JPH0259003B2 - - Google Patents

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a rolling mill position control method,
In particular, the present invention relates to a control method for improving the thickness accuracy of rolled materials.

[Prior Art] FIG. 8 is a typical example of a roll position control system for a rolling mill, and shows a rolling mill equipped with a control device of the Bisla control system. In the figure, 1 is a rolling mill, 2 and 3 are work rolls, 4 and 5 are backing rolls, 6 is a ram of a reduction cylinder 7, 8 is a load cell, 9 is a displacement meter of the ram 6, and 10 is an arithmetic and control unit.

In Fig. 8, if the command value is X ₀ , the rigidity coefficient of the rolling mill (generally called mill constant) is K _n , the load constant is C, the displacement of the ram is X, and the rolling load is P, then in Bisla control, X= The displacement X of the ram 6 can be obtained from the relationship: X ₀ +C/K _n P (1). That is,
The ram 6 is displaced by C/K _n P greater than the command value X ₀ to compensate for the expansion of the roll gap 11 due to the elongation of the rolling mill 1 and eliminate plate thickness fluctuations. Furthermore, when C=1, it is possible to completely compensate for the influence of elongation of the rolling mill 1, and it is said that the ultimate goal of the Bisla control system can be achieved.

FIG. 9 shows an example of the configuration of a control device for a conventional Visla control system, and the area within the two-dot chain line indicates the arithmetic and control device 10 of FIG. In this control system, first, before rolling, a command value X ₀ of the ram position is given to the comparator 12 in accordance with rolling conditions such as the finished thickness of the rolled material S. This value X ₀ is compared with the ram displacement X detected by the displacement meter 9 in the comparator 13, and the deviation Δε=X ₀ −X is obtained. Δε is appropriately operationally amplified by an operational amplifier 14 and becomes an input to a servo valve 15. The servo valve 15 controls the piping 1 depending on the magnitude and direction of the input.
6 to control the amount of pressure fluid flowing into and out of the compression cylinder 7. As a result, the ram 6 in the cylinder 7 moves up and down, the backing roll 5 and work roll 3 supported thereon move, and the roll gap 11 is changed. The amount of change in roll gap 11 is ram 6
The displacement is detected by the displacement meter 9 in the cylinder 7. When the output X of the displacement meter 9 matches the command value X ₀ ,
Since the deviation Δε becomes zero and the input to the servo valve 15 becomes zero, the flow of fluid into and out of the cylinder 7 is stopped, and the ram 6 stands still at that position. That is,
The initial roll gap is set.

At the start of rolling, when the rolled material S is caught in the roll gap 11 and a rolling load P is generated, it is detected by the load cell 8, and its output is sent to the lock-on circuit 1.
7. The lock-on circuit 17 is, for example,
As shown in FIG. 10, after the rolled material S is bitten, the switch 19 is turned on at an appropriate timing, and the rolling load P ₀ at that time is stored in the memory 20 and output to the comparator 21. do. In other words, it provides a reference value for the rolling load for Bisla control. Thereafter, the computing unit 21 detects the output P ₀ of the memory unit 20 from time to time, computes the difference in rolling load, and calculates the variation ΔP in the rolling load. This ΔP is caused by disturbances such as changes in the thickness of the inlet side and temperature unevenness of the rolled material S. As a result, the rolling mill 1 is deflected, the roll gap 11 is varied, and the thickness of the outlet side of the rolled material S is changed.
In order to compensate for this, the calculator 18 multiplies the elongation by 1/K _n to determine the elongation of the rolling mill 1, further multiplies it by a preset constant C to determine the correction amount ΔX, and sends it to the comparator 12. Add. This becomes the target value for position control of the ram 6, and the ram 6 moves by ΔX as described above. Therefore, changes in roll gap caused by elongation of the rolling mill are compensated by a predetermined amount.

In this way, in the Bisla control system, changes in the roll gap 11 during rolling, which cannot be directly detected, are indirectly determined by measuring the rolling load, and the ram 6 of the rolling cylinder is adjusted to compensate for the changes in the roll gap 11 during rolling. moving position.

By the way, in order for this bisla control to function effectively, it is necessary that the relationship between the elongation of the rolling mill and the rolling load at that time be an ideal proportional relationship (linear) as shown in FIG. In this way, the load fluctuation ΔP caused by, for example, a change in the entrance plate thickness can be detected by the load cell 8 and divided by the mill constant K _n determined in advance to determine the elongation ΔX of the rolling mill.

However, in reality, the mill constant K _n is not always a constant value, and especially when the diameter of a multi-high rolling mill with a large number of rolls or rolling rolls is reduced, the mill constant K n is The elongation of a rolling mill and rolling load often do not have a constant proportional relationship.

