KR101749018B1 - Flatness control device - Google Patents

Flatness control device Download PDF

Info

Publication number
KR101749018B1
KR101749018B1 KR1020177000992A KR20177000992A KR101749018B1 KR 101749018 B1 KR101749018 B1 KR 101749018B1 KR 1020177000992 A KR1020177000992 A KR 1020177000992A KR 20177000992 A KR20177000992 A KR 20177000992A KR 101749018 B1 KR101749018 B1 KR 101749018B1
Authority
KR
South Korea
Prior art keywords
flatness
value
coefficient
influence
influence coefficient
Prior art date
Application number
KR1020177000992A
Other languages
Korean (ko)
Other versions
KR20170018419A (en
Inventor
토시히로 니이
나오히로 쿠보
Original Assignee
도시바 미쓰비시덴키 산교시스템 가부시키가이샤
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 도시바 미쓰비시덴키 산교시스템 가부시키가이샤 filed Critical 도시바 미쓰비시덴키 산교시스템 가부시키가이샤
Priority to PCT/JP2014/075488 priority Critical patent/WO2016046945A1/en
Publication of KR20170018419A publication Critical patent/KR20170018419A/en
Application granted granted Critical
Publication of KR101749018B1 publication Critical patent/KR101749018B1/en

Links

Images

Classifications

    • BPERFORMING OPERATIONS; TRANSPORTING
    • B21MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
    • B21BROLLING OF METAL
    • B21B37/00Control devices or methods specially adapted for metal-rolling mills or the work produced thereby
    • B21B37/28Control of flatness or profile during rolling of strip, sheets or plates
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B21MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
    • B21BROLLING OF METAL
    • B21B38/00Methods or devices for measuring, detecting or monitoring specially adapted for metal-rolling mills, e.g. position detection, inspection of the product
    • B21B38/02Methods or devices for measuring, detecting or monitoring specially adapted for metal-rolling mills, e.g. position detection, inspection of the product for measuring flatness or profile of strips
    • GPHYSICS
    • G05CONTROLLING; REGULATING
    • G05BCONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B13/00Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B21MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
    • B21BROLLING OF METAL
    • B21B2263/00Shape of product
    • B21B2263/04Flatness

Abstract

A flatness control apparatus according to the present invention is a flatness control apparatus according to the present invention is a flatness control apparatus including a shape measuring system for measuring a flatness, a flatness target value setting apparatus for setting a target value of flatness, an operation amount calculating apparatus for calculating an operation amount of each actuator, A monitoring device for monitoring whether or not the amount of change is larger than a predetermined flatness threshold value and a monitoring device for monitoring whether the amount of change in the actual value of the flatness degree and the actual value of the operation amount of each actuator are correlated with each other when the variation amount of the actual value of the flatness exceeds the flatness threshold value A flatness influence coefficient arithmetic unit for calculating the same value, a flatness calculating unit for calculating a current learned value based on the current value and the previous learned value, and setting the current learned value to the above- An influence coefficient learning value calculating device and a flatness influence coefficient learning value storing device for storing the learning value of each influence coefficient The features.

