JP2003245706A - Method for re-pass scheduling under rolling in a plurality of pass rollings - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method for re-pass scheduling under rolling in a plurality of pass rollings, where flatness and plate thickness accuracy can be improved without changing the number of passes in an initial pass schedule regardless of the existence or the kind of a shape control function. <P>SOLUTION: Using a measured value about the result of the (n-1)-th pass such as a plate thickness on an exit side, the estimated amount δ (ΔCr<SB>i</SB>) of variation to the variation value (ΔCr<SB>i</SB>) of plate crown ratio before a re- calculation on and after the n-th pass is calculated based on the deviation between the measured value and its target value set before the re-calculation. A correction amount Δh<SB>i</SB>in plate-thickness setting to the target plate thickness h<SB>i</SB>on the exit side of each pass, which ranges from the n-th pass to the pass one pass before the last pass in the initial pass schedule X preset, is calculated so that the variation amount δ (ΔCr<SB>i</SB>) of plate crown ratio is minimized by performing a calculation based on the estimation model of the variation amount δ(ΔCr<SB>i</SB>). <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION During rolling when a steel sheet is manufactured by performing reverse rolling over a plurality of rolling passes, the rolling pass schedule is recalculated and changed for the remaining rolling passes that have not yet been rolled. The present invention relates to a repath scheduling method.

[0002]

2. Description of the Related Art In the case of thick plate rolling in which a reverse rolling mill performs rolling over a plurality of rolling passes, conditions on the rolling mill side (allowable rolling load, allowable torque, etc.) and conditions on the rolled material side (plate Before the rolling, a pass schedule calculation for determining the number of rolling passes and the amount of reduction in each rolling pass is performed according to the thickness, strip width, allowable flatness, etc.). In this pass schedule calculation, a load prediction model is used to calculate the rolling load based on the specifications of the rolling mill (roll diameter, mill constant, etc.) and the material conditions to be rolled (sheet thickness, sheet width, temperature, etc.). Used. However, during actual rolling, various disturbances and model errors exist, so that a load prediction error occurs, which causes deterioration of the plate thickness accuracy and flatness. Therefore, in multi-pass rolling, the parameters in the model formula are learned based on the measured values such as the rolling load results for the rolling passes that have already been completed to correct the load prediction model, and the passes for the remaining rolling passes are corrected. Recalculation of the schedule (repass scheduling) is also performed.

As a re-pass scheduling method in the prior art, for example, a method described in Japanese Patent Application No. 7-54811 is known. In this method, when performing reverse rolling according to a predetermined initial pass schedule, the first pass of rolling is the initial setting of the shape control function according to the target rolling load obtained from the load prediction formula and the actual rolling during rolling. Rolling is performed while correcting the rolling position for matching the load with the target rolling load and learning the parameters of the load prediction formula. Then, for the second and subsequent passes, correction of the pass schedule up to the final pass based on the load prediction formula learned in the previous pass, and the initial value of the shape control function according to the target rolling load obtained from the load prediction formula Rolling is performed while repeating setting, correction of the rolling position for matching the actual rolling load with the target rolling load, and learning of parameters in the load prediction formula during or immediately after rolling of each pass.

However, according to the above method, since the initial value of the shape control function is set, it cannot be applied when the rolling mill does not have the shape control function, and the thermal crown control is used as the shape control function. In the case where an actuator having a long time constant is provided, it is extremely difficult to change the setting from a path on the way and control. Then, in the case of a rolling mill that does not have the shape control function, or when the limit value is reached even if the shape control function is provided (that is, the shape control actuator has already reached the limit of the control range in the initial pass schedule). If the shape control function cannot be used because the control capability is exceeded), even if the above method is applied, if the recalculation is performed with the same pass schedule as the initial pass schedule, the load prediction error Since the target rolling load is changed by modifying the above, there may occur a case where the flatness allowable condition cannot be secured. Therefore, it is necessary to increase or decrease the number of paths in the modification of the path schedule up to the final path based on the learned load prediction formula.

As described above, in the re-pass scheduling, if the number of passes increases or decreases with respect to the initial pass schedule, the rolling direction of the final pass differs depending on whether the number of passes is odd or even. It will also occur if it does not match the path end direction. Therefore, after the rolling is completed, time is wasted for sending the plate in the opposite direction again. Also, when hot rolling (for thick plate rolling, etc.), the change in the number of passes (mainly longer) not only lowers the rolling productivity but also lowers the temperature due to the longer rolling time. Will also occur. When such a temperature drop occurs, the final finishing temperature, which has an important influence on the material properties, may not be maintained. For this reason,
In rolling mills that do not have the shape control function, and in situations where the shape control function has reached its limit, re-pass scheduling can be performed without changing the number of passes in the initial pass schedule. It is necessary to secure the final finishing temperature conditions. Then, after that, it is necessary to correct the load prediction model error and perform repass scheduling so as to satisfy the target values of the flatness, the final plate thickness, and the final plate crown.

[0006]

SUMMARY OF THE INVENTION In view of the above-mentioned circumstances, the present invention has made it possible to improve the flatness and the thickness accuracy without changing the number of passes in the initial pass schedule regardless of the presence or type of the shape control function. It is an object of the present invention to provide a re-pass scheduling method during rolling in multi-pass rolling, which can improve the error and reduce the error of the plate crown.

[0007]

In order to solve the above-mentioned problems, a repath scheduling method according to claim 1,
During rolling when manufacturing a steel sheet by performing reverse rolling over multiple rolling passes, before the n-th rolling, which is one of the second or subsequent rolling passes, the pass schedule already set , A re-pass scheduling method of re-calculating and changing the pass schedule of the n-th pass and thereafter based on the rolling record up to the (n-1) th pass.
Using the measured value of at least one of the delivery side plate thickness of the pass, the delivery side plate temperature of the n−1th pass, and the rolling load from the first pass to the n−1th pass, the nth pass and thereafter. The predicted change amount for the value before recalculation of the plate crown ratio change in each of the, the respective target value and each measured value before the recalculation of the exit side plate thickness or the exit side plate temperature or the rolling load Find based on deviation,
A calculation is performed based on a prediction model of the change amount of the plate crown ratio change, and a pass immediately before the last pass from the n-th pass in the initial pass schedule preset at the start of reverse rolling so as to minimize the change amount. It is characterized in that the plate thickness setting correction amount for the target plate thickness on the exit side of each path is calculated.

According to this structure, the change in the strip crown ratio in each pass after the nth pass, which was calculated before the recalculation, is re-calculated with respect to the exit side plate thickness or the exit side plate temperature or the rolling load in the pass in which rolling has already been completed. The amount of change in strip crown ratio change in each pass after the n-th pass is calculated by using a calculation model that predicts how much it will change depending on the deviation between each target value before calculation and each measured value in the rolling results. In order to minimize, the plate thickness setting correction amount for the target plate thickness on the exit side of each pass from the nth pass to the pass immediately before the final pass at the time of the initial pass schedule is calculated. Therefore, even if the rolling mill does not have the shape control function, or if the rolling mill has the shape control function and is applied in a situation where it reaches the limit, the reverse rolling is performed. In the middle of the process, the rolling history can be learned to correct the load prediction model error and re-pass scheduling for improving the flatness can be performed without changing the total number of passes in the initial pass schedule.
Since the load prediction accuracy is improved, the final plate thickness accuracy can be improved, and the error from the target value of the plate crown can be reduced. Further, since the flatness is improved, it is possible to reduce the need for performing shape correction with a leveler or the like after the completion of rolling, so that the processing load in the refining step can be reduced.

According to a second aspect of the present invention, there is provided the repath scheduling method according to the first aspect, wherein the prediction model is a model calculated by linearly approximating the variation by the deviations. It is characterized in that an evaluation function including a function obtained by squaring is assumed, and the plate thickness setting correction amount is calculated from the condition that minimizes this evaluation function.

According to this configuration, the deviation between the actual rolling and the target value thereof can be used to formulate the change amount of the change of the strip crown ratio in the actual multiple-pass rolling as a minute change, and linearly By approximating accurately,
In addition, it is possible to obtain a simple prediction model of the amount of change in the plate crown ratio change. Then, by setting the evaluation function obtained by squaring this prediction model, it is possible to easily calculate the plate thickness setting correction amount that minimizes the change amount of the plate crown ratio change.

According to a third aspect of the present invention, there is provided the re-pass scheduling method according to the second aspect, wherein the evaluation function includes a weighting function having respective adjustment parameters for weighting according to a change in plate crown ratio of each path, Roll diameter, the coefficient of shape change of each pass obtained as a function of the plate thickness and the plate width of the rolled material, and the shape dead zone threshold of each pass obtained as a function of the plate thickness and the plate width of the rolled material, The weighting function is obtained by setting each of the adjustment parameters so as to be proportional to a value obtained by squaring a value obtained by dividing the shape change coefficient by the shape dead zone threshold value. .

With this configuration, it is possible to set the change amount of the plate crown ratio change to be smaller for the path having the larger shape change coefficient, and similarly, the change of the plate crown ratio change for the path having the smaller shape dead zone threshold value. The amount can be set to be small. That is, it is possible to provide a plate thickness setting correction amount that can minimize the amount of change in the plate crown ratio change according to the conditions of each pass to be re-scheduled.

Repath scheduling according to claim 4.
Method is to perform reverse rolling over multiple rolling passes
During the rolling process for manufacturing steel sheets by the second pass or later
Has already been set before the nth pass, which is the rolling pass of
The rolling schedule of the n-1th pass against the existing pass schedule.
Recalculate the pass schedule after the nth pass based on the record
The re-pass scheduling method is to change the
Initial pass schedule preset at the start of berth rolling
For the total number m of paths in
Sheet crown in the i-th pass, which is either rolling pass
Ratio change ΔCr _iI-pass outgoing goal before recalculation
Plate thickness h_iMatrix J whose element is the rate of change
Calculated according to the following formula, and
Target value before recalculation, and the actual thickness of the inlet side plate of the nth pass
Deviation ΔH from the measured value of performance_nAnd obtain this ΔH_nAnd the board
Raun ratio change ΔCr_iEntry thickness H of the nth pass of_nAgainst
Vector A obtained with the product of the change rate_HWhen,
Change in plate crown ratio ΔCr_iWeighting according to
Adjustment parameter w for_iA matrix whose elements are
W and W are calculated according to the following formulas, respectively, and J, A and W are
Using the i-pass output side target thickness h before recalculation_iAgainst
Plate thickness setting correction amount Δh_iTo calculate using
Is characterized by.

[0014]

[Equation 4]

According to this configuration, the change in the plate crown ratio in each pass after the nth pass, which was calculated before the recalculation, is the target value before the recalculation of the exit side plate thickness of the (n-1) th pass and the rolling thereof. Use a calculation model that linearly approximates and predicts how much further changes will occur depending on the deviation from the actual measured values, and minimize the amount of change in the plate crown ratio change in each pass after the nth pass. , The plate thickness setting correction amount for the target plate thickness on the output side of each pass from the nth pass to the m-1th pass at the time of the initial pass schedule is calculated. Therefore, even if the rolling mill does not have the shape control function, or if the rolling mill has the shape control function and is applied in a situation where it reaches the limit, the reverse rolling is performed. In the middle of the process, the rolling history can be learned to correct the load prediction model error and re-pass scheduling for improving the flatness can be performed without changing the total number of passes in the initial pass schedule. Since the load prediction accuracy is improved, the final plate thickness accuracy can also be improved, and the error from the target value of the plate crown can be reduced. Further, by using the above model, it is possible to easily and accurately calculate the plate thickness setting correction amount that minimizes the change amount of the plate crown ratio change.

According to a fifth aspect of the present invention, in the repass scheduling method according to the fourth aspect, when calculating each element of the matrix J, the measured values of rolling load results from the first pass to the (n-1) th pass are calculated. It is characterized by using a weight prediction formula which is learned and corrected by using the formula.

According to this configuration, each element of the matrix J can be calculated using the load prediction formula with higher accuracy, and the prediction calculation accuracy in repath scheduling can be improved.

Repath scheduling according to claim 6.
The method according to claim 4 or 5, wherein the matrix
Element w of W_iOf the roll diameter, strip thickness and strip width
Shape change coefficient ζ in the i-th pass obtained as a function_i
And the i-parameter obtained as a function of the thickness and width of the rolled material.
Shape dead zone threshold η_iAnd using w_i= Ζ _i ²
/ Η_i ²It is characterized by giving as.

According to this configuration, it is possible to set the change amount of the plate crown ratio change to be smaller for the path having the larger shape change coefficient, and similarly, the change of the plate crown ratio change for the path having the smaller shape dead zone threshold value. The amount can be set to be small. That is, it is possible to provide a plate thickness setting correction amount that can minimize the amount of change in the plate crown ratio change according to the conditions of each pass to be re-scheduled.

A repass scheduling method according to a seventh aspect is the repass scheduling method according to any one of the fourth to sixth aspects, wherein the rolling load from the first pass to the n-1th pass is a target value in each pass before recalculation. And a deviation ΔP _k from the measured rolling load actual value of the k-th rolling pass which is one of the first to n−1th rolling passes, and this ΔP _k and the plate crown ratio change ΔCr _{i are obtained.} Characterized in that a vector A _P obtained in consideration of the term of the sum of the product of the change of the k-th pass to the rolling load is also calculated according to the following formula, and this vector A _P is used instead of the vector A _H. To do.

[0021]

[Equation 5]

According to this construction, the rolling load of the k-th pass is actually measured.
Deviation ΔP from each measurement value _kCalculated in consideration of
Matrix vector A_PBy calculating using
Plate thickness setting correction to minimize the amount of change in plate crown ratio change
The quantity can be calculated more accurately.

According to a seventh aspect of the present invention, there is provided the repass scheduling method according to the seventh aspect, wherein the target value before the recalculation of the (n−1) th pass outlet plate temperature and the (n−1) th pass outlet plate temperature result. The deviation ΔT _n-1 from the measured value is calculated and this ΔT
A vector _AT obtained by considering the product term of _n-1 and the rate of change of the plate crown ratio change ΔCr _i with respect to the _i- th pass outlet plate temperature is calculated according to the following equation, and this vector A _T is calculated as follows. It is characterized in that it is used in place of A _P.

[0024]

[Equation 6]

According to this configuration, the deviation from the measured value of the exit side plate temperature on the n-1th pass (the entrance side plate temperature deviation on the nth pass) Δ
By using the matrix A _T calculated in consideration of T _n-1 , it is possible to more accurately calculate the plate thickness setting correction amount that minimizes the change amount of the plate crown ratio change.

According to a ninth aspect of the present invention, in the repass scheduling method according to the eighth aspect, the ΔH _n , ΔP _k , and ΔT _n-1 are calculated before rolling for each n-th pass, and the shape dead zone threshold is calculated. When η _i exceeds a predetermined value, the pass schedule for the nth pass and thereafter is recalculated.

According to this structure, since the pass schedules for the nth pass and thereafter are recalculated only when the shape dead zone threshold is exceeded, repass scheduling can be performed efficiently.

[0028]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below. INDUSTRIAL APPLICABILITY The present invention is preferably applied to thick plate rolling or the like during hot reverse rolling over a plurality of rolling passes to produce a steel sheet, particularly during finish rolling, which is the final stage of steel sheet production. The re-pass scheduling method of It should be noted that the method other than the thick plate rolling can be applied.

Of the re-path scheduling method according to the present embodiment, first, the re-path scheduling method 1 according to the first embodiment will be described with reference to the flowchart of FIG. In FIG. 1, before starting reverse rolling, first, similar to the conventional pass schedule calculation, the initial pass schedule calculation that is the basis of the re-pass schedule calculation is performed (X), and the target plate thickness on each pass exit side is calculated. Calculated. Then, the re-pass scheduling method 1 is performed according to the initial pass schedule X (conventional pass schedule calculation), after the reverse rolling is started, and before the rolling of any of the second and subsequent passes during the rolling ( A
1-11). First, the initial pass schedule calculation X will be briefly described.

A flow chart for explaining the initial pass schedule calculation X is shown in FIG. In FIG. 5, first, when the product plate thickness (final pass delivery side plate thickness) h _out is given (X
1) Assuming a rolling reduction r _i of each pass as an initial value for the pass schedule calculation (X2). Then, a one-dimensional difference calculation in the plate thickness direction is performed using a calculation model that considers the roll contact condition and the air-cooling condition between passes, and the temperature T _i in each pass is predicted (X3). Based on this, the rolling load etc. of each pass is calculated. First, the calculation is performed from the final rolling pass (m pass) (X4), and the previous pass is calculated while successively moving to the upstream side (X5). . In each pass, the deformation resistance during rolling given as a temperature function is calculated, and the rolling load P
Predictively calculate _i (X6). The elastic deflection of the rolling roll is calculated based on this predicted rolling load P _i (X7). Then, based on this roll deflection calculation result, the plate (rolled material)
Then, the change ΔCr of the plate crown ratio is calculated. It is judged whether or not the plate crown ratio change ΔCr is smaller than a predetermined value, and if it is larger than the predetermined value, the entrance side plate thickness is corrected to be thinner so that the rolling reduction of the pass is lightened (X9). Further, it is judged whether or not the predicted rolling load and the predicted rolling torque exceed the load and torque constraint conditions (mechanical specifications) of the rolling mill, and if they exceed, the reduction rate of the pass is reduced. Yes (X10). As long as the entrance side plate thickness of the pass is smaller than the finish entrance side plate thickness (plate thickness before finish rolling), the calculation is not completed for all rolling passes, so these calculations (X5 to 10) are repeated (X11). . When the calculation is completed for all the passes and the constraint condition is satisfied, the average plate temperature T _n and the average delivery side plate thickness h _{out, n} in all the passes (for example, nth pass) are determined (X13). . If the constraint condition is not satisfied, the initial rolling reduction is changed and the calculation is performed again (X12).

Initial path schedule calculation X described above
When is finished, according to this initial pass schedule X,
Rolling starts as shown in FIG. First, based on the schedule of the first pass, the roll gap of the rolling mill is set (A1), and the temperature T ₀ at the start of rolling is measured (A1).
2). Then, rolling starts (A3). During the rolling of the first pass, the actual rolling load is measured, and the outlet plate thickness is calculated (measurement of the gauge meter plate thickness) based on the gauge meter formula (A4). Here, the gauge meter plate thickness is a gauge meter type that considers the rolling mill as a spring,
It is a calculated value (measured value) of the actual output side plate thickness calculated from the actual rolling load based on the mill constant, the rolling position information, and the like. Learning correction of the load prediction model (load prediction formula) is performed based on the rolling load and the delivery side plate thickness (gauge meter plate thickness) measured by this rolling record (A7). This learning correction is
For example, it is performed by changing the parameters in the load prediction formula to values or the like calculated back from the actual rolling results and modifying the load prediction formula. When this learning is performed, it is next determined whether or not the re-pass schedule calculation is performed (A5). The processing when it is determined that the re-pass schedule calculation is necessary will be described later, but when the re-pass schedule calculation is determined to be unnecessary, the process continues to the second pass, and the same processing is performed for the second pass. (A1 to A5). These steps (A1 to A5) are performed until the completion of rolling of the previous pass of the final pass, and when the processing of the final pass is completed (A6),
The rolling is finished (A12).

Next, the processing when it is judged that the re-pass schedule calculation is necessary during the whole rolling process described above will be explained. First, the total number of passes preset in the initial pass schedule calculation X is set to m (that is, the final pass is the m-th pass) before the nth pass, which is one of the second and subsequent passes. It is determined that the repass schedule is calculated (A5). In this case, n-
Based on the rolling results of the first pass, the pass schedule for the nth pass and thereafter will be recalculated and changed. In this repass schedule calculation, it is possible to improve the plate thickness accuracy and flatness by correcting the load prediction model error by learning the rolling results without changing the total number of passes m. And the plate thickness setting correction amount Δh _{i with} respect to the target plate thickness h _i in the initial pass schedule X from n to m−1 passes.
Is to give. The subscript i is from the n path to m-
It indicates that the value is for the i-th pass, which is one of the rolling passes up to the first pass.

In calculating the plate thickness setting correction amount Δh _i , the measured value of the output side plate thickness of the n−1th pass (gauge meter plate thickness) h _{n-1 # act} (the input side plate thickness of the nth pass H _{n # act} ) can also be used. Then, the predicted change amount δ (ΔCr _i ) with respect to the value ΔCr _i before the recalculation of the plate crown ratio change in the i-th pass is calculated as the target plate thickness h _n−1 on the exit side of the n−1 pass and its actual measured value H _{n. Deviation from #act} ΔH _n
Change amount δ (ΔC
r _i ) is calculated based on the prediction model, and the variation δ
The plate thickness setting correction amount Δh _i is calculated so as to minimize (ΔCr _i ). As the prediction model of the change amount δ (ΔCr _i ), a model in which the change amount δ (ΔCr _i ) is linearly approximated by the deviation ΔH _n is used, and this prediction model is obtained by squaring. Assuming an evaluation function φ including a given function, the plate thickness setting correction amount Δh _i is calculated from the condition that minimizes this evaluation function. Hereinafter, the concept of setting the variation amount δ (ΔCr _i ) and the evaluation function φ will be described, and a specific method of calculating the plate thickness setting correction amount Δh _i will be described.

In general, the irregularity of flatness during rolling in plate rolling is the difference between the plate crown Cr _out on the rolling mill exit side and the plate crown Cr _in on the rolling mill entrance side formed by roll bending due to rolling load. represented by the strip crown is divided by the side plate thickness h out ratio change _{ΔCr = (Cr ou t -Cr in} ) / h. At this time, the factors that determine the output side plate crown Cr _out are the rolling load P, the plate width W, and the input side plate crown Cr _in . On the other hand, the rolling load P is represented by the rolling exit side plate thickness h, the entrance side plate thickness H, the deformation resistance k of the material to be rolled, and the like. Further, the deformation resistance k has temperature dependence and is a function of the temperature T of the material to be rolled. Therefore, in general, the change in plate crown ratio ΔCr
Can be expressed as ΔCr = func (h, H, T, W, Cr _in ).

The amount of change in the plate crown ratio change ΔCr during the actual multi-pass rolling can be Taylor-developed on the assumption of a minute change.
The change amount Δδ (ΔCr _i ) of r _i can be formally expressed as follows.

[0036]

[Equation 7]

Here, h _j and T _j respectively represent the plate thickness on the exit side of the j path. Further, only the first pass is described by using the entrance side plate thickness H _1, and after the second pass, since the exit side plate thickness of the preceding pass is the entrance side plate thickness as it is, it is indicated by using the exit side plate thickness of the previous pass. ing.

Usually, in the initial pass schedule calculation, the plate temperature and the delivery side plate thickness in the final rolling pass satisfy predetermined values, and the flatness is also satisfied (that is, the predicted shape in each pass). Where the predicted shape = plate crown ratio change ΔCr _i × shape change coefficient ζ _i <shape dead zone threshold η _i
Have been decided to satisfy the relationship. For this reason,
When a deviation from the target value occurs with respect to the plate thickness or temperature in each pass, the plate crown ratio change changes according to the above equation (1). Therefore, the change amount Δδ of the plate crown ratio change ΔCr _i with respect to these target values
In order to minimize (ΔCr _i ) and improve the flatness, the change amount Δδ (ΔCr _i ) is expressed by the following equation (2).
If the evaluation function φ obtained by squaring is defined, the processing becomes easy.

[0039]

[Equation 8]

Here, w _i is a weighting function having each adjustment parameter for weighting according to the plate crown ratio change in each pass. This will be described later.

The evaluation function φ shown in the above equation (2) is minimized.
Thus, the output side target plate thickness h of the i-th pass _iThickness setting for
Correction amount Δh_iIs given, the flatness on the output side of each i-path
Can be re-scheduled while satisfying
It becomes Noh.

Here, when approximation is performed by ignoring the term relating to the temperature deviation in the equation (1), the evaluation function φ can be expressed in a matrix as the following equation (3).

[0043]

[Equation 9]

However, J, Δh, A, and W are given below, respectively.

[0045]

[Equation 10]

Incidentally, the superscript T means a transposed matrix, and the abbreviation symbol (.) Is used in the notation in the expressions J, A, and W. Since h _m + Δh _m needs to match the product plate thickness h _m , the above equation is derived by imposing the condition of Δh _m = 0.

The condition for minimizing the evaluation function φ is φ /
Since it is given by Δh = 0, it can be calculated based on the following equation (4).

[0048]

[Equation 11]

Therefore, the plate thickness setting correction amount Δh that minimizes the evaluation function φ (that is, makes the flatness good)
Is given by the following equation (5).

[0050]

[Equation 12]

Further, with respect to the matrix W which is the above-mentioned weighting function, in order to perform weighting according to the change of the plate crown ratio of each path i, the elements w _{i of the} matrix W are
The shape change coefficient ζ _i in the i-th pass obtained as a function of the roll diameter, the plate thickness and the plate width of the rolled material, and the shape dead zone threshold η in the i-th pass obtained as a function of the plate thickness and the plate width of the rolled material It is desirable to give w _i = ζ _i ² / η _i ² using _i and. Thus, the change amount of the plate crown ratio change can be set to be smaller for the path having the larger shape change coefficient, and similarly, the change amount of the plate crown ratio change can be set to be smaller for the path having the smaller shape dead zone threshold. Can be set to.

The above equation (5) is an equation for obtaining the plate thickness setting correction amount Δh in each pass when the entrance-side plate thickness deviation ΔH ₁ of the first pass exists with respect to the initial pass schedule calculation. . Therefore, when rolling is performed up to the (n-1) th pass, the deviation between the actual output plate thickness and the target plate thickness at the (n-1) th pass, that is, the entrance-side plate thickness deviation ΔH _n of the nth pass is used. And m−1 from the n path that minimizes the evaluation function φ
The plate thickness setting correction amount Δh up to the pass can be calculated by the following equation (6).

[0053]

[Equation 13]

In the equation (6), the vector A in the equation (5) is defined as the vector A _H.

Here, returning to FIG. 1 again, the case where the re-pass schedule calculation is performed for the pass schedules of the n-th and subsequent passes after the completion of rolling to the (n-1) th pass will be described. n
When it is judged that re-pass schedule calculation is to be performed at a stage before the n-th rolling is completed after the rolling up to the -1st pass (A6), first, the target before the re-calculation of the entrance side plate thickness of the n-th pass The deviation ΔH _n between the measured value and the measured value of the nth pass entrance side plate thickness is obtained (A8). The target of the crown ratio change ΔCr _i , which is each element of the matrix J in the formulas (A7) and (6), is calculated using the load prediction formula learned and corrected by the rolling load results from the first pass to the n−1th pass. The rate of change {∂ΔCr _j / ∂h _i } with respect to the plate thickness h _i is calculated (A9). Further, all matrices or vectors required for the calculation of the expression (6), J, A _H , and W are calculated (A10). Then, using these calculated J, A _H , and W, the plate thickness setting correction amount Δh from the nth pass to the m−1th pass is calculated by the equation (6) (A11). This allows
The re-pass schedule calculation before the n-th rolling is completed, and the rolling from the n-th pass to the m-th pass is continued based on the re-calculated pass schedule (A1 to A12).

The above is the repath scheduling method 1 according to the first embodiment. Next, a repath scheduling method 2 according to the second embodiment will be described. Figure 2
2 shows a flowchart for explaining the repath scheduling method 2. This re-path scheduling method 2 is
Case of repass scheduling 1 and arithmetic processing A8 and A
It is the same except for 10. Therefore, the description of the overlapping portions will be omitted and only different portions will be described.

When it is decided to perform the re-pass schedule calculation (A5), in the re-pass scheduling method 2, in addition to the deviation ΔH _{n from the} target plate thickness on the n-pass entry side, the first pass through the n−1 pass The deviation ΔP _k between the target load and the actual load for the rolling load up to the eye is calculated (A8
2). Here, one of the first to (n-1) th rolling passes is referred to as a k pass. Then, in order to consider the deviation of the initial plate crown Cr _in due to the actual load, the following vector A _P is calculated instead of the vector A _{H in} the repass scheduling method 1 (A102).

[0058]

[Equation 14]

Therefore, Δh, which is the plate thickness setting correction amount from the nth pass to the m−1th pass, is Δh = − (J ^T · W ·
J) would be given by ^{-1 · J T · W · A} P. According to this re-pass scheduling method 2, it is possible to more accurately calculate the plate thickness setting correction amount Δh _i that minimizes the change amount δ (ΔCr _i ) of the plate crown ratio change ΔCr _i .

Next, a repath scheduling method 3 according to the third embodiment will be described. FIG. 3 shows a flowchart for explaining the repath scheduling method 3.
This re-pass scheduling method 3 is the same as the re-pass scheduling 1 except for the arithmetic processes A8 and A10. Therefore, the description of the overlapping parts will be omitted and only different parts will be described.

Decide to perform re-pass schedule calculation
If (A5), the re-path scheduling method 3
Is the deviation ΔH from the target plate thickness on the n-pass entry side._n,as well as
Eyes on rolling load from the 1st pass to the n-1th pass
Deviation ΔP between standard load and actual load _kIn addition to the n-1 pass
Target value and its actual value before recalculation of the outlet plate temperature of
Deviation ΔT from plate temperature_n-1(A83). And
Vector A in repass scheduling method 1_Hof
The following vector A instead_TIs calculated (A103).

[0062]

[Equation 15]

Therefore, Δh, which is the plate thickness setting correction amount from the nth pass to the m−1th pass, is Δh = − (J ^T · W ·
J) ^-1 · J ^T · W · A _T will be given. According to this re-pass scheduling method 3, it is possible to accurately calculate the plate thickness setting correction amount Δh _i that minimizes the change amount δ (ΔCr _i ) of the plate crown ratio change ΔCr _i .

Finally, a repath scheduling method 4 according to the fourth embodiment will be described. FIG. 4 shows a flowchart for explaining the repath scheduling method 4. The re-pass scheduling method 4 is the same as the re-pass scheduling 1 except for the arithmetic processes A5, A8 and A10, and therefore the description of the overlapping parts will be omitted.
Only different parts will be described.

This repath scheduling method 4 is
After measuring the rolling load and exit side plate thickness during n-1 pass rolling
(A4), the bias described in the repath scheduling method 3
Difference ΔH _n, ΔP_k, ΔT_n-1Is calculated (A5
4). Then, we explain in Re-path Scheduling Method 3.
Vector A_TIs calculated (A55). In addition,
Tor A_TAll the element components of_iBeyond
It is judged whether or not (A56), and if it exceeds this,
Recalculating the pass schedule for the nth pass and beyond
(A9, A104, A11). In this case,
Vector A_THas already been calculated (A55),
In the arithmetic processing A104, the matrices J and W
Only the calculation needs to be done.

According to this method, the pass schedules after the nth pass are recalculated only when the shape dead zone threshold value η _i is exceeded, so that the re-pass scheduling can be performed efficiently.

The repath scheduling method (1 to 4) according to the present embodiment has been described above. Among these, an example to which the repath scheduling method 1 shown in FIG. 1 is applied will be described.

FIGS. 6 and 7 show the case where the initial path schedule calculation X is performed and the total number of paths is 11 and the repath scheduling method 1 is applied. The invention is a result of comparison between the invention and the case where the invention is not applied (prior art).
All the results are shown for the case where the deviation (corresponding to ΔH _n in the description of the embodiment) of the entrance side plate thickness of the fifth pass (corresponding to the n path in the description of the embodiment) occurs, and FIG. The rolling load calculation value in each pass is compared with FIG. 7 in which the plate crown ratio change (flatness) calculation value in each pass is compared. As can be seen from FIG. 6, in the conventional technique (marked with Δ), a large load fluctuation occurs after the fifth pass, whereas in the present invention (marked with ○), the load fluctuation can be smoothed. There is. Further, as can be seen from FIG. 7, the plate thickness deviation ΔH ₅ on the entry side of the 5th pass causes a drastic change in the crown ratio change after the 5th pass in the conventional technique (marked with Δ). On the other hand, in the present invention (marked with □), it can be seen that the fluctuation of the change in the crown ratio also changes smoothly. In addition, when rolling was actually carried out based on this re-pass scheduling method 1, the flatness and strip thickness accuracy were improved without changing the total number of passes of the initial pass schedule, and the error of strip crown was also small. I was able to confirm that I can do it.

The above is the description of the present embodiment. In addition,
The embodiment is not limited to the above-described embodiments, and may be modified as follows, for example.

(1) In this embodiment, the gauge meter plate thickness is used as the measured value of the output plate thickness in the (n-1) th pass, but it is not necessarily limited to the case where the gauge meter plate thickness is used. You may use the measured value with a plate thickness meter.

(2) In the re-pass scheduling method 4, all of the deviations ΔH _n , ΔP _k , and ΔT _n-1 are calculated (A5
4) Although it is determined whether or not the repass schedule calculation is performed using the vector A _T (A55), this need not be the case. For example, only the deviation ΔH _n , or only the deviations ΔH _n and ΔP _k may be calculated, and the vector A _H or A _P may be calculated to determine whether or not to perform the repath scheduling calculation.

[0072]

According to the first aspect of the present invention, the change in the strip crown ratio in each pass after the nth pass calculated before the recalculation is caused by the exit side plate thickness or the exit side plate temperature in the pass in which rolling has already been completed, or Using a calculation model that predicts how much further changes will occur depending on the deviation between each target value before recalculation of the rolling load and each measured value in the rolling results, the sheet crown in each pass after the nth pass The plate thickness setting correction amount for the target plate thickness on the output side of each pass from the nth pass to the pass immediately before the final pass at the time of the initial pass schedule is calculated so as to minimize the change amount of the ratio change. Therefore, even if the rolling mill does not have the shape control function, or if the rolling mill has the shape control function and is applied in a situation where it reaches the limit, the reverse rolling is performed. In the middle of the process, the rolling history can be learned to correct the load prediction model error and re-pass scheduling for improving the flatness can be performed without changing the total number of passes in the initial pass schedule. Since the load prediction accuracy is improved, the final plate thickness accuracy can be improved, and the error from the target value of the plate crown can be reduced. Further, since the flatness is improved, it is possible to reduce the need for performing shape correction with a leveler or the like after the completion of rolling, so that the processing load in the refining step can be reduced.

According to the second aspect of the invention, attention is paid to the fact that the deviation between the actual rolling and the target value thereof can be formulated as a minute change using the deviation between the actual rolling and the target value during the multi-pass rolling. By linearly approximating, it is possible to obtain an accurate and simple prediction model of the amount of change in the plate crown ratio change. And this prediction model is
By setting the evaluation function obtained by the multiplication, it is possible to easily calculate the plate thickness setting correction amount that minimizes the change amount of the plate crown ratio change.

According to the third aspect of the present invention, it is possible to set such that the change amount of the plate crown ratio change becomes smaller for the path having the larger shape change coefficient, and similarly, the plate crown ratio becomes smaller for the path having the smaller shape dead zone threshold value. It can be set so that the amount of change is small. That is, it is possible to provide a plate thickness setting correction amount that can minimize the amount of change in the plate crown ratio change according to the conditions of each pass to be re-scheduled.

According to the invention of claim 4, the change of the plate crown ratio in each pass after the nth pass calculated before the recalculation is the target value before the recalculation of the exit side plate thickness of the (n-1) th pass. And a deviation from the measured value in the rolling record, a calculation model that linearly approximates and predicts how much change will be used to minimize the amount of change in the plate crown ratio change in each pass after the nth pass As described above, the plate thickness setting correction amount for the target plate thickness on the output side of each pass from the nth pass to the m-1th pass at the time of the initial pass schedule is calculated. Therefore, even if the rolling mill does not have the shape control function, or if the rolling mill has the shape control function and is applied in a situation where it reaches the limit, the reverse rolling is performed. In the middle of the process, the rolling history can be learned to correct the load prediction model error and re-pass scheduling for improving the flatness can be performed without changing the total number of passes in the initial pass schedule.
Since the load prediction accuracy is improved, the final plate thickness accuracy can be improved, and the error from the target value of the plate crown can be reduced. Further, by using the above model, it is possible to easily and accurately calculate the plate thickness setting correction amount that minimizes the change amount of the plate crown ratio change.

According to the fifth aspect of the present invention, each element of the matrix J can be calculated by using the load prediction formula with higher accuracy, and the prediction calculation accuracy in the repath scheduling can be improved.

According to the sixth aspect of the present invention, it is possible to set so that the change amount of the plate crown ratio change becomes smaller as the shape change coefficient becomes larger, and similarly, the plate crown ratio becomes smaller as the shape dead zone threshold becomes smaller. It can be set so that the amount of change is small. That is, it is possible to provide a plate thickness setting correction amount that can minimize the amount of change in the plate crown ratio change according to the conditions of each pass to be re-scheduled.

According to the invention of claim 7, the sheet crown ratio change is calculated by using the matrix vector A _P calculated in consideration of the respective deviations ΔP _k from the measured values of the rolling load results of the k-th pass. The plate thickness setting correction amount that minimizes the change amount of can be calculated more accurately.

According to the invention of claim 8, the matrix A _T calculated in consideration of the deviation from the measured value of the output side plate temperature of the _n- 1th pass (the input side plate temperature deviation of the nth pass) ΔT _n-1 By using, it is possible to more accurately calculate the plate thickness setting correction amount that minimizes the change amount of the plate crown ratio change.

According to the ninth aspect of the present invention, since the pass schedules for the nth pass and thereafter are recalculated only when the shape dead zone threshold is exceeded, repass scheduling can be performed efficiently.

[Brief description of drawings]

FIG. 1 is a flowchart illustrating a repath scheduling method 1 according to this embodiment.

FIG. 2 is a flowchart illustrating a repath scheduling method 2 according to this embodiment.

FIG. 3 is a flowchart illustrating a repath scheduling method 3 according to the present embodiment.

FIG. 4 is a flowchart illustrating a repath scheduling method 4 according to this embodiment.

FIG. 5 is a flowchart illustrating an initial path schedule calculation method.

FIG. 6 is a diagram showing a result of applying the present embodiment,
It is a figure showing rolling load in each pass.

FIG. 7 is a diagram showing a result of applying the present embodiment,
It is the figure which showed the plate crown ratio change in each pass.

[Explanation of symbols]

A1-A12, A54-56, A82, 83, 102-
104 Re-pass Schedule Calculation Flow X Initial Path Schedule Calculation X1-13 Initial Path Schedule Calculation Flow

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Sadao Morimoto             2222 Ikeda Pioneering, Ikeda, Onoue-cho, Kakogawa-shi, Hyogo             Address 1 Kakogawa Research Site, Kobe Steel, Ltd.             Within the ward (72) Inventor Hideto Fujiuchi             2222 Ikeda Pioneering, Ikeda, Onoue-cho, Kakogawa-shi, Hyogo             Address 1 Kakogawa Research Site, Kobe Steel, Ltd.             Within the ward F-term (reference) 4E024 AA03 AA07 AA08 BB07 CC02                       EE02 GG05

Claims (9)

[Claims] 1. It is already set during rolling when manufacturing a steel sheet by performing reverse rolling over a plurality of rolling passes and before rolling of the nth pass, which is one of the second and subsequent passes. A re-pass scheduling method of recalculating and changing the pass schedule of the n-th pass and thereafter based on the rolling record up to the (n-1) th pass with respect to the pass schedule. The plate in each pass after the n-th pass using the measured value of at least one of the thickness, the exit-side plate temperature of the n-1-th pass, and the rolling load from the first pass to the n-1-th pass. A predicted change amount of the crown ratio change with respect to the value before recalculation is obtained based on the deviation between each target value and each measured value before the recalculation of the outlet side plate thickness or the outlet side plate temperature or the load. , The plate A calculation is performed based on a prediction model of a change amount of the change of the rune ratio, and from the nth pass to the pass immediately before the final pass in the initial pass schedule preset at the start of reverse rolling so as to minimize the change amount. A re-pass scheduling method characterized by calculating a plate thickness setting correction amount for a target plate thickness on the exit side of each pass. 2. The prediction model according to claim 1, wherein:
It is a model in which the amount of change is linearly approximated by each of the deviations and is calculated, and an evaluation function including a function obtained by squaring this prediction model is assumed, and the condition is minimized from the condition described above. A re-pass scheduling method characterized by calculating a plate thickness setting correction amount. 3. The evaluation function according to claim 2, wherein the evaluation function includes a weighting function having adjustment parameters for weighting according to a change in plate crown ratio of each pass, and a roll diameter, a plate thickness of a material to be rolled, and The shape change coefficient of each pass obtained as a function of the plate width, and the shape dead zone threshold value of each path obtained as a function of the plate thickness and plate width of the material to be rolled are calculated, and the shape change coefficient is the shape dead zone threshold value. A repath scheduling method, wherein each of the adjustment parameters is set so as to be proportional to a value obtained by further squaring a value obtained by dividing, and the weighting function is obtained. 4. The rolling is set during rolling when a steel sheet is manufactured by performing reverse rolling over a plurality of rolling passes and before the n-th rolling, which is one of the second and subsequent rolling passes. A re-pass scheduling method for re-calculating and changing the pass schedule of the n-th pass and thereafter based on the rolling record of the (n-1) th pass with respect to the pass schedule, wherein an initial pass set in advance at the start of reverse rolling. For the total number of passes in the schedule m, the plate thickness ratio change ΔCr _i in the i-th pass, which is one of the rolling passes from the n-th pass to the m−1-th pass, the target thickness of the i-pass exit side before recalculation Matrix J whose element is the rate of change with respect to h _i
Is calculated according to the following equation, and the deviation Δ between the target value before recalculation of the n-th entry side thickness and the measured value of the n-th entry side thickness is
H _n is obtained, and this ΔH _n and the plate crown ratio change ΔCr
_The vector A _H obtained by using the product of the rate of change of _i with respect to the entrance side plate thickness H _n of the n-th pass and the adjustment parameter w _i for weighting according to the plate crown ratio change ΔCr _i are used as elements. Matrix W is calculated according to the following formulas, and using J, A, and W, the plate thickness setting correction amount Δh _i for the i-pass exit side target plate thickness h _i before recalculation is calculated using the following formulas. A repath scheduling method characterized by the above. [Equation 1] 5. The matrix J according to claim 4,
A re-pass scheduling method characterized by using a load prediction formula learned and corrected by using the measured values of the rolling load results from the first pass to the (n-1) th pass when calculating each element of. 6. The shape change coefficient ζ _i in the i-th pass obtained as a function of the roll diameter, the plate thickness and the plate width of the material to be rolled, and the element w _{i of the} matrix W, as defined in claim 4 or 5, Using the shape dead zone threshold η _i in the i-th pass, which is obtained as a function of the plate thickness and the plate width of the rolled material, w _i
The re-path scheduling method is characterized by giving as ζ _i ² / η _i ² . 7. The method according to claim 4, wherein the first
The target value in each pass before recalculation of the rolling load from the pass to the n−1th pass and the rolling load actual result of the kth pass, which is one of the first to n−1th rolling passes The deviation ΔP _k from each measured value is calculated, and this ΔP _k
And a vector A _P obtained by also taking into consideration the term of the sum of the product of the change in the plate crown ratio change ΔCr _i with respect to the rolling load at the k-th pass are calculated according to the following equation, and this vector A _P is calculated according to the vector A _H. A repath scheduling method characterized by being used in place of [Equation 2] 8. The deviation ΔT _n-1 between the target value before the recalculation of the outlet side plate temperature of the n−1th pass and the measured value of the outlet side plate temperature of the n−1th pass according to claim 7. The vector A _T obtained by taking into consideration the term of the product of this ΔT _n−1 and the change rate of the plate crown ratio change ΔCr _i with respect to the _i- th exit side plate temperature is calculated according to the following formula, and this vector A _T
Is used in place of the vector A _P , the re-path scheduling method. [Equation 3] 9. The method according to claim 8, wherein the ΔH _n , ΔP _k , and ΔT _n-1 are calculated before rolling for each n-th pass.
A re-pass scheduling method characterized in that when the shape dead zone threshold η _i exceeds a predetermined value, the pass schedules for the nth pass and thereafter are recalculated.