JPH0899103A - Method for controlling sheet crown and shape in sheet rolling - Google Patents

Abstract

PURPOSE: To exclude change by manual intervention of an operator and set a crown controller with a high precisin by determining the amount of change of a crown controller corresponding to the change of rolling conditions between a preceding rolled stock and the succeeding rolled stock and taking the set value of the crown controller of the succeeding rolled stock as a value that is actual value of the preceding rolled stock that the amount of change is added. CONSTITUTION: In a tandem mill which consists of seven rolling mills 1, a roll bending device 2 and roll shifting device 3 are provided as a crown controller on a 3rd to a 7th stand, sheet crown/shape control is executed and a sheet crown measuring device 4 is provided on the outlet side of the 7th stand. And, in the crown control, the bender load of the roll bending device 2 is controlled and the bender load FWSETN of the succeeding rolled stock is determined by the equation I. Where, FWAN<-1> : the actual bender load of the preceding rolled stock, FWAN: the bender load of the next rolled stock of a predicting model, FWN<-1> : the bender load of the preceding rolled stock of the predicting model.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a strip crown and a shape control method in strip rolling.

[0002]

2. Description of the Related Art In plate rolling, a plate crown and shape control (a plate crown is a deviation of a plate thickness between a central portion and an end portion in the plate width direction, for securing product tolerance and stabilizing a threading, The shape is expressed by the steepness η = δ / L using the wave height δ in the plate longitudinal direction and the pitch L of the wave.). As a device for controlling the plate crown / shape, there is a crown control device. As the crown control device, various types such as a roll shift device that axially shifts the rolling roll, a roll bending device that applies a roll bending force to both ends of the rolling roll, and a roll cross angle control device that changes the cross angle of the rolling roll. There is one.

In order to control the strip crown / shape of the rolled material by the crown controller, it is necessary to grasp the relationship between the set amount of the crown controller and the strip crown / shape of the rolled material. As this relationship, for example, various models in which the deformation of the rolled material and the deformation of the rolling roll are formulated have been proposed, as described in JP-A-59-130614.

Further, in order to improve the accuracy of predicting the plate crown / shape by the model, the thermal crown generated by the temperature distribution in the axial direction of the roll and the wear amount distribution caused by the slip between the rolled material and the roll are superposed. As a result, the roll profile is also predicted.
Hereinafter, a model for performing the prediction of rolled material deformation, roll deformation, and roll profile together will be referred to as a plate crown / shape prediction model, which is abbreviated as “prediction model”.

Conventionally, based on this prediction model, rolled material ratio crown C (ratio crown C = (plate width central part plate thickness H
c − within the rolled material shape limit, which is a flat region between the stretched side where the strip edge thickness He) / the strip width central portion thickness Hc) becomes smaller and the selvage side where the rolled material ratio crown increases, Calculate the set amount of the crown control device such as shift position, bender load, roll cross angle, etc. so that the plate crown becomes the target,
This calculated value was directly used as the set value of the crown control device.

[0006] However, since the prediction model has a plate crown and rolled material shape and a roll profile prediction error,
Final stand exit side plate crown (hereinafter referred to as "product crown"
Sometimes called. ) Is used to statistically correct the parameters of the prediction model using the difference between the actual value and the predicted value.
This error has been corrected by using so-called roll profile learning or manual intervention learning as described in JP-A-59-215205.

In the roll profile learning described in the above publication, the difference between the actual value of the plate crown or rolled material shape and the calculated value calculated by the prediction model using the actual rolling conditions is caused by the roll profile estimation error. As a result, the calculation of the next rolled material is corrected. Further, with the manual intervention learning, a roll bending device, a crown control device such as a roll cloth device that can be set and changed during rolling, the operator manually intervenes by observing the shape of the rolled material,
The difference between the optimized actual value and the set value calculated by the prediction model using the actual rolling conditions is considered to be due to the roll profile estimation error, and the calculation of the next rolled material is corrected.

That is, the roll profile learning and the manual intervention learning are performed to compensate for the roll profile prediction error.

[0009]

In order to perform the roll profile learning, it is necessary to provide a plate crown or a shape measuring device in order to measure the actual value of the plate crown or the shape. In a tandem rolling mill in which a plurality of rolling mills are arranged in series, it is very expensive to measure the plate crown or rolled material shape by providing a plate crown or shape measuring device on the exit side of all the stands, which is very expensive. Not.

