JPH06122009A - Method for controlling shape of rolled stock - Google Patents

Abstract

PURPOSE:To provide a method for controlling the shape of a rolled stock by which a proper manipulated variable is easily set. CONSTITUTION:In the method for controlling the shape of the rolled stock by which the shape such as sheet crown and flatness of a steel sheet 3 is controlled with a rolling mill 2 having a control operation terminal, manipulated variable is determined by a multiple regression model of given rolling conditions, the rolling of the steel sheet 3 is executed by giving this determined manipulated variable to the control operation terminal of the rolling mill 2 and the coefficients of the multiple regression model are successively corrected by successive least squares method based on the obtained actual data of rolling.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for controlling the shape of a rolled material by controlling the shape such as the plate crown and flatness of a steel sheet by a rolling mill having a control operation end.

[0002]

2. Description of the Related Art In a rolling machine for rolling a steel sheet to be rolled by, for example, two rolling rolls, when the two rolling rolls are crossed with each other, the plate crown and the flatness after rolling are the cross angles of the rolls. It is known to change depending on the value of. For such a rolling mill,
There is a method of controlling the plate crown and flatness by adjusting the cross angle. The conventional cross angle setting calculation method will be described below.

In the case of continuously rolling a steel sheet by a rolling mill composed of m stands, the combination of the cross angles of the rolling mills is set so that the plate crown and the flatness of the steel sheet after the final rolling have the target values. Set. Assuming that only the plate crown and flatness on the exit side of the final stand can be detected, the cross angle setting calculation is formulated as follows. Objective function Clast = C ^* (Target plate crown) (1) Constraint condition | λ _i | ≦ λ _i max (2) β _i −ζ _i α _i ≦ C _i −η _i (C _i-1 −λ _i-1 h _i-1 ) ≦ β _i (1 ≦ i ≦ m) (3) where α _i = (L _B ² tan ² θ _maxi ) / (2D _W ) ... (4) λ _i : Flatness of rolled steel plate θ _i : Cross angle of i-th rolling mill θ _maxi : Maximum cross angle of i-th rolling mill C _i : i-th rolling Exit side plate crown Clast: Plate crown after final rolling β _i : Cross angle 0 (deg), entrance side plate crown 0 (μ
The i rolling mill delivery side crown D _W when the m): work roll diameter L _B: roll barrel length zeta _i, eta _i: correction factor satisfying such a constraint, and the following conditions {(Clast-C ^* ) ² + Σλ _i ² } → min ... (5) obtains a C _i by the optimization method such as quadratic programming that satisfies, _{further, ζ i C si = β i} -C i ... (6) C si = (L B 2 tan ² θ _i ) / (2D _W ) ... (7) C _si : A combination of the cross angles θ _i obtained by the equivalent mechanical crown by the roll cross of the i-th rolling mill is set as an appropriate combination.

[0004]

The calculation by the above method is complicated, and in particular, the parameters represented by ζ _i and η _i contained in the equation (3) must be adjusted. It took a lot of work to do. In addition, adjustment of these parameters was manually performed offline until an arbitrary number of actual data could be collected. The multiple regression model that has been used as a conventional operation amount setting model must provide actual data of operation amounts for all rolling conditions in order to set appropriate operation amounts for various rolling conditions. , It took a lot of time and effort to collect the result data, and the adjustment was performed offline. An object of the present invention is to provide a shape control method for a rolled material that can easily set an appropriate operation amount.

[0005]

[Means for Solving the Problems] In order to solve the above problems and achieve the object, the following measures were taken. In the shape control method of the rolled material that controls the shape such as the plate crown and flatness of the steel plate by the rolling mill having the control operation end,

An operation amount is obtained by a multiple regression model of given rolling conditions, the obtained operation amount is given to the control operation end of the rolling mill to cause the steel sheet to be rolled, and based on the obtained rolling performance data. The coefficient of the multiple regression model is sequentially corrected by the recursive least squares method.

That is, according to the present invention, the coefficient a _j (j = j =
1, 2, ..., N) are sequentially modified according to the recursive least squares method.

[0008]

As a result of taking the above-mentioned means, the following effects occur. As an example, when the manipulated variable is the cross angle θ, the physical property value indicating N rolling conditions z = (

It is assumed that M sets of sets (z, θ) of p ₁ , p ₂ , p _N ) and their corresponding manipulated variables have been obtained. At this time, the multiple regression model uses the i-th data (i = 1, 2,
..., M), the following relationship holds.

