JP3396774B2 - Shape control method - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a hot-rolling plant for steel sheets, which is formed into a desired shape when the hot-rolled steel sheet passes through a finish rolling mill located on the most downstream side of the rolling process. It relates to a control method.

[0002]

2. Description of the Related Art Shape control of a hot-rolled steel sheet into a target shape from the tip to the tail of a coil is performed by using a bender installed in a finishing rolling mill located on the most downstream side of the rolling process.
This is an important control for maintaining the quality of the steel sheet. FIG. 7 is a view showing a shape control device for a general finish rolling mill. In the figure, 20 is a hot rolled steel sheet, 21 is a flatness meter for measuring the flatness, 22 is a backup roll, and 23 is a hot rolled steel sheet 20.
Work roll with a bender for controlling the shape of the work roll, 24 is the final stand of the finishing mill, 25 is the winder, 2
6 is the controller actually running the vendor, 27
Is a calculator for calculating the amount of movement.

Next, a conventional shape control method will be described with reference to FIG. When a plate bites into each stand in the process of rolling the hot-rolled steel sheet by the finish rolling mill, in the process of rolling the plate, the shape of the plate is controlled to the shape at the time of biting (hereinafter, FF shape control and This control is performed from the time the plate bites into each stand until the plate comes out of the stand. By this control, the hot rolled steel sheet can maintain a constant shape from the tip to the tail. This F
The F shape control is control for suppressing a change in flatness due to a change in rolling load generated in AGC.

Here, the operating mechanism of the FF shape control will be described. The rolling load applied to the hot rolled steel sheet is represented by the sum of the true rolling load F _R and the bender force F _B. F = F _R + F _B (1) Further, the true rolling load fluctuation ΔF _R and the bending force fluctuation ΔF _B
If in the flatness is Δλ changed, varying proportions of flatness when the true rolling load is changed, that is, when the sensitivity to true rolling load flatness ∂λ / ∂F _R, vendors force has changed If the flatness change ratio of the flatness, that is, the sensitivity of the flatness to the bending force is ∂λ / ∂F _B , Δλ = (∂λ / ∂F _R ) ・ ΔF _R + (∂λ / ∂F _B ) ・ΔF _B (2) Here, since the flatness is constant and Δλ = 0, The sensitivity of flatness to the true rolling load ∂λ / ∂
F _R, the sensitivity ∂λ / ∂F _B for vendors force of flatness has become a normal specific layer data, so that the value also changes if Kaware size or the like of the rolling material.

Assuming that the flatness is not constant, the following equation holds between the true rolling load fluctuation ΔF _R and the bending force fluctuation ΔF _{B. ΔF R = α · ΔF B ···} (4) Equation _{(3), F R -F RO} = K · (F B -F BO) from equation _{(4) F R -F RX =} α · (F B - F _BX ) (5) Is established. Here, F _RO is the lock-on value of the true rolling load, F _RX is the value after the change of the true rolling load, F _BO is the lock-on value of the bender force, and F _BX is the value after the change of the bender force. . From the formula (6), the vendor correction amount Z is Here, since F _R = F−F _B , The bender force for correcting the flatness changed due to the change of the rolling load and the bender force of each stand is calculated by the formula (8). And the output P of the actual vendor power
Is a value obtained by multiplying the expression (8) by the gain K _F , P = K _F · Z (9) (K _F is a gain, which is also a stratified value). The above calculation logic is executed at regular intervals from each stand-on (usually the last half 3 stands) to the stand-off, absorbs the flatness disorder due to AGC, etc., and changes the shape of the hot-rolled steel sheet from the tip to the tail. Keep constant.

In addition to the FF shape control that makes the shape of the hot rolled steel sheet constant from the tip to the tail end, there is FBK shape control that makes the shape of the hot rolled steel sheet a target flatness from the tip to the tail end. This is a control for making the shape from the tip to the tail end constant by the FF shape control and then making it the target shape, and is usually carried out at the final stand of the finish rolling mill. The operation timing of the FBK shape control is from the flatness meter ON to the down coiler OFF, and during that time, it is executed in parallel with the FF shape control.