Therefore, in practice, the plate crown measuring device is often provided only on the exit side of the final stand, which is directly connected to the product accuracy. In this case, roll profile learning is possible for the plate crown in the final stand, but roll profile learning cannot be performed for the intermediate stands because the actual values of the plate crown and shape cannot be measured. Therefore, it is highly possible that the predicted value of the plate crown and shape at the intermediate stand is different from the actual value.

Further, when the sheet thickness is thin, it can be considered that most of the ratio crown change ΔC (change of the ratio crown C on the inlet side and the outlet side) at the time of rolling is changed to the elongation strain difference.
The shape between the stands (steepness η) is η = (2 / π) · √
It is expressed as (ΔC). Therefore, even when the ratio crown change ΔC is as small as 0.1%, the steepness η becomes 2%, and a considerably large shape defect occurs. That is,
It can be seen that in the rolled material shape prediction, the required prediction accuracy is much more severe than that of the sheet crown prediction.

Therefore, in the single-stand rolling mill, plate crown measuring devices are provided on the inlet side and the outlet side of the rolling mill, or
In the tandem rolling mill, by providing the plate crown measuring device on the exit side of all the stands, roll profile learning can be performed in all passes, and even if the plate crown can be predicted accurately, it is difficult to accurately predict the shape.

When the rolled material shape is measured and the roll profile learning is performed, it is necessary to convert between the rolled material shape and the ratio crown change ΔC twice by using a predictive model as follows. That is, a roll profile is calculated after converting the rolled material shape prediction error into a ratio crown change using a prediction model, and further, a ratio crown change is calculated using the predicted rolling condition and roll profile of the next rolled material, The change in the ratio crown is converted into the shape of the rolled material.

Therefore, as long as there is a shape prediction error in the prediction model, even if the roll profile learning is performed, the shape prediction error does not become sufficiently small. Therefore, with respect to the shape, the effect of the roll profile learning is small and it is not actually performed. It was normal. The same applies to the manual intervention learning. It is necessary to convert the manual intervention amount of the crown controller into a roll profile and convert the ratio crown change ΔC calculated using this roll profile into a rolled material shape using a prediction model. Therefore, it was difficult to improve the prediction accuracy of the rolled material shape.

As described in detail above, by providing a plate crown measuring device on the exit side of the final stand, the parameters of the prediction model are corrected or the parameters of the prediction model are corrected using a method such as roll profile learning. It is relatively easy to match the predicted value of the product crown with the actual measured value by using the modified prediction model, but it is difficult to accurately predict the shape of each stand.

Therefore, not only for the final stand, but for rolled materials such as ultra-thin sheets where the shape of rolled material at each stand is important, even if a prediction model whose accuracy is improved by roll profile learning is used, It is difficult to use the calculated setting value of the crown control device "as it is", and it is the current situation that the operator observes the shape of the rolled material and manually changes the crown control device.

As described above, when the operator changes the set value based on his experience and intuition, the plate crown deviates from the target value, etc.
There is a problem that the product accuracy may deteriorate.
Therefore, it is an object of the present invention to provide a strip crown and shape control method in strip rolling, which can set the crown control device with high accuracy by eliminating the manual intervention change of the operator.

[0018]

In order to achieve the above object, the present invention takes the following means. That is, a feature of the present invention is that in order to obtain a target plate crown / shape by using the crown control device, in a plate crown / shape control method for controlling the crown control device, a pre-rolled material and a next rolling material are used. The change amount of the crown control device corresponding to the change of rolling conditions with the strip is calculated using the plate crown / shape prediction model, and the set value of the crown control device for the next rolled strip is added to the previous rolled strip actual value by adding the change amount. There is a point to be a value.

The crown control device may be a roll bending device that applies a roll bending force to both ends of a pair of rolling rolls. In that case, the control of the crown control device means the bending load of the roll bending device. Bender load FWSET ^N in the next rolled material

[0020]

[Formula 3] FWSET ^N = FWA ^N-1 + FW ^N -FW ^N-1 where FWA ^N-1 : Actual bender load in the pre-rolled material FW ^N : Bender load for the next rolling material predicted rolling condition by the prediction model FW ^{N- 1} : Previous rolling material performance based on a predictive model It can be obtained from the vendor load for the rolling conditions.

Bender load FW of the next rolled material
SET ^N is

[0022]

[Formula 4] FWSET ^N = FWA ^N-1 + ΔFW However, FWA ^N-1 : Actual bender load in the pre-rolled material ΔFW: Bender correction amount.

