KR100425601B1 - Real-time simulator of shape control system for rolling mills - Google Patents

Abstract

The present invention relates to an apparatus capable of simulating the shape control characteristics of a rolling mill. The present invention is configured by connecting a PLC (Programmable Logic Controller) and a PC (Personal Computer), and can perform a real-time simulation of the performance of the shape controller, change in exit shape for roll gap change, and removal performance for various shape disturbances. Relates to a device.

Description

Real-time simulator of shape control system for rolling mills}

The present invention relates to an apparatus capable of simulating the shape control characteristics of a rolling mill, and more particularly, to an apparatus capable of simulating a change in exit shape for a roll gap change and a removal performance for various shape disturbances in real time. In other words, the device of the present invention was developed for the purpose of reducing the risk through online test before applying the control parameters to the actual shape control system, developing new shape control logic, and rolling operation and maintenance education.

So far, shape-related simulators have been steady-state static rolling analysis programs based on the theory of tack. In these simulators, the elastic deformation of the roll at steady state, the pressure distribution between the roll and the strip, the pressure distribution between the roll and the roll, and the exit thickness profile can be simulated. Such a simulator can be used to determine the specifications of the mill according to the rolling conditions, and to calculate the exit thickness profile according to the rolling conditions.

However, in such a simulator, there was a problem that simulation over time, such as performance review of the shape controller, change in exit shape for roll gap change, and removal performance for various shape disturbances, was impossible.

An object of the present invention is to provide a device capable of real-time simulation of the performance of the shape controller, the change in exit shape for the roll gap change, and the removal performance for various shape disturbances by connecting the PLC and the PC.

1 is a diagram schematically showing a shape measuring apparatus.

Fig. 2 is a diagram showing types of shape defects and stress distributions according to them.

3 is a view schematically showing a continuous rolling mill.

4 is a diagram schematically showing a configuration of a shape actuator.

5 is a flowchart showing a simulator program structure in a PC.

6 is a flowchart illustrating a mill shape model.

7 is a diagram illustrating shape control logic.

8 is a diagram illustrating a + skewing (asymmetrical reduction) direction.

9 is a diagram illustrating a + bending direction.

10 is a graph showing a shape change by + skewing.

It is a graph which shows the shape change by + work roll bending.

It is a graph which shows the shape change by + intermediate roll bending.

FIG. 13 is a graph showing a change in shape of region # 14 when cooling is turned on. FIG.

14 is a graph showing the evaluation parameters according to the skew control gain variation.

15 is a graph showing evaluation parameters according to work roll bending control gain variations.

16 is a graph showing the evaluation parameters according to the variation of the mid-roll bending control gain.

17 is a graph showing a shape error before adjusting the skew control gain.

18 is a graph showing a shape error after adjusting the skew control gain.

<Explanation of symbols for the main parts of the drawings>

1: shape measuring instrument 2: shape measuring signal correction unit

3: target shape setting unit 4: control coefficient calculating unit

5: Skewing Proportional-Integral Control 6: Work Roll Bending Proportional-Integral Control

7: middle roll bending proportional-integral control

8: mechanical shape control model calculation unit

9 cooling control 10 shape error (F (i))

11: control coefficient (C) 12: cooling error

13: Shape actuator input signal

According to the present invention for achieving the above object, the mill-shaped model unit for outputting the width direction shape distribution to the shape actuator input constituting the rolling mill; A shape controller for calculating the shape actuator input with a shape distribution value calculated at the mill shape model unit through at least one shape control variable; A control signal manipulation unit which inputs the shape control variable to the shape controller; And a communication unit which transmits an output from the mill shape model unit and a signal input from the control signal manipulation unit to the shape controller, and enables the shape control unit to exchange data with a central processing unit (CPU) mounted on a computer. Provided is a shape control simulation apparatus comprising: a.

In the following, a shape control simulation apparatus according to the present invention will be described in detail.

1. Definition of shape

In rolled products, the terms profile and shape are widely used. Profile refers to the thickness distribution in the strip width direction. If the thickness distribution is not constant in the width direction, it may cause product defects. Assuming that there is no change in the strip width direction length during rolling, the thickness reduction rate distribution in the width direction is represented by the elongation distribution. In addition, when the strip is tensioned, the elongation distribution appears as a stress distribution. This width direction stress distribution is defined as a shape. Stress, which is a measurement unit of shape, is widely used in terms of I unit, which is an elongation unit, by Equation (1).

The shape measuring device has a non-contact type and a contact type, and FIG. 1 shows a contact shape measuring device. Geometry is an important parameter that affects the quality of the final product and the rolling process.

E is a Young's ratio, Δσ _i is a stress deviation in the i-th region, and ΔI _i is an elongation deviation in the i-th region, respectively.

Shape defects occur when the profile of the entry strip and the roll gap do not coincide, as in FIG. FIG. 2 (a) shows the relationship between the roll and the side material when asymmetrical edge wave shape defect occurs, the stress distribution measured by the shape measuring instrument, and the shape defect of the plate. In FIG. 2B, the relationship between the roll and the side material when the center wave shape defect occurs, the stress distribution measured by the shape measuring instrument at this time, and the shape defect of the plate are shown. Therefore, it is necessary to control the shape in order to adjust the roll gap profile to produce a flat plate which does not wave in the longitudinal direction.

2. Continuous mill equipment structure

The continuous rolling mill of this cold rolling mill is a continuous rolling mill which consists of five stands, Comprising: 4Hi mill is a # 1-# 4 stand, and 6Hi mill is a # 5 stand. Shape control is performed only in the last five stands, the configuration of which is shown in FIG.

The shape measurement is performed by ABB's contact shape measuring device, which measures the tension distribution in the width direction of the strip, a strip scanner that can measure the rolling direction of the strip, and the MP200 PLC system for signal processing and shape control. It consists of.

Shape correction corrects asymmetrical shape through skew adjustment of left and right of backup roll, and uses symmetrical shape by using upper and lower intermediate roll and work roll bender. I am correcting. And, the local residual shape error that cannot be corrected through mechanical shape control is controlled by local roll cooling by spot cooling to the work roll (see FIG. 4).

3. Configuration of the shape simulator

The simulator is composed of a mill shape model that outputs the width direction shape distribution to the shape actuator input, and a controller that calculates the shape actuator input using the shape distribution values calculated from the mill shape model.

The shape controller is ABB's MP200 PLC, which is used in the real industry. The mill shape model was programmed with LabVIEW, a graphical program language, and the PC model used a Pentium II 300MHz Dual CPU and 256MB of memory.

5 and 6 show a program structure implemented in the shape simulator and the PC. If you look at the structure of the program, it can be divided into Mill Model, Virtual Keyboard, and Communication. First, the wheat model part consists of the information of the mill, shape disturbance setting, shape calculation, storage and loading of simulation results, and report preparation. Secondly, the virtual keyboard is a part that can operate the MP200PLC. It is possible to turn on / off each shape control, turn on / off each nozzle of the coolant, adjust the screen of the display unit provided by the MP200, and various parameters in the MP200 PLC. It is possible to adjust at. Finally, the communication part is a UDP communication program that can send and receive data from the mill model and virtual keyboard and the data from the MP200 PLC.

4. Wheat shape model

The shape change in the shape control system is determined by the mismatch between the entry strip profile and the roll gap profile. In order to obtain the shape output value analytically, the roll elastic deformation analysis of the 6Hi stand is required, and the side profile and tension distribution of the strip should be measured. However, although the # 5 stand entry profile and tension distribution are not measurable at present and the simulation method is capable of steady-state analysis, it is suitable to evaluate the dynamic behavior of the shape according to the control input and performance evaluation of the control system. not. Therefore, in order to implement this, a skew (asymmetrical reduction) and a bending model [where the exit shape change for the unit shape control input (Skew, Bender) is expressed as a function of the strip width direction] are obtained through actual tests, and the entrance strip profile and the roll gap are obtained. Shape changes caused by inconsistencies with the profile are treated as disturbances to form a mill shape model.

That is, the mill shape model implemented in the PC is a shape disturbance model that generates a shape disturbance as shown in FIG. 6, a shape actuator having the output of the MP200 controller for controlling the shape, and a skewing and bending showing the shape change by the output of the shape actuator. A model and a cooling model which show the shape change by cooling.

4.1. Shape actuator

Skew, WR (Work Roll) vendors, and IMR (Intermediate Roll) vendors modeled the first order delay functions as shown in Equation 2.

In addition, the above continuous system function was discretized and programmed using the Trapezoid rule (a formula used to change the continuous system transfer function to a discrete system transfer function).

Where T is the actuator's time constant, t _p is the sampling time, r (k) is the current input value, r (k-1) is the input value one step back, y (k) is the current output value, and y (k-1 Denotes the output value one step before.

4.2. Feed delay function

There is a time delay between the output of the shape actuator and the output of the shape measuring machine as shown in Equation 4, and the delay time is determined by the rolling speed.

Where y (k) is a new output value, r (k-n) is an n step delayed input value, and n is a delay integer, respectively.

4.3. Skewing, bend model

The shape output relation function according to the input of Skew, WR, and IMR bender applies step input to each actuator of the real rolling mill, and calculates the shape change values before and after the curve fitting method. Skewing, bending model as shown in Equation (5) is obtained through mathematical method represented by exponential function or multidimensional function. The actual test results are shown in Section 6.1.

Here, f _{SK, WB, and IB} are skewing, WR bending, IMR bending models, and x represents a normalized strip width.

4.4. Cooling model

Spot coolant for local cooling of the work roll can see the local shape change in the corresponding area of the rolling mill exit shape meter with each nozzle on / off. The cooling model is shown in equation (6).

Where N is the stress distribution by cooling, I is the unit vector, Y _i (k) is the stress of the region, Amp is the magnitude of the stress, damp is the damping ratio, and t _p is the sampling time.

4.5. Shape disturbance model

Shape disturbances can be made by varying the shape of each x term in time to sinusoidal wave shape, step shape, ramp shape as shown in equation (7).

5. Shape control logic

The plate flatness signal measured by the shape measuring device 1 performs the calculation of the deviation value of the signal measured in each area of the shape measuring device in the shape measuring part 2 and corrects it according to the measurement conditions, thereby setting the target shape setting part 3. Calculate the shape error, which is the difference from the target flatness set in. The shape error F (i) 10 may be expressed as a function of each shape control actuator model as shown in Equation (8). Here, the shape control actuator model is a value obtained by experimenting with a function measured by a shape measuring instrument when symmetrical reduction is applied to the work roll bender, the intermediate roll bender, and asymmetric reduction is applied to the backup roll.

Where F (i) is the flatness error value in the i-th region, Ψ ₁ (i) is the asymmetric reduction control model, Ψ ₂ (i) is the work roll bender control model, Ψ ₃ (i) is the intermediate roll bender control The model, i, represents the number of divided regions in the shape measuring instrument, respectively.

The problem solving method of shape error control in Equation 8 is to find the coefficient value C of each control model. The least-squares method is used as a mathematical method to find the coefficients. Equation 8 is expressed as a determinant.

F = A C

In Equation 9, the coefficient value C of each control model can be obtained by matrix calculation as follows.

The coefficient values C1, C2, and C3 of each control model obtained by the least-squares calculation unit 4 enter the inputs of the proportional-integral control units 5, 6, and 7 respectively, and proportional-integral control such that C1, C2, and C3 converge to zero. To give the input value 13 of each shape actuator.

Here, the meaning of the coefficient value C (11) of the control model is as follows. A large value of C means a large difference between the target shape and the measured shape, and a small value of C means that the target shape and the measured shape are similar. This is calculated by the mechanical shape control model calculation unit 8. The error 12 which cannot be separated as a work roll bender model, an intermediate roll bender model, and an asymmetrical reduction model is obtained as shown in Equation 11, and in order to compensate for the error value, the error of each area is determined by the cooling control calculation unit 9. Determine on / off to be zero.

6. Wheat Shape Characterization

6.1. Skew, shape influenced by vendor

In order to obtain the skewing, WR bending, and IMR bending models, the change of the shape output by the input of the shape actuator was examined through the actual test on the stand # 5 of the continuous rolling mill. Mill specifications of the # 5 stand are shown in Table 1 and the test was performed as follows.

division BUR IMR WR Early crown 5/100 48/100 Flat roll Barrel length 1425 1425 1425 Barrel diameter 1360/1285 490/450 390/350 Distance between journals 2340 2360.15 2160.05 Journal diameter - 304.9 220.66

When steady rolling speed, that is, change of shape output and change of shape control value were small, step input was applied to each actuator, and each model was derived by using curve fitting method.

8 and 9 show the direction of the shape actuator. In skew, (+) means the work side (WS) side is smaller than the gap on the drive side (DS) side, and (-) sign means the opposite. In flex, the plus sign means opening both journals on the roll, and the minus sign means the opposite.

FIG. 10 shows the change in shape in I units when a Skew step input of +0.03 mm was given for a strip of width 908.50 mm and exit thickness 0.239 mm. If the roll gap of the WS is 0.03 mm smaller than the DS of the roll, the I unit of the WS of the strip is 2.4 I units larger than the DS of the roll. The skew model is expressed as Equation 12 when the shape change value thus obtained is expressed as a shape change value of 0.01 mm, and the unit is expressed as a stress unit applied to an actual program using Equation 1. SK in FIG. 10 means Skewng.

11 and 12 show the shape change in I units when + 10% WR bending step input and + 15% IMR bending step input were given for strips of width 967.50 mm and exit thickness 0.203 mm, respectively. In the case of the WR bend, the center of the strip is about 16 I unit relatively larger than the edge part, and in the case of IMR bend, the center part is about 14 I unit larger than the edge part. The WR and IMR bending models are the same as in Equation 13 and Equation 14. 11 and 12, WB means Work roll Bending and IB means Intermediate roll Bending.

6.2. Shape influence due to cooling

FIG. 13 shows that for a coil having a strip thickness of 0.254 mm, a specific area of coolant is manually opened for a period of time during a rolling operation, measured in the area of the corresponding stressometer (region 14 of the shaper, see FIG. 1). Shape values are expressed in I units. The results of the function fitting of the measured shape values as an exponential function are shown in equations (15) and (16) for the case of cooling on / off. In this equation, the shape change is rapidly changed to about 5 I unit within 5 seconds, after which the change is small, especially when the response is faster than off when cooling on.

Therefore, the above two equations can be applied to the cooling model corresponding to each area in the width direction of the strip 0.254 mm thick.

7. Simulation results and application

7.1. Shape disturbance elimination performance by control gain adjustment

In this section, we simulate the stepped disturbance used in the previous section to compare the control performance according to the adjustment of the shape control gain. Simulation conditions are as follows.

That is, the shape actuator time constant should be 200 msec, the simulation period 50 msec, the strip width and thickness should be 967.5 mm, 0.203 mm, and the target shape should be flat.

First, in order to examine the skew control performance, simulation was performed while increasing the control gain when asymmetric disturbance of the x component was applied. As a result, the disturbance removal response improves as the gain increases, but it can be seen that overshoot occurs. In the shape control system of the actual continuous rolling mill, the skew control gain is set to 0.15, but by increasing it to 0.3, it can be seen from the simulation result of FIG. 14 that the asymmetric disturbance can be quickly removed from about 25 seconds to 10 seconds.

To investigate the control performance of WR and IMR flexures, simulations were performed with increasing control gains when symmetric disturbances of x2 components were applied. As a result, the gains of the WR and IMR vendors currently applied to the actual system are judged to be appropriate, and if the WR vendor gain is 0.5 or more and the IMR vendor gain is 0.3 or more, the system may be unstable. The result was unknown. 15 and 16 show evaluation parameters of WR bending and IMR bending according to control gain variations, respectively.

7.2. Application case

Based on the simulation results in the previous section, by adjusting the skew gain for a coil with a width of 793.5 mm and an exit thickness of 0.324 mm, the removal time of the asymmetric disturbance component appearing at the beginning of the coil can be reduced.

FIG. 17 shows that the asymmetric disturbance removal time is 12 seconds when 0.15 before the skew gain adjustment, and FIG. 18 shows that the disturbance removal time is reduced to 8 seconds after the gain adjustment.

According to the simulation apparatus according to the present invention, it is possible to simulate the performance of the shape controller, the change in the exit shape for the roll gap change, the removal performance for various shape disturbances, and the like according to the time change.

Claims (4)

A mill shape model unit configured to output a width direction shape distribution to a shape actuator input constituting the rolling mill; A shape controller for calculating the shape actuator input with a shape distribution value calculated at the mill shape model unit through at least one shape control variable; A control signal manipulation unit which inputs the shape control variable to the shape controller; And A communication unit which transmits the output from the mill shape model unit and the signal input from the control signal manipulation unit to the shape controller, and enables data exchange with the central processing unit (CPU) mounted on a computer. Shape control simulation apparatus characterized by including. The method of claim 1, The mill shape model unit, At least one shape actuator receiving the output of the shape controller to perform a predetermined control action; Delay means for inputting an output of the at least one shape actuator and a delay calculation result of the rolling speed of the rolling mill; At least one control model receiving the output of the delay means to generate a control model corresponding to the shape control variable; And And a deviation calculator configured to receive the output of the control model and calculate a deviation of each result to output a desired shape. The method of claim 2, The at least one control model is a shape control simulation device, characterized in that obtained through the test of the actual rolling mill. The method of claim 3, wherein The at least one control model comprises asymmetric skewing and at least one bending model generated in a plurality of rolls constituting the rolling mill, The asymmetrical rolling and bending model, For the asymmetrical reduction (SK), the work roll bending (WB) and the middle roll bending (IB), the shape change values before and after inputting the shape control variable to the shape actuator are obtained through a curve fitting method as shown below. Shape control simulation apparatus, characterized in that the loss. (Where x represents the normalized strip width of the rolled product.)