JP2651741B2 - Edge drop control method in strip rolling - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to edge drop control in sheet rolling.

[Conventional technology]

In cold rolling, continuous rolling is performed by installing a multiple-pass multiple rolling mill on a rolling line and performing edge drop control for high-speed rolling of a rolled material and improvement of edge drop. In such equipment, a target edge drop of the material at the final position of the rolling pass is determined as the target value of the edge drop.

The edge drop control performs preset control to output set values before material rolling.The edge drop in the final pass becomes the target edge drop, and the rolling is performed as much as possible on the upstream side of the rolling line of a multi-pass multi-rolling mill and as little as possible. Rolling by adjusting the work roll direction shift (hereinafter referred to as “work roll shift”) of the work roll, whose roll barrel end is tapered (hereinafter referred to as “taper”), is optimal for both quality assurance and production cost reduction. is there.

In order to satisfy such a condition, it is required to obtain and control the work roll shift amount of each pass from the line upstream. In the edge drop control, the edge drop at the final pass exit position over the entire length of the material matches the target edge drop, taking into account errors in the preset control, fluctuations in the edge drop of the sheet before rolling, and fluctuations in the rolling conditions. Therefore, it is necessary to adjust a work roll shift amount in a multi-pass multiple rolling mill during material rolling.

Conventionally, as for this control, as disclosed in Japanese Patent Application Laid-Open No. 60-12213, feedback control for actually measuring the edge drop at the exit side of the final pass and adjusting the work roll shift amount of the pass on the upstream side is known. Proposed.
In Japanese Patent No. 901438, the edge drop in the longitudinal direction of the plate before rolling is measured in advance, and feedforward control is performed based on the measured value in the rolling line, and the edge drop of the plate at the exit side of the final pass is measured. Although learning is to be performed using values, specific details of the learning method are not described.

[Problems to be solved by the invention]

In control during material rolling with edge drop control,
In actual rolling, the time delay of control due to the material transfer time between the actual measured value of the edge drop in the final pass and the work roll shift position of each pass due to the conditions of the plate before rolling is inevitable, and the edge drop is set to the target value. It takes a certain amount of time to match, and control overshoot and hunting are likely to occur.

The present invention is not the feedback control of the deviation from the edge drop target value during the same lot material rolling,
The feed edge control based on the edge drop of the sheet before rolling makes the material edge drop in the final pass coincide with the target edge drop, and the control model (edge drop deviation model) coefficient value is set to the edge drop of the sheet before rolling and the value of each pass. By successively correcting in a feedback manner based on the actual measured values of the work roll shift amount and the edge drop on the exit side of the final pass,
Provided is an edge drop control method in sheet rolling that realizes control having both feedforward and feedback double-sided functions.

[Means for Solving the Problems] The present invention relates to an edge profile of a plate by a multi-pass multi-rolling mill having a mechanism for shifting a work roll whose tapered end is tapered in a body axis direction when rolling a steel material. In the control, when the target edge profile of the sheet at the exit side of the final pass is determined, the edge profile is represented by at least two or more parameters, the deviation from the target value is set as a parameter, and this is set as a dependent variable. A mathematical model is prepared in advance with the edge profile parameters and the shift position of the shift mechanism of the multi-pass multiple rolling mill as independent variables, and during the material rolling, the edge profile of the sheet before rolling is measured, and the edge profile is determined based on the measured values. The edge profile on the exit side of the final pass matches the target edge profile using the drop deviation model The feed roll control is performed by adjusting the shift in the width direction of the work roll in the upstream pass from time to time, and the edge drop of the sheet before rolling and the edge on the exit side of the final pass when the material passes through the final pass. A rolled steel sheet, wherein the model coefficients are successively corrected based on the measured values of the drop and the shift position of the work roll in the width direction in each pass, and the calculation error of the model is converged to substantially zero in a feedback manner. This is an edge drop control method in rolling.

(Operation)

One embodiment of a multi-pass multiple rolling mill to which the present invention is directed,
FIG. 1 shows an outline of a system configuration of the edge drop control.

Reference numeral 7 in FIG. 1 indicates a multi-pass multiple rolling mill.
Is a steel sheet rolled by the rolling mill. The starting point of the line is S
The point and the end point are O point, 5 is an edge drop detector for measuring the thickness distribution in the strip width direction on the entry side of the rolling mill, and 8 is an edge drop detector for measuring the thickness distribution in the strip width direction on the exit side of the rolling mill. An edge drop detector 6 is a strip width detector on the rolling mill entry side. O
The steel plate 1 at the point is given a target value for edge drop.