[0023]

As the crown control device in the present invention, a device such as a roll bending device, a shift device, a cross angle control device or the like can be used, but in the explanation of this operation, the roll bending device according to claim 2 will be described for convenience. Those using will be described. Further, in the present invention, any prediction model may be used, but for simplicity of description, the following formula is used.

[0024]

[Equation 5] C = B ₁ · Cin + B ₂ · FW + α (1) λ = ξ (C−Cin) where C: output side ratio crown Cin: input side ratio crown λ: output side rolled material strain difference ξ : Shape change coefficient FW: Bender load B ₁ : Plate crown influence coefficient B ₂ : Bender load influence coefficient α: Other influence terms First, the setting of the crown control device corresponding to the change of rolling conditions between the pre-rolled material and the next rolled material. The amount of change is calculated using a prediction model. Specifically, using the prediction model, FW ^N and F
W ^N-1, and the difference (FW ^N -FW ^N-1 ) is calculated as the setting change amount.

That is, the actual rolling conditions of the pre-rolled material (N-1th) are input to the prediction model, and each stand bender load F required for the actual strip crown C ^{N-1 of the} pre-rolled material is input.
Find W ^{N- 1} . Similarly, the predicted rolling condition of the next rolled material (Nth) is input to the prediction model to obtain each stand bender load FW ^N required for the next rolled material.

Next, the set value of the crown control device for the next rolled material is set to a value obtained by adding the setting change amount to the previous rolled material actual value. Specifically, obtains the set value FWSET ^N vendor load in following rolled material, as ^{FWSET N = FWA N-1 +} FW N -FW N-1 ...... (2).

That is, by not using FW ^N obtained by predicting the bender load as it is, but using FW SET ^N which is a modified version of FW ^N , good strip crown / shape control of rolled material can be performed. Intervention changes can be eliminated. In the present invention, at the time of collecting the actual rolling data of the pre-rolled material, it is considered that the shape of the rolled material was good by optimizing the crown control device, and there is a prediction error in the prediction model. However, it is assumed that it is possible to predict to some extent relative changes in rolling conditions, and it is difficult to accurately predict the inter-stand shape, but it is possible to predict the final stand exit side plate crown almost accurately. Has been done.

Note that the bender correction amount Δ for each stand is equal to the difference between the target plate crown and the predicted plate crown.
By using FW, instead of the above equation (2), the following equation
(3) can be used. FWSET ^N = FWA ^N-1 + ΔFW (3) In the formula (2), the change of the rolling condition is
Crown control other than roll bending device such as roll shift position change amount and roll cross angle change amount in addition to changes in rolled material dimensions, steel type, target plate crown, etc., and roll crown changes such as thermal crown and wear amount. Although it is assumed that the setting change amount of the device is also included, the above formula (3)
In the simple formula, the setting change amount of the crown control device other than the roll bending device cannot be included.

[0029]

The following is a general example applicable to the tandem rolling mill in which the type of crown control device is different depending on the stand, and one example of the crown control device is the same for all stands. Two simple examples applicable to the case will be described.

As shown in FIG. 1, in a tandem rolling mill consisting of seven rolling mills 1, a roll bending device 2 and a roll shift device 3 are provided as a crown control device.
7 to 7 stands for plate crown / shape control, and the final stand (7th stand) is provided with a plate crown measuring device 4 on its exit side. Although any plate crown / shape prediction model may be used, the following prediction model formula is used.

[0031]

[Equation 6] C _i = B _1i · Cin _i + B _2i · FW _i + α (1) ′ λ _i = ξ _i (C _i −C in _i ) where C _{i is} the i-th stand output side ratio crown C in _i : I-th stand entry side ratio crown λ _i : i-th stand exit side rolled material elongation strain difference ξ _i : i-th stand shape change coefficient FW _i : i-th stand bender load B _1i : i-th stand plate crown influence coefficient B _2i : I-th stand bender load influence coefficient α: Other influence terms The influence coefficient is a coefficient indicating how much the input side ratio crown or the output side ratio crown changes with respect to the bender load. -It is obtained by calculation such as FEM.

Assuming that the shape of the rolled material entry side is flat, by adding the front stand exit side elongation strain difference λ _i-1 to the front stand exit side ratio crown C _i-1 , The ratio crown Cin _i can be expressed as Cin _i = C _i-1 −λ _i-1 .