Specifically, the FBK shape control will be described with reference to FIG. This hot rolled steel sheet is rolled at a finish rolling mill, passing through the tip flatness meter 21 of the plate, after came up the flatness signal, calculates a difference between the signal and the target flatness lambda ₀ . Δλ = λ ₀ −λ (10) Next, when the vendor changes the bender force correction amount Y that eliminates the difference between the actual flatness λ calculated by the equation (10) and the target flatness λ _0. Of the flatness, that is, flatness sensitivity ∂λ / ∂P _N by the vendor. Where Y is the bender pressure of the final stand. With this, the theoretical value of the bender pressure value at the final stand for controlling the shape of the hot-rolled steel sheet to the target flatness is calculated, and the output value ΔP _N of the pressure fluctuation to the actual plant is calculated by the gain in Equation (11). K _F multiplied by _{ΔP N = K F · Y ···} (12) value is output to the plant. Here, flatness sensitivity ∂λ / ∂P _N and gain K _F according to the vendor are stratified values. The above control is performed at the finishing stand, and is executed at regular intervals from the flatness meter ON to the Down Coiler ON, and the flatness that becomes constant by the FF shape control is the target flatness. To

Conventionally, the sensitivity used in the FF shape control or the FBK shape control is calculated by experiments or simulations, and the sensitivity is controlled by being placed on an online system. Also, the gain is
It was impossible to change the sensitivity and gain without manual intervention by changing the gain manually while watching the flatness online. This can be judged from Japanese Patent Laid-Open No. 59-47006, in which the vendor output value is not gained and the gain needs to be adjusted online. Further, there is a control method for automatically correcting the sensitivity by providing a correction coefficient α to the sensitivity as in Japanese Patent Laid-Open No. 9-174128, but this also has only the correction item as the correction item, is not versatile and is suitable for fine control. Not not.

[0009]

An FF shape in which a hot-rolled steel sheet is formed into a constant shape from the tip to the tail of a coil by using a bender installed in a finishing rolling mill located on the most downstream side of the rolling process. In the control and the FBK shape control for making a target shape, conventionally, the sensitivity and gain used there have been experimentally obtained. However, it is necessary to accurately set the sensitivity and parameter for each stratum, which causes a problem in time.

The present invention has been made in order to solve the above-mentioned problems, and from the tip end to the tail end of the coil,
FF shape control for constant shape and FB for target shape
The purpose of the K shape control is to enable the sensitivity and gain used there to be automatically adjusted during operation.

[0011]

A shape control method according to the present invention is an FF shape control method for making a shape of a hot-rolled steel sheet constant by a bender installed in a finishing rolling mill. After rolling N rolls in the same layer with a certain gain K _F up to the tail end, online regression analysis is performed on N rolls to determine the flatness sensitivity of the vendor, and the flatness sensitivity obtained is the stratified value as online data. The flatness sensitivity is updated to improve the flatness accuracy.

Further, there is provided a FF shape control method in which the shape of the hot-rolled steel sheet is made constant by a bender installed in the finishing rolling mill, and a certain gain K _F is applied from the front end to the tail end of the hot-rolled steel sheet.
After rolling N rolls in N, online regression analysis is performed on N rolls to obtain the flatness sensitivity of the vendor, and the flatness sensitivity, which is a stratified value, is updated using the obtained flatness sensitivity as online data, and the determination is made. The sensitivity is multiplied by the gain K _F to carry out rolling, the standard deviation of the flatness signal detected in the width direction is calculated, and the gain K _F is determined so as to minimize it.

Further, there is provided an FBK shape control method in which the shape of the hot-rolled steel sheet is set to a target value by a bender installed in the finishing rolling mill, and the hot-rolled steel sheet turns on the flatness meter and then turns on the winder. In the last stand until a certain gain K
After rolling N strips from the tip to the tail of the hot rolled steel sheet at _F ,
Online regression analysis is performed on N final stands to determine the flatness sensitivity of the vendor, and the flatness sensitivity, which is a stratified value, is updated using the obtained flatness sensitivity as online data to improve the flatness accuracy. I tried to go.