Description

{FLATNESS CONTROL DEVICE}
The present invention relates to a flatness control apparatus.
BACKGROUND ART Conventionally, a rolling mill for rolling a rolled material such as a metal is known. In the rolling machine, the flatness control for flat rolling the rolled material is performed. Hereinafter, the flatness control will be described.
In the flatness control, the deviation between the actual value (actual value) of the flatness of the rolled material detected by the shape meter provided in the rolling mill and the target value of the flatness is calculated. Then, the manipulated variables of the actuators provided in the rolling mill are calculated such that the deviation is minimized. The calculated manipulated variables are transmitted to the control devices of the respective actuators. This is repeated at a constant control cycle to suppress deviation of the actual value of the flatness from the target value over the entire length of the rolled material.
Further, Patent Document 1 discloses learning control using the actual values of the flatness and the manipulated variables of the respective actuators. By performing the learning control, the precision of the flatness control of the rolled material can be improved.
Japanese Patent Application Laid-Open No. 9-174128
However, in the learning control disclosed in Patent Document 1, the actual value of the accurate and stable flatness and the manipulated variable of each actuator may not be obtained due to the influence of coolant or disturbance included in the signal . As a result, the learning accuracy may be lowered.
An object of the present invention is to provide a flatness control device capable of suppressing the influence of coolant and disturbance included in a signal and performing learning control with high precision.
A first aspect of the present invention is a flatness control apparatus for achieving the above object,
A flatness control device for controlling a flatness in a width direction of a rolled material provided in a rolling process for rolling a rolled material to a desired product by operating a plurality of actuators,
A shape measuring device for measuring the flatness of each of the plurality of measurement positions set in the width direction of the rolled material,
A flatness target value setting device for setting a target value of the flatness at each of the measurement positions,
Wherein the amount of change in flatness at each measurement position when each of the actuators is manipulated is represented by a polynomial expression in which each of the measurement positions is a variable and each term of the polynomial expression indicates an effect Based on the deviation between the actual value of the flatness at each measurement position and the target value of the flatness at each measurement position, using the flatness influence coefficient model multiplied by the coefficient (influence coefficient) An operation amount calculating device for calculating an operation amount of each of the actuators,
A monitoring device for monitoring whether or not the amount of change in the actual value of the flatness at each of the measurement positions is larger than a predetermined flatness threshold;
Wherein when the change amount of the actual value of the flatness at each of the measurement positions exceeds the flatness threshold value, the change amount of the actual value of the flatness at each measurement position is correlated with the actual value of the manipulated variable of each actuator, Wow,
The respective influence coefficients of the flatness influence coefficient model are identified based on the change amounts of the actual values of the flatness at the respective measurement positions read from the storage device and the actual values of the manipulated variables of the actuators, A flatness influence coefficient arithmetic unit for calculating a fixed value,
Calculating current learning value of each influence coefficient on the basis of the same value of the current time of each influence coefficient and the previous learning value of each influence coefficient, A flatness influence coefficient learning value computing device set for the manipulated variable computing device;
And a flatness degree-of-influence coefficient learning value storage device for storing the learning value of each influence coefficient.
According to the present invention, since the performance data to be used for the learning control can be appropriately selected, the influence of the coolant and the influence of the disturbance can be reduced. As a result, the flatness prediction accuracy improves.
BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a diagram showing a configuration of a system according to a first embodiment; Fig.
2 is a view showing a monitoring procedure of the manipulated variable monitoring apparatus according to the first embodiment;
3 is a diagram showing a configuration of a system according to a second embodiment;
Embodiment Mode 1.
[System configuration]
In Embodiment 1, a single rolling mill equipped with actuators such as work roll (WR) bending, intermediate roll (IMR) bending, IMR shift, leveling and the like is controlled. Here, WR bending is an actuator for correcting looseness of a work roll by the force of oil pressure, IMR bending is an actuator for correcting slack of an intermediate roll by the force of hydraulic pressure, And the leveling is an actuator provided for correcting the rolling material to meander or the shape to be disturbed.
1 is a diagram showing a configuration of a system according to a first embodiment. 1, a rolling mill 1 is shown. The rolling mill 1 is provided with an actuator 5 such as WR bending, IMR bending, IMR shift and leveling as described above. The rolling mill 1 rolls the rolled material 2 in the direction of the arrow 3. On the outlet side of the rolling mill 1, there is provided a forming system 4. The form system 4 has sensor rolls at each of a plurality of measurement positions set at predetermined intervals in the width direction of the rolled material 2. [ The actual value of the flatness of the rolled material 2 is measured for each of the plurality of sensor rolls.