[0010]

[Equation 1] From equation (1) for M actual data, A _M X _M = B _M (9) Here,

[0011]

[Equation 2] If M> N is satisfied, solving for X yields X _M = (A _M ^T A _M ) ^-1 (A _M ^T B _M ) ... (10) (However, A _M ^T Is the transposed matrix of A _M ) The recursive least squares method expresses a matrix vector created based on M actual values as follows.

[0012]

[Equation 3]

[0013]

[Equation 4] Also,

[0014]

[Equation 5] far. The expression (8) corresponds to the bottom row of the matrix A. From these, X _{M + 1} = X _M −P _{M + 1} [z _{M + 1} z _{M + 1} ^T X _M −z _{M + 1} θ _{M + 1} ² ] (14) P _{M + 1} = P _M −P _M z _{M + 1} (1 + z _{M + 1} ^T P _M z _{M + 1} ) ^-1 z _{M + 1} ^T P _M (15) is derived.

As described above, when new rolling performance data (z _M , θ _M ) is obtained from the equations (14) and (15), the coefficient a _j of the setting calculation model is corrected and the appropriate operation is sequentially performed. You can set the amount.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a procedure flow chart showing an embodiment of a manipulated variable setting calculation method according to the present invention. As shown in FIG. 1, the operation amount setting calculation unit 1 obtains the shape control operation amount by a multiple regression model (setting calculation model) of given rolling conditions. The calculated operation amount is given to the control operation end of the rolling mill 2 which is the control target. By doing so, the steel plate 3 is rolled into a predetermined state by the rolling mill 2.

On the basis of the obtained rolling performance data, the learning unit 4 performs the calculation by the recursive least squares method, and sequentially corrects the setting calculation model given to the manipulated variable setting calculation unit 1.

As the rolling mill 2 in this embodiment, rolling rolls 2a and 2b are crossed and rolled as shown in FIG.
The PC (Pair Cross) mill shown in is applied, and the tandem rolling mill composed of rolling mills of several stands is used. The present calculation method will be specifically described below by taking the case where the cross angle of the PC mill is used as an operation amount as an example. Here, an example of the following cross angle setting calculation model is shown as the formula (8) in the PC mill. θ ² = A ₁ B + a ₂ C + a ₃ ε + a ₄ H ₁ + a ₅ H ₂ + a ₆ F (16) θ: Cross angle B: Strip width C: Strip crown after final rolling ε: Flatness after final rolling H: Inlet strip Thickness h: Delivery side plate thickness F: Rolling load a _{1 to} a ₆ : Coefficient

Preliminarily M rolling conditions and actual data (B, C, ε, H, h, etc.) of cross angles at that time (when rolling).
F, θ) and (M> N), the coefficient a _{j is calculated} from the equation (10).
(J = 1, 2, ..., 6) is calculated. Using the equation (16), which is the cross angle setting calculation model equation determined in this way, for the rolling conditions (B, H, h, F) set at that time and the plate crown and flatness of the rolled steel plate, A target value (C, ε) is given and the cross angle at that time is calculated. Rolling is performed with the calculated cross angle and rolling conditions as set values, and the actual data (B, C, ε, H,
h, F, θ), the coefficient a _{j of the} calculation model is (14),
Correct according to equation (15).

That is, in the present embodiment, the rolled steel plate is set to have a target crown and flatness of 1
An appropriate amount of operation is set and calculated from the rolling conditions by sequentially learning the setting calculation model from these actual values for each rolling operation. The present invention is not limited to the above-mentioned embodiments, and it goes without saying that various modifications can be made without departing from the gist of the present invention.

[0021]

According to the shape control method for rolled material of the present invention,
Since the operation amount setting calculation model is corrected each time rolling is performed, it is not necessary to collect a large amount of actual data in advance, and it is possible to provide a shape control method for a rolled material that can easily set an appropriate operation amount.

[Brief description of drawings]

FIG. 1 is a flowchart showing a control procedure of an embodiment of the present invention.

FIG. 2 is a diagram showing a rolling state by a rolling mill applied to the embodiment.

[Explanation of symbols]

1 ... manipulated variable setting calculation unit 2 ... rolling mill 2
a, 2b ... rolling roll 3 ... steel plate 4 ... learning unit

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁵ Identification code Office reference number FI technical display location G05B 13/04 9131-3H

Claims (1)

[Claims] 1. A shape control method for a rolled material in which the shape such as a plate crown and flatness of a steel sheet is controlled by a rolling mill having a control operation end, and an operation amount is obtained by a multiple regression model of given rolling conditions. The calculated operation amount is given to the control operation end of the rolling mill to perform the rolling of the steel sheet, and the coefficient of the multiple regression model is sequentially corrected by the successive least squares method based on the obtained rolling record data. A shape control method for a rolled material, comprising: