JPH11290920A - Method for controlling flying change of rolling mill and device therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To elevate the yield of rolled stocks by elevating dimensional accuracy before and behind the flying changing point of the rolled stock. SOLUTION: Rolling load is calculated with a rolling load calculating part 18 based on actual value data which is acquired and processed by an actual value acquiring processing part 17. Based on the rolling load acquired by the rolling load acquiring part 16 and rolling load calculated by the rolling load calculating part 18, a learning factor is calculated by a learning factor calculation part 19. The optimum learning factor is determined from the latest learning factor calculated and the past learning factor by a learning part 20. When the thickness of the rolled stock (a) is changed by the change of a manufacturing specification during rolling the rolled stock (a), each changing amount of the screw-down location of each stand 1-7 is determined by a screw-down location changing amount calculating part 21 based on the learning factor which is previously learned in a learning part 20. Each screw-down location of each stand is adjusted by each calculated changing amount with a screw-down location adjusting device 22.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a running change control method and a running change control device for a rolling mill having at least one stand.

[0002]

2. Description of the Related Art In recent years, in order to improve the yield and quality of rolled materials, complete continuous rolling in which a plurality of rolled materials are joined and continuously rolled has been performed. In this method, in order to manufacture a plurality of types of products from the joined rolled material, a running distance change control for changing a manufacturing specification (a thickness, a width, a crown, and the like) is performed while the rolled material is running.

[0003] In this running change control, taking the thickness control as an example, the rolling load at the tail end of the preceding material at the running change point and the leading end of the succeeding material at the running change point is predicted. Calculate the required rolling position using a known gauge meter formula,
It is common to change the difference between the rolling positions when the joining point passes through a rolling mill (for example, “theory and practice of sheet rolling”
(Published by The Iron and Steel Institute of Japan in September 1984) No. 5.6
chapter).

[0004] In the above running distance change control, in order to improve the dimensional accuracy before and after the joining point, a model formula for correcting the parameters of the rolling load prediction formula based on the actual data of the rolled material at the time of rolling load prediction is used. In many cases, learning control is also performed. In cold rolling where full continuous rolling is first performed, it is common practice to perform learning control based on the actual data of the leading edge of the succeeding material at the change point between running distances of the already rolled material. (For example, the invention described in JP-A-6-15318).

[0005]

However, when the above-mentioned method is applied to the change of the running distance in hot rolling, which has been developed in recent years, when the preceding material tail end of the running change point and the trailing material front end of the running change point are changed. In addition, the difference in the level of the rolling temperature at the entry side of the rolling mill deteriorates the prediction accuracy of the rolling load, and thus has a disadvantage that the dimensional accuracy before and after the joining point is not always good.
For example, as in the invention described in Japanese Patent Application Laid-Open No. 9-192714, inconvenience occurs when a preceding material and a following material are joined on the entrance side of a finishing mill.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a method and an apparatus for controlling a running distance change of a rolling mill in which the dimensional accuracy before and after a change point of the rolling material is improved to improve the yield of the rolling material. Is to provide.

[0007]

In order to solve the above problems and to achieve the object of the present invention, according to the first aspect of the present invention, there is provided a method for controlling a change in running distance of a rolling mill having at least one stand. The data on the rolling before and after the point where the production specification of the rolled material changes during the run passes through the stand is acquired, learning is performed based on the acquired data, and a learning coefficient is obtained in advance. When the manufacturing specification of the rolling mill is changed, a running distance change control method for a rolling mill is provided which predicts a rolling load before and after a running distance change point of the rolled material based on the obtained learning coefficient.

According to a second aspect of the present invention, in the method for controlling a running distance change of a rolling mill according to the first aspect, when the running distance change point of the rolled material arrives at the stand, the predicted rolling load is reduced. The pressure of the stand was adjusted based on this.

Further, the present invention also provides the following measures. That is, according to the third aspect of the invention,
In a running change control device for a rolling mill having at least one stand and capable of adjusting a reduction of the stand during hot rolling, a running change point at which a manufacturing specification of a rolled material is changed passes through the stand. Data acquisition means for acquiring data on rolling before and after rolling, learning coefficient calculation means for learning in advance based on the data acquired by the data acquisition means to obtain a learning coefficient, and changing the manufacturing specifications of the rolling mill. A rolling load estimating means for predicting a rolling load before and after a running change point of the rolled material based on a learning coefficient obtained by the learning coefficient calculating means; Is provided.

