JPH03106507A - Control method for flatness of rolling mill - Google Patents

Abstract

PURPOSE:To improve the control accuracy of flatness by extracting a single component in lower order with the use of a lower order single component extraction factor in the case of the output of an actuator for controlling a high order component reaching a saturated state and performing the control of a plate by only an actuator for control ling a lower order component. CONSTITUTION:The shape of a rolling stock is detected by detecting the tension distri bution in the width direction by each zone Z-P - ZP found in the width direction of the rolling stock. The detected values of each zone are respectively multiplied by prefound lower order component extraction factor and higher order component extrac tion factor, the respective total sums are found, and the lower order component extrac tion factor and higher order component extraction factor in plate shapes are extracted. In order to obtain the aiming plate shape based thereon, an actuator for controlling the lower order component of a roll shifter, etc., and actuator for controlling the high order component of a work roll bender, etc., are controlled. Also in the case of the output of the high order component controlling actuator being saturated, the lower order single component in place of the lower order and higher order components is extracted by using the lower order single component extraction factor to control the actuator for controlling the lower order component. The control accuracy of flatness can thus be improved.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention mainly relates to a method for controlling the flatness of a strip in a strip rolling mill or the like.

(Prior Art) Various methods have been proposed to control this type of flatness. For example, the tension distribution S, is measured for each of a plurality of zones set in the width direction of the strip being rolled, and the tension distribution S is superimposed on this. The optimal parabolic shape (low-order component quadratic curve) for shape control is determined based on the sum of the values multiplied by the coefficient W1, and the roll bender is controlled based on this, as well as residual components (high-order components, etc.)
This can be corrected by using a spot coolant that sprays water or other coolant onto the strip surface in spots (Japanese Patent Publication No. 62-27882), or by detecting the shape of the strip during rolling and detecting low-order function components. There is a method in which the former is corrected by operating a roll shifter, and the latter is corrected by operating a work roll bender (JP-A-62-230).
Publication No. 412).

[Problem to be solved by the invention]

In the former method, low-order components are extracted and actively controlled by a roll shifter, but high-order components are treated as residual components and controlled by spot coolant. However, there is a limit to the ability of spot coolant to control higher-order components, and sufficient flatness cannot be obtained.

The latter method also uses low-order components of the strip shape. The high-order components are separated and the low-order components are controlled by a roll shifter, and the high-order components are controlled by a work roll bender. However, even if the work roll bender output reaches saturation during control, the roll shifter and work roll Since the roll shifter operates independently of the vendor, there is a problem in that the roll shifter continues to control only low-order components, and high-order components are no longer controlled.

The present invention has been made in view of the above circumstances, and its purpose is to extract a low-order single component from the plate shape when the output of the actuator for high-order component control reaches a saturated state, and to reduce the An object of the present invention is to provide a flatness control method for controlling an actuator for controlling a next component.

[Means to solve the problem]

The flatness control method according to the present invention detects the shape in each of a plurality of zones defined in the width direction of the rolled material during rolling, and adds predetermined low-order component extraction factors and high-order components to the shape detection value of each zone. The low-order components and high-order components of the plate shape are extracted by multiplying each extraction factor and calculating the sum of each, and based on this, the actuator for controlling the low-order components is used to obtain the target plate shape. The actuator for high-order component control is controlled, and when the output of the actuator for high-order component control reaches a saturated state, the shape detection value of each zone is multiplied by a predetermined low-order single component extraction factor, and the sum of the The present invention is characterized in that a plate-shaped low-order single component is extracted to replace the low-order component and high-order component, and the low-order component control actuator is controlled based on this.

[Effect]

According to the present invention, when the output of the high-order component control actuator reaches a saturated state, the low-order single component is replaced with the low-order component and the high-order component by using the predetermined low-order single component extraction factor. Based on this, flatness control is performed using only the low-order component control actuator. [Examples] Hereinafter, the present invention will be specifically explained based on drawings showing examples thereof.

FIG. 1 is a block diagram of an apparatus for carrying out the flatness control method for a rolling mill according to the present invention, and 1 in the figure indicates each zone Zi to Zp defined in the width direction of the strip ST during rolling.
It shows shape sensors A-, ~A, arranged in (
2p+t: Number of shape sensors in the board width direction). The detection values detected by these shape sensors A-P-Ap are output to the low-order component extraction section 1l and the high-order component extraction section 21. The low-order component extracting unit 11 individually guides the detected shapes S, ~S, for each shape sensor A1-A, for each zone Z-, ~ZP, to the multipliers E1-E, and the multiplier E- , ~E, each detected value 31~S, and the number of superimposed prisoners W-'l! for low-order component extraction manually input from each of a plurality of setters. ), W old model 1...
w,′! 2, W. Multiply all of these to give each zone 2−,
-2, multiplication results are added at an addition point 12 to obtain a plate-shaped low-order component X ("1", which is output to the subtracter 13. The above-mentioned W91...
In addition to the convolution factors for low-order component extraction such as
! ! +WVゝ, and both have switches st+, respectively.
,,~SW+,p selectively multipliers E-,-E
The work is now being done manually by P. Each switch SW. ，
1,~sti. ,, is the superposition factor for low-order component extraction during steady operation of the rolling mill, and the superposition factor for low-order single component extraction when the output of the work roll bender, etc., which is an actuator for controlling high-order components, reaches a saturated state. The convolution factor of multiplier E
Switching control is performed so that the output is output to -P-EP.

The subtractor 13 subtracts the extracted low-order component from the low-order component of the target shape obtained in advance and outputs the deviation, that is, the shape deviation low-order component to the gain setters 14a and 14b. Gain setter 14a. In 14b, a proportional gain Kpt and an integral gain K1 are set according to the rolling speed ■? It has been done,
The gain setter 14a sets a proportional gain K to the shape deviation low-order component.
The gain setter 14b performs an integral operation on the value obtained by multiplying the shape deviation low-order component by the integral gain K, and outputs it to the summation point 15, and outputs it to the summation point 1.
In step 5, a signal corresponding to the sum of both values is output as a control signal to a low-order component control actuator such as a roll shifter to drive and control the actuator.

On the other hand, the higher-order component extraction unit 21 detects each of the shape sensors A-P to AP in each of the zones Z-2 to Z2 {if! S-p
～SP is individually guided to multipliers F-P～F,-, and each detected value S F """ S F
and the superposition factors for high-order component extraction input from multiple setting devices w-(i'', w-', ?mi1...w,'yH, W
+<,'', and add the multiplication results for each zone Zi to Z at the addition point 22 to obtain the higher-order component x(2) of the plate shape, which is sent to the subtracter 23. The inputs from each setter connected to the multipliers Fi to FF include the above-mentioned superposition factors for extracting higher-order components such as W-(,Z), etc. and "0'°," and both are selectively input to the multipliers F-P to Fp by switches swz, -r to SW2, P, respectively. Each switch SWz. -r~SW2, P is the superimposition factor for extracting higher-order components during steady operation of the rolling mill, and is "0'" when the output of the work roll bender, etc., which is an actuator for controlling higher-order components, reaches a saturated state. Multiplier F
Switching control is performed to output to -, ~F, and so on.

The subtracter 23 subtracts the extracted high-order component Only the high-order components are set by the gain setter 24a. 24
Output to b. Gain setter 24a. 24b, the proportional gain KP. according to the rolling speed ■. integral gain K
．． The gain setter 24a integrates the value obtained by multiplying the shape deviation high-order component by the proportional gain K, and the gain setter 24b integrates the value obtained by multiplying the shape deviation high-order component by the integral gain K. The calculations are performed and outputted to the summing point 25, and at the summing point 25, a signal corresponding to the sum of both values is outputted as a control signal to an actuator for high-order component control such as a work roll bender, and this is driven. It is meant to be controlled. Further, when the output of the high-order component control actuator reaches a saturated state, the output of the summing point 25 is held unchanged.

