JPH0470083B2

JPH0470083B2 -

Info

Publication number: JPH0470083B2
Application number: JP58162781A
Authority: JP
Inventors: Kozo Nakai
Original assignee: Hitachi Ltd
Current assignee: Hitachi Ltd
Priority date: 1983-09-05
Filing date: 1983-09-05
Publication date: 1992-11-10
Also published as: JPS6054217A

Description

[Brief explanation of the drawing]

第１図は６段式圧延機の構成図、第２図は形状
パラメータ表示を示す図、第３図は本発明の実施
例を示す構成図、第４図は処理フローチヤートで
ある。１……６段式圧延機、２……形状検出器、３…
…板厚検出器、８……プリセツト装置、９……適
応修正装置。 FIG. 1 is a block diagram of a six-high rolling mill, FIG. 2 is a diagram showing shape parameter display, FIG. 3 is a block diagram showing an embodiment of the present invention, and FIG. 4 is a processing flow chart. 1...6-high rolling mill, 2...shape detector, 3...
... Plate thickness detector, 8... Preset device, 9... Adaptive correction device.

Claims

[Scope of Claims] 1. In a rolling control device that has a shape detector and a detector that detects strip thickness and forms a plate material of a predetermined thickness, the actual rolled shape measured by the shape detector has a target shape. The first step to calculate the variation
Determine the influence coefficient of the above shape variation using the means of , the measured value of the roll bending force during actual rolling, the measured value of the plate thickness, and the change in the actual rolled position relative to the target rolled position of the measured rolled position value. . A rolling control device, characterized in that a second means is provided for determining a shape control value including a rolling position setting value for the next rolling from the influence coefficient. [Scope of Claim] [Field of Application of the Invention] The present invention relates to a rolling control device, and particularly to a rolling control device that can easily perform constant shape control. BACKGROUND OF THE INVENTION A setup apparatus for a conventional six-high rolling mill is shown in FIG. In the figure, F _W is the set value of the roll bending pressure of the work roll, F _I is the set value of the roll bending pressure of the intermediate roll, and δ is the set value of the position of the intermediate roll. If shape parameters, which are a known display method as shown in FIG. 2, are used in this setup device, the control variables F _W , shown in FIG.
The relationship between F _I and δ is given by C _e C _q _Co = f _e1 f _e2 f _e3 f _q1 f _q2 f _q3 f _o1 f _o2 f _o3 F _W F _I 1 ...(1). Here, f _1j is δ, rolling load P,
As a function of the plate width b, the rolling load P is expressed by the linear formula,
δ and b can be expressed by quadratic equations. Therefore, conventional rolling control first calculates P=k b from the given rolling specifications (inlet side plate thickness H, plate width b, outlet target plate thickness, material steel type, target shape _e , _q , _o, _etc. ) The rolling load is calculated from the known rolling load formula: √'(-)T( _tb , _tf ) _Qp (μ)...(2) R'=R(1+ _CHP /H-h).
Here, R is the work roll radius, t _b and t _f are the rear and front tensions, T is the tension correction term, Q is the Masatsu coefficient correction term, and C _H is the Hitchikotsuk constant. Next, from the above equation (1) that satisfies the target shape, _W ,
Once F _I is determined, the reduction setting value S can be determined from the well-known gauge meter formula: S=-P-(F _W +F _I )/K (3). Here, K is the spring constant δ
is a function of In this equation (3), the zero point is set to 0 for simplicity. In this way, the various setting values that satisfy the target shape and the target exit side plate thickness at the center of the plate are expressed by formula (1).
It is determined by equations (2) and (3). Now, if we write equations (1) and (3) in matrix representation, we get becomes. As is clear from equation (4), the controlled amount h is influenced by the set values F _W , _F _I , _δ , and S as the controlled variables of the rolling mill. On the other hand, the shape parameters C _e , C _q , and _Co are not affected by the rolling reduction setting value S at all. However, according to the inventors' experiments, F _W ,
It has become clear that, while F _I and δ have almost no effect on h, the shape parameter changes due to fluctuations in the rolling reduction setting value S are considerably affected.
This is a traditional AGC (Automatic Gauge)
It is clear from the fact that the shape changes by dynamically changing the rolling reduction (Control). JP-A-55-68110 discloses that the actual values of the intermediate roll position and work roll bending pressure during the preceding rolling are detected, and the actual values are used to determine the intermediate roll position and work roll bending pressure during the next rolling. An initial setting method for a six-layer rolling mill is disclosed that corrects the determined value of . However, this publication also does not mention or take into consideration the influence of the rolling position error on the shape or how to correct this. In this way, in the conventional rolling control device,
When controlling the shape, the rolling set value S was not controlled, so the shape could not be controlled to a constant shape. [Object of the Invention] An object of the present invention is to provide a rolling control device that can perform constant shape control. [Summary of the Invention] The present invention provides that the error in roll bending force calculated using a known model to achieve the target shape has a small effect on the center plate thickness, and the error in the roll bending force calculated using the known model to achieve the target thickness. Taking advantage of the fact that the rolled position error has a large effect on the shape, we can calculate the variation of the actual rolled shape with respect to the target shape measured by a shape detector and the measured value of the roll bending force during actual rolling. The influence coefficient of the shape variation is calculated from the measured value of plate thickness and the change in the actual rolling position relative to the target rolling position measured value, and the rolling position for the next rolling is determined based on this influence coefficient. By controlling and determining the dewing force, constant shape control is performed. [Embodiments of the Invention] Examples of the present invention will be described below. FIG. 3 shows an embodiment of the invention. In the figure, 1 is a six-high rolling mill, 2 is a shape detector, 3 is a plate thickness detector, 4 is an intermediate roll shifter, and 5 is a 6-high rolling mill.
6 is a work roll bender, 7 is an intermediate roll bender, 8 is a preset device, and 9 is an adaptive correction device. Presetting device 8 for each rolling coil
Given the rolling specifications (inlet thickness H, target outlet thickness h, width b, target shape C _e , C _q , Co _etc. ),
Calculate the set values F _W , F _I , δ, and S using the following equation (5),
Setting values are given to the intermediate roll shifter 4, reduction feeder 5, work roll bender 6, and intermediate roll bender 7, respectively. Equation (5) ignores the influence of Equation (4 ₎ in the conventional device on the shape parameters C _e , C _q , _{and Co} of the rolling position S, whereas _Equation (5) The influence on _Co is determined as unknown coefficients α ₁ , α ₂ , α ₃ and adaptively corrected. For example, the target shape parameters F _W ^* , F _I *, δ ^* ^, S * such that C _e = C _q = _Co = 0 h = h ^* are obtained using equations (1), (2), and (3 ⁾ . When measured in actual rolling, ΔC _e , ΔC _q , and ΔC _o are detected as fluctuations from the target values. Now, if we expand equation (1) around the target value and summarize only the first-order fluctuations,
It becomes as shown in equation (6). ΔC _e ΔC _q ΔC _o = ∂C _e ／∂F _W ∂C _e ／∂F _I ∂C _e ／∂δ ∂C _q ／∂F _W ∂C _q ／∂F _I ∂C _q ／∂δ ∂C _o ／∂F _W ∂C _o ／∂F _I ∂C _o ／∂δΔF _W ΔF _I Δδ ……(6) Here, ∂C _e /∂F _W is f _e1 from equation (1), and for the others, (1 ) can be calculated from the formula. Therefore, by calculating and correcting ΔF _W , ΔF _I , and Δδ from equation (6) from the actually measured ΔC _e , ΔC _q , and ΔC _o , it should be possible to make ΔC _e = ΔC _q = ΔC _o = 0, but the actual result at this time is The values F _WA , F _IA , and Δ _A are usually F _WA ≠ F _W ^* , F _IA ≠ F _I ^* , δ _A ≠ δ ^* (7). On the other hand, regarding plate thickness, AGCetc is working during steady rolling, so naturally h _A = h ^* , but normally S _A ≠ S ^* . This is due to zero point error etc. That is, this S _A −
Because the influence of S ^* = ΔS on ΔC _e , ΔC _q , and ΔC _o is ignored, an unreasonable condition like equation (7) occurs. Therefore, we expand equation (5) in the same way and first obtain (8).
Get the formula. ΔC _e ΔC _q ΔC _o Δh＝∂C _e ／∂F _W ∂C _q ／∂F _W ∂C _o ／∂F _W ∂h／∂F _W ∂C _e ／∂F _I ∂C _q ／∂F _I ∂ C _o ／∂F _W ∂h／∂F _I ∂C _e ／∂δ ∂C _q ／∂δ ∂C _o ／∂δ ∂h／∂S ∂C _e ／∂S ∂C _q ／∂S ∂C _o /∂S ∂h／∂SΔF _W ΔF _I Δδ ΔS ...(8) Here, ∂C _e /∂S=α ₁ , ∂C _q /∂S=α ₂ , ∂C _o /∂S
= α ₃ , and the others can be found from equation (5). As mentioned above, the effects of ΔF _W , ΔF _I , and Δδ on Δh are almost negligible, and from the steady state actual data, ΔC _eA = (∂C _e /∂F _W )ΔF _WA + (∂C _e /∂ F _W ) ΔF _IA + (∂C _e / ∂δ) Δδ _A + α ₁ ΔS _A ΔC _qA = (∂C _q / ∂F _W ) ΔF _WA + (∂C _q / ∂F _W ) ΔF _IA + (∂C _q /∂δ)δ _A +α ₂ ΔS _A ΔC _oA = (∂C _o /∂F _W )ΔF _WA + (∂C _o /∂F _W )ΔF _IA + (∂C _o /∂δ)δ _A +α ₃ Find the unknown coefficients α ₁ , α ₂ , α ₃ from ΔS _A (9). These α ₁ , α ₂ , and α ₃ are used in the above equation (5) during the next setting calculation. Further, the influence coefficient ∂C _e /∂F _W . When the material is being rolled and in steady state, the detection results of the shape detector 2 and plate thickness detector 3 h _A ,
