JPH0491811A - Method for controlling edge drop in rolling plate - Google Patents

Abstract

PURPOSE:To prevent an edge drop control from overshooting and huntering by conducting a feedforward control based on an edge drop of a plate before rolling and correcting a coefficientue value of a control model successively by feedback. CONSTITUTION:An target edge profile of the outlet side of the final pass of rolling is determined and a deviation from this target value is made a parameter or a dependent variable. Further, a mathematical model where an edge profile parameter before rolling and a shift position of a shift mechanism are independent variables is made previously. While material is rolled, feed forward control of a shift in the direction of the width of a work roll in the upstream side pass is conducted based on measured values of the edge profile of the plate before rolling. The model coefficient is corrected successively based on measured values of the edge drop before rolling, on the outlet side of the final pass and at a shift position in the direction of the work roll width of each pass to converge the calculation error of the model actually into zero by feedback.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to edge drop control in plate rolling.

[Conventional technology]

In continuous cold rolling of plates, a multi-pass rolling mill is installed on the rolling line to control edge drops in order to roll the rolled material at high speed and improve edge drop. In such equipment, the target edge drop of the material at the final position of the rolling pass is determined as the target edge drop value.

Edge drop control performs Bricent control that outputs the set value before rolling the material, and in the final pass, the J position is ``J Tsubu? On the flow side of the rolling line, there is also a work roll with a tapered body end (hereinafter referred to as a t-dever) in as few rolling mills as possible.
1. Adjust the lower crawl shift). Rolling is the most suitable method from both the viewpoint of quality assurance and production cost reduction3.
, In order not to satisfy such conditions, the workpiece τ of each path of the line ``-flow'' is determined and controlled by the Lenoft amount.
9.1. In engineering and sojtrosop control, it is necessary to deal with errors in briscent control, VJ:, fluctuations in edge drop of colored plates, fluctuations in rolling conditions, etc., and to control the total length of the material. Nick ID at the final pass exit position - Lucif l
-a needs to be adjusted.

Conventionally, regarding this control, Japanese Patent Application Laid-Open No. 1986-12
As disclosed in Publication No. 213, the godge tip on the exit side of the final pass was actually measured, and the "No."BuJ'!;ll's Ent "Shi...
The plate is placed in the rolling line at 1., = -C, and forward control is performed based on the measured value. (Ding/Non) 'O Tsubu actual measured value is used for 4 learning steps (,, is it?
Regarding the braking during monthly charge rolling using edge drop control,
In actual rolling, depending on the condition of the plate before rolling, it is possible to avoid the control time delay due to the measured value of edge drop in the final pass and the material movement time LJ between the workpiece shift positions in each pass. It takes some time to match the target value.

In addition, control overshoot and hunting are likely to occur.The present invention does not perform feedback control of the deviation from the edge drop target value during rolling of the same lot material, but rather performs feedforward control based on the edge drop of the plate before rolling. Through control, the material edge drop in the final pass is brought to the target edge drop, and the control model (
Edge Drop Deviation Model) By sequentially correcting the coefficient value in a feedback manner based on the edge drop of the plate before rolling, the work roll shift amount of each pass, and the actual value of each edge drop at the exit side of the final pass, the feed An edge drop control method in plate rolling that realizes control with both forward and feedback functions is provided.

[Means to solve the problem]

The present invention is aimed at controlling the edge profile of a plate by a multi-pass rolling mill having a mechanism for shifting the work roll in the direction of the axis of the work roll with a tapered roll body end when rolling a steel material. When a target edge profile is determined, the edge profile is expressed by at least two or more parameters, and the deviation from this target value is used as a parameter, which is used as a dependent variable, and the edge profile parameters before rolling and multiple passes and multiple juice rolling are A mathematical model is created in advance with the shift position of the machine's shift mechanism as the q variable, and during material-1 rolling, the edge profile of the plate before rolling is actually measured, and based on this measured value, the errani-no-trot deviation is calculated. Using the model, feedforward control is performed by momentarily adjusting the width direction shift of the work roll in the downstream pass so that the edge profile on the final lot exit side matches the target edge profile. The edge drop of the plate before rolling and the edge drop on the exit side of the final pass when passing through the pass,
In plate rolling of rolled steel plate material, the model coefficients are successively corrected based on the actual measured value of the width direction shift position of the work roll in each pass, and the calculation error of the model is converged to substantially 0 in a feedback manner. This is an edge drop control method.

[For production]

An overview of one aspect of a multi-pass multi-rolling mill and the system configuration of edge drop control, which is the subject of the present invention, is shown in Fig. 1.
,3 1st [to]! 7 in ： indicates a multi-bus multi-rolling mill, 1 is the steel plate rolled by the rolling mill, 1 is the starting point of the line, 8 is the ending point, and 5 is (]-: rolling. 8 is an edge drop detector for measuring the thickness distribution in the width direction of the strip on the rolling mill side; 8 is a J-entrobe detector for measuring the thickness distribution in the width direction of the strip at the exit side of the rolling mill; i is the strip width detector at the entry side of the rolling mill, the steel plate 1 at point 0 is given the edge drop 1] target value λ, radii: , 1 is I] Babas In the example, there is a multi-rolling mill with 5 baths).
J Functional Work [1-2 and H-Fujishi Full: (
Equipped with 1, edged o Tsubu's change is -ra 1
When the plate on the inlet side of the rolling mill is controlled by 0; The workpiece of each Kawaishi rolling machine is getting bigger〕
- control the shift position Col Bee, 1-Event operation (, “
- Set to 14 pressures for each C through cylinder 1 to roller 10'4
The ends of this binding material are reduced in size and the Elzno drop of the rolled plate is reduced.

[Example] In the back control to the word and the Noy 1 in the same lot, the actual measured values of the Tososhito'IJ knob of the rolling plate at the 8 points shown in Fig. 1 are used to 3. By adjusting the feed forward amount of the workpiece [1-L/·ft, the edge drop at the 0 point is controlled to the target edge F (T:/B), and the actual edge drop value at the 0 point and edge trod deviation model are Due to the error with the old calculated value,
In intra-lot control, in which the model coefficients are corrected sequentially in a feedback manner, the edge drop of the embroider plate [7 Tsuzu Kaaruko 2] is a factor that greatly influences the edge drop on the final U -, the horn drop deviation model is expressed as 1, Δi), (
t ) i11+ a2 ” Δl−, (t ) −a, i,,
#ΔSw, (t')Δ1. ','',) 2 (t
) b, t b2・1Δr2(t)−b3・Δr2' +
(+) +b,・△SW2 (t) ΔD3(t) C1+ C:, 2・Δr−3(t,'+ ±
-C3・ΔSshield't(t)4C1・△S1%'z
(t) + C5・Δ5IVa(tJ here CΔD 1(
t)=D +(t)-1)Δ11)x(t)=D
z(tL-D 2Δ d) 5(t)== o
an) -〇.

△r i (t) = V l(i > −rΔd′2(
t)=r z(tL-r2Δr' :+(t)=
r a(t)-1”.

△Sw+(t)=5w1(t)-SWI”△Swz(
t)=Sw2(t)-3W2”ΔSty, (t)=5
w3(t) -3w+"where, ,D, and , are the edge drop amounts that the board passed through the final pass exit side at time t, and in the case of edge 1 heropob control by the first bus, D i is Also, in the case of the control up to the second bus I)
and I-)2 for control of up to the third bus r]■
), 1 and ri.

Subscripts 1, 2 and 3 indicate control parts in the board width direction2,
In the case of control using the first pass, a point 15 mm in the width direction from the plate end is set as 1, and in the case of control up to the second bus, a point 30 mm from the plate end is set as 1. The point 15mm from the plate end is set as 2 and 1, and in the case of control up to the third pass, the point 3 is set from the plate end.
0rnm point 1. The point 20 mm from the edge of the plate is 2 degrees, and the point 15 mm from the edge of the plate is 3.

These control parts vary depending on the material, construction conditions, etc. , , r2 and `` are the edge drops of the plate at the entrance of the rolling mill at time t, and the subscripts 1.2 and 3 are as above, , D, and .

This is the same as shown in . That is, the strip widthwise drop control location on the exit side of the rolling mill must also be a measurement point at the same location in the strip width direction on the entry side of the rolling mill.

D I + '5j 2 and Kan3 are D,, D2 and D2) 1] target value of edge drop of 3, 5, '1.
■”′.

and 1′,′ are used in edge drop prediction model calculation before material rolling in preset control.

This is the esonto tip of the plate before rolling, and the subscript l.