The above-mentioned low-order component extraction factor W3 + 1...W,
j l l Number of prisoners for higher-order component extraction W (jl ・・・WJ
2) is the coefficient of each quadratic function and octadic function when the plate shape detected in each detected zone 2-, ~ZP is expressed most accurately by, for example, a quadratic function, an octadic function, and a linear combination of constants. This corresponds to That is, for example, if this linear combination equation is expressed as the following (Equation 11), the coefficient α of the quadratic term that minimizes the value J of Equation (2), which is the sum of the squared errors at the position of each sensor, and the 8th-order term The coefficient β is a convolution factor for extraction of the low-order component and high-order component of the plate shape, respectively.

The numbers of prisoners α and β are shown below (3). It is expressed by equation (4).

(3), W in formulas (4), +11. W, (Zl
are given by the following equations (5) and (6), respectively.

(Margins below) pp However, i=-P, ...5P However, 2P+1: Number of shape sensors in the board width direction i: Shape detection sensor number SA: Shape sensor detection value δ,: Constant However, δt: Constant Each sensor position γ that minimizes the value of the sum of squared errors in equation (8) is the extraction factor for low-order single components. This number of prisoners T is expressed by the following equation (9). W, +3) in equation (9) is as follows (1 (given by equation 11. The low-order single component extraction factor W1, which is used in place of Then, P i = - P - Pi = - P z P factor W!'), Wilt, Wp) The values differ depending on the plate width, and the specific values are shown in Table 1. ．． Table 1 shows that the board width is 1400mm. 1
For three types of strips of 100 m and 600 m, each zone Z. ~Zl:I+ZO
~Z1. , Z0 to ZS (Table 1 shows specific values for each zone from the center of the strip to one end). It shows the coefficient value to be multiplied by ~A. P Figure 2 is a graph showing the results of flatness control using the method of the present invention and conventional method 1.2, where the horizontal axis represents the position in the width direction of the strip, and the vertical axis represents the difference in unevenness of the plate shape. It is shown.

FIG. 2(a) shows the case of the method of the present invention, and FIG. 2(b) and (c) show the case of the conventional method 1. 2 shows the results obtained based on 2. As is clear from this graph, the conventional method l
Except for the above, the method of the present invention and conventional method 2 provide substantially the same overall flat shape. On the other hand, in a situation where the output of the actuator for controlling higher-order components is saturated, the method of the present invention suppresses the shape deformation at both ends of the strip/drop to a smaller extent than in conventional method 2. I understand.

〔Effect of the invention〕

As described above, in the method of the present invention, not only during steady-state operation but also when the output of the actuator for controlling the high-order component reaches a saturated state, the low-order individual component of the plate shape is extracted and the actuator for controlling the low-order shape is used only. Since it is planned to be controlled by
The present invention has excellent effects such as low-order component control being performed over the entire width of the plate, avoiding an uncontrollable state, and greatly improving flatness control accuracy.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an apparatus for carrying out the method of the present invention, and FIG. 2 is a graph showing the results of a comparative test between the method of the present invention and a conventional method.

Claims (1)

[Claims] 1. During rolling, the shape of the rolled material is detected by detecting the tension distribution in the width direction in each of multiple zones defined in the width direction of the rolled material, and lower-order components determined in advance are extracted from the detected values in each zone. an actuator for controlling low-order components to obtain a target plate shape based on the multiplication of the factors and the high-order component extraction factors, the respective sums thereof, and the extraction of low-order components and high-order components of the plate shape; Controls the actuator for controlling higher-order components,
In addition, when the output of the actuator for high-order component control reaches a saturated state, the shape detection value of each zone is multiplied by a predetermined low-order single component extraction factor, the sum is calculated, and the low-order component and high-order component are extracted. 1. A flatness control method for a rolling mill, characterized by extracting a low-order single component of a plate shape in place of the flatness of the plate, and controlling an actuator for controlling the low-order component based on the extracted low-order single component.