JPH11104726A - Correction method for roll thermal expansion prediction model - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve prediction accuracy of a roll thermal expansion prediction model by learning. SOLUTION: The amount of change ΔΔG(t)=ΔG(t)-ΔG(t-Δt) is obtained using a gauge meter error G(Δt) which is the difference between the exit side plate thickness HG(t) calculated in terms the gauge meter formula to be applied for the time (t) and actually measured plate thickness HI(t). The amount of change Δu of the thermal expansion calculated value: u(t) per roll radius in roll center position nc is calculated using roll thermal expansion prediction model: Δu=u(nc, t)-u(nc, t-Δt). When a roll thermal expansion prediction model is corrected, using the amount of change ΔΔG(t) and Δu, the roll thermal expansion quantity u(x, t+Δt) of the axial direction coordinates is corrected by the formula: u'(x, t+Δt)=u(x, t+Δt)×(Δu+α.ΔΔG/4)/u.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for correcting a roll thermal expansion prediction model, and more particularly to a method for improving the accuracy of a roll thermal expansion prediction model in the case where a rolled material is continuously hot-rolled, thereby achieving good quality rolling. The present invention relates to a method for correcting a roll thermal expansion prediction model, which is preferably applied when a material is stably supplied.

[0002]

2. Description of the Related Art For example, in a rolling operation using a plate crown prediction model such as hot finish rolling, a demand for a thickness accuracy of a rolled material in a sheet width direction is becoming stricter year by year. It has become extremely important to improve the accuracy of the data.

[0003] Generally, when predicting the strip crown of a rolled material, the strip crown is predicted by a simulation such as a division model or a finite element method. In order to achieve this, a regression equation of the calculation result of the strict calculation or a simplified model is used.

The prediction model of the sheet crown can be classified as follows when focusing on the physical meaning.

(1) Calculation model of roll deformation of rolling mill (2) Prediction model of roll profile (3) Calculation model of deformation characteristics of rolled material

Among them, the roll profile prediction model is divided into a roll thermal expansion amount prediction model and a roll wear prediction model. Generally, the roll thermal expansion by the former is based on heat conduction from a rolled material during rolling and plastic working. Fever,
From the heat balance such as frictional heat, the roll wear due to the latter is determined based on the roll diameter, the roll material, the rolling extension, the rolling load, and the like.

In these models, it is difficult to adjust various coefficients, physical property values, and the like to values suitable for actual rolling, and errors due to regression and simplification of calculation results are included. Since a predicted error occurs in the calculated sheet crown, correction (learning) based on the rolling performance is necessary.

As an example of such a prediction model learning method, for example, as disclosed in Japanese Patent Application Laid-Open No. 8-243624, a plate thickness calculated using a gauge meter formula and a plate thickness actually measured are disclosed. There is a learning method performed using the thickness. This learning method is for the i-th rolling stand
The difference ΔH between the gauge meter plate thickness and the actual plate thickness given by the following equation (1) is determined, and the ΔH and the estimated value of the thermal expansion amount u _t (c) in the axial center obtained from the roll thermal expansion prediction model are used. , The correction coefficient A _i is obtained from the following equation (2),
Using a learning coefficient Am _i (= α · A _i ) obtained by multiplying the coefficient by, for example, an adjustment coefficient α, an estimated value u of the thermal expansion amount (width direction distribution) with respect to the axial coordinate: x by the equation (3).
_This is a learning method for correcting _t (x).

H _Gi = S _i + P _i / M _i (1) where H _Gi : gauge meter plate thickness S _i : roll interval P _i : actual rolling load M _i : mill stiffness coefficient A _i = (ut _{) (c) i + ΔH / 2} ) / u t (c) i ... (2) u 't (x) i = Am i · u t (x) i ... (3)

In the above publication, the following equation is adopted as a roll thermal expansion prediction model. This model formula is obtained by improving a known model in which the roll has a two-layer structure including a core material and an outer layer material and has one average temperature in the roll radial direction.

U _t = {(θ _m1 −θ ₀ ) β ₁ r ¹² + (θ _m2 −θ ₀ ) β ₂ (R ² −r ¹² )} / R (4) where θ _m1 : core material Average temperature of the outer layer θ _m2 : average temperature of the outer layer material θ ₀ : initial roll temperature β ₁ : linear expansion coefficient of the core material β ₂ : linear expansion coefficient of the outer layer material R: roll radius r ₁ : roll core material radius

[0012]

However, the gauge meter type is based on a theoretical model of the change in the roll gap, and the deflection of the BUR (back-up roll), the flatness between the BUR-WR (work roll), WR-material flatness, deformation of the housing, oil film thickness of the BUR bearing, thermal expansion of the roll, and calculation formula of the roll wear are configured as elements.