2 and 3 are the same as the description of the subscripts above. S
v, S%I1. and 3w3 represent the tapered work roll shift position in the nominal pass from the first stream side of the rolling mill, and subscripts 1, 2, and 3 are bus numbers.

Swl”, Sw Bi and Swi.1 is the taper work roll shift position for each bus determined by preset calculation. a1+ az+ a3+ bl+
b2+ b,++1), ,C: ,C2,C,+ ,
Ca and C5 are model coefficients.

In this example, l'), , D, and I) 3 (7)
An edge drop deviation model is constructed using three parameters. This deviation model is based on the edge drop of the plate before rolling. About the last bus exit, ]-tsu, ;5'g
The position of the work roll shift is also set individually.

Accordingly, Fig. 1 shows a machine detector 8 for measuring the plate width distribution in the direction of the plate width at the exit side of the bus at the end of the small bus.
5. Control Method for Controlling Both Ends of the Board Since the control method is the same for both ends, the following explanation will focus on one control method. Also, the edge drop deviation model ΔI)+(1:), ΔD 2(t> and ΔLl, (
Since ΔN), (t) and ΔD3(t,) in 11) are the same as the control method for ΔD+(t) described below, here we will describe the method for sequentially correcting the model parameters of ΔD, (t). decrease.

In the edge drop deviation model △D + (t), the King's Del coefficient a (+ a 2 and a are successively corrected during rolling, so they are hereinafter expressed as al * ai * ai... During intra-lot control is the value of 1, Sw: ° and a 1. + of the material 1'4 before rolling.
It is necessary to determine the initial values of a 21 a , , h-, and they are determined as follows. , Swi.1'' is the workpiece D-wheel shift position setting value determined by the preset control calculation, and I'' is the edge of the plate on the inlet side of the rolling mill used for the preset control calculation. The initial value of 21 is (], and the initial values of 1, a, , a are buttons, ,
How does r change?
Find y) using the following formula, a, l! ” a I) 1.
aIT(1',, =V 1. Sw, =Sw,'''
) a , =aD 1"' a N"("l = r' 1 , Sw ]
”-Sw °) The value of the above influence coefficient is 1゛,
1) The amount of change ΔI] due to Δr 1 and ΔSw is calculated using the edge drop prediction model 9. During rolling of the same lot, the rolling mill is When the edge drop of the board at the entrance side is actually measured 1, and the edge drop at that measurement point is I'lft>, the edge drop of the board at the exit side of the final bus coincides with D 1 (t) or the target edge drop officer. The work roll shift Sw+(t) is determined by the following equation using the edge drop deviation model of equation (1).

! nv, (tL=sw1　゛　al　+a.

(r”, (t) −U, ')
I 2'',, f2) The value of Swi(t) is
Material 1+ rolling machine, K side eso, 1-・L]・Knob detector 5
Taking into account the travel time to the first bus (the first bus) and the slowness of the work roll shift,
Materials or workpieces [7] When passing the bus subject to software operation,
1. Work at appropriate timing so that the work roll shift on the bus to be operated by software is Sw,N). While performing foot-forward control of the tip of the plate at the exit side of the UF rolling mill (J), the estimated model coefficient values Rii1A knee, vertical -H+, - are sequentially estimated using the following formula, a (t) - a, (t-1)→-p(t,)・x(1
)・ε(1)... (3) Here a (t): [a t(t), a,
−(tL ”,,,,i,,,(,,t)]”x(t
)=(1, ΔI'i(t,), ΔSiv+(t)J
1ε (1) = ΔL) +(t)
a (L-1)1 ・ x (t) where t is time (Zumbling period unit), Δi-+m, ΔS
w, (口, ΔI) l (t) is the actual value of the material that passed through the outlet side of the Tsukinobu machine at each time t, P (t) ( -g
Using multiplication, ``(1-find white), this successive 伽I]- is the error of the Ko-Sontrosov polarizer model, the same 7 Soto-Ten, Enchu...:〕iE: Change in the sword condition And for various disturbances, etc. It is difficult to measure the value during rolling under rolling conditions by performing the control operation every time from the start of rolling to the end of rolling. It is possible to use the influence coefficient for the control amount calculation (-del) to improve adaptability even to various factors (for example, heat crown at zero), and it is always possible to use the Good person: Tsushido: ・B control or 1-1J ability.