Fig. 1 shows a flatness control device 6 for controlling the actuator 5 of the rolling mill 1. The flatness control device 6 includes a shape measuring device 4, an operation amount calculating device 7, and a flatness target value setting device 8. The manipulated variable computing device 7 is an apparatus for computing the manipulated variable of the actuator 5. [ The flatness target value setting device 8 is a device for setting a target value indicating the flatness at each measurement position of the target shape after rolling of the rolled material 2.
(Normal control) is performed by using the constituent device of the flatness control device 6 described above. Hereinafter, the normal control according to the first embodiment will be described.
The form system 4 transmits the actual value of the measured flatness to the manipulated variable computing device 7. The manipulated variable computing device 7 calculates the deviation between the target value output from the flatness target value setting device 8 and the actual value of the flatness at every control cycle. Then, the manipulated variable computing device 7 computes the manipulated variable of the actuator 5 using a flatness influence coefficient model, which will be described later, so that the deviation is minimized. The actuator 5 is operated on the basis of the manipulated variable calculated by the manipulated variable computing device 7. [
In Embodiment 1, in addition to the above-described normal control, learning control of the flatness is performed. Hereinafter, the learning control will be described.
The flatness degree influence coefficient arithmetic unit 9, the flatness degree influence coefficient learning value arithmetic unit 10, the flatness degree influence coefficient learning value storage unit 11 and the information collecting unit 20 are connected to the flatness control unit 6 . The information collecting device 20 is constituted by an operation amount monitoring device 12 and a variation storage device 13. The flatness degree influence coefficient calculating device 9 is an apparatus for identifying the learning coefficient of the flatness influence coefficient model using an evaluation function to be described later. The flatness-based influence coefficient learning-value calculation device 10 is an apparatus for calculating learning values of influence coefficients. The flatness degree-of-influence coefficient learning value storage device 11 is a device for storing the learning value calculated by the flatness-based influence coefficient learning coefficient arithmetic unit 10. [ The information collecting apparatus 20 is a device for sorting actual values used for learning of the flatness influence coefficient model.
The actual value of the flatness measured by the profile 4 and the actual value of the manipulated variable of the actuator 5 are inputted to the flatness control device 6. [ In the flatness control apparatus 6, the flatness coefficient of influence calculating apparatus 9 calculates the difference between the actual value of the flatness before the operation of the actuator 5 and the actual value of the flatness after the operation by using the evaluation function described later (Hereinafter referred to as a change amount of the actual value of the flatness) and the actual value of the manipulated variable of the actuator 5, the learning coefficient of the flatness influence coefficient model is identified. The value of this identified learning coefficient is called the same value. The flatness influence coefficient arithmetic unit 9 transfers the calculated dynamic coefficient of the learning coefficient to the flatness influence coefficient learning value arithmetic unit 10.
The flatness degree-of-influence coefficient learning-value calculation device 10 calculates the flatness-degree-of-influence coefficient learning coefficient computing device 10 based on the difference between the previous learning value sent from the flatness-effect-coefficient learning value storage device 11 and the learning coefficient transmitted from the flatness- The learning value of this time is calculated. In the flatness degree-of-influence coefficient learning-value calculating apparatus 10, the current learning value is obtained from the average or weighted average of the learning value of the learning coefficient and the previous learning value.
The flatness degree-of-influence coefficient learning-value calculation device 10 transmits the current learned value to the manipulated variable computing device 7 and the flatness-influenced coefficient learning value storage device 11. The flatness degree-of-influence coefficient learning value storage device 11 stores the current learning value transmitted from the flatness-based influence coefficient learning value computing device 10. [ The flatness degree-of-influence coefficient learning value storage device 11 stores the received learning value of this time in a learning table which is layered for each steel type, plate thickness, and plate width, for example. The manipulated variable computing device 7 performs the flatness control using the received learning value of the current time for the flatness influence coefficient model.
The details of the evaluation function stored in the flatness influence coefficient calculation device 9 and the flatness influence coefficient model stored in the manipulated variable calculation device 7 will be described in detail below.
[Flatness Influence Factor Model]
The flatness influence coefficient model stored in the manipulated variable computing device 7 is expressed by the following equations (1) to (4). In the following equations, the flatness influence coefficient model of WR bending is represented by equation (1), the IMR bending flatness influence coefficient model is represented by equation (2), the IMR shift flatness influence coefficient model is represented by equation (3) The influence coefficient model is expressed by Equation (4).