According to a fourth aspect of the present invention, in the running distance change control device for a rolling mill according to the third aspect, when a running change point of a rolled material according to a change in manufacturing specifications arrives at the stand, A means for adjusting the rolling reduction of the stand based on the rolling load predicted by the rolling load predicting means is provided.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described below with reference to the drawings. The running change control method and the running change control device of the rolling mill of the present invention,
A description will be given of a case where the present invention is applied to a change in a rolling position (change in a sheet thickness) in complete continuous rolling of a finishing hot rolling mill having a plurality of stands.

As shown in FIG. 1, the finishing hot rolling mill 8 comprises:
First to seventh stands 7 are provided. Each of the first to seventh stands 7 includes a pair of work rolls 9 sandwiching the rolled material a, and backup rolls 10 for backing up the pair of work rolls 9, respectively.

On the entry side of the first stand 1, an entry side thickness gauge 11 for measuring the entry side thickness of the rolled material a and a thermometer 14 for measuring the entry side temperature of the rolled material a are provided. I have. Also,
On the output side of the seventh stand 7, an output side thickness gauge 12 for measuring the thickness of the output side of the rolled material a and a thermometer 15 for measuring the temperature of the output side of the rolled material a are installed. Further, each of the first to seventh stands 7 is provided with a load meter 13 for measuring the rolling load of the rolled material a.

Next, an embodiment of a running distance change control device for a rolling mill of the present invention applied to the finishing hot rolling mill 8 configured as described above will be described. The running distance change control device 30 of this embodiment is actually constituted by a computer and realizes its control by software, and is constituted by the following components having various functions for the realization. . That is, the running distance change control device 30 is configured as shown in FIG.
As shown in the figure, the actual measurement value acquisition unit 16 and the actual value acquisition processing unit 1
7, a rolling load calculating unit 18, a learning coefficient calculating unit 19, a learning unit 20, and a rolling position change amount calculating unit 21.

The actual measurement value acquiring section 16 acquires a rolling load as an actual measurement value measured by each load cell 13 installed in the first to seventh stands 7. The actual value acquisition processing unit 17 can acquire each measured plate thickness of the entrance side thickness gauge 11 and the exit side thickness gauge 12, and each measured temperature of the entrance side thermometer 14 and the exit side thermometer 15. Further, the actual value acquisition processing unit 17 determines the sheet thickness, rolled material temperature, roll speed, etc. before and after the running sheet thickness change point a1 of the rolled material a passes through each of the first to seventh stands 7. The actual value data relating to the rolling is obtained.

The rolling load calculating section 18 calculates the rolling load based on the actual value data acquired by the actual value acquiring section 17 as described later. The learning coefficient calculating unit 19 calculates a learning coefficient based on the rolling load acquired by the actual measurement value acquiring unit 16 and the rolling load calculated by the rolling load calculating unit 18 as described later. . The learning unit 20 performs processing for obtaining an optimal learning coefficient from the latest learning coefficient calculated by the learning coefficient calculation unit 19 and the past learning coefficient.

The rolling position change amount calculating unit 21 uses a learning coefficient previously learned by the learning unit 20 when the thickness of the rolled material a is changed due to a change in manufacturing specifications during rolling of the rolled material a. Thus, as will be described later, the amounts of change in the pressing positions of the first to seventh stands 7 to 7 are determined. The rolling position adjustment device 22 is configured to change the first stand to the seventh stand 7 by the change amount calculated by the rolling position change amount calculating unit 21.
Each of the rolling positions is adjusted.

In the running distance change control device 30 according to the embodiment having such a configuration, for example, during rolling, the production specification (sheet thickness)
Each time there is a change, as shown in FIG. 2, data on the rolling before and after the running thickness change point a1 of the rolled material a passes through the first to seventh stands 7 are obtained,
Since a learning coefficient is obtained in advance by performing learning based on the acquired data, a processing procedure for obtaining the learning coefficient will be described with reference to FIGS.

In the following description, in the case where the running thickness is changed at the joint of the rolled material a, that is, the running distance change point a1 of the rolled material a is determined by the preceding material and the following material. This is the case of a junction point, and it is assumed that a draft schedule (a target value at each stand) is set in advance.

If there is a change in the thickness of the rolled material a, the actual value acquisition processing unit 17 determines that the joining point of the rolled material a is the plate before and after passing through each of the first to seventh stands 7. The actual value data related to rolling such as thickness, rolled material temperature, and roll speed is obtained (step S1). However, since a thickness gauge or a thermometer is not arranged between the stands, the thickness of the rolled material a in a portion that cannot be directly obtained is determined by the measurement thickness of the entrance thickness gauge 11 and the exit thickness gauge 12. The actual value acquisition processing unit 17 obtains the actual value acquisition processing unit 17 based on the thickness, and calculates the temperature of the rolled material a in the part that cannot be directly acquired based on the measured temperatures of the entrance thermometer 14 and the exit thermometer 15. Asks.

The rolling load calculation unit 18 determines that the joining point of the rolled material a acquired by the actual value acquisition processing unit 17 is the first stand 1
Based on each rolling data just before passing through each stand of the seventh stand 7, the rolling load P _Ci of each stand just before the passing is calculated by the formula (1) (step S).
2). Similarly, the rolling load calculating unit 18 calculates the rolling data a based on the rolling data immediately after the joining point of the rolled material a acquired by the actual value acquiring processing unit 17 has passed through each of the first to seventh stands 7. Then, the rolling loads P _Di of the first stand 7 to the seventh stand 7 immediately after the passage are calculated by the formula (2) (step S2).

P _Ci = b × k _mi × O _pi × ｛R _i ′ (H _i −h _i )｝ ^1/2 (1) P _Di = b × k _mi × O _pi × ｛R _i ′ (H _i −h _i )｝ ^1/2 (2) i (subscript): number of the first to seventh stands b: plate width k _mi : average deformation resistance O _pi : rolling force function R _i ′ : flat roll radius H _i: thickness at entrance side (actual value) h _i: output mean deformation resistance k _mi of the side plate pressure (actual value) above approximated by the element as shown in (3), the above rolling force function O _Pi is determined by an element as shown in equation (4), and the above flat roll radius R
_i ′ is obtained by equation (5).

[0023] _{k mi = g (H i,} h i, T i, V Ri, m, n _{other) ··· (3) O Pi =} f (H i, h i, R ') ··· (4 ) R _i ′ = R {1 + [(C · P _i ) / b (H _i −h _i )]} (5) m and n in the equation (3) are constants, and T _i is the temperature of the rolled material. , V _Ri are rolling speeds, and C in equation (5) is a constant.

The actual measurement value acquiring section 16 is installed in the first stand 7 to the seventh stand 7 at a timing before and after the joining point of the rolled material a passes through each of the first stand 7 to the seventh stand 7. The rolling load of the rolled material a measured by each load cell 13 is obtained. Each rolling load obtained at the timing immediately before the joining point of the rolled material a passes through each stand is _defined as P _Ei, and each rolling load obtained at the timing immediately after the joining point of the rolled material a passes through each stand is represented by P _Ei. P _Fi is set (step S3). This step S
The processing of 3 may be performed before the processing of steps S1 and S2.

The learning coefficient calculating section 19 includes an actual measured value acquiring section 16.
Based on the actual measured values P _Ei , P _Fi of the rolling loads acquired by the rolling load calculating unit 18 and the rolling loads P _Ci , P _Di calculated by the rolling load calculating unit 18, an instantaneous (temporary) ) learning coefficients Z _Pi 'learning coefficient Z _Qi' (6) (7) obtained by equation (step S4).

Z _Pi ′ = P _Ei / P _Ci (6) Z _Qi ′ = P _Fi / P _Di (7) The instantaneous learning coefficients Z _Pi ′ and Z _Qi ′ thus obtained
Since it is considered that there is a deviation from the true value (optimum value), it is preferable to obtain the true value by learning. Therefore, the learning unit 17 uses the learning coefficients Z _Pi ′ and Z _Qi ′ obtained by the learning coefficient calculation unit 19 to calculate, for example, (8) and (9)
Learning is performed by the equation, and the learning coefficients Z _Pi and Z _Qi are sequentially updated (step S5).

Z _Pi = Z _Pi-1 + α (Z _Pi '-Z _Pi-1 ) (8) Z _Qi = Z _Qi-1 + α (Z _Qi ' -Z _Qi-1 ) ... (9) In the equation (8), Z _Pi ′ on the right side is a learning coefficient calculated by the learning coefficient calculating unit 19, and Z _{Pi-1 on the} right side is a learning coefficient Z _Pi on the left side in the previous equation (8). . Similarly, (9)
Z _Qi ′ on the right side of the equation is the latest learning coefficient calculated by the learning coefficient calculation unit 19, and Z _{Qi-1 on the} right side is the learning coefficient Z _Qi on the left side of the previous equation (9). Also, (8)
Α in the expression (9) is a numerical value in the range of 0 to 1.

