Patent Application
20120101625
The present invention relates to a control method for the meniscus of a continuous casting mold,
A control method of this kind is known, for example, from U.S. Pat. No. 5,921,313 A. The known control method only has one single oscillating compensator. In this case, the sum of the interference frequency components is identical to the sole interference frequency component determined.
The various embodiments disclosed herein also relates to a computer program, which comprises a machine code, which can be implemented directly by a control device for a continuous casting machine and the execution of which by the control device causes the control device to control the meniscus of a continuous casting mold of the continuous casting machine according to a control method of this kind.
The various embodiments disclosed herein also relates to a control device for a continuous casting machine, which is embodied in such a way that, in operation, it executes a control method of this kind.
Finally, the various embodiments disclosed herein relates to a continuous casting machine, which is controlled by a control device of this kind.
During continuous casting, the cast strand is withdrawn from the continuous casting mold while the core of the strand is still liquid. When the strand has emerged from the continuous casting mold, the strand is guided and supported over roll pairs to support the strand shell against the metallostatic pressure of the core. The support prevents inter alia bulging of the cast strand on the broad side of the strand. The spacing of the rolls, which support the strand at the same point on both sides, must correspond to the desired strand thickness.
After emerging from the continuous casting mold, the cast strand is actively and/or passively cooled. The cooling causes the strand thickness to shrink. For this reason, the rolls supporting the cast strand at the same point on both sides must have the correct spacing from each other. Until complete solidification, also known as the crater end, the cast strand has not completely solidified. Therefore, it has a liquid core. Therefore, uneven impacts on the strand as it passes through the roll pairs exert an effect on the meniscus. However, for various reasons, for example due to the risk of casting powder being drawn into the surface of the strand, meniscus level fluctuations should be avoided where possible.
Fluctuations in the shell thickness that develop in the continuous casting mold can result in the occurrence of so-called “unsteady bulging” when passing through the roll pairs. The “bulging” is caused when a point with impaired shell thickness passes through different roll pairs one after the other and the meniscus therefore undergoes cyclical changes. Since, when viewed in the direction of transport of the strand, the roll pairs generally have constant spacing from one another and the withdrawal speed at which the strand is withdrawn from the continuous casting mold is constant, “unsteady bulging” results in periodic changes in the meniscus level. Consequently, oscillations with a constant frequency form in the meniscus.
The control method known from U.S. Pat. No. 5,921,313 A has the object of overcoming meniscus fluctuations of this kind. The known control method already works very well. In particular, it enables the meniscus to be regulated precisely to a few millimeters.
From the specialist article “Suppression of Periodic Disturbances in Continuous Casting using an Internal Model Predictor” by C. Furtmueller and E. Gruenbacher, IEEE International Conference on Control Applications, Munich, Germany, Oct. 4-6, 2006, pp. 1764 to 1769, a control method is known for the meniscus of a continuous casting mold in which the inflow of liquid metal into the continuous casting mold is set by means of a closure device and the partially solidified metal strand is withdrawn from the continuous casting mold by means of a withdrawal device. A measured actual value of the meniscus is fed to a meniscus controller, which determines a target position for the closure device on the basis of the actual value and a corresponding target value. The motor currents from drives of the withdrawal device are subjected to a frequency analysis. The components of a fundamental frequency and its harmonic frequencies are used to determine a disturbance variable compensation value, which is connected to the output signal of the meniscus controller. The closure device is controlled according to the output signal of the meniscus controller corrected in this manner.
According to various embodiments, opportunities for achieving even more precise control can be provided.
