills.
As an explanatory variable of the regression device, there is a regression device constructed by machine learning using at least one operation parameter group of the operation parameter group related to the steelmaking process of the material to be rolled and the operation parameter group related to the heating process, and the regression A prediction unit that predicts the camber amount of the material to be rolled on the entry side of the rolling apparatus using a vessel, and a prediction unit.
A calculation unit that calculates the rolling leveling amount of each rolling mill based on the predicted camber amount of the material to be rolled.
A control device for a rolling apparatus provided with a control unit for controlling each rolling mill based on the calculated rolling leveling amount of each rolling mill. The rolling according to claim 8, wherein the prediction unit constructs the regressionr using silicon equivalent as an operation parameter related to the steelmaking process and the soaking temperature in the tropical furnace as an operation parameter related to the heating process. Device control device. The prediction unit sets the carbon equivalent of the material to be rolled, the silicon equivalent of the material to be rolled, the number of the casting machine from which the material to be rolled is refined and cast, as the operation parameters related to the steelmaking process, and averages them as the operation parameters related to the heating process. The control device for a rolling mill according to claim 8 or 9, wherein the recirculator is constructed using the tropical furnace time and the heating furnace extraction temperature. The rolling apparatus rolls the material to be rolled, which has been subjected to width reduction using the width reduction device in the width reduction process after the heating step.
The prediction unit uses an equation constructed by machine learning using at least one operation parameter from each of the operation parameter group related to the steelmaking process, the operation parameter group related to the heating process, and the operation parameter group related to the width rolling process. The control device for a rolling apparatus according to claim 8, further comprising predicting the camber amount of the material to be rolled on the entry side of the rolling apparatus. The prediction unit constructs the regressionr using silicon equivalent as an operation parameter related to the steelmaking process, soaking time in a tropical furnace as an operation parameter related to the heating process, and width rolling amount as an operation parameter related to the width rolling process. 11. The control device for a rolling mill according to claim 11. The prediction unit sets the carbon equivalent of the material to be rolled, the silicon equivalent of the material to be rolled, the number of the casting machine from which the material to be rolled is refined and cast, as the operation parameters related to the steelmaking process, and averages them as the operation parameters related to the heating process. Tropical furnace time, heating furnace extraction temperature, and operating parameters related to the width reduction process include width reduction amount, mold usage amount, plate width of the material to be rolled before width reduction, and plate of the material to be rolled before width reduction. The control device for a rolling mill according to claim 11 or 12, wherein the returner is constructed using the thickness. A setting unit for setting the leveling amount of the rolling mill according to the rolling down leveling amount calculated by the calculation unit is further provided.
The calculation unit includes an influence coefficient representing the relationship between the camber amount of the material to be rolled predicted by the prediction unit and the change in the camber amount on the exit side of each rolling mill with respect to the operation of the rolling leveling amount in each rolling mill. To calculate the amount of rolling down leveling that can control the camber amount of the material to be rolled on the exit side of the final rolling pass of the rolling apparatus to a target value.
The control device for a rolling apparatus according to any one of claims 8 to 13, wherein the control unit controls each rolling mill based on a reduction leveling amount set by the setting unit. .. A method for manufacturing a steel sheet, which comprises a step of manufacturing a steel sheet by using the method for controlling a rolling apparatus according to any one of claims 1 to 7.

Images