Describing the equation (8) specifically, when the first learning coefficient is obtained by the learning coefficient calculating section 19, the first learning coefficient is substituted into Z _Pi ′ on the right side. = 1 and there is no previous data to be substituted for Z _Qi-1 on the right side, so that the learning coefficient (Z _{Pi on the} left side) of the learning unit 20 is the first learning coefficient Z
_Pi '. On the other hand, when the learning coefficient Z _Pi ′ for the second time or later is obtained by the learning coefficient calculating unit 19, the obtained learning coefficient is substituted into Z _Pi ′ on the right side of the equation (8), and Z _Pi ′ on the right side is obtained.
_The learning coefficient (Z _{Pi on the} left side) previously learned is substituted for _Pi-1 . At this time, if the value of α is appropriately small, the learning coefficients Z _Pi and Z _Qi converge to optimal values. If it is desired to converge the learning coefficients Z _Pi and Z _Qi in a short time, α is set to a value close to 1 at the beginning of learning, and is set to approach 0 as learning progresses.

The learning coefficients Z _Pi and Z _Qi learned in advance by the learning section 20 in this way are _calculated each time the sheet thickness is changed at the joining point due to a change in the manufacturing specification during the running of the rolled material a. As shown in FIG. 3, the roll-down position change amount calculation unit 21 is used to obtain the change amounts of the roll-down positions of the first stand to the seventh stand 7. Therefore, a processing procedure in which the roll-down position change amount calculating section 21 calculates the change amount of the roll-down position will be described below with reference to FIG.

First, the rolling position change amount calculating section 21 calculates a predicted value P _Ai of each rolling load of the first to seventh stands 7 at the tail end of the preceding material and the leading end of the following material (the next material). By using the learning coefficients Z _Qi and Z _Pi learned in advance as described above, the predicted values P _Bi of the respective rolling loads of the first stand 7 to the seventh stand 7 in (10) are obtained.
It is obtained by equation (11) (step S11).

P _Ai = Z _Qi × b × k _mi × O _Pi × {R _i ′ (H _i −h _i )} ^1/2 (10) P _Bi = Z _Pi × b × k _mi × O _Pi × {R _i ′ (H _i −h _i )} ^1/2 (11) (10) In order to obtain the predicted values P _Ai and P _Bi of the rolling load by the formulas (11), it is necessary to use each stand. The target value (draft schedule) is determined in advance.

Next, the rolling position change amount calculating section 21 calculates
Predicted rolling load value P_Ai, P_BiUsing the preceding material
Prediction of the rolling position at the tail end (the position of the axis of the work roll)
Value S _AiAnd the predicted value S of the rolling position at the tip of the following material_BiWhen
Is obtained by the following equations (12) and (13) (step S
12).

S_Ai= H_Ai− ｛(P_Ai+ 2F_Ai) / K｝ + S_O... (12) S_Bi= H_Bi− ｛(P_Bi+ 2F_Bi) / K｝ + S_O... (13) K: Mill stiffness coefficient F_Ai, F_Bi: Roll bending force S_O: Roll gap zero correction term Further, the rolling position change amount calculating section 21 calculates the calculated rolling reduction.
Predicted position value S_Ai, S _BiUsing the first stand 1-
7 Roll-down position of stand 7 (position of work roll axis)
Change amount ΔS_iIs obtained by equation (14) (step
S13).

ΔS _i = S _Bi −S _AI (14) The first to seventh stands 7 determined in this way.
The change amount ΔS _i of the rolling position is supplied to the rolling position adjusting device 22. When the rolling material a reaches the stand at the joining point, the rolling position adjusting device 22 lowers the rolling position of each stand by the changed amount. Change the position.

Next, an example in which the off-gauge length is compared between this embodiment and the conventional method will be described with reference to FIG. This data shows that the preceding material is 2.3m thick
m, the board width is 1020 mm, the following material has a board thickness of 2.0 mm, and the board width is 1000 mm.
It is the average value of the case. According to this result, it is understood that the off-gauge length is reduced by half in the method of this embodiment as compared with the conventional method.

As described above, according to this embodiment, when there is a change in the sheet thickness at the joint of the rolled material, the data on the rolling before and after the joint passes through each stand is acquired. Learning is performed based on the acquired data to obtain a learning coefficient, and when there is a change in the sheet thickness at the joining point of the rolled material, the learning coefficient before and after the joining point is determined based on the obtained learning coefficient. The rolling load at each stand was predicted, and the amount of change in the rolling position of the roll of each stand was determined based on the predicted value. For this reason, the prediction accuracy of the rolling load is improved, and as a result, the dimensional accuracy before and after the joining point is improved, and the yield of the rolled material is improved.