Further, it is an FBK shape control method in which the shape of the hot-rolled steel sheet is set to a target value by a bender installed in the finish rolling mill, and the hot-rolled steel sheet turns on the flatness meter and then turns on the winder. In the last stand until a certain gain K
After rolling N strips from the tip to the tail of the hot rolled steel sheet at _F ,
Online regression analysis is performed on N final stands to determine the vendor flatness sensitivity, and the flatness sensitivity that is the stratified value is updated using the obtained flatness sensitivity as online data.
After the flatness accuracy is improved and the flatness sensitivity is finally determined, a gain K _B is applied to the determined sensitivity and rolling is performed,
By increasing or decreasing the gain K _B for each rolling, the overshoot amount and the undershoot amount of the flatness, which is the output signal, are determined, and the flatness accuracy is improved.

Further, there is provided an FF shape control method in which the shape of the hot-rolled steel sheet is made constant by a bender installed in the finishing rolling mill, and a certain gain K _F is applied from the front end to the tail end of the hot-rolled steel sheet.
After rolling N rolls in N, the flatness sensitivity is obtained by performing an online regression analysis with N rolls, and the flatness sensitivity, which is a stratified value, is updated using the obtained flatness sensitivity as online data to improve accuracy. After that, the determined sensitivity is multiplied by a gain K _F to carry out rolling, the standard deviation of the flatness signals adjacent in the width direction is calculated, and the gain K _F is determined so that it becomes the minimum. , while fixing the flatness sensitivity and gain K _F, overshoot of flatness, which is the output signal by raising or lowering the gain K _B using the flatness of hot rolled steel sheet to FBK shape control to the value of the target for each rolling The amount of undershoot and the amount of undershoot are determined to improve the flatness accuracy.

Further, the flatness sensitivity of the vendor is obtained by repeatedly performing online regression analysis for every N lines, and the flatness sensitivity, which is a stratified value, is updated by using the obtained flatness sensitivity as online data to obtain the flatness accuracy. I tried to raise it.

Further, obtains a flatness sensitivity vendor online regression analysis with N present, the obtained flatness sensitivity updated flatness sensitivity is stratification value as an online data, 1 to N ₀ knots of After rolling, the flatness sensitivity of the vendor is obtained by performing online regression analysis on the N-th, which is the track record of the 2nd to N ₀ + 1th rolling, and the flatness sensitivity of the vendor is obtained, and the flatness sensitivity that is the stratified value is updated using the obtained flatness sensitivity as online data. Then, the flatness accuracy is increased.

Further, the FF shape control is performed so that the shape of the hot-rolled steel sheet is made constant by a bender installed in the finishing rolling mill, and N is obtained at a certain gain K _F from the tip to the tail end of the hot-rolled steel sheet.
After performing the main rolling, online regression analysis was performed on N rolls to determine the flatness sensitivity of the vendor, and the flatness sensitivity, which is a stratified value, was updated using the obtained flatness sensitivity as online data, and the flatness sensitivity was determined. The sensitivity is multiplied by the gain K _F to carry out rolling, the average error of the flatness signal detected in the width direction is calculated, and the gain K _F is determined so as to minimize it.

[0019]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiment 1. Embodiment 1 of the present invention will be described with reference to the drawings. In FIG. 1, 20 is a hot rolled steel sheet, 21 is a flatness meter for measuring the flatness, 22 is a backup roll, 23 is a work roll in which a bender for controlling the shape of the hot rolled steel sheet 20 is installed, and 24 is finishing. The last stand of the rolling mill, 25 is a winding machine, 26 is a controller that actually moves the bender, and 27 is a computer that calculates the amount of movement thereof. The calculator 27 performs calculations such as FF sensitivities 1 and 4, rolling results 2, FF shape control 3 and the like.

[0020] Sensitivity ∂λ / ∂F to the true rolling load flatness using the shape of a plate when rolling the plate in FF shape control to maintain a constant shape _R, sensitivity to vendors force of flatness ∂λ / ∂ F _B and gain K _F are set to constant empirical values, and FF
Bring shape control online. In that case, the FBK shape control is off-line. Then, a certain number N ₀
The main rolling is performed using the same material for each layer. At this time, the flatness, the true rolling load, and the bender force have lock-on values of λ ₀ ,
_{Let's call} it F _RO and F _BO .

[0021] After N ₀ this rolling, N ₀ duty of lambda ₀ in the computer 27, F _RO, using F _BO, performing a multiple regression calculation by the following formula, the sensitivity to the true rolling load flatness ∂ λ / ∂
F _R, to calculate the sensitivity ∂λ / ∂F _B for vendors force of flatness. Δλ = (∂λ / ∂F _R ) · ΔF _R + (∂λ / ∂F _B ) · ΔF _B (13) Then, this is repeated M times, which is the same layer,
Outliers in the data averages all data without putting the sensitivity (∂λ / ∂F _{R) R,} determining (∂λ / ∂F _{B) R.} When the sensitivities are determined by the equations (14) and (15), they are put into online data, and both FF shape control and FBK shape control are brought online and reflected in the next material.

When the above control method is adopted, the portion where the sensitivity is conventionally determined experimentally and the online sensitivity is manually replaced is automated, and the time required to obtain a highly accurate product is shortened. It will be.

Embodiment 2. Next, a second embodiment will be described with reference to FIG. 2 has the same configuration as that of FIG. 1, but the computer 27 performs the FF gain calculation 7 to obtain the FF gain 6. Sensitivity ∂λ / ∂F _R to the true rolling load flatness using the shape of a plate when rolling the plate in FF shape control to a constant shape, the sensitivity ∂λ / ∂F _B for vendors force flatness Using the value obtained above, the gain K _F is set to a constant empirical value, and the FF shape control is made online. In that case, the FBK shape control is off-line. Then, while rolling the material of the same layer, FF
The ASC gain is changed in steps of M between 0.1 and 1.5. And for each FF ASC gain,
For adjacent flatness signals λ _N and λ _{N + 1} , Was calculated, σ (K _F) optimum FF a K _F which is minimized
The FF shape control and FBK shape control are made online with the ASC gain and reflected in the next material.

When the above control method is adopted, the part where the sensitivity is experimentally determined and the FF ASC gain is manually replaced in the past is automated, and the time required to obtain a highly accurate product is shortened. Will be.

Embodiment 3. The third embodiment will be described with reference to FIG. In the figure, 8 and 9 are FBK sensitivity, and 10 is FBK shape control, which is calculated by the computer 27. The flatness sensitivity ∂λ / ∂P _N to the amount of bender force correction used in the FBK shape control to make the shape of the hot-rolled steel sheet a target flatness from the tip to the tail is set to a constant empirical value.
Bring K-shape control online. Then, a certain number N ₀ of rolling is performed using the same material for each layer. At this time, the lock-on value of the bender force correction value is set to Y ₀ .

[0026] After N ₀ this rolling, using the Y ₀ of N ₀ duty in the computer 27 calculates the flatness sensitivity ∂λ / ∂P _N for vendors force correction amount by performing a multiple regression calculation by the following formula . Δλ = ∂λ / ∂P _R · ΔP _R + ∂λ / ∂P _B · ΔP _B (17) Then, this is repeated M times, which is the same stratum,
All values are averaged without including any abnormal values in the data, and the sensitivities (∂λ / ∂P _N ) _R and (∂λ / ∂P _B ) _R are determined. When the sensitivities are determined by the equations (18) and (19), they are put in online data and reflected in the next material.

When the above control method is adopted, the portion where the sensitivity is conventionally determined by experiment and the online sensitivity is manually replaced is automated, and the time required to obtain a highly accurate product is shortened. It will be.

Fourth Embodiment The fourth embodiment will be described with reference to FIG. Although FIG. 4 has the same configuration as FIG. 3, the computer 27 performs the FBK shape control gain calculation 12 to obtain the FBK shape control gain 11. The flatness sensitivity ∂λ / ∂P _N with respect to the bender force correction amount used in the FBK shape control that makes the shape of the hot-rolled steel sheet a target flatness from the tip to the tail is obtained by the procedure of the third embodiment. After that, the FBK ASC gain is increased from 0.1 to 1.3 by 0.1 in increments of one coil in each layer. At that time, if the FBK ASC gain is too high,
The output signal overshoots, that is, a phenomenon in which the target flatness is greatly exceeded and the signal convergence rate is extremely deteriorated. If the FBK ASC gain is too low, the output signal undershoots, that is, the phenomenon that the output signal converges to the target flatness for a long time occurs.

