EP0691895B1 - Method and device for regulating the molten metal level in a mould of a continuous metal casting machine - Google Patents

Abstract

PCT No. PCT/FR94/00292 Sec. 371 Date Oct. 23, 1995 Sec. 102(e) Date Oct. 23, 1995 PCT Filed Mar. 17, 1994 PCT Pub. No. WO94/22618 PCT Pub. Date Oct. 13, 1994The subject of the invention is a method for regulating the level of the meniscus (13) of the liquid metal in a mold (5) of a machine for the continuous casting of metals, according to which method the electrical signals supplied by at least one pair of sensors (17, 18) overhanging said meniscus are picked up, said signals being a function of the respective distances (h1, h2) between said sensors and said meniscus, these two signals are combined so as to obtain a single signal representing an imaginary level of said meniscus and said signal is sent to means (15, 24) for controlling a device (14) for regulating the flow rate of metal penetrating the mold, so that said control means actuate said device so as to bring said imaginary level of said meniscus back to a predetermined set value (h), wherein each signal coming from said sensors is conditioned, eliminating therefrom the oscillations having both a frequency greater than a threshold (F) and an amplitude less than a threshold (D). The invention also relates to a mode of combining said signals and a device for implementing said method.

Description

The invention relates to the field of continuous casting of metals, in particular steel. More specifically, it relates to the regulation of the level of liquid metal present in a continuous casting mold.

In a continuous steel casting installation, the molten metal flowing from the ladle first passes through an intermediate container, called a distributor. One of the roles of the distributor is to orient the liquid metal towards the single oscillating ingot mold or, more generally, the multiple oscillating ingot molds of the continuous casting machine, in which the solidification of steel products begins (slabs, blooms or billets). . Above each ingot mold, the metal flows out of the distributor through an outlet orifice, and thus forms a pouring jet which penetrates into the ingot mold by passing through the meniscus, i.e. the surface of the liquid metal. present in the mold. On its path between the distributor and the ingot mold, the pouring jet is confined in a tube made of refractory material, called a pouring nozzle. The upper end of the nozzle is fixed to the bottom of the distributor, while its lower end passes through the meniscus and plunges into the liquid metal. The purpose of the nozzle is to protect the jet of liquid metal against its oxidation by the atmosphere, to prevent that during its crossing of the meniscus, the jet does not carry with it part of the covering slag which covers the meniscus, this which would deteriorate the cleanliness of the cast product, and finally to impose on the flows of liquid metal in the mold a configuration favorable to a satisfactory solidification of the product. As such, its lower end may include a multiplicity of lateral orifices (or gills), each oriented towards one or the other of the faces of the mold.

One of the essential parameters in obtaining a healthy product is the stability of the level of the meniscus in the mold. If this stability is not ensured in a way satisfactory, the solidification of the product takes place under excessively variable conditions. We can thus end up with a solidified thickness of the product locally too small, hence a risk of more or less significant tears da solidified skin. At best, we end up with a product of poor surface quality; at worst, liquid metal can flow through the tears (a phenomenon called "breakthrough"), causing the casting to stop and serious damage to the machine. The average level of the meniscus is conditioned by the flow of steel flowing out of the distributor and by the speed at which the product being solidified is extracted from the mold. It is generally with a stopper rod, the conical nose of which more or less closes the outlet of the distributor, that the flow of liquid steel entering the ingot mold is regulated. Even if it is desired to maintain this flow rate at a constant value, it is necessary to vary the position of the stopper nose to take account of the gradual or instantaneous changes in the other casting parameters. These changes can be, for example, a variation in the height of the metal in the distributor, the progressive wear of the gills of the nozzle, or their plugging by non-metallic inclusions, or their sudden uncorking if these inclusions come to come off from the walls. To obtain satisfactory regulation of the level of liquid metal in the ingot mold, it is essential to use an automatic system which controls the position of the stopper rod. It moves it according to the results of a comparison between the level of the meniscus desired and that actually measured. This level measurement is usually obtained using a single optical or inductive sensor. It provides an electrical signal which, after processing, is used to control the position of the stopper rod.

It is in the case of continuous slab flows that the problem of regulating the level of the meniscus is most complex. Indeed, these ingot molds are long and narrow, and at a given time, the level fluctuations of the meniscus can be very uneven from one region to another of the mold. The indications given by a single sensor are therefore not necessarily representative of fluctuations in the level of the meniscus. On the other hand, on these machines, the lower end of the nozzle most often comprises two diametrically opposite openings which each orient a fraction of the metal jet towards one of the small faces of the mold. However, these two openings do not become blocked or necessarily widen identically during the entire casting. The flows in the ingot mold can therefore evolve asymmetrically, and the undulations which affect the meniscus then have very different configurations on either side of the nozzle at a given instant. In particular when one of the openings suddenly opens, if this opening takes place on the side of the nozzle where the sensor is located, this one attributes to the corresponding disturbance an exaggerated importance compared to the real evolution of the average level of the meniscus that it provokes. Conversely, if the unclogging takes place on the side opposite to that on which the sensor is located, the latter does not detect the disturbance at the instant when it occurs, or only in a very attenuated manner. In both cases, the stopper cannot be ordered in the most appropriate way to react to this event.

It has been proposed (see document JP 02 137655) to use for this purpose not one, but two sensors each located on either side of the nozzle, and moving along the longitudinal axis of the mold. The casting speed is controlled according to the simple difference between the signals delivered by each of the sensors. If this constitutes progress compared to the configuration with a single sensor, such a device is still insufficient to provide satisfactory consideration (neither overestimated nor underestimated) of all disturbances of the meniscus.

The aim of the invention is to propose a method for regulating the level of liquid metal which takes account of local disturbances of the meniscus by correctly estimating their real influence on the average level of liquid metal in ingot mold, and which makes it possible to appreciably reduce the amplitude of the fluctuations in the level of the meniscus harmful to the quality of the slabs, and this taking into account the entire meniscus.

To this end, the subject of the invention is a method for regulating the level of the meniscus of liquid metal in an ingot mold of a continuous metal casting machine, according to which the electrical signals supplied by at least one