[Equation 1]
Figure 112017003900510-pct00001
[Equation 2]
Figure 112017003900510-pct00002
[Equation 3]
Figure 112017003900510-pct00003
[Equation 4]
Figure 112017003900510-pct00004
Here, each of the above-mentioned Expressions (1) to (4) will be described below.
[Equation 5]
x i : Reference position in i when standardizing the width from -1 to 1 [-]
[Equation 6]
 i: Number of each sensor roll (i = 1, 2, 3, ... N)
[Equation 7]
Figure 112017003900510-pct00005
: WR bending
Flatness Influence Factor Model [I-unit / (kN / chock)]
[Equation 8]
Figure 112017003900510-pct00006
: IMR bending
Flatness Influence Factor Model [I-unit / (kN / chock)]
[Equation 9]
Figure 112017003900510-pct00007
: IMR Shift
Flatness Influence Factor Model [I-unit / mm]
[Equation 10]
Figure 112017003900510-pct00008
: Leveling
Flatness Influence Factor Model [I-unit / mm]
[Equation 11]
α WRB2 : WR Bending Flatness Influence Factor
The coefficient of the second term of the model [I-unit / (kN / chock)]
[Equation 12]
α WRB4 : WR Bending Flatness Influence Factor
The fourth-order coefficient of the model [I-unit / (kN / chock)]
[Equation 13]
α IRB2 : IWR Bending Flatness Influence Factor
The coefficient of the second term of the model [I-unit / (kN / chock)]
[Equation 14]
α IRB4 : IWR Bending Flatness Influence Factor
The fourth-order coefficient of the model [I-unit / (kN / chock)]
[Equation 15]
α IRS2 : Factor of second order term of IWR shift flatness influence coefficient model [I-unit / mm]
[Equation 16]
α IRS4 : IWR Shift flatness coefficient Factor of the fourth order term of the model [I-unit / mm]
[Equation 17]
α LVL1 : coefficient of 1st order term of the leveling flatness influence coefficient model [I-unit / mm]
[Equation 18]
α LVL3 : coefficient of third order of leveling flatness influence coefficient model [I-unit / mm]
[Expression 19]
α WRB6 : WR Bending Flatness Influence Factor Model
Coefficient of sixth order [I-unit / (kN / chock)]
[Equation 20]
α IRB6 : IWR Bending Flatness Influence Factor Model
Coefficient of sixth order [I-unit / (kN / chock)]
[Equation 21]
α IRS6 : Factor of sixth order term of IWR shift flatness influence coefficient model [I-unit / mm]
[Equation 22]
α LVL5 : coefficient of 5th order of leveling flatness influence coefficient model [I-unit / mm]
Further, the coefficients of the above equations 11 to 22 are fixed values.
[Equation 23]
Z WRB2 : WR Bending Flatness Influence Factor Model
Second-order learning coefficient [I-unit / (kN / chock)]
[Equation 24]
Z WRB4 : WR Bending Flatness Influence Factor Model
The learning coefficient of the fourth term [I-unit / (kN / chock)]
[Equation 25]
Z IRB2 : IWR Bending Flatness Influence Factor Model
Second-order learning coefficient [I-unit / (kN / chock)]
[Equation 26]
Z IRB4 : IWR Bending Flatness Influence Factor Model
The learning coefficient of the fourth term [I-unit / (kN / chock)]
[Equation 27]
Z IRS2 : IWR shift flatness coefficient of the learning coefficient of the second term of the influence coefficient model [I-unit / mm]
[Equation 28]
Z IRS4 : IWR Shift Plane Influence Factor Learning coefficient of the fourth term of the model [I-unit / mm]
[Equation 29]
Z LVL1 : Learning coefficient of first order term of leveling flatness influence coefficient model [I-unit / mm]
[Equation 30]
Z LVL3 : Learning coefficient of third-order term of leveling flatness influence coefficient model [I-unit / mm]
[Equation 31]
Z WRB6 : WR Bending Flatness Influence Factor Model
Learning factor of 6th order [I-unit / (kN / chock)]
[Equation 32]
Z IRB6 : IWR Bending Flatness Influence Factor Model
Learning factor of 6th order [I-unit / (kN / chock)]
[Equation 33]
Z IRS6 : Learning factor of sixth order term of IWR shift flatness influence coefficient model [I-unit / mm]
[Equation 34]
Z LVL5 : Learning coefficient of the 5th order of the leveling flatness influence coefficient model [I-unit / mm]
The learning coefficients in the above Equations (23) to (34) are variables. The coefficient obtained by multiplying the coefficients of the above-mentioned expressions 11 to 22 by the learning coefficients of the expressions 23 to 34 is the influence coefficient. For example, the influence coefficient of the quadratic term of the flatness degree influence coefficient model of WR bending shown in equation (1) is Z WRB2 a WRB2 .
[Evaluation function]
The evaluation function stored in the flatness degree influence coefficient calculating device 9 is represented by the following equation (5). The flatness influence coefficient arithmetic unit 9 calculates the same value of each learning coefficient so that the following evaluation function is minimized.
[Equation 35]
Figure 112017003900510-pct00009
[Equation 36]
n : Actual value of the nth set [-]
[Equation 37]
Figure 112017003900510-pct00010
: Flatness value at position (i) [I-unit]
[Expression 38]
Figure 112017003900510-pct00011
: Performance WR bending [kN / chock]
[Equation 39]
Figure 112017003900510-pct00012
: Performance IWR bending [kN / chock]
[Equation 40]
Figure 112017003900510-pct00013
: Performance IWR Shift [mm]
[Equation 41]
Figure 112017003900510-pct00014
: Performance IWR Leveling [mm]
[Equation 42]
Figure 112017003900510-pct00015
: WR Bending Flatness Influence Factor Model of Second Order
The dynamic coefficient of learning coefficient [I-unit / (kN / chock)]
[Equation 43]
Figure 112017003900510-pct00016
: WR Bending Flatness Influence Factor Model 4th order
The dynamic coefficient of learning coefficient [I-unit / (kN / chock)]
[Equation 44]
Figure 112017003900510-pct00017
: WR Bending Flatness Influence Factor Model 6th order
The dynamic coefficient of learning coefficient [I-unit / (kN / chock)]
[Equation 45]
Figure 112017003900510-pct00018
: IWR Bending Flatness Influence Factor Model Second-Order
The dynamic coefficient of learning coefficient [I-unit / (kN / chock)]
[Equation 46]
Figure 112017003900510-pct00019
: IWR Bending Flatness Influence Factor Model 4th order
The dynamic coefficient of learning coefficient [I-unit / (kN / chock)]
[Equation 47]
Figure 112017003900510-pct00020