To set the mill constant of the Bisla control system in such a state, the curve showing the relationship between the elongation of the rolling mill and the rolling load shown in FIG. There are possible ways.

[Problems to be solved by the invention] However, in the above method, when the same disturbance occurs, the mill constant changes depending on whether the steady rolling load is low or high, so the same correction coefficient C is set. However, there was a problem that the correction amount ΔX was different and the thickness accuracy on the exit side varied depending on the conditions.

[Means for Solving the Problems] The present invention provides a rolling mill using a Visla control system that controls the exit plate thickness to a constant value, by multiplying the value obtained by dividing the rolling load by a mill constant by an appropriate coefficient. When determining the amount of correction by control, the Mill constant is set as the value of the maximum slope part of the Mill constant curve obtained by actual measurement, and the difference between the reciprocal of the maximum Mill constant value and the reciprocal of the actual Mill constant is calculated. The rolling force is multiplied by the rolling load and added to the correction amount for Bisla control, the obtained correction amount is added to an external command value, and the rolling mill position is controlled based on the added value. It is.

[Function] Therefore, regardless of the operating conditions of the rolling mill, that is, the rolling load, a predetermined control effect can always be obtained, and the exit side plate thickness deviation can be eliminated.

[Example] In order to explain the basic principle of the present invention, the Mill constant curve as shown in Fig. 12 is changed to a constant value of K _n1 in the low load region and K _n2 in the high load region as shown in Fig. 2. Approximate with two straight lines.

In the conventional Bisla control system shown in Fig. 9, when the mill constant is K _n1 (when P<P ₁ ), the outlet plate thickness h ₁ is calculated by setting the lock-on value of the rolling load to P ₀ at that time, h ₁ = S ₀ + P ₀ It is given by /K _n1 +ΔP/K _n1 −CΔP/K _n1 (2). Here, S ₀ is the initial roll gap, P ₀ /K _n1 is the steady elongation, and the normal ram position command value X ₀ is set in consideration of this elongation.
That is, the roll gap S ₀ is given so that S ₀ =h ₁₀ −P ₀ /K _n1 (3). In other words, the elongation of the rolling mill due to the load P ₀ is predicted and the roll gap is narrowed in advance. _h10 is the target plate thickness.

In equation (2), ΔP/K _n1 is the plate thickness deviation caused by changes in the entrance plate thickness, hardness differences in the rolled material, etc.
C・ΔP/K _n1 is the correction amount due to the Bisla control mentioned earlier, and rearranging the third and fourth terms of equation (2) gives Δh ₁ =1−C/K _n1 ΔP …(4), This corresponds to the apparent Mill constant set to Ke=K _n /1-C (5) by control. Since the Mill constant can be arbitrarily set by changing the value of C, this is sometimes referred to as Mill constant variable control.

After the rolled material is caught in the roll gap, it is locked on at a certain timing and the bisler control is performed, and the thickness deviation on the exit side is given by equation (4).

Similarly, when the Mill constant is K _n2 , Δh ₂ =1−C/K _n2 ΔP (6).

By the way, the load constant C can be expressed as equation (4) for the disturbance ΔP caused by changes in the plate thickness on the entry side, differences in hardness of the rolled material, etc.
If they are set to the same value in equation (6), Δh ₁ >Δh ₂ because K _n1 <K _n2 . That is, in a region where the rolling load is low, the plate thickness deviation increases with respect to the same disturbance as the mill constant decreases. this is,
This also poses a problem in terms of product quality.

In order to solve this problem, _{the load} for Bisler control when _{K n1 is} Just adjust the constant C. If the load constants at K _n1 and K _n2 are C ₁ and C ₂ respectively, then from equations (4) and (6), 1-C ₁ /K _n1 ΔP=1-C ₂ /K _n2 ΔP …( 7) From this, we obtain C ₁ = 1-K _n1 /K _n2 (1-C ₂ )...(8). That is, if the load constant C ₁ is selected as shown in equation (8), the plate thickness deviation will be the same value as Δh ₁ =Δh ₂ in the region of Mill constant K _n1 and in the region of K _n2 in response to disturbance ΔP. Substituting equation (8) into equation (4) and rearranging, Δh ₁ = ΔP/K _n1 + (-1/K _n1 + 1/K _n2 −C ₂ 1/K _n2
) ΔP… (9) The first term is the elongation of the rolling mill due to the actual disturbance ΔP, and the term in the box is derived from C ₁ .
That is, this part is configured in the control device as hardware using an electronic circuit or software using a computer.

