JP2016074025A - Control method and control device of rolling machine - Google Patents

Abstract

PROBLEM TO BE SOLVED: To control a roling machine based on an estimation result, by highly accurately estimating a rolling state of a rolled material even when the number of unknown rolling characteristic value is more than the number of equation constituted of a rolling model.SOLUTION: A control method of a rolling machine on the present invention includes a construction step of constructing the optimization problem composed of an evaluation function composed of a rolling model for predicting a rolling phenomenon of a rolled material S in a plurality of roling stands F1-F7 and a constraint expression on a variation in the rolling model, an estimation step of estimating an unknown variation included in the rolling model by solving the optimization problem constructed in the construction step by substituting a measured value in a measurable variation in the rolling model and a control step of controlling the rolling machine by using the unknown variation estimated in the estimation step.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a rolling mill control method and a control apparatus for estimating a rolling state of a material to be rolled and controlling the rolling mill based on the estimation result.

  In general, in a rolling facility, a material to be rolled is continuously rolled hot or cold using a plurality of rolling stands arranged along the conveying direction of the material to be rolled. In such a rolling facility, in order to produce a rolled product of a desired size, shape, and material, initial setting of the operation amount such as the rolling roll opening degree, the rolling roll speed, and the cooling water amount is performed before rolling, When the control amount such as the thickness, width, rolling load, and temperature of the material to be rolled can be measured after rolling is started, the feedback that controls the operation amount based on the deviation between the actual value of the control amount and the target value Control is taking place.

  Here, in the initial setting of the rolling roll opening and the rolling roll speed, the rolling load and the advanced rate (from the ratio of the speed of the material to be rolled and the peripheral speed of the rolling roll on the rolling stand exit side based on the given rolling conditions. 1) and a setting calculation using a rolling model that predicts the flat roll radius and the like. This rolling model includes variables that cannot be directly measured, such as the deformation resistance of the material to be rolled and the friction coefficient between the material to be rolled and the rolling roll.

  In addition, although the rolling load can be measured with high accuracy, it is difficult to directly measure the advanced rate. For this reason, it is difficult to improve the prediction accuracy of the rolling model by combining the predicted value and the actual value by rolling. As a result, there is a discrepancy between the predicted value and the actual value by the rolling model, and the setting calculation value becomes an inappropriate value, so that the dimensions of the rolled product are out of the tolerance and the yield is lowered. When the material passes through the rolling roll, trouble may occur.

  Therefore, various learning is performed in an actual rolling facility in order to match the predicted value and the actual value of the rolling model. Specifically, in the setting calculation, the friction coefficient between the material to be rolled and the rolling roll is generally given as a fixed value, and the setting calculation is generally performed using the predicted value of the rolling model. However, the prediction error of the rolling model is unavoidable, and the actual value of the rolling load is different from the predicted value of the rolling model.

  Therefore, the ratio between the actual value and the predicted value of the rolling load is determined for each material to be rolled, and the value obtained by exponentially smoothing this is used as a learning coefficient. Improving load prediction accuracy has been performed. In addition, the deformation resistance between the material to be rolled and the rolling roll is calculated from the actual value of the rolling load using the rolling model, and the ratio between the predicted value and the calculated value of the deformation resistance based on the rolling model is used as the learning coefficient. There is also.

  However, since the coefficient of friction between the material to be rolled and the rolling roll is a fixed value, it does not always reflect the actual situation. In addition, the rolling load prediction error includes the influence of deformation resistance prediction error, etc., but since the learning coefficients described above are collectively represented by one coefficient, each error factor is the rolling load. This is not reflected correctly in the predicted value, and causes a setting calculation error.

  From such a background, Patent Document 1 discloses that the advanced rate is measured using the load cell on signal of the rolling stand in the upstream rolling stand of the hot finish rolling mill, the measured value of the advanced rate, the actual value of the rolling load, A method for calculating the deformation resistance of the material to be rolled and the friction coefficient between the rolling roll and the material to be rolled has been proposed. Patent Document 2 proposes a method for estimating a rolling characteristic value that cannot be directly measured by solving a rolling equation constructed based on measurement data. Specifically, in Patent Document 2, a numerical expression indicating that the mass flow in each rolling stand of a cold tandem rolling mill is constant, a rolling load prediction expression, and an advanced rate prediction expression are combined, and some of them A method for estimating the value of a variable by solving the variable (rolling characteristic value) as an unknown is described.