Since all of these elements constituting the gage meter equation each have a prediction error, other prediction equations (elements) may have a larger prediction error than the prediction equation of the thermal expansion amount. . Therefore, even if learning is performed using the absolute amount of the prediction error by the gauge meter method as in the conventional method disclosed in the above publication, the prediction accuracy of the roll thermal expansion amount, that is, the prediction accuracy of the roll profile is not necessarily improved. There was a problem that it was not possible.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned conventional problems, and it is an object of the present invention to provide a method for correcting a roll thermal expansion prediction model which can improve the prediction accuracy of a roll thermal expansion prediction model by learning. I do.

[0015]

SUMMARY OF THE INVENTION The present invention provides a roll thermal expansion prediction model used for predicting a sheet crown in hot rolling, with respect to time: t, and an outlet side sheet thickness calculated using a gauge meter formula: H _G (t) and the measured sheet thickness: H _J (t) and the difference in a gauge meter _{error: ΔG (t) = H G} (t) -H J (t) ... (5) gauge determined using the Amount of change in meter error: ΔΔG (t) = ΔG (t) −ΔG (t−Δt) (6) and a thermal expansion per roll radius at a roll center position: nc calculated using a roll thermal expansion prediction model. Calculated value: change amount of u (t): Δu = u (nc, t) −u (nc, t−Δt) (7) The above problem is solved by correcting based on the relationship: is there.

According to the present invention, in the method for correcting a roll thermal expansion prediction model, the roll thermal expansion amount u (x, t + Δt) for the roll axis direction coordinate x at the time t + Δt calculated by the roll thermal expansion prediction model is calculated. The following equation: u '(x, t + Δt) = u (x, t + Δt) × (Δu + α · ΔΔG / 4) / Δu (8) (u ′ (x, t + Δt): Corrected roll thermal expansion amount , Α: adjustment coefficient), so that the prediction accuracy of the roll thermal expansion amount can be reliably improved.

The present invention has been made based on the knowledge obtained by examining in detail the prediction accuracy of the roll thermal expansion prediction model constituting the roll profile prediction model.

This finding will be described below with reference to an example of a seven-stand hot finishing mill as shown in FIG.

FIG. 1 shows the number of rolled materials constituting a rolling cycle actually rolled to examine the prediction accuracy of the above model, and the width and thickness of the rolled material. This rolling cycle is performed by joining sheet bars before finish rolling.
From the coils, eight coils are used as one unit, and finish rolling is performed without any interval between the coils in the same unit.

FIG. 2 shows the calculated value of the thermal expansion amount at the center in the roll width direction of each coil in the fourth stand F4 when the hot finishing mill is rolled according to the rolling cycle shown in FIG. Indicates a gauge meter error which is a value obtained by subtracting the gauge meter plate thickness from the measured plate thickness at the stand exit side.

Incidentally, for the sake of convenience, the sign of this gauge meter error is reversed because it takes a difference opposite to the expression (5). In FIG. 2, the thermal expansion amount is shown by multiplying the value per radius by four so as to correspond to the gauge meter error. The outline of the roll thermal expansion prediction model used for calculating the thermal expansion amount is described below. .

That is, in obtaining the roll temperature distribution, the following basic equation of heat conduction, which is obtained by averaging the circumferential temperature, heat input, and heat removal conditions of the roll and obtaining an axially symmetric two-dimensional problem, is obtained by difference approximation. Solve, calculate radial and axial temperature distributions.

(1 / r) (∂θ / ∂r) + ∂ ² θ / ∂r ² + ∂ ² θ / ∂x ² = (ρc / κ) (∂θ / ∂t) (9) , Θ: roll temperature, r: radial coordinate, x: axial coordinate, t: time, ρ: specific gravity, c: specific heat, κ: thermal conductivity.

From the obtained temperature distribution θ (x, r), the following (1)
The thermal expansion distribution u (x) in the roll axis direction was calculated by the equation (0), and the equation (10) was used as the thermal expansion prediction model.

[0025]

(Equation 1) Here, β: coefficient of thermal expansion, R: roll radius.

From FIG. 2 above, it can be seen that the thermal expansion fluctuates between the units, but that the amount of thermal expansion continues to increase while rolling the same unit. It can be seen that the gauge meter error decreases during rolling.

Then, it was examined whether or not there was a relationship between the gauge meter error and the calculated value of the thermal expansion.
The result is shown in FIG. 4, and no correlation was found between the two. However, when the relationship between the amount of change in the calculated value of the thermal expansion from the first coil of each unit and the amount of change in the gauge meter error was examined, a high correlation was found between the two as shown in FIG. It was found that there was.

From the above findings, the above equation (6) is compared with the conventional method in which learning is performed by the above equation (3) using the correction coefficient calculated based on the gauge meter error ΔH itself by the above equation (2). Of the gauge meter error given by
It is clear that the method of the present invention in which learning is performed by the above equation (8) using ΔG is superior.

[0029]

Embodiments of the present invention will be described below in detail with reference to the drawings.

FIG. 6 shows a schematic configuration of a hot finishing mill comprising seven stands applied to one embodiment of the present invention.

In this embodiment, a plurality of sheet widths and sheet thicknesses as shown in FIG. 7 are provided by the above-mentioned finishing mill in which sheet thickness gauges are installed on the exit sides of the stands from the fourth stand F4 to the seventh stand F7. And the fourth stand F4 to the seventh stand F7.
The learning of the thermal expansion model of the roll was performed by the method of the present invention using the above equation (8) and the conventional method using the above equation (3). However, the roll thermal expansion prediction model used is the above equation (10).

The learning is performed for each of the fourth stand F4 to the seventh stand F7 in a state where the rolling reaches the steady rolling 8 seconds after the leading end of the rolled material passes through the seventh stand F7 for each rolled material. The measured values of the thickness of the delivery side plate at the stand, the thickness of the gauge meter, and the calculated value of the thermal crown by the roll thermal expansion prediction model were used. At that time, the adjustment coefficient α was calculated as 0.8.

After the completion of the rolling, the roll of the fourth stand F4 was pulled out from the stand, the distribution of the thermal expansion in the width direction (thermal crown) was measured, and the measured value was compared with the calculation results of the method of the present invention and the conventional method. Is shown in FIG. From FIG. 8, it can be seen that the thermal expansion amount of the roll can be more accurately predicted over the entire width direction in the method of the present invention than in the conventional method.

FIG. 9 shows the calculated values of the thermal expansion using the roll thermal expansion models learned according to the present invention and the conventional method, and the plate on the exit side of the seventh stand, which is obtained from the calculated values. 25mm from edge
3 shows a prediction error of a calculated value of the plate crown with respect to the actually measured plate crown at the position of. From FIG. 9, it can be seen that according to the present invention, it is possible to more accurately predict the crown of the plate as compared with the conventional method.

As described above in detail, according to the present invention, since the roll thermal expansion amount can be predicted with high accuracy, the roll profile can be predicted with high accuracy, and therefore, a desired good plate profile (plate crown) can be obtained. Can be obtained.

Although the present invention has been described in detail, the present invention is not limited to the above-described embodiment, and can be variously modified without departing from the gist thereof.

For example, the roll thermal expansion prediction model is not limited to the one used in the above embodiment, and the method of the present invention can be applied to any model formula.

[0038]

As described above, according to the present invention,
The prediction accuracy of the roll thermal expansion prediction model can be reliably improved by learning.

[Brief description of the drawings]

FIG. 1 is a diagram showing a rolling cycle used in rolling in which the grounds of the invention were found.

FIG. 2 is a diagram showing a change in the amount of thermal expansion with the progress of rolling.

FIG. 3 is a diagram showing a transition of a gauge meter error with progress of rolling.

FIG. 4 is a diagram showing a relationship between a gauge meter error and a calculated value of a thermal expansion amount.

FIG. 5 is a graph showing a relationship between a change amount of a gauge meter error and a change amount of a calculated value of a thermal expansion amount.

FIG. 6 is an explanatory view schematically showing a hot finishing mill applied to an embodiment according to the present invention.

FIG. 7 is a diagram showing sheet thickness and sheet width of rolled materials constituting a rolling cycle rolled in the embodiment.

FIG. 8 is a diagram showing a measurement result of a thermal expansion amount and a learning result of a roll thermal expansion prediction model according to a conventional method and the present invention method.

FIG. 9 is a diagram showing a prediction error of a plate crown with respect to an actually measured plate crown calculated using the thermal expansion amount prediction results according to the conventional method and the present invention method.

[Explanation of symbols]

 F1 to F7: Stand TM: Thickness gauge

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁶ Identification code FI G05D 5/02 B21B 37/02 BBM (72) Inventor Katsuhiro Takebayashi 1 Kawasaki-cho, Chuo-ku, Chiba-shi, Chiba Chiba Pref. Inside the steelworks (72) Inventor Hiroshi Shiomi 1st Kawasaki-cho, Chuo-ku, Chiba-shi, Chiba Kawasaki Steel Co., Ltd.Chiba Works

Claims (2)

[Claims] 1. A roll of thermal expansion prediction model for use in predicting the strip crown in the hot rolling, the time: For t, exit side thickness is calculated using the gauge meter equation: H _G (t) and the actually measured plate thickness: H _J (t) and the gauge meter error is the _{difference: ΔG (t) = H G} (t) -H J (t) the amount of change in the gauge meter error is determined using: ΔΔG (t) = ΔG (t) −ΔG (t−Δt) and the calculated value of thermal expansion per roll radius at the roll center position: nc, calculated by using the roll thermal expansion prediction model: u (t): Δu = u (nc , t) -u (nc, t-.DELTA.t). A method for correcting a roll thermal expansion prediction model, comprising: 2. The method according to claim 1, wherein t + Δt calculated by the roll thermal expansion prediction model.
The roll thermal expansion amount u (x, t + Δt) with respect to the roll axis direction coordinate x at the time is expressed by the following equation: u ′ (x, t + Δt) = u (x, t + Δt) × (Δu + α · ΔΔG / 4) / A correction method for a roll thermal expansion prediction model, wherein the correction is performed by Δu (u ′ (x, t + Δt): corrected roll thermal expansion amount, α: adjustment coefficient).