According to an embodiment, in a control method for the meniscus of a continuous casting mold,—the inflow of liquid metal into the continuous casting mold is set by means of a closure device and the partially solidified metal strand is withdrawn from the continuous casting mold by means of a withdrawal device,—a measured actual value of the meniscus is fed to a meniscus controller, which determines a target position for the closure device on the basis of the actual value and a corresponding target value,—the measured actual value of the meniscus is fed to a disturbance variable compensator,—the target position for the closure device, a target position for the closure device corrected by a disturbance variable compensation value, an actual position of the closure device or an actual position of the closure device corrected by the disturbance variable compensation value are further fed to the disturbance variable compensator,—the disturbance variable compensator determines the disturbance variable compensation value on the basis of values fed to it,—the target position corrected by the disturbance variable compensation value is fed to the closure device,—wherein the disturbance variable compensator comprises a model of the continuous casting mold, by means of which the disturbance variable compensator determines an expected value for the meniscus on the basis of a model input value,—wherein the disturbance variable compensator comprises a number of oscillating compensators, by means of which the disturbance variable compensator determines an interference frequency component on the basis of a difference between the actual value and the expected value each relative to a related interference frequency,—wherein the sum of the interference frequency components corresponds to the disturbance variable compensation value,—wherein the model input value is determined by the relationship i=p′+z′ wherein p′ is the uncorrected target or actual position of the closure device and z′ is a jump compensation value,—and wherein the disturbance variable compensator comprises a jump determiner, by means of which the disturbance variable compensator determines the jump compensation value by integrating the difference between actual value and expected value.
According to a further embodiment,—the model of the continuous casting mold consists of a series connection of a model integrator with a model delay element, each oscillating compensator consists of a series connection of two oscillating integrators and the jump determiner consists of an individual jump integrator,—as the respective input value
According to a further embodiment, the adaptation factors can be determined in such a way that the poles of the transmission function determined by the model of the continuous casting mold fulfill the following conditions:—for each interference frequency, a pair of conjugate complex poles is formed, whose real parts are smaller than zero and whose imaginary parts are equal to an angular interference frequency defined by the respective interference frequency,—three real poles are formed, which are all smaller than zero. According to a further embodiment, the adaptation factors can be determined in such a way that the real parts of the conjugate-complex poles, relative to the respective angular interference frequency, are between −0.3 and −0.1. According to a further embodiment, the adaptation factors can be determined in such a way that the real poles are all smaller than −2.0. According to a further embodiment, the adaptation factors can be determined in such a way that the real poles differ from one another in pairs. According to a further embodiment, the adaptation factors can be determined in such a way that one of the real poles is between −2.5 and −3.5, one is between −3.5 and −4.5 and one is between −4.5 and −5.5.
According to a further embodiment, the number of oscillating compensators can be greater than one. According to a further embodiment, the target position for the closure device or the target position for the closure device corrected by the disturbance variable compensation value can be fed to the disturbance variable compensator, but not the actual position of the closure device or the actual position of the closure device corrected by the disturbance variable compensation value.
According to another embodiment, a computer program may comprise a machine code that can be executed directly by a control device for a continuous casting machine and the execution of which by the control device causes the control device to control the meniscus of a continuous casting mold of the continuous casting machine according to a control method as described above.
According to a further embodiment of the computer program, the program can be stored on a data medium in machine-readable form. According to a further embodiment of the computer program, the data medium can be a component of the control device.
According to another embodiment, a control device for a continuous casting machine can be embodied in such a way that, in operation, it executes a control method as described above.
According to yet another embodiment, a continuous casting machine can be controlled by a control device as described above.
Further advantages and details are disclosed in the following description and exemplary embodiments in conjunction with the drawings, which show:
According to various embodiments, a control method of the type mentioned in the introduction can be provided in such a way
i=p′+z′
In an embodiment, it is provided
The different adaptation factors can be determined as required. In experiments, good results can be achieved if the adaptation factors are determined in such a way that the poles of the transmission function determined by the model of the continuous casting mold fulfill the following conditions:
In an embodiment, it is also provided that the adaptation factors are determined in such a way that the real parts of the conjugate-complex poles, relative to the respective angular interference frequency, are between −0.3 and −0.1. In particular a value of about −0.2 is desirable. Good damping properties were achieved with values of this kind in experiments.
Preferably, the adaptation factors are determined in such a way that the real poles are all smaller than −2.0. In this case, the control method still works reliably and stably even if the model of the continuous casting mold is only a very imprecise model of the real continuous casting mold.
Particularly good results can also be achieved if the adaptation factors are determined in such a way that the real poles differ from one another in pairs.
Obviously, the last two measures named (real poles smaller than −2.0 and differing from one another in pairs) can be combined with each other. Optimum results are achieved when the real poles are −3.0, −4.0 and −5.0, in each case +/−0.5.
The number of oscillating compensators is preferably greater than one. This makes it possible to compensate for more than one “bulging-oscillation”.
It is also preferable for the target position for the closure device or the target position for the closure device corrected by the disturbance variable compensation value to be fed to the disturbance variable compensator, but not the actual position of the closure device or the actual position of the closure device corrected by the disturbance variable compensation value. This produces better results.
According to further embodiments, a computer program of the type mentioned in the introduction can be provided such that when executed causes the control device to control the meniscus of the continuous casting mold according to a control method according to various embodiments. The computer program can, for example, be stored on a data medium in machine-readable form. The data medium can in particular be a component of the control device.
According to further embodiments, a control device for a continuous casting machine can be embodied in such a way that, in operation, it executes a control method according to various embodiments. Finally, according to yet further embodiments a continuous casting machine can be controlled by a control device according to various embodiments.
According to
The liquid metal 3 in the continuous casting mold 1 is cooled by means of cooling devices so that a strand shell 5 is formed. However, the core 6 of the metal strand 7 is still liquid. It only solidifies later. The cooling devices are not shown in
The meniscus 9 of the liquid metal 3 in the continuous casting mold 1 should be kept as constant as possible. A withdrawal speed v, at which the partially solidified metal strand 7 is withdrawn from the continuous casting mold 1, is generally constant. Therefore—both in the prior art and in the various embodiments—the position of the closure device 4 is tracked in order to set the inflow of the liquid metal 3 in the continuous casting mold 1 in such a way that the meniscus 9 is kept as constant as possible.
An actual value hG of the meniscus 9 is acquired by means of a corresponding measuring device 10 (known per se). The actual value hG is fed to a control device 11 for the continuous casting machine. The control device 11 uses a control method, which will be explained in more detail below, to determine a target position p* to be adopted by the closure device 4. The closure device 4 is then controlled accordingly by the control device 11. Generally, the control device 11 issues a corresponding control signal to an adjusting device 12 for the closure device 4. The adjusting device 12 can, for example, be a hydraulic cylinder unit.
Generally, a corresponding measuring device 13 (known per se) determines an actual position p of the closure device 4 and feeds it to the control device 11. Therefore, there is usually closed loop control of the closure position. Alternatively, open loop control would also be possible.
The control device 11 is embodied in such a way that, in operation, it executes a control method according to various embodiments. Generally, the mode of operation of the control device 11 is determined by a computer program 14 with which the control device 11 is programmed. To this end, the computer program 14 is stored inside the control device 11 in a data medium 15, for example a flash EPROM. Obviously, it is stored in machine-readable form.
The computer program 14 can be fed to the control device 11 via a mobile data medium 16, for example a USB memory stick (shown) or an SD storage card (not shown). Obviously, the computer program 14 is also stored in machine-readable form on the mobile data medium 16. Alternatively, it is possible for the computer program 14 to be fed to the control device 11 via a computer network link or a programming unit.
The computer program 14 comprises a machine code 17 that can be executed directly by the control device 11. The execution of the machine code 17 by the control device 11 causes the control device 11 to control the meniscus 9 of the continuous casting mold 1 according to a control method according to various embodiments. This control method is explained in more detail in the following in conjunction with
According to
The target position p* for the closure device 4 is fed to the closure device 4. However, prior to this, the target position p* is corrected by a disturbance variable compensation value z.
As already mentioned, the setting of the closure device 4 is controlled by closed loop control. In this case, which is depicted in
p*−z
is fed to a position controller 19, to which, in addition, the actual position p of the closure device 4 is also fed. The position controller 19 can, for example, be embodied as a P controller.
Due to the inflow of the liquid metal 3 set thereby, the actual position p of the closure device 4 acts on the meniscus 9 itself. The actual value hG of the meniscus 9 is acquired and, as already mentioned, fed to the meniscus controller 18.
The continuous casting mold 1 can be exposed to disturbance variables which influence the meniscus 9. A disturbance variable compensator 20 is provided to compensate the disturbance variables. The measured actual value hG of the meniscus 9 and a further variable are fed to the disturbance variable compensator 20.
According to
The determination of the disturbance variable compensation value z using (inter alia) the corrected or uncorrected target position p*−z or p* of the closure device 4 may be preferred for the purposes of the various embodiments. Alternatively, the actual position p or the actual position p−z of the closure device 4 corrected by the disturbance variable compensation value z can be fed to the disturbance variable compensator 20. These alternatives are also shown by dashed lines in
The structure and mode of operation of the disturbance variable compensator 20 are explained in more detail in the following in conjunction with
According to
I=p′+z′
is fed to the model 21. In the above relationship, p′ is the uncorrected target position p* of the closure device 4, that is the output signal from the meniscus controller 18. If the actual position p of the closure device 4 were fed to the disturbance variable compensator 20 instead of the target position p*, in the above relationship, the value p would have to be used instead of the value p*. z′ is a jump compensation value.
The jump compensation value z′ is determined by the disturbance variable compensator 20 by means of a jump determiner 22, which is also a component of the disturbance variable compensator 20. According to
According to
The minimum number of oscillating compensators 23 is one. In this case, only one single frequency disturbance proportion zS is compensated. Alternatively, the number of oscillating compensators 23 can be greater than one. In this case, the corresponding interference frequency component zS is determined for each oscillating compensator 23 each with its own interference frequency fS.
The output signals zS from the oscillating compensators 23 are summated in a nodal point 24, the result of which corresponds to the disturbance variable compensation value z. In the case of only one single oscillation compensator 23, obviously no summation is necessary, since, in this case, the sum total is identical to the single summand.
In an embodiment of the disturbance variable compensator 20—see FIG. 4—the model 21 of the continuous casting mold 1 consists of an integrator 25 and a time-delay element 26, which, according to the depiction in
The model integrator 25 comprises an integration time constant T1, the model delay element 26 a delay time constant T2. The time constants T1, T2 are determined in such a way that they describe the real continuous casting mold 1 as realistically as possible.
A value
m=V·i+h1·e
is fed to the model integrator 25 as an input signal m. V is an amplification factor. i is the model input value already mentioned. e is the difference which has also already been mentioned. h1 is an adaptation factor.
The model integrator 25 supplies an output signal I. The output signal I is corrected in a nodal point 27 by a value
h2·e
and then fed to the model delay element 27 as its input signal. h2 is a further adaptation factor.
The variables I and h2·e fed to the nodal point 27 are summated in the nodal point. This results from the fact that the two input signals I, h2·e of the nodal point 27 are not provided with minus signs on the input side of the nodal point 27.
The adaptation factors h1 and h2 are related to the model 21 of the continuous casting mold 1. Therefore, in the following, they are referred to as model adaptation factors h1, h2.
The oscillating compensators 23 essentially have the same structure. Therefore, in the following only one of the oscillating compensators 23 will be described in detail, namely the upper oscillating compensator 23 shown in
According to
The two integrators 28, 29 are described in the following as oscillating integrators 28, 29 since they are components of the corresponding oscillation compensator 23. The supplement “oscillating” serves solely to indicate the association of these two integrators 28, 29 to the respective oscillating compensator 23. No further significance is attached to the supplement “oscillating”.
The oscillating integrators 28, 29 have an integration time constant a. The integration time constant a amounts to
fS is the respective interference frequency to be compensated. The interference frequency fS must be known in advance.
According to
s1=h3·e−S2
is fed to the front oscillation generator 28 as input value s1. The value
s2=h4·e+S1
is fed to the back oscillation generator 29 as input value s2. S1 and S2 are the output signals of the front and of the back oscillation generator 28, 29. h3 and h4 are adaptation factors. Due to their association with the respective oscillating compensator 23, they are referred to in the following as oscillation adaptation factors h3, h4.
The jump determiner 22 consists of a single integrator 30, which due to its association with jump determiner 22, is referred to in the following as a jump integrator 30. It is fed a value
s3=h5·e,
wherein h5 is an adaptation factor, in the following referred to as a jump adaptation factor.
As already mentioned, there can be a plurality of oscillating compensators 23. In this case, the oscillation adaptation factors h3, h4 of the individual oscillating compensators 23 are independent of each other. In addition, the integration time constants a of all the oscillating compensators 23 are different from one another.
To determine the adaptation factors h1 to h5, that is the model adaptation factors h1, h2, of the jump adaptation factor h5 and for each oscillating compensator 23 of the two respective oscillation adaptation factors h3, h4, preferably the transmission function of the system shown in
Now, the desired zero settings are specified for the denominator polynomial, that is the desired poles of the transmission function. This produces an equation system, in which only the adaptation factors h1 to h5 are unknown. The equations of the equation system are independent of one another. Their number conforms to the number of adaptation factors h1 to h5. The equation system may, therefore, be used to determine the adaptation factors h1 to h5 unequivocally.
Preferably, the desired poles are specified as follows: for each interference frequency fS to be compensated, a pair of conjugate-complex poles is specified. The imaginary parts of the respective pole pair are equal to +/−2πfS. As already mentioned, fS is the interference frequency fS to be compensated. The imaginary parts are, therefore (in terms of value) equal to the corresponding angular interference frequency ωS. The real parts of the respective pole pair are smaller than zero.
The three further poles are preferably all real and smaller than zero, that is negative.
If the model time constants T1, T2 model the real continuous casting mold 1 well, the real parts of the conjugate-complex poles and the real poles are variable within wide limits, without this impairing the quality of the control method. However, frequently, the correct model time constants T1, T2 can only be roughly estimated. Nevertheless, the control quality is good if the real parts of the conjugate-complex poles and the real poles fulfill specific criteria.
The stability of the control method can, for example, be increased if the real parts of the conjugate-complex poles lie between −0.1 times and −0.3 times the corresponding angular interference frequency ωS. In experiments, it has been found to be particularly advantageous for the real parts to be approximately equal to −0.2 times the corresponding angular interference frequency ωS.
It has also been found to be advantageous for the real poles all to be smaller than −2.0 or to differ from one another in pairs. It is even better for both criteria to be met. Particularly good results are achieved if one of the real poles lies at −3.0, one at −4.0 and one at −5.0 (in each case +/−0.5, preferably +/−0.2).
On the other hand,
It was mentioned above that the interference frequencies fS to be compensated must be known in advance. The interference frequencies fS can, for example, be determined by evaluating the time characteristic of the actual value p of the meniscus 9 in
The above specification serves exclusively to explain the present invention. The scope of protection of the present invention should, however, be determined exclusively by the appended claims.
| Number | Date | Country | Kind |
|---|---|---|---|
| 09163538.3 | Jun 2009 | EP | regional |
This application is a U.S. National Stage Application of International Application No. PCT/EP2010/056151 filed May 6, 2010, which designates the United States of America, and claims priority to EP Patent Application No. 09163538.3 filed Jun. 24, 2009. The contents of which are hereby incorporated by reference in their entirety.
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/EP2010/056151 | 5/6/2010 | WO | 00 | 12/23/2011 |