: IWR Bending Flatness Influence Factor Model Second-Order
The dynamic coefficient of learning coefficient [I-unit / (kN / chock)]
[Equation 48]
Figure 112017003900510-pct00021
: IWR shift flatness coefficient of influence model second order
Dynamic constant of learning coefficient [I-unit / mm]
[Equation 49]
Figure 112017003900510-pct00022
: IWR shift flatness coefficient of influence model 4th order
Dynamic constant of learning coefficient [I-unit / mm]
[Equation 50]
Figure 112017003900510-pct00023
: IWR shift flatness coefficient of influence model 6th order
Dynamic constant of learning coefficient [I-unit / mm]
[Equation 51]
Figure 112017003900510-pct00024
: Leveling flatness coefficient of influence
Dynamic constant of learning coefficient [I-unit / mm]
[Equation 52]
Figure 112017003900510-pct00025
: Leveling flatness coefficient of influence
Dynamic constant of learning coefficient [I-unit / mm]
[Equation 53]
Figure 112017003900510-pct00026
: Leveling flatness coefficient of influence
Dynamic constant of learning coefficient [I-unit / mm]
However, when the actual value of the flatness and the manipulated variable of the actuator 5 are used in the learning control, when the actual value of the flatness includes the influence of the coolant or the disturbance, the prediction error of the flatness becomes large and stable control becomes difficult There is a concern that
Therefore, in Embodiment 1, in order to suitably select the actual value used for the learning control, the actual value of the flatness and the manipulated variable of the actuator 5 are received for each control period, and the variation amount of the actual value of the flatness and the actuator 5) exceeds the threshold value. Hereinafter, a determination routine performed in the first embodiment will be described with reference to FIG.
[Judgment routine]
2 is a decision routine executed by the manipulated variable monitor 12. First, the manipulated variable monitoring device 12 determines whether the learning flag is ON (S100). When it is determined that the learning flag is not turned ON, the manipulated variable monitoring device 12 ends this routine.
On the other hand, in S100, when it is determined that the learning flag is ON, the manipulated variable monitoring device 12 measures the actual time of the time and the actual value of the manipulated variable of the actuator 5 (S110).
Next, the manipulated variable monitoring device 12 calculates the actual value of the elapsed time? T from the time when S110 is executed, the change amount of the actual value of the flatness, and the manipulated variable of the actuator 5 (S120).
Next, the manipulated variable monitoring device 12 determines whether the elapsed time? T is longer than a predetermined time? T UL (S130). If the elapsed time? T is equal to or less than the predetermined time? T UL , the present routine returns to the starting point.
On the other hand, when the elapsed time? T is longer than the predetermined time? T UL , the average value of the absolute values of the variations in the flatness at the positions of the sensor rollers in the width direction of the shape system 4 is smaller than the predetermined threshold? LL ) (S140). When the average value of the absolute values of the variations in the flatness at the positions of the sensor rolls in the width direction of the shape system 4 is equal to or smaller than the predetermined threshold value? LL , the influence of the coolant, disturbance, (S120) of the actual value of the elapsed time? T, the change amount of the actual value of the flatness, and the manipulated variable of the actuator 5 is retried again.
On the other hand, when the average value of the absolute values of the variations in the flatness at the positions of the sensor rolls in the width direction of the form system 4 is larger than the predetermined threshold value? LL , the manipulated variables of the respective actuators 5 are determined in advance It is judged whether or not it is smaller than the threshold value (S150, S170, S190, S210). The actuators 5 having the manipulated variables smaller than the threshold value are replaced with zero, and the actuators 5 having the manipulated variable equal to or larger than the threshold value are replaced with the manipulated variables (S160, S180, S200, S220). Thereafter, the amount of change in the actual value of the flatness is simultaneously transmitted to the change amount storage device 13 (S230).
The change amount storage device 13 stores the actual values of the received flatness and the manipulated variable of the actuator 5 up to the maximum M sets. The change amount storage device 13 stores the M sets of data and then transmits the M sets of data to the flatness influence coefficient arithmetic device 9. [ Then, every time one set of data is updated, the M sets of data are transmitted to the flatness influence coefficient arithmetic unit 9. When the type of steel is changed, the M sets of data are all deleted.
Normally, measures for raising the degree of approximation function are performed in order to increase the approximation accuracy. However, if the degree is too high, there is a possibility that the influence of coolant and disturbance such as disturbance may be modeled as an influence coefficient of the actuator 5. However, in the present invention, since the performance data used for learning control can be appropriately selected, the influence of the coolant and the influence of the disturbance can be reduced, and the approximate function can be set to a higher order (higher order). This improves the flatness prediction accuracy.
Further, as a modification of the flatness influence coefficient model, the following lower formulas (6) to (9) may be used.
[Equation 54]
Figure 112017003900510-pct00027
[Equation 55]
Figure 112017003900510-pct00028
[Equation 56]
Figure 112017003900510-pct00029
[Equation 57]
Figure 112017003900510-pct00030
Furthermore, in Embodiment 1, the flatness coefficient of impact calculation device 9 identifies the learning coefficient of the influence coefficient model, but this is not limitative. For example, the flatness coefficient of influence calculating device 9 may identify the influence coefficient of the influence coefficient model. This is also true in the second embodiment which will be described later.
Embodiment 2:
3 is a diagram showing the configuration of the system according to the second embodiment. Embodiment 2 is the same as Embodiment 1 except that the flatness degree-of-influence coefficient coefficient arithmetic unit 14 obtains the actual value of the flatness from the form-setting unit 4, the target value from the flatness- And controls the flatness influence coefficient arithmetic unit 9, as shown in FIG. Only operations different from those of the first embodiment will be described below.
The flatness coefficient of influence coefficient arithmetic unit 14 receives the actual value of the flatness for each control period from the form system 4 and the target value from the flatness target value setting device 8. When the average value of the absolute value of the deviation between the actual value of the flatness and the target value of the flatness does not monotonically increase for a predetermined time, the flatness degree-of-influence coefficient coefficient arithmetic unit 14 calculates the equation (1) 4), the same coefficient of learning coefficient of 5th order and learning coefficient of 6th order is set to zero. Then, the flatness degree-of-influence coefficient-order arithmetic unit 14 identifies the same value of the learning coefficients of the first-order, second-order, third-order, and fourth-order terms. On the other hand, when the average value of the absolute values of the deviations between the actual values of the flatness and the target values of the flatness increases monotonously, the flatness coefficient of influence coefficient arithmetic unit 14 calculates the coefficients of the first, second, third, Identify the variance of the learning coefficients of the first and sixth terms. The flatness degree-of-influence coefficient coefficient arithmetic unit 14 is also capable of calculating the degree of flatness coefficient of influence of the following types of materials from the following materials: first grade, second grade, third grade, fourth grade, fifth grade , The same value of learning coefficient of sixth order is identified.
In the second embodiment, it is evaluated whether or not the average value of the absolute value of the deviation between the actual value of the flatness and the target value of the flatness increases monotonically. However, the present invention is not limited to this. For example, it may be evaluated whether the actual values of the manipulated variables of WR bending and IMR bending are diverging. When the actual values of the manipulated variables of the WR bending and the IMR bending do not diverge, the same value of the learning coefficient of the fifth-order coefficient and the learning coefficient of the sixth-order coefficient of the above equations (1) to (4) Then, the flatness degree-of-influence coefficient-order arithmetic unit 14 identifies the same value of the learning coefficients of the first-order, second-order, third-order, and fourth-order terms. On the other hand, when the actual values of the manipulated variables of the WR bending and the IMR bending diverge, the flatness coefficient of influence coefficient arithmetic unit 14 calculates the coefficients of learning coefficients of the first order, second order, third order, fourth order, fifth order and sixth order Identify this politics.
Generally, if the degree of the polynomial expression of the flatness influence coefficient model is set too high, the influence of the coolant and the disturbance may be learned. However, as described above, only when the order of the flatness influence coefficient model is not appropriate, By using the degree, it is possible to select the order of the optimal approximation function without unnecessarily increasing the degree, and the accuracy of the flatness control can be improved.
The flatness degree influence coefficient model of the actuator 5 acting on the target component of the flatness is the sixth order polynomial and the flatness degree influence coefficient model of the actuator 5 acting on the asymmetry component is the fifth order polynomial. It is not limited. Thus, the learning of the efficient and high-precision flatness influence coefficient model can be executed.
In the above description, the actuator 5 has been described as WR bending, IMR bending, IMR shift and leveling, but it may be combined with other actuators such as VC roll, WR shift, and the like. Further, the present invention can be applied to all rolling mills including a hot rolling mill, a cold rolling mill, a tandem mill, and the like provided with a forming system 4.
According to the present invention, since the influence of the coolant and the disturbance can be reduced by executing the learning when the amount of change in the degree of flatness and the amount of manipulation of the actuator 5 are larger than a preset threshold within a preset time, The order of the approximate function of the influence coefficient model can be increased, and the accuracy of the flatness prediction accuracy can be improved. Further, even when the flatness degree influence coefficient has a high dimensional component due to various rolling conditions and characteristics of the rolled material, it is possible to learn with an optimal approximation function, so that the flatness prediction accuracy can be improved.
1: rolling mill
2: rolled material
3: rolling direction
4:
5: Actuator
6: Flatness control device
7:
8: Flatness target setting device
9: Flatness coefficient of influence coefficient
10: flatness coefficient of influence learning value calculating device
11: flatness coefficient of influence learning value storage device
12: Operation volume monitoring device
13: Variation memory
14: Flatness coefficient of influence coefficient arithmetic unit
20: Information collecting device

Claims (4)

  1. A flatness control device for controlling a flatness in a width direction of a rolled material provided in a rolling process for rolling a rolled material to a desired product by operating a plurality of actuators,
    A shape measuring device for measuring the flatness of each of the plurality of measurement positions set in the width direction of the rolled material,
    A flatness target value setting device for setting a target value of the flatness at each of the measurement positions,
    Wherein the amount of change in flatness at each measurement position when each of the actuators is manipulated is represented by a polynomial expression in which each of the measurement positions is a variable and each term of the polynomial expression indicates an effect Based on the deviation between the actual value of the flatness at each of the measurement positions and the target value of the flatness at the respective measurement positions using the flatness influence coefficient model multiplied by the coefficient, An operation amount calculating device for calculating an operation amount of the actuator,
    A monitoring device for monitoring whether or not the amount of change in the actual value of the flatness at each of the measurement positions is larger than a predetermined flatness threshold;
    Wherein when the change amount of the actual value of the flatness at each of the measurement positions exceeds the flatness threshold value, the change amount of the actual value of the flatness at each measurement position is correlated with the actual value of the manipulated variable of each actuator, Wow,
    The influence coefficients of the flatness influence coefficient model are identified on the basis of the change amounts of the actual values of the flatness at the respective measurement positions read from the storage device and the actual values of the manipulated variables of the actuators, A flatness influence coefficient arithmetic unit,
    Calculating current learning values of the respective influence coefficients on the basis of the current value of the current time of each influence coefficient and the previous learned value of each influence coefficient and setting the current learning value of each influence coefficient to the manipulated variable computing device A flatness influence coefficient learning value computing device to be set,
    And a flatness degree-of-influence coefficient learning value storage device for storing the learning value of each influence coefficient.
  2. The method according to claim 1,
    Wherein the monitoring device monitors whether an actual value of the manipulated variable of each of the actuators is larger than a predetermined manipulated variable threshold value,
    Wherein the storage device replaces the actual value of the stored manipulated variable with zero for an actuator whose actual value of the manipulated variable does not exceed the manipulated variable threshold value.
  3. 3. The method according to claim 1 or 2,
    The flatness degree influence coefficient coefficient order calculating means for calculating a flatness degree of influence coefficient coefficient order which changes the order of the flatness degree influence coefficient model in accordance with a change in the average value of the absolute value of the deviation between the actual value of the flatness degree at each measurement position and the target value of the flatness degree at each measurement position Wherein the apparatus further comprises an apparatus.
  4. 3. The method according to claim 1 or 2,
    Further comprising a flatness degree influence coefficient degree arithmetic unit for changing the degree of the flatness degree influence coefficient model in response to a divergence situation of an actual value of an operation amount of each of the actuators.
KR1020177000992A 2014-09-25 2014-09-25 Flatness control device KR101749018B1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/JP2014/075488 WO2016046945A1 (en) 2014-09-25 2014-09-25 Flatness control device