The FBK ASC gain is set to 0.1 to 1.
When 3 is changed in 0.1 steps, the output signal changes from undershoot to overshoot FBK ASC
Because there is a gain, the FBK A with the best gain there
It means SC gain. If you want to produce the shape with higher accuracy, FBKAS in 10 ^-2 steps.
The best FB by changing the C gain and using the same method as above.
Search for the K ASC gain.

When the above control method is adopted, the part where the sensitivity is experimentally determined and the online gain is replaced manually is automated, and the time required to obtain a highly accurate product is shortened. It will be.

Embodiment 5. The fifth embodiment will be described with reference to FIG. Sensitivity to true rolling load flatness using the shape of the hot-rolled steel sheet during the rolling hot-rolled steel sheet in FF shape control to maintain a constant shape ∂λ / ∂F _R, sensitivity to vendors force flatness ∂Ramuda / determined value in the manner of the first embodiment of ∂F _B, and the gain K _F to a certain experience, FF
Bring shape control online. Next, the FF ASC gain is changed in steps of M between 0.1 and 1.5 while rolling the material for each layer. And each FF ASC
With gain, for adjacent flatness signals λ _N , λ _{N + 1} , Was calculated, sigma a (K _F) is minimized K _F and optimum FF ASC gains in the M, applied to the online data.

[0032] In the above operations, after determining the sensitivity ∂λ / ∂F _R to the true rolling load flatness, sensitivity ∂λ / ∂F _B for vendors force of flatness, the FF ASC gain K _F, FBK ASC Gain from 0.1 to 1.3 is 0.
In each step, the coils of the same layer are used to raise them one by one. At that time, if the FBK ASC gain is too high, the output signal overshoots, that is, a phenomenon in which the convergence rate of the signal is extremely deteriorated by significantly exceeding the target flatness occurs. Also,
If the FBK ASC gain is too low, the output signal undershoots, that is, the time for the output signal to converge to the target flatness increases.

The FBK ASC gain is set to 0.1 to 1.
When 3 is changed in 0.1 steps, the output signal changes from undershoot to overshoot FBK ASC
Because there is a gain, the FBK A with the best gain there
It means SC gain. If you want to obtain the shape with higher accuracy, FBKA in 10 ^-2 steps.
Change the SC gain and use the same method as before to obtain the best F
Just look for the BK ASC gain.

When the above-mentioned control method is adopted, the part where the sensitivity is experimentally determined and the online gain is manually replaced is automated, and the time required to obtain a highly accurate product is shortened. It will be.

Sixth Embodiment FIG. 5 is a diagram for explaining the sixth embodiment. In the figure, 1,2,3 ··· N ₀ ··· shows the rolling number. In the sixth embodiment, the flatness sensitivity of the vendor is obtained by performing online regression analysis using the equation (13) on the first N ₀ lines, and the flatness sensitivity, which is a stratified value, is updated.
Next, after rolling N ₀ + 1st to N ₀ rolls, the flatness sensitivity of the vendor is obtained by similarly performing online regression analysis using Equation (13), and the flatness sensitivity, which is a stratified value, is updated.
Similarly, online regression analysis is performed for each N ₀ lines to obtain the flatness sensitivity of the vendor, and the flatness sensitivity, which is a stratified value, is updated.

By adopting the above control method, the flatness sensitivity, which is a stratified value, can be automatically updated over a long period of time, and the stratified value can be adjusted with high accuracy.

Embodiment 7. FIG. 6 is a diagram for explaining the seventh embodiment. In the figure, 1,2,3 ··· N ₀ ··· shows the rolling number. In the seventh embodiment, the flatness sensitivity of the vendor is obtained by performing online regression analysis using the equation (13) on the first N ₀ lines, and the flatness sensitivity, which is a stratified value, is updated.
After the 1st to N _0th rolling, the online flattening sensitivity of the bender is calculated by online regression analysis using the equation (13) with the N ₀ which is the actual result of the 2nd to N ₀ + 1st rolling, and the flatness which is the stratified value. Update degree sensitivity. Similarly, N is rolled every time one roll is rolled.
The flatness sensitivity of the vendor is obtained by performing an online regression analysis with ₀ line, and the flatness sensitivity, which is a stratified value, is updated.