There are n-pass (five-pass in the embodiment) multi-rolling mills on the line, and each rolling mill has a tapered end of the roll body, the details of which are shown in FIG. Roll 2 and backup roll 3 are provided. The edge drop is improved by the shift operation of the work roll 2, and the operation amount is controlled by the control computer 9 and the controller 10.

When the edge drop of the plate on the entry side of the rolling mill is large, the work roll shift position of each rolling mill is calculated by the control computer so that the contact length of the tapered portion of the work roll becomes larger in the upstream rolling mill, and the controller 10 is used. Through each rolling mill. As a result, the edge drop of the rolled sheet is reduced due to the reduced reduction of the edge of the material.

〔Example〕

In the feed-for-work and feedback control within the same lot, by using the measured value of the edge drop of the sheet before rolling at the point S shown in FIG. Then, the edge drop at the point O is controlled to the target edge drop, and the model coefficient is successively corrected in a feedback manner based on the error between the measured value of the edge drop at the point O and the value calculated by the edge drop deviation model.

In the in-lot control, a factor that greatly affects the edge drop on the exit side of the final pass is the edge drop of the sheet before rolling, and is expressed as the following equation as an edge drop deviation model.

Wherein _{ΔD 1 (t) = D 1} (t) - 1 ΔD 2 (t) = D 2 (t) - 2 ΔD 3 (t) = D 3 (t) - 3 ΔΓ 1 (t) = Γ 1 ( ^{_{t) -Γ 1 * ΔΓ 2 (}} t) = Γ 2 (t) -Γ 2 * ΔΓ 3 (t) = Γ 3 (t) -Γ 3 * ΔSw 1 (t) = Sw 1 (t) -Sw 1 ^* ΔSw ₂ (t) = Sw ₂ (t) −Sw ₂ ^* ΔSw ₃ (t) = Sw ₃ (t) −Sw ₃ ^* However, D ₁ , D ₂ and D ₃ indicate the last pass exit side at time t. Is the edge drop amount that has passed, D ₁ in the case of edge drop control by the first pass, D ₁ and D ₂ in the case of the control up to the second pass, and D _{1 in} the case of control up to the third pass. , using D ₂ and D _3.

The suffixes 1, 2 and 3 indicate control portions in the plate width direction.
The 15mm point is set to 1, and in the case of control up to the second pass, the point 30mm from the plate end is set to 1, and the point 15mm from the plate end is set to 2, in the case of control up to the third pass, 30mm from the plate end. 1 point, 20 from plate edge
The mm point is 2 and the point 15 mm from the edge of the plate is 3.

These control parts vary depending on the material, rolling conditions, and the like. Γ ₁ , Γ ₂ and _{３ 3} are edge drops of the sheet at the time t at the entrance to the rolling mill, and the subscripts 1, 2 and 3 are
This is the same as that described above for D ₁ , D ₂ and D ₃ . That is, the edge drop control portion in the strip width direction on the exit side of the rolling mill must be a measurement point at the same portion in the strip width direction on the entry side of the rolling mill.

_{1, 2} and ₃ is the target value of edge drop of D _1, D ₂ and D _{3. １ 1} ^* , Γ ₂ ^* and _{３ 3} ^* are
This is the edge drop of the sheet before rolling used in the edge drop prediction model calculation before the material rolling in the preset control, and the suffixes 1, 2 and 3 are the same as those described above. Sw ₁ , Sw ₂ and Sw ₃ represent the tapered work roll shift position in each pass from the rolling mill upstream side,
Subscripts 1, 2 and 3 are pass numbers. Sw ₁ ^* , Sw ₂ ^* and
Sw ₃ ^* is the tapered work roll shift position in each pass determined by the preset calculation. a ₁ , a ₂ , a ₃ , b ₁ , b ₂ , b ₃ ,
b ₄ , C ₁ , C ₂ , C ₃ , C ₄ and C ₅ are model coefficients.

In the present embodiment, an edge drop deviation model is constituted by three parameters D ₁ , D ₂ and D ₃ . The deviation model, because the edge drop of the sheet before rolling is not necessarily the same amount of edge drops at both ends in the sheet width direction,
The final pass exit side edge drop is determined for each of the drive side and the operation side of the rolling mill, and the setting of the work roll shift position is also set independently. Therefore, edge drop detectors 8 for measuring the thickness distribution in the width direction at the exit side of the final pass shown in FIG. 1 are installed at both ends of the plate.

As the edge drop control method is performed in the same manner on both side ends, one of the control methods will be described below. The formula (1) edge drop deviation model [Delta] D ₁ at (t), [Delta] D ₂ (t) and [Delta] D ₃ [Delta] D ₂ in (t)
(T) and [Delta] D ₃ (t) is the same as the method of controlling the [Delta] D ₁ (t) described below, is described here successive correction method of model parameters [Delta] D ₁ (t).

The model coefficients a ₁ , a ₂ and a ₃ in the edge drop deviation model ΔD ₁ (t) are successively modified during rolling, and are hereinafter referred to as a ₁ , a ₂ and a ₃ . In the in-lot control, it is necessary to determine the values of _{１ 1} ^* , Sw ₁ ^* and the initial values of a ₁ , a ₂ , a ₃ before rolling the material of the lot, and are determined as follows.

Sw ₁ ^* is the work roll shift position set value obtained by the preset control calculation, and _{１ 1} ^* is the edge drop of the plate on the rolling mill entry side used for the preset control calculation. The initial value of a ₁ is 0, and the initial values of a ₂ and a ₃ are Sw ₁ and Γ ₁ respectively.
The impact coefficient on the change of the D _1, calculated by the following equation. _{a 2 = ∂D 1 / ∂Γ 1} (Γ 1 = Γ 1 *, Sw 1 = Sw 1 *) a 3 = ∂D 1 / ∂Γ 1 (Γ 1 = Γ 1 *, Sw 1 = Sw 1 *) the value of the impact factor, minimal change in the gamma ₁ and Sw ₁ delta
The variation ΔD ₁ of D ₁ due to Γ ₁ and ΔSw ₁ is obtained by calculating using an edge drop prediction model.

During the same lot rolling, the edge drop of the plate at the entry side of the rolling mill is actually measured in synchronization with a predetermined sampling, and when the edge drop at the actually measured point is Γ ₁ (t), the edge drop of the plate at the exit side of the final pass is obtained. Work roll shift Sw such that D ₁ (t) matches target edge drop _1
1 (t) is obtained by the following equation using the edge drop deviation model of equation (1).

Sw ₁ (t) = Sw ₁ ^* − {a ₁ + a ₂ • (Γ ₁ (t) −Γ ₁ ^* )} / a
₃ (2) The value of Sw ₁ (t) takes into account the travel time between the edge drop detector 5 on the entry side of the rolling mill and the upstream pass (first pass in the above example) and the work roll shift operation delay. When the material passes through the work roll shift operation target path,
By setting the work roll shift operation target path at an appropriate timing so that the work roll shift in the work roll shift operation target path is Sw ₁ (t), the feed forward of the edge drop of the plate at the rolling mill exit side is performed. Perform control.

On the other hand, the model coefficient estimation values a ₁ , a ₂ , and a ₃ are sequentially estimated by the following equations. a (t) = a (t -1) + p (t) · x (t) · ε (t) ...... (3) where a (t) = _{[a 1 (t), a 2} (t), a ₃ (t)] ^T x (t) = [1, ΔΓ ₁ (t), ΔSw ₁ (t)] ^T ε (t) = ΔD ₁ (t) −a (t−1) ^T × x (t Where t is time (unit of sampling cycle), ΔΓ
₁ (t), ΔSw ₁ (t), and ΔD ₁ (t) are actual values relating to the edge drop of the material that has passed the rolling mill at the time t. P (t) ∈R ^{3 × 3} is a corrected gain matrix, and its value is obtained by using a known weighted sequential least squares method.

This successive correction corrects the error of the edge drop deviation model,
This is performed to adjust the edge drop at point O in FIG. 1 to the target edge drop in a feedback manner in response to changes in rolling conditions and various disturbances during the same lot rolling, and from the start of rolling to the end of rolling. Through each control operation to control the adaptive edge drop influence coefficient for factors that are difficult to measure during rolling (for example, heat crown of rolls) among rolling conditions. It can be used for a quantity calculation model and always enables accurate edge drop control.

Further, in a normal rolling state by using this method, in order to perform self-learning of P (t) during rolling,
It is not necessary to create a mathematical model by extensive data analysis.