Furthermore, in the normal rolling state by using this Rt, self-learning of P(1) was performed during rolling.
, 'm, human-independent data analysis (, - creation of a mathematical model, etc. is no longer necessary.) At the start of learning, a corrective action with a large influence coefficient is activated, or as learning progresses. It is possible to make the mouth correction movement small and to make the mouth correction motion small. Work opening - Lua "-) To (7,) change and F; [□' I:, U
Modify the parameters of the knob prediction model to reduce the size of the mouth to 5tIII and reduce the time to 1t.
．．・1・Ll-) Actual measurement value and -7-ri of the 18th P [7-
Take in the actual measured value 6 of the 7 feet, L ノ':' l・口・I
Step deviation (input the data into the Del formula and calculate the final output side of Eff. [Lodge (ΔI)1)d),
)l or (), end the process of performing the setting output J'
do. If ΔD is not 0, the work [l
~ Run: Find the 7th position, strange! It! Output the amount. Next, the edge drop deviation actual value at the final bus exit side is taken in, and the parameters of the edge drop deviation model are updated.

. Figure 4 shows the controlled forming during rolling of the material as shown in Figure 1.
In response to the change in '
While operating the TI control, the final bus exit edge drop size (1) was able to satisfy the target esocitro knob, that is, within the dashed-dotted line by 1 minute - 4. , In the above explanation, the first bus $IJ Hi on Tsuchiura Bus
From this, we have described the case in which the target, machining, and torque requirements are satisfied at the exit side of the final bus. If the formula (1
) is used to operate the two-crawl shift of the second bus or third pass in the current bus. In this embodiment, when edge drop control is performed using multiple buses, the contact length between the plate and a part of the taper of the work roll is increased on the upstream pass side.

1. Effects of the Invention] As detailed above, according to the present invention, during plate rolling, 2. In addition to conventional feedback control, 1. control during material rolling is performed. q By adjusting the work roll shift position on the valley bus using the actual measured value of the edge drop of the board before rolling, the feedforward curve is brought to the target value. Regarding the coefficients of the edge drop deviation model used for this calculation p, Material 1-il.
In addition to determining the initial value using the edge prediction model for grinding, and during the material immediate rolling, the actual value of the no-cool shift amount of the final bus exit side [J knob and δ bus] and the By sequentially modifying the model coefficients using the edge 1-1'' of the plate, the calculation error can be reduced to virtually zero through feedback. Through the above process, the control during material rolling can be made as fast as possible. It has both word function and buid pack function,
It is possible to deal with various disturbances and always maintain the target edge drop on the final bus exit side over the entire length of the board.

Figure 2 is a diagram showing the configuration of the work rolls in a multi-roll mill, Figure 3 is a flowchart showing the procedure for changing the work roll shift amount and modifying the parameters of the Endkot tip deviation model, Figure 4 The figure is a drawing showing the control characteristics during rolling of the material.

1...Rolled m plate, 2...Tapered work roll, 3
... Backup roll, 5... Entry side plate edge drop detector, 6... Plate width detector, 7... Multiple pass multi-rolling machine, 8... Output side plate edge drop detector, 9.
...Control computer, IO...Controller agent Patent attorney Masaaki Akizawa Mitsunobu 1 name, 2 figures, 3 figures

Claims (1)

[Claims] When rolling steel materials, the target edge profile of the plate at the exit side of the final pass is determined by controlling the edge profile of the plate using a multi-pass rolling mill that has a shift mechanism in the direction of the axis of the work roll whose roll body end is tapered. is determined, the edge profile is expressed by at least two or more parameters, the deviation from this target value is used as a parameter, and this is used as a dependent variable, and the edge profile parameter before rolling and the shift mechanism of the multi-pass rolling mill are expressed. A mathematical model with the shift position as an independent variable is created in advance, and the edge profile before rolling is actually measured during material rolling. Based on this measured value, an edge drop deviation model is used to set the edge profile at the exit side of the final pass as the target. Feedforward control is achieved by momentarily adjusting the widthwise shift of the work roll in the upstream pass to match the edge profile, as well as the edge drop of the plate before rolling and the final pass as the material passes through the final pass. A steel plate characterized in that model coefficients are sequentially corrected based on actual measured values of the edge drop on the exit side and the shift position in the width direction of the work roll in each pass, and the calculation error of the model is converged to substantially 0 in a feedback manner. Edge drop control method in plate rolling of rolled material.