In addition to C _eA , C _qA , _{and CoA} , actual rolling results F _WA , F _IA ,
δ _A and S _A are collected by the adaptive correction device 9, and the unknown coefficients α ₁ , α ₂ , α ₃ are determined by the above equation (9), and from now on, they are used for the next material setting calculation using the exponential smoothing method. It is sent to the preset device 8. FIG. 4 shows a processing flowchart in the preset device 8 and adaptive correction device 9. In the figure, calculation of the set values for the relevant material is normally performed during the rolling of the previous material or at the latest before the end of the rolling of the previous material. b. Take in the material steel type Ω. Next, in step 101,
The rolling load P P = P (H, h, b, k, t _f , t _b ) for achieving the target plate thickness is calculated, and f _ij (δ, P, b) is calculated. next,
In step 102, the target shape parameters C _e ,
From C _q , _Co , and target plate thickness h, the optimal F _W , F _I , and S are determined from the above equation (5). Next, in step 103, it is determined whether or not the previous material has been rolled. When the rolling of the previous material has been completed, in step 104, set values F _W , F _I , and S are output to the rolling mill. Next, in step 105, it is determined whether actual data has been taken in during rolling or steady rolling of the material. Note that the unknown parameter α _i (i
=1, 2, 3), start with the initial value = 0 and find the optimal F _W , F _I , and S. In step 105, when the loading of the rolling results of the material is completed, the processing of the preset device 8 is completed. Next, in step 106, the differences between the target shape parameters and the actual detection results ΔC _eA , ΔC _qA , ΔC _oA and
Difference between setting F _W , F _I , δ, S and actual value ΔF _WA , ΔF _I
_A , Δδ _A , ΔS _A are determined, and in step 107, (9)
Find α ₁ , α ₂ , and α ₃ based on the formula. Next step
In step 108, the next adaptive correction value is calculated from the previous value of adaptive correction amount α _i ^-1 and α _i = α _i ^-1 + g (α _i - α _i ^-1 ), and in step 109, it is used for the calculation for the next material. α _i
(α ₁ , α ₂ , α ₃ ) is stored. In this way, the rolling specifications (ex
target plate thickness, target shape, steel type, entrance plate thickness, plate width, etc.)
is stored in a predetermined storage area of the computer in advance, and the computer starts operating after being matched with the actual material. For example, matching is performed by an operator inputting the coil number currently being rolled or scheduled for next rolling into a computer, and the computer can learn the rolling specifications of the material from the coil number of the actual material and perform setting calculation processing. becomes. Once the matching is done, automatic signals from the plant (rolling machine kami, blank signal) are sent from the plant.
etc), it is now possible for the computer to automatically track the materials to be rolled one after another. (Naturally, if there is a change in the scheduled rolling order, the operator must make corrections to the computer.) The present invention learns the rolling results of material A and corrects the setting value calculation model for material B, and from B to C and C. This is related to adaptive correction from material to material, such as from D to D..., and is not a feed back control in the rolling length direction of the material during rolling. In FIG. 3, the preset device 8 calculates the rolling reduction setting value S and the roll bending value (F _W , F _I ) that will meet the rolling specifications of the material for the rolling mill before the first rolling material is rolled. The adaptive correction device 9 has a function of detecting errors in the model calculated by the preset device 8 based on the rolling results of the first material and correcting the set value calculation model for the material to be rolled next time. If a model error is detected, that error is directly included in the next calculation formula.
Feedback usually causes a handing phenomenon, so the exponential smoothing gain is made small. [Effects of the Invention] As explained above, according to the present invention, the set value of the rolling position for the next rolling is determined using the change in the actual rolling position with respect to the target rolling position during the actual rolling. Even when rolling plate materials having a mixture of various sizes, it is possible to perform constant shape control.