By adopting the above control method, the flatness sensitivity, which is the stratified value, can be automatically updated every time after the N ₀ + 1st rolling, and the stratification can be performed without increasing the load of regression analysis. The value can be adjusted with high precision.

Embodiment 8. Embodiment 8 will be described with reference to FIG. Sensitivity to true rolling load flatness using the shape of the hot-rolled steel sheet during the rolling hot-rolled steel sheet in FF shape control to maintain a constant shape ∂λ / ∂F _R, sensitivity to vendors force flatness ∂Ramuda / using the value determined for ∂F _B above, and the gain K _F to a certain experience, it is online FF shape control. In that case, the FBK shape control is off-line. Then, during rolling of the material for each layer, F
The F ASC gain is changed between 0.1 and 1.5 in M steps. Then, for each FF ASC gain, with respect to the flatness signal drive side, center, and work side signals λ _D , λ _C , and λ _W in the width direction, Was calculated, Δ (K _F) optimum FF a K _F which is minimized
The FF shape control and FBK shape control are made online with the ASC gain and reflected in the next material.

When the above control method is adopted, the part where the sensitivity is experimentally determined and the FF ASC gain is manually replaced in the past is automated, and the time required to obtain a highly accurate product is shortened. Will be.

[0041]

As described above, according to the present invention, the flatness can be made constant from the leading end to the tail end of the steel sheet with high accuracy and in a short time, or the flatness can be hit to the target value. In addition, conventionally, the part where the sensitivity is determined experimentally and the online sensitivity is manually replaced will be automated, and the time required to obtain a highly accurate product will be shortened.

Further, conventionally, the part where the sensitivity is experimentally determined and the FF ASC gain is replaced manually is automated, and the time required to obtain a highly accurate product is shortened.

Further, conventionally, the part where the sensitivity is experimentally determined and the online gain is replaced manually will be automated, and the time required to obtain a highly accurate product will be shortened.

Further, the flatness sensitivity, which is a stratified value, can be automatically updated over a long period of time, and the stratified value can be adjusted with high accuracy.

The flatness sensitivity, which is a stratified value, is N ₀ +1.
It can be automatically updated every time after the actual rolling, and the stratified value can be adjusted with high accuracy without increasing the load of regression analysis.

[Brief description of drawings]

FIG. 1 is a configuration diagram for explaining a shape control method according to a first embodiment, a fifth embodiment, a sixth embodiment, and a seventh embodiment of the present invention.

FIG. 2 is a configuration diagram for explaining a shape control method according to second and eighth embodiments of the present invention.

FIG. 3 is a configuration diagram for explaining a shape control method according to a third embodiment of the present invention.

FIG. 4 is a configuration diagram for explaining a shape control method according to a fourth embodiment of the present invention.

FIG. 5 is a diagram for explaining a shape control method according to a sixth embodiment of the present invention.

FIG. 6 is a diagram for explaining a shape control method according to a seventh embodiment of the present invention.

FIG. 7 is a configuration diagram showing a shape control device of a general finish rolling mill.

[Explanation of symbols]

20 rolled steel plate, 21 flatness meter, 22 backup roll, 23 work roll, 24 finishing mill final stand, 25 winder, 26 controller, 27
calculator.

─────────────────────────────────────────────────── --Continued from the front page (56) References JP-A-5-208204 (JP, A) JP-A-4-100624 (JP, A) JP-A-2-37906 (JP, A) JP-A-63- 264208 (JP, A) JP 10-263658 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) B21B 37/00-37/78

Claims (8)

(57) [Claims] 1. A FF shape control method for making a shape of a hot-rolled steel sheet constant by a bender installed in a finishing rolling mill, the hot-rolled steel sheet having a certain gain K _F from the front end to the tail end in the same layer. After rolling N rolls, online regression analysis is performed on N rolls to obtain the flatness sensitivity of the vendor, and the flatness sensitivity, which is a stratified value, is updated using the obtained flatness sensitivity as online data to obtain the flatness accuracy. A shape control method characterized in that the shape is raised. 2. A FF shape control method for making a shape of a hot rolled steel sheet constant by a bender installed in a finishing rolling mill, wherein N is obtained from a tip to a tail edge of the hot rolled steel sheet with a certain gain K _F.
After the main rolling, online regression analysis is performed on N rolls to determine the flatness sensitivity of the vendor, and the flatness sensitivity, which is a stratified value, is updated using the obtained flatness sensitivity as online data.
Rolling is performed by multiplying the determined sensitivity by the gain K _F ,
A shape control method, characterized in that a standard deviation of a flatness signal detected in the width direction is calculated, and a gain K _F is determined so as to minimize the standard deviation. 3. An FBK shape control method in which the shape of a hot-rolled steel sheet is set to a target value by a bender installed in a finishing rolling mill, and the hot-rolled steel sheet turns on a flatness meter and then turns on a winder. In the meantime, in the final stand, N rolls are rolled from the front end to the tail end of the hot-rolled steel sheet with a certain gain K _F , and then an N-line online regression analysis is performed for each final stand to determine the flatness sensitivity of the vendor. A shape control method, characterized in that the flatness sensitivity, which is a stratified value, is updated by using the obtained flatness sensitivity as online data to improve the flatness accuracy. 4. An FBK shape control method in which a shape of a hot rolled steel sheet is set to a target value by a bender installed in a finish rolling mill, and the hot rolled steel sheet turns on a flatness meter and then turns on a winder. In the meantime, in the final stand, after rolling N rolls from the front end to the tail end of the hot rolled steel sheet with a certain gain K _F , online regression analysis is performed with N rolls for each final stand to obtain the vendor flatness sensitivity. , The flatness sensitivity, which is a stratified value, is updated by using the obtained flatness sensitivity as online data, the flatness accuracy is improved, and the flatness sensitivity is finally determined. Then, the gain K is added to the determined sensitivity. It is characterized by improving the flatness accuracy by determining the overshoot amount and the undershoot amount of the flatness, which is an output signal, by rolling over _B and rolling up and down the gain K _B for each rolling. Shape control method. 5. A FF shape control method for making a shape of a hot-rolled steel sheet constant by a bender installed in a finishing rolling mill, wherein N is obtained from a tip to a tail end of the hot-rolled steel sheet at a certain gain K _F.
After performing main rolling, online regression analysis is performed on N rolls to obtain flatness sensitivity, and the obtained flatness sensitivity is used as online data to update the flatness sensitivity, which is a stratified value, and after the accuracy is improved, Rolling is performed by multiplying the determined sensitivity by a gain K _F , the standard deviation of flatness signals adjacent in the width direction is calculated, and the gain K _F is determined so as to minimize it. With the temperature sensitivity and the gain K _F fixed, the gain K _B used for the FBK shape control that sets the flatness of the hot-rolled steel sheet to a target value is increased or decreased for each rolling, and the overshoot amount of the flatness, which is the output signal, A shape control method characterized in that the amount of undershoot is judged to improve the flatness accuracy. 6. The flatness sensitivity of a vendor is obtained by repeatedly performing online regression analysis for every N lines, and the flatness sensitivity, which is a stratified value, is updated by using the obtained flatness sensitivity as online data to obtain the flatness accuracy. The shape control method according to claim 1, wherein the shape control method is performed so that the shape is raised. 7. The flatness sensitivity of the vendor is obtained by performing online regression analysis with N lines, and the flatness sensitivity, which is a stratified value, is updated using the obtained flatness sensitivity as online data, and then 1
After rolling N ₀ -th from online regression analysis with N present a performance from 2 N ₀ +1 -th rolling seeking flatness sensitivity of vendors is the stratification values flatness sensitivity obtained as an online data The shape control method according to claim 1, wherein the flatness sensitivity is updated to increase the flatness accuracy. 8. An FF shape control for making a shape of a hot rolled steel sheet constant by a bender installed in a finishing rolling mill,
After rolling N rolls with a certain gain K _F from the front end to the tail end of the hot-rolled steel sheet, online regression analysis is performed on the N rolls to obtain the flatness sensitivity of the vendor, and the flatness sensitivity obtained is used as online data for layering. The flatness sensitivity, which is another value, is updated, the determined sensitivity is multiplied by the gain K _F , rolling is performed, the average error of the flatness signals detected in the width direction is calculated, and it is minimized. A shape control method, characterized in that the gain K _F is determined as follows.