JP-A-4-284909 JP-A-4-344812

  However, in the method described in Patent Document 1, the friction coefficient between the material to be rolled and the rolling roll is a common value regardless of the rolling stand. Generally, the coefficient of friction between the material to be rolled and the rolling roll changes according to the roughness of the rolling roll, the lubrication state, and the like, and should be set for each rolling stand. For this reason, according to the method described in Patent Document 1, there is a possibility that a deviation occurs between the predicted value and the actual value by the rolling model, and the set calculation value becomes an inappropriate value.

  On the other hand, in Patent Document 2, rolling is performed when the number of simultaneous equations and the number of unknown rolling characteristic values are equal, and when the number of simultaneous equations is greater than the number of unknown rolling characteristic values (dominant decision system). A characteristic value calculation method is described. However, in actual rolling equipment, there are many cases where sensors for measuring rolling characteristic values are not installed for economic and technical reasons, and thus the number of unknown rolling characteristic values increases.

  In addition, the deformation resistance of the material to be rolled, which is treated as known in the examples described in Patent Document 2, greatly affects the rolling load and the advanced rate, but depends on the temperature and components of the material to be rolled, the rolling history so far It fluctuates greatly. For this reason, the deformation resistance of the material to be rolled should be treated as an unknown rolling characteristic value.

  Furthermore, although the coefficient of friction between the material to be rolled and the rolling roll affects the rolling load and the advanced rate, it is a physical quantity that cannot be directly measured, and a predictive model with sufficient accuracy cannot be obtained. For this reason, it is necessary to treat the friction coefficient as an unknown rolling characteristic value.

  For these reasons, in an actual rolling facility, the number of unknown rolling characteristic values is greater (underdetermined system) than the number of simultaneous equations. However, Patent Document 2 does not disclose or suggest a rolling property value calculation method that is effective when the number of unknown rolling property values is larger than the number of equations to be simultaneously provided.

  The present invention has been made in view of the above problems, and its purpose is to accurately determine the rolling state of the material to be rolled even when the number of unknown rolling characteristic values is larger than the number of equations constructed from the rolling model. An object of the present invention is to provide a rolling mill control method and a control apparatus capable of performing high estimation and controlling the rolling mill based on the estimation result.

  A rolling mill control method according to the present invention is a rolling mill control method including a plurality of rolling stands arranged in a conveyance direction of a material to be rolled, and predicts a rolling phenomenon of the material to be rolled in the plurality of rolling stands. A construction step for constructing an optimization problem consisting of an evaluation function comprising a rolling model and a constraint equation relating to a variable in the rolling model; and construction in the construction step by substituting measurement values into measurable variables in the rolling model. An estimation step for estimating an unknown variable included in the rolling model by solving the optimization problem, and a control step for controlling the rolling mill using the unknown variable estimated in the estimation step. And

  In the control method of a rolling mill according to the present invention, in the above invention, the construction step includes an evaluation function comprising a rolling model for predicting rolling load and advance rate in a plurality of rolling stands, a rolling roll in each rolling stand, and a material to be rolled. An optimization problem consisting of a constraint equation relating to the friction coefficient between the rolling material and the estimation step comprising at least measuring the thickness of the material to be rolled on the entry side and the exit side of each rolling stand and each rolling By inputting the measured value of the rolling roll speed of the stand into the rolling model to obtain the estimated value of the rolling load and the advanced rate, and solving the optimization problem using these estimated value and at least the measured value of the rolling load. Including a step of estimating a deformation resistance of the material to be rolled in each rolling stand, a friction coefficient between the rolling roll and the material to be rolled, and an advanced rate. That.

  A control device for a rolling mill according to the present invention is a control device for a rolling mill that includes a plurality of rolling stands arranged in a conveyance direction of the material to be rolled, and predicts a rolling phenomenon of the material to be rolled in the plurality of rolling stands. An optimization problem consisting of an evaluation function consisting of a rolling model and a constraint equation relating to a variable in the rolling model is constructed, and an optimization constructed in the construction step by substituting measured values into measurable variables in the rolling model It comprises an estimation means for estimating an unknown variable included in the rolling model by solving a problem, and a control means for controlling the rolling mill using the unknown variable estimated by the estimation means.

  According to the control method and control device of a rolling mill according to the present invention, even when the number of unknown rolling characteristic values is larger than the number of equations configured from the rolling model, the rolling state of the material to be rolled is accurately estimated, The rolling mill can be controlled based on the estimation result.

FIG. 1 is a schematic diagram showing a configuration of a hot finish rolling mill to which a rolling mill control method according to an embodiment of the present invention is applied. FIG. 2 is a schematic diagram for explaining the relationship between the minimum value of the evaluation function and the value of the friction coefficient. FIG. 3 is a diagram showing an example of estimating the rolling state of the material to be rolled. FIG. 4 is a diagram illustrating an example of estimating the rolling state of the material to be rolled.

  Hereinafter, with reference to drawings, the control method of the rolling mill which is one embodiment of the present invention is explained.

[Configuration of hot finishing mill]
First, the configuration of a hot finish rolling mill to which a rolling mill control method according to an embodiment of the present invention is applied will be described with reference to FIG. FIG. 1 is a schematic diagram showing a configuration of a hot finish rolling mill to which a rolling mill control method according to an embodiment of the present invention is applied.

  As shown in FIG. 1, a hot finish rolling mill 1 to which a rolling mill control method according to an embodiment of the present invention is applied includes seven rolling stands F1 to F7. The rolling position of the rolling rolls constituting each rolling stand can be controlled by the rolling devices 2a to 2g. Each rolling roll is driven to rotate by mill motors 3a to 3g, and the rotation speed can be controlled by a mill motor speed control system (not shown).

  The rolling load of the rolling roll can be measured by load detectors 4a to 4g configured by a load cell or the like. The material S to be rolled passes continuously through the rolling stands F1 to F7, and is rolled to the target plate thickness on the exit side of the final rolling stand F7 by being rolled down by each rolling stand. Between the rolling stands, loopers 5a to 5f for supporting the material to be rolled S from below and applying tension to the material to be rolled S during rolling according to a looper control system (not shown) are arranged. Further, on the exit side of the rolling stands F4 to F7, plate thickness meters 6a to 6d for measuring the plate thickness of the material S to be rolled are arranged.

[Rolling mill control method]
Next, the control method of the rolling mill which is one Embodiment of this invention is demonstrated.

  In the rolling mill control method according to an embodiment of the present invention, the rolling state of the material to be rolled S is accurately estimated even when the number of unknowns is larger than the number of equations configured from the rolling model, and based on the estimation result. The hot finish rolling mill 1 is controlled. Below, the rolling state estimation method in the present embodiment will be described in detail.

In the present embodiment, an information processing apparatus such as a process computer estimates the rolling state of the material to be rolled S using a plurality of rolling models that represent the relationship between the variables related to the rolling state of the material to be rolled S (details). (See Theory and Practice of Sheet Rolling, Japan Iron and Steel Institute, 1984). Specifically, the rolling load prediction model is expressed by the following mathematical formula (1). Here, in Equation (1), i denotes a rolling stand number (i = 1 to 7), _{P i} is the rolling load at the i-th rolling stand, _{H i} is the material to be rolled in the i-th rolling stand entry side The thickness of S (incoming side thickness), h _i is the thickness (outside thickness) of the material S to be rolled on the i-th rolling stand outlet side, V _Ri is the rolling roll speed of the i-th rolling stand, and t _bi is The rear tension of the i-th rolling stand, t _fi is the forward tension of the i-th rolling stand, K _i is the deformation resistance of the material S to be rolled in the i-th rolling stand, and μ _i is the rolling roll in the i-th rolling stand. The coefficient of friction with the material to be rolled S is shown.

The advanced rate prediction model is expressed by the following mathematical formula (2). Here, in Equation (2), f _i denotes the forward slip of the rolling material S in the i-th rolling stand. Moreover, in Formula (2), _Rdi shows the flat roll radius (effective roll radius when a roll deform | transforms with a rolling load) in an i-th rolling stand, and it is like Formula (3) shown below. expressed.

  Further, in the steady state of the hot finish rolling mill 1, since the mass flow constant law is established in each rolling stand, the following formula (4) is established. Here, in Formula (4), C represents a fixed value that does not depend on the rolling stand.

In the present embodiment, the above mathematical formulas (1) to (4) are provided as many as the number of rolling stands, the deformation resistance K _{i of} the material to be rolled S in each rolling stand, and the friction between the rolling roll and the material to be rolled S. The coefficient μ _i and the advance rate f _i are obtained as unknowns. As specific examples, the deformation resistance K _i (i = 5 to 7) of the material to be rolled S in the three rolling stands F5 to F7 shown in FIG. 1 and the friction coefficient μ _i (i between the rolling roll and the material to be rolled S). = 5-7), and shows a case of obtaining a forward slip _f i (i = 5-7) below.

In this case, the measurable variables are the actual measurement value P _mi (i = 5 to 7) of the rolling load P _i by the load detectors 4e to 4g, and the actual measurement value H _mi of the entry side plate thickness H _i by the plate thickness gauges 6a to 6c. (I = 5 to 7), measured value h _mi (i = 5 to 7) of the delivery side plate thickness h _i by the plate thickness gauges 6b to 6d, measured value V _Rmi of the rolling roll speed V _Ri by a mill motor speed control system (not shown). (i = 5 to 7), the measured value _t bmi backward tension _{t bi} by looper control system (not shown) (i = 5 to 7), and the measured value _t fmi forward tension _{t fi} by looper control system (not shown) (i = 5-7). For this reason, the equations shown in the following mathematical formulas (5) to (8) are obtained. Incidentally, the flat roll radius R _di in Equation (6) below because expressed as Equation (3) Equation below from (9) is substituted for equation (9) into equation (3).

Here, the variables to be estimated are the deformation resistance K _i (i = 5 to 7) of the material to be rolled S in the rolling stands F5 to F7, and the friction coefficient μ _i (i = 5) between the rolling roll and the material to be rolled S. ˜7) and the advanced rate f _i (i = 5 to 7) in total. For this reason, since the number of equations becomes an underdetermined system with fewer than the number of variables, the equations cannot be solved as they are. Therefore, in this embodiment, a constraint condition is given to the variable. Specifically, in this embodiment, as shown in the following formula (10), the value of the friction coefficient μ _i between the rolling roll and the material to be rolled S is that of the hot finish rolling mill 1 shown in FIG. There is a constraint that it is equal to the empirical value μ _si used for the setting calculation.

Then, using the above formulas (5) to (10), construct an optimization problem with constraints that minimizes the evaluation function J shown in the following formula (11), and solve the constructed optimization problems with constraints To determine the deformation resistance K _i (i = 5 to 7) and the advanced rate f _i (i = 5 to 7). Here, in Equation (11), α and β represent weights.

In order to solve the optimization problem with constraints, equation (10) is substituted into equations (5) and (6), and deformation resistance K _i (i = 5 to 7) that minimizes evaluation function J and What is necessary is just to obtain | _{require the} advanced rate fi (i = 5-7). Here, since the constraint condition shown in the equation (10) is eliminated by substituting it into the evaluation function J shown in the equation (11), it becomes an optimization problem without a constraint condition and can be easily solved by a known method. it can.

In addition, it is difficult to measure the friction coefficient μ _i between the rolling roll and the material to be rolled S in the hot finish rolling mill 1, and there is no accurate friction coefficient prediction model. The value of the friction coefficient μ _si includes an error. For this reason, there are few cases where the constraint condition shown in Expression (10) is strictly established, and it is considered that the estimation accuracy of other variables is improved by appropriately changing the friction coefficient value μ _si . That is, as shown in FIG. 2, it is considered that the friction coefficient value μ ₀ that minimizes the evaluation function J is not necessarily a fixed value μ _si but changes within a predetermined range (minimum value μ _simin to maximum value μ _simax ). It is done.

Therefore, to set the range as shown in Equation (12) below the value mu _si friction coefficient used in the setting calculation, the following ranges of values mu _si friction coefficients shown in equation (12) as a constraint condition The deformation resistance K _i (i = 5 to 7) and the advanced rate f _i (i = 5 to 7) are obtained by solving the optimization problem with constraints that minimizes the evaluation function J shown in Equation (13). May be. Here, in Expression (13), γ represents a weight.

  Thus, according to this embodiment, the rolling state of the material to be rolled S that cannot be directly measured even when the number of unknowns is larger than the number of equations formed from the rolling model is accurately estimated, and the estimation result The hot finish rolling mill 1 can be controlled based on the above. In particular, according to the present embodiment, the deformation resistance of the material to be rolled S and the friction coefficient between the rolling roll and the material to be rolled S, which were impossible with the prior art, can be obtained separately based on the rolling data. . For this reason, since the estimated value can be used without assuming the friction coefficient between the rolling roll and the material to be rolled S, the learning accuracy of the rolling model can be improved.

  In the present embodiment, the deformation resistance of the material to be rolled S, the friction coefficient between the rolling roll and the material to be rolled S, and the advanced rate were estimated. Can also be estimated. Thereby, the rolling precision of the to-be-rolled material S can be improved by comprising a feedforward control system or a feedback control system using the estimated plate | board thickness. Further, it is possible to estimate the tension from the rolling load or the rolling torque by using the influence of the tension of the material S to be rolled between the rolling stands on the rolling load or the rolling torque, and to perform the tension control based on the estimated tension. .

In this example, the deformation resistance of the material S to be rolled in the rolling stands F5 to F7 using time series data of the inlet side plate thickness H _i , the outlet side plate thickness h _i and the rolling roll speed V _{Ri in} the rolling stands F5 to F7. K _i (i = 5 to 7), friction coefficient μ _i (i = 5 to 7) between the rolling roll and the material to be rolled S, and advanced rate f _i (i = 5 to 7) were estimated. The estimation results in the rolling stand F5 are shown in FIGS. In this example, for the sake of simplicity, the influence of the rear tension and the front tension on the rolling load was ignored.

FIG. 3 shows the estimation result of the rolling state of the material to be rolled S using the above formula (11), and the friction coefficient μ _si between the rolling roll and the material to be rolled S is a value used for setting calculation. 22 is fixed. FIG. 4 shows the estimation result of the rolling state of the material to be rolled S using Equation (13), and the friction coefficient μ _si between the rolling roll and the material to be rolled S slightly fluctuates, and the average value is calculated by setting. It is slightly larger than the used value of 0.22.

The variation of the deformation resistance K _i and advanced rate f _i of the rolled material S shown in FIG. 4 is slightly larger than in the case shown in FIG. However, in any case, the deformation resistance K _i of the material S to be rolled shows a correlation with the plate thickness and the rolling load, and it is shown that the change in the deformation resistance K _{i in} the longitudinal direction causes the variation of the plate thickness and the load. Has been. Also, forward slip f ₇ in the rolling stand F7, it was confirmed that good agreement when compared with the forward slip of the rolling stand F7 obtained from the peripheral speed of the pinch rolls of the coiler.

  In this embodiment, the constraint that the friction coefficient between the rolling roll and the material S to be rolled is equal to the value used for the setting calculation is imposed, but instead, the friction coefficient obtained from the following formula (14). May be used as a constraint. Here, in Equation (14), a and b are variables. In Equation (14), the friction coefficient is assumed to change linearly in the hot finish rolling mill, and instead of obtaining the friction coefficient of each rolling stand, the linear change gradient a and intercept b are obtained.

  The embodiment to which the invention made by the present inventors is applied has been described above, but the present invention is not limited by the description and the drawings that constitute a part of the disclosure of the present invention. That is, other embodiments, examples, operational techniques, and the like made by those skilled in the art based on this embodiment are all included in the scope of the present invention.

DESCRIPTION OF SYMBOLS 1 Hot finishing rolling mill 2a-2g Reduction device 3a-3g Mill motor 4a-4g Load detector 5a-5f Looper 6a-6d Thickness gauge F1-F7 Rolling stand

Claims (3)

A control method for a rolling mill comprising a plurality of rolling stands arranged in the conveying direction of a material to be rolled,
A construction step for constructing an optimization problem consisting of an evaluation function comprising a rolling model for predicting a rolling phenomenon of a material to be rolled in a plurality of rolling stands and a constraint equation relating to a variable in the rolling model;
An estimation step for estimating an unknown variable included in the rolling model by substituting a measured value into a measurable variable in the rolling model and solving the optimization problem constructed in the construction step;
A control step of controlling the rolling mill using the unknown variable estimated in the estimation step;
A control method for a rolling mill, comprising: The construction step is an optimization problem consisting of an evaluation function comprising a rolling model for predicting rolling load and advance rate in a plurality of rolling stands and a constraint equation regarding a friction coefficient between a rolling roll and a material to be rolled in each rolling stand. Including the steps of building
The estimation step inputs at least the measured value of the material to be rolled on the entry side and the exit side of each rolling stand and the measured value of the rolling roll speed of each rolling stand into the rolling model to input the rolling load and the advance rate. By solving these optimization values and at least the measurement value of the rolling load, the deformation resistance of the material to be rolled in each rolling stand, the difference between the rolling roll and the material to be rolled is obtained. The rolling mill control method according to claim 1, further comprising a step of estimating a friction coefficient and an advanced rate. A control device for a rolling mill comprising a plurality of rolling stands arranged in the conveying direction of the material to be rolled,
Establish an optimization problem consisting of an evaluation function consisting of a rolling model that predicts the rolling phenomenon of the material to be rolled in multiple rolling stands and a constraint equation related to variables in the rolling model, and measure the measurable variables in the rolling model By estimating the unknown variable included in the rolling model by substituting a value and solving the optimization problem constructed in the construction step,
Control means for controlling the rolling mill using the unknown variable estimated by the estimation means;
A control device for a rolling mill, comprising:

Images