Publications (2)

Publication Number Publication Date
KR20170018419A KR20170018419A (en) 2017-02-17
KR101749018B1 true KR101749018B1 (en) 2017-06-19

Family

ID=55580505

Family Applications (1)

Application Number Title Priority Date Filing Date
KR1020177000992A KR101749018B1 (en) 2014-09-25 2014-09-25 Flatness control device

Country Status (4)

Country Link
JP (1) JP6229799B2 (en)
KR (1) KR101749018B1 (en)
CN (1) CN106457325B (en)
WO (1) WO2016046945A1 (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3461567A1 (en) * 2017-10-02 2019-04-03 Primetals Technologies Germany GmbH Flatness control with optimiser
JP6784253B2 (en) * 2017-11-22 2020-11-11 東芝三菱電機産業システム株式会社 Shape control device for cluster rolling mill
WO2020129173A1 (en) * 2018-12-19 2020-06-25 東芝三菱電機産業システム株式会社 Learning control device for rolling process
TWI675708B (en) * 2019-01-28 2019-11-01 中國鋼鐵股份有限公司 Hot rolling system and method for steel strip
JP2021115589A (en) 2020-01-24 2021-08-10 東芝三菱電機産業システム株式会社 Rolled material shape control system

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0513730B2 (en) * 1985-03-15 1993-02-23 Mitsubishi Jukogyo Kk
JP2588233B2 (en) * 1988-02-15 1997-03-05 株式会社東芝 Rolled material flatness control device
JPH05119806A (en) * 1991-10-25 1993-05-18 Toshiba Corp Flatness controller
JPH09174128A (en) 1995-12-26 1997-07-08 Kawasaki Steel Corp Method for controlling shape of rolled stock
SE529074C2 (en) * 2005-06-08 2007-04-24 Abb Ab Method and apparatus for optimizing flatness control when rolling a belt
PT2505276E (en) * 2011-03-28 2013-12-05 Abb Research Ltd Method of flatness control for rolling a strip and control therefor
CN102658298B (en) * 2012-04-29 2014-07-23 北京科技大学 Plate-shape quality online judgment method applicable to hot-rolled thin strip steel
CN103949481B (en) * 2014-04-23 2016-01-13 北京科技大学 Take into account the flatness Discrete control method of Hot Rolling Strip stability and quality

Also Published As

Publication number Publication date
KR20170018419A (en) 2017-02-17
WO2016046945A1 (en) 2016-03-31
JPWO2016046945A1 (en) 2017-04-27
JP6229799B2 (en) 2017-11-15
CN106457325B (en) 2018-06-29
CN106457325A (en) 2017-02-22

Similar Documents

Publication Publication Date Title
KR101749018B1 (en) Flatness control device
KR101617375B1 (en) Feedforward thickness control method for performance of cold rolling mill
JP5759206B2 (en) Learning coefficient controller
JP4988171B2 (en) Rolling mill control device
JP2006224154A (en) Method for controlling thickness in tandem mill
KR101832653B1 (en) Rolling process learning control device
KR20040043911A (en) Apparatus and method for prediction rolling force using neural network
JP4983589B2 (en) Control device for cold continuous rolling equipment
JP3223856B2 (en) Rolling mill control method and rolling mill control device
JP6315818B2 (en) Control device and control method for tandem rolling mill
JP2005152919A (en) Correction method of plate crown estimation model
JP6020263B2 (en) Rolling load learning control device and learning control method, and metal plate manufacturing method using the same
KR100929015B1 (en) Prediction of rolling load by calibrating plasticity factor of rolled material
JP2004255436A (en) Method for controlling sheet crown in tandem hot rolling mill
JP2588233B2 (en) Rolled material flatness control device
JP6057774B2 (en) Identification method of mill elongation formula in rolling mill
KR102002237B1 (en) Plate width control device of rolled material
KR100851200B1 (en) Method for forecasting roll force of roughing mill considering width reduction
JP3949597B2 (en) Method for preventing plate bending at the tail end of material to be rolled in tandem rolling mill
KR100641755B1 (en) Control apparatus for width margin in hot strip mill and its method
KR100929013B1 (en) Roll gap setting method of rolling mill by automatic prediction of mill constant and gauge meter error
JP2019055415A (en) Plate crown control method, plate crown control device, and steel plate manufacturing method
JP5565214B2 (en) Thickness control method of rolling mill
JP5970969B2 (en) Shape control apparatus and shape control method for rolling mill
KR950013243B1 (en) Control method and apparatus for rolling mill

Legal Events

Date Code Title Description
A201 Request for examination
E701 Decision to grant or registration of patent right
GRNT Written decision to grant