A correction operation with a large influence coefficient works at the start of learning due to the variable gain. However, the correction operation can be reduced as the learning progresses, and the learning can be speeded up and stability can be promoted.

FIG. 3 shows a procedure for changing the work roll shift at time t during rolling and correcting the parameters of the edge drop deviation model according to the embodiment.

At time t, the measured value of the edge drop of the sheet before rolling and the measured value of the work roll shift of the first pass are taken in,
By inputting the data into the edge drop deviation model formula, an edge drop (ΔD ₁ ) at the exit side of the final pass is obtained, and ΔD ₁ is 0.
If it is, the process ends without performing the setting output. ΔD ₁ is 0
If not, the work roll shift position is obtained from equation (2), and the change amount is output. Next, the measured value of the edge drop at the output side of the final pass is taken in, and the parameters of the edge drop deviation model are updated.

FIG. 4 illustrates a control result during material rolling.
By manipulating the work roll shift position Sw in response to a change in edge drop Γ of pre-rolling of the plate in FIG. 1 S point, the final pass exit side edge drop D ₁ is sufficiently satisfy the target edge drop words in dashed line The results have been obtained.

In the above description, the case where the target edge drop is satisfied on the exit side of the final pass by the control of the first pass in the upstream pass has been described, but the material of the steel material or the edge drop of the plate on the entry side of the rolling mill is described. When the target edge drop is not satisfied depending on the size, the work roll shift of the second pass or the third pass in the upstream pass is operated using ΔD ₂ or ΔD ₃ in the equation (1).
In the present embodiment, when performing edge drop control by a plurality of passes, the contact length between the tapered portion of the work roll and the plate is increased toward the upstream pass side.

〔The invention's effect〕

As described in detail above, according to the present invention, in plate rolling, not only the conventional feedback control in the control during material rolling, but also the work in each pass using the measured edge drop value of the plate before rolling. By adjusting the roll shift position, the edge drop in the final pass is made to match the target value in a feedforward manner. Regarding the coefficients of the edge drop deviation model used for this calculation,
The initial value is obtained by an edge drop prediction model before material rolling, and during material rolling, the model coefficient is calculated using the edge drop on the exit side of the final pass, the measured value of the work roll shift amount of each pass, and the edge drop of the sheet before rolling. Are successively corrected to make the calculation error substantially zero in a feedback manner.

By the above processing, the control during material rolling is provided with both the feed forward function and the feedback function, and it is necessary to always hold the target edge drop on the final pass exit side edge drop over the entire length of the plate in response to various disturbances. Becomes possible.

[Brief description of the drawings]

FIG. 1 is a diagram showing an outline of a rolling line configuration and a system configuration of a multi-pass multiple rolling mill, FIG. 2 is a configuration diagram of a work roll in a multiple rolling mill, and FIG. FIG. 4 is a flowchart showing a procedure for correcting the parameters of the deviation model, and FIG. 4 is a drawing showing control characteristics during material rolling. 1 ... rolled steel plate, 2 ... taper work roll, 3 ...
Backup roll, 5 ... Entry side plate edge drop detector, 6 ... Plate width detector, 7 ... Multi-pass multiple rolling mill, 8
… Outboard edge drop detector, 9… Control computer, 10… Controller

Claims (1)

(57) [Claims] When rolling a steel material, in the edge profile control of a plate by a multi-pass multi-rolling mill having a mechanism for shifting a work roll having a tapered end to a body axis in a roll direction, a roll at a final pass exit side is provided. When a target edge profile of the plate is determined, the edge profile is represented by at least two or more parameters, and a deviation from the target value is used as a parameter, which is used as a dependent variable. A mathematical model with the shift position of the shift mechanism of the machine as an independent variable is created in advance, the edge profile before rolling is actually measured during material rolling, and the final pass exit side is determined using an edge drop deviation model based on the measured values. So that the edge profile matches the target edge profile. Feed forward control is performed by adjusting the widthwise shift of the roll every moment, and when the material passes through the final pass, the edge drop of the plate before rolling and the edge drop on the exit side of the final pass, and the work roll of each pass, An edge drop control method in sheet rolling of a rolled steel sheet, wherein a model coefficient is sequentially corrected based on an actually measured value of a shift position in a width direction, and a calculation error of the model is converged to substantially zero in a feedback manner.