Also, as can be seen from equation (9), in the Bisla control device of the present invention, the elongation of the rolling mill (the first term of equation (9)) caused by the mill constant K _n1 is completely corrected by control ( (1/K _n2 −
C ₂ /K _n2 )・ΔP is created and added.
This is simply a realization of equation (6) through control.

This corresponds to linearizing the mill constant as shown in FIG. 2 as shown in FIG. It is characterized by removing thickness deviation. Therefore, for disturbances of the same magnitude, a constant control effect can always be obtained in the entire region used.

Up to now, in order to explain the principle of the present invention, the Mill constant curve of an actual machine as shown in FIG. 12 has been simplified as shown in FIG. 2. It was also found that in order to realize the present invention, the relationship expressed by equation (9) should be configured in the control device.

In order to implement the present invention with respect to the mill constant curve as shown in FIG. 12 in an actual machine, the relationship between the elongation X of the rolling mill and the generated load P must be The reciprocal of the natural Mill constant (K _n in Figure 4a), which is not linear, is determined by taking P on the horizontal axis, as shown in b, and is stored in a storage device. Then, read out each corresponding P and multiply by ΔP obtained by the circuit shown in FIG. 10 to create ΔP/K _n . This becomes the first term in the bracket of equation (9). In addition, the mill constant in the region where the rolling load is high and the mill constant is a constant value (P>P ₁ ) is
Assuming K _nx , it is sufficient to set the reciprocal of this value in the gain circuit and input ΔP into it to create ΔP/K _nx . This becomes the second term in the bracket of equation (9). In the same way, further multiply by the correction gain C,
The third term in the bracket of equation (9) can be created.

The above description will be explained with reference to the block diagram of the embodiment shown in FIG. FIG. 1 shows an example of a control device for carrying out the method of the present invention, and the same parts as in FIG. 9 are given the same reference numerals. The method of setting the initial roll gap in this control device is as already described in FIG. 9, and the difference from the conventional control device is the configuration of the bisler control system that functions during rolling.

At the start of rolling, when the rolled material S is caught in the roll gap 11 and a rolling load P is generated, it is detected by the load cell 8 (see FIG. 8) and input to the lock-on circuit 17. The lock-on circuit 17 is the first
0, it stores the rolling load at an appropriate timing, provides a reference value for Bisla control, and calculates and outputs the load fluctuation ΔP. This ΔP is input to the gain circuit 22 and the memory/calculator 23.
The gain circuit 22 multiplies ΔP by 1/K _nx and outputs the result.

The output of the gain circuit 22 is input to an adder/subtractor 25 and also to another gain circuit 24, multiplied by a load constant C, and input to the adder/subtracter 25. On the other hand, ΔP input to the memory/calculator 23 is multiplied by the value of 1/K _n read out according to the value of the rolling load P detected by the load cell 8,
The signal is input to the adder/subtracter 25.

As a result of the above, the adder/subtractor 25 adds ΔX=1/K _n ΔP−1/K _nx ΔP+C/K _nx ΔP (10) to the comparator 12 as a correction amount. This corresponds to the term inside the box in equation (9). Note that the signs of equations (9) and (10) are opposite to each other. Equation (9) represents the plate thickness deviation (+ increases) as a result of control, whereas (10) This is because the equation ) represents the target value of the ram position.
The ram 6 moves by ΔX according to the target value for position control of the ram 6 expressed by equation (10). Therefore, as mentioned above, the Mill constant is corrected to be K _nx in the entire load range, and then the Bisler control functions and the plate thickness disturbance is removed.

In the present invention, the Mill constant apparently becomes Ke=K _nx /1-C (11) through control.

FIG. 5 is a diagram for explaining another embodiment of the present invention. When the rolling load P is less than P ₁ , the mill constant of the rolling mill has nonlinearity as shown by K _n in the diagram. . This has already been shown in FIG.

In the first embodiment of the present invention shown in FIG. 1, this natural Mill constant is linearized as K _nx by control so that it becomes a constant Mill constant for all load changes. However, in practice, it is not necessary to linearize by using the Mill constant of load P ₁ or less as K _nx ; for example, as shown in FIG. 5, it is sufficiently effective to increase K _n appropriately. For that purpose, (9)
It is sufficient to appropriately correct the first and second terms in the brackets of the equation. That is, for example, by introducing the parameter α (0<α<1), equation (9) can be changed to Δh ₁ =ΔP/K _n1 −α(1/K _n1 −1/K _n2 )ΔP−C ₂ ΔP
Replace with /K _n2 …(12). In equation (9), from the third term on the right side, Δh ₁ ′ = ΔP/K _n1 + (-1/K _n1 + 1/K _n2 ) ΔP = ΔP/K _n2 … (13) In other words, the Mill constant is expressed as K in the entire region. _n2 , and by making a correction using the fourth term on the right side, the Mill constant is set as ΔP/Δh=K _n2 /1−C (14).

On the other hand, in equation (12), from the third term on the right side, Δh ₁ ′=ΔP/K _n1 −α (1/K _n1 −1/K _n2 ) ΔP = 1−α/K _n1 ΔP+α/K _n2 ΔP Therefore, ΔP/Δh ₁ ′=1/1−α/K _n1 +α/K _n2 (15). This consists of two springs as shown in Figure 6.
This is equivalent to coupling K _n1 /1-α and K _n2 /α.

When α=0, Mill constant control in this embodiment is not performed at all, and when α=1, the first
This corresponds to the completely straightened case of the illustrated embodiment.
By setting α to an appropriate value between 0 and 1, as shown in Figure 5, an appropriate Mill constant K _n (i) (i = 1, 2, 3, etc.) with P < P ₁ can be obtained. Since (K _n (i)<K _{nx )} , a correction is made to this using the fourth term on the right side of equation (12).

Based on the above explanation, a specific example of the configuration of the control device is shown in FIG. Basic functions are the first
It is similar to the figure, but the difference is that the gain circuit 22,
This is the point where the output of the memory/arithmetic unit 23 is input to the adder/subtractor 26, and the output of the adder/subtractor 26 is input to the gain circuit 27 to produce α(1/K _n -1/K _nx )ΔP. Therefore, the correction amount ΔX is ΔX=α(1/K _n −1/K _nx )ΔP+C/K _nx ΔP (16).

Note that the control method of the present invention can be implemented by either hardware using an electronic circuit or software using a computer. It goes without saying that a combination of the two can also be implemented. Further, in the embodiment, a method (lock-on method) of giving the target value of the load by storing the initial rolling load was shown, but this value may also be given by creating a prediction formula and calculating it.

[Effects of the Invention] As explained above, according to the present invention, the Mill constant can be controlled to a constant value or a substantially constant value over the entire range, so that the Bisler constant can be set to a constant value or a substantially constant value over the entire range. By performing control, it is possible to eliminate the exit plate thickness deviation, and therefore,
A constant control effect can always be obtained over the entire range of operating conditions of the rolling mill.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an apparatus for carrying out the method of the present invention, FIG. 2 is a simplified diagram of an actual Mill constant curve, and FIG. 3 is a block diagram showing an apparatus for implementing the method of the present invention. Figures 4a and 4b are diagrams for explaining the control contents of the present invention, respectively, and Figure 5 is a diagram for explaining another embodiment of the present invention. Fig. 6 is an explanatory diagram of control contents in another embodiment of the present invention, Fig. 7 is a control block diagram in another embodiment of the present invention, and Fig. 8 is Bisler control. Fig. 9 is a control block diagram of the Bisla control system, Fig. 10 is an explanatory drawing of the lock-on circuit, and Fig. 11 shows the relationship between elongation and load in an ideal rolling mill. 12 are diagrams showing the relationship between elongation and load in an actual rolling mill. 1 is a rolling mill, 6 is a ram, 7 is a reduction cylinder, 8
is a load cell, 9 is a displacement meter, 10 is an arithmetic and control unit, 17 is a lock-on circuit, 22 is a gain circuit,
23 is a memory/calculator, 24 is a gain circuit, 25,
26 is an adder/subtractor, and 27 is a gain circuit.

Claims (1)

[Scope of Claims] 1. In a Visla control type rolling mill that controls the exit plate thickness to a constant value, when determining the correction amount by Bisla control by multiplying the value obtained by dividing the rolling load by the mill constant by an appropriate coefficient, The above correction amount is determined by using the mill constant as the value of the maximum slope part of the mill constant curve obtained by actual measurement, the obtained correction amount is added to the external command value, and the rolling mill is adjusted based on the added value. A rolling mill position control method characterized by position control. 2. In a Visla control type rolling mill that controls the exit plate thickness to a constant value, when calculating the correction amount by Bisla control by multiplying the value obtained by dividing the rolling load by the mill constant by an appropriate coefficient, the mill constant is calculated by actual measurement. Find the above correction amount as the value of the maximum slope part of the Mill constant curve obtained by Find the second correction amount by multiplying by A method for controlling a rolling mill position, the method comprising controlling the rolling position of a rolling mill. 3. The rolling mill rolling position control method according to claim 2, wherein the second correction amount is multiplied by an appropriate coefficient of 0 or more and 1 or less.