First, as a general example, a general example will be described in which an appropriate set value can be calculated even if there are plural types of crown control devices by setting the crown control device for each stand. In this general example, first, the actual rolling conditions of the pre-rolled material (N-1th) are input to the prediction model, and each stand bender load required for the actual strip crown C _i ^{N-1 of} the pre-rolled material is input. Calculate FW _i ^N-1 . At this time, in the prediction model, there are innumerable solutions for each stand bender load necessary for the actual plate crown of the pre-rolled material, but each stand outlet side elongation strain difference is constant, or each stand ratio crown change is constant. If some assumption is made such as, it is limited to one type.

Similarly, the predicted rolling conditions of the next rolled material (Nth) are input to the predictive model, and each stand bender load FW _i ^N required for the next rolled material is obtained. In this case as well, there are innumerable combinations of required stand bender loads for the target plate crown of the next rolled material, but some assumptions such as a constant difference in elongation strain at each stand or a constant change in each stand ratio crown are used. If placed, it is limited to one type.

Next, the difference between the calculated FW _i ^N and FW _i ^N-1 is calculated to obtain the rolled material actual vendor load FWA _i.
Set the deviation to ^N-1 and set it. Bender load FWSET _i
^N is calculated as FWSET _i ^N = FWA _i ^N-1 + FW _i ^N −FW _i ^N-1 (2) ′. The actual value FWA _i ^N-1 is stored in the computer of the controller as the actual rolled material actual data.

Next, the obtained bender load FWSET
_It is verified whether _i ^N is appropriate as the set load. That is, the predicted plate crown CPP of the next rolled material is determined by the prediction model using the bender load FWSET _i ^N. If this predicted plate crown CPP is within the allowable range of the target plate crown, this FWSET _i ^N is used as it is as the set value.

When the predicted plate crown CPP is out of the target plate crown allowable range, the bender load FWSE is set.
Modify T _i ^N. That is, the difference Δ between the predicted plate crown and the upper limit value CU or the lower limit value CL of the target plate crown allowable range
A uniform bender load ΔFW for all stands equivalent to C ₇
Calculated by the prediction model and corrected evenly for all stands. Note that the method of determining ΔFW conforms to the simple example described below, and thus the description thereof is omitted here.

When CL ≦ CPP ≦ CU, FWSET _i ^N is used as the set bender load in the next rolled material, and when CPP <CL or CPP> CL, the set bender load in the next rolled material is FWSET _i ^N + ΔFW. To use. Next, a simple example will be described. After setting the other crown control device (for example, work roll shift device) that is difficult to change during rolling, only the setting calculation for one type of crown control device such as roll bending device that can be changed during rolling When performing, the following simple method is effective.

First, the predicted rolling condition of the next rolled material (Nth) and the actual bender load FW of the previously rolled material (N-1th).
Using A _i ^N-1, determined as a provisional prediction strip crown CP ^N from said predictive model follows.

[0040]

[Equation 7] C ₀ = (Rough rolling mill output side ratio crown) λ ₀ = 0 C ₁ = B ₁₁ · C ₀ + B ₂₁ · FWA ₁ ^N-1 + α λ ₁ = ξ ₁ (C ₁ −C ₀ ) ·・ C _i = B _1i・ Cin _i + B _2i・ FWA _i ^N-1 + α λ _i = ξ _i (C _i −Cin _i ) ・ ・ C ₇ = B ₁₇・ Cin ₆ + B ₂₇・ FWA ₇ ^N-1 + α CP here ^{_{N = C 7 · H 7,}} H 7 is a thickness at delivery side of the seventh stand.

Next, a bender correction amount ΔFW that is uniform for each stand and corresponds to the difference between the target plate crown CMK ^N and the provisional predicted plate crown CP ^N is determined as follows. However,
When there is a stand that wants to suppress the fluctuation of the bender load, the coefficient β (0 <β <1) is multiplied in the calculation of the stand. Here, only the final seventh stand, which is directly connected to the product shape, is calculated by multiplying it by the coefficient β that suppresses the vendor fluctuation amount. When the amount of change in the crowns on the outlet side and the inlet side of each stand and the difference in elongation strain due to the bender correction amount ΔFW is represented by Δ,
The following formula holds.

[0042]

[Expression 8] ΔC ₃ = B ₂₃ · ΔFW ΔC ₄ = B ₁₄ · ΔCin ₄ + B ₂₄ · ΔFW, ΔCin ₄ = ΔC ₃ −Δλ ₃ ΔC ₅ = B ₁₅ · ΔCin ₅ + B ₂₅ · ΔFW, ΔCin ₅ = ΔC ₄ -Δλ ₄ ΔC ₆ = B ₁₆ · ΔCin ₆ + B ₂₆ · ΔFW, ΔCin ₆ = ΔC ₅ −Δλ ₅ ΔC ₇ = B ₁₇ · ΔCin ₇ + B ₂₇ βΔFW, ΔCin ₇ = ΔC ₆ −Δλ ₆ ΔC ₇ = CMK ^N − using the formula CP ^N above formula and _{λ i = ξ i (C i} -Cin i), ΔCin i, erasing the Δλ _i, ΔFW is represented as follows.

[0043]

[Formula 9] ΔFW = (CMK ^N −C P ^N ) / B ₁₇ / (B
₂₃ Y ₆ Y ₅ X ₄ X ₃ + B ₂₄ Y ₆ Y ₅ X ₄ + B ₂₅ Y ₆ Y ₅ +
B ₂₆ Y ₆ + B ₂₇ β / B ₁₇ ) Here, X _i = 1−ξ _i and Y _i = B _1i X _i + ξ _i .

Therefore, the set bender load FWSET _i ^N for the next rolled material is FWSET _i ^N = FWA _i ^N-1 + ΔFW. However, for the 7th stand, FWSET ₇ ^N = FWA ₇ ^N-1 + β · ΔFW.

2 and 3 show the conventional method and the case where 19 rolled materials with a plate thickness of less than 2 mm were continuously rolled, which was difficult to use the bender load set value as it was in the conventional method. The results of setting the bender using the plate crown and the shape control method of the present invention shown in the simplified example are shown. As shown in FIG. 3, in the conventional method, the bender load set value obtained by the prediction model is significantly higher than the actual value (the operator manually changes the set value), but it is set by using the method of the present invention. The value is almost the same as the actual value. Therefore, no manual intervention change is required. Then, as shown in FIG. 2, since the actual plate crown substantially matches the target plate crown, it can be understood that the product crown substantially matches the target crown when the bender load is set using this method.

The present invention is not limited to the above-mentioned embodiment, and the same setting can be made for a crown control device such as a roll cross angle adjusting device other than the roll bending device. Further, although the crown control device of the tandem rolling mill has been described, the same setting can be performed in a single-stand rolling mill, a reverse rolling mill, and the like.

[0047]

According to the present invention, even if there is a certain degree of prediction error in the prediction model itself, the crown control device set based on this model can be set in a good range, so that manual intervention changes can be made. It can be eliminated and the product accuracy can be improved.

[Brief description of drawings]

FIG. 1 is a configuration diagram of a tandem rolling mill for carrying out the method of the present invention.

FIG. 2 is a graph showing a transition between a target crown and an actual crown.

FIG. 3 is a graph showing an example of each stand bender load of the conventional example and the present invention.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification number Office reference number FI Technical indication location 8315-4E B21B 37/00 116 B 8315-4E 117 B (72) Inventor Hiroyuki Hasegawa Kakogawa City, Hyogo Prefecture Kanazawa 1st Kamigawa Steel Works Kakogawa Works

Claims (3)

[Claims] 1. A plate crown / shape control method for controlling a crown control device to obtain a target plate crown / shape by using the crown control device, wherein a rolling condition change between a pre-rolled material and a next rolled material is performed. The amount of change in the crown control device corresponding to is calculated using the plate crown / shape prediction model, and the set value of the crown control device in the next rolled material is set to the value obtained by adding the above-mentioned amount of change to the previous rolled material actual value. A plate crown and a shape control method in a characteristic plate rolling. 2. The crown control device is a roll bending device that applies a roll bending force to both ends of a pair of rolling rolls. The control of the crown control device is to control a bender load of the roll bending device. The bender load FWSET ^N in the rolled material is calculated by the following formula: FWSET ^N = FWA ^N-1 + FW ^N −FW ^N-1 where FWA ^N-1 : Actual bender load in the pre-rolled material FW ^N : Next by the prediction model The shape control method in plate rolling according to claim 1, wherein the bender load for predicted rolling conditions for rolled material FW ^N-1 : Bender load for actual rolling conditions for pre-rolled material by a prediction model. 3. The crown control device is a roll bending device that applies a roll bending force to both ends of a pair of rolling rolls, and the control of the crown control device is to control a bender load of the roll bending device. Claims characterized in that the bender load FWSET ^N in the rolled material is obtained by the following formula: FWSET ^N = FWA ^N-1 + ΔFW, where FWA ^N-1 is the actual bender load in the previous rolled material ΔFW: Bender correction amount Item 2. A shape control method in strip rolling according to Item 1.