Here, the data acquisition means according to the third aspect comprises an actual measurement value acquisition section 16, an actual value acquisition processing section 17,
And the rolling load calculating unit 18, the learning coefficient calculating unit corresponds to the learning coefficient calculating unit 19 and the learning unit 20, and the rolling load predicting unit corresponds to the rolling position change amount calculating unit 21.

In the above embodiment, the thickness control in the running change control is mainly described. However, the predicted value of the rolling load of the present invention can be applied to other width control, crown / shape control, and the like. Therefore, the effect of improving the accuracy of the thickness, crown and shape can be obtained.

Further, in the above-described embodiment, the change point a1 in the running thickness of the rolled material a has been described as the joining point of the rolled material a. No need to be.

[0040]

As described above, according to the first and third aspects of the present invention, data relating to rolling before and after passing between the stands at the point where the running specification at which the manufacturing specification of the rolled material is changed is acquired. Then, a learning coefficient is obtained in advance by performing learning based on the obtained data, and when there is a change in the manufacturing specification, before and after the running distance change point of the rolled material based on the obtained learning coefficient. The rolling load was predicted. For this reason, the prediction accuracy of the rolling load is improved,
As a result, the dimensional accuracy of the rolled material before and after the change in running distance is improved, and the yield of the rolled material can be improved.

Further, in the invention according to claims 2 and 4, when the point of change in running distance of the rolled material arrives at the stand, the rolling reduction of the stand is adjusted based on the predicted rolling load. did. For this reason, the prediction accuracy of the rolling load is improved, and as a result, the dimensional accuracy of the rolled material before and after the change in running distance is improved, and the yield of the rolled material is improved.

[Brief description of the drawings]

FIG. 1 is a block diagram of an apparatus according to an embodiment in a case where the present invention is applied to a change in a rolling position of a complete continuous rolling of a finishing hot rolling mill having a plurality of stands.

FIG. 2 is a flowchart illustrating a data processing procedure for obtaining a learning coefficient.

FIG. 3 is a flowchart illustrating a processing procedure of data for obtaining a change amount of a rolling-down position;

FIG. 4 is a diagram showing a comparative example of an off-gauge length between this embodiment and a conventional method.

[Explanation of symbols]

 a Rolled material 1-7 Stand 11,12 Thickness gauge 13 Load cell 14,15 Thermometer 16 Actual measurement value acquisition unit 17 Actual value acquisition processing unit 18 Rolling load calculation unit 19 Learning coefficient calculation unit 20 Learning unit 21 Change amount of rolling position Calculation unit 22 Roll-down position adjustment device

Claims (4)

[Claims] 1. A running change control method for a rolling mill having at least one stand, wherein data on rolling before and after a running change point at which a manufacturing specification of a rolled material is changed passes through the stand is acquired. A learning coefficient is obtained in advance by performing learning based on the acquired data, and when a manufacturing specification of a rolling mill is changed, before and after a change point between running distances of the rolled material based on the obtained learning coefficient. A method for controlling a change in running distance of a rolling mill, wherein a rolling load of the rolling mill is predicted. 2. The rolling machine according to claim 1, wherein when the change point between the running distances of the rolled material arrives at the stand, the rolling reduction of the stand is adjusted based on the predicted rolling load.
3. A method for controlling a change in running distance of a rolling mill according to claim 1. 3. A running distance change control device for a rolling mill having at least one stand and capable of adjusting a reduction in the height of said stand during hot rolling, wherein a manufacturing specification of a rolled material is changed. Data acquisition means for acquiring data relating to rolling before and after passing through the stand, learning coefficient calculation means for learning in advance based on the data acquired by the data acquisition means to obtain a learning coefficient, and manufacturing a rolling mill. When the specification is changed, rolling load prediction means for predicting a rolling load before and after a change point between running distances of the rolled material based on a learning coefficient obtained by the learning coefficient calculation means, A change control device for the running distance of a rolling mill. 4. A means for adjusting a rolling reduction of the stand based on a rolling load predicted by the rolling load predicting means when a change point between running distances of a rolled material according to a change in manufacturing specifications arrives at the stand. The running distance change control device for a rolling mill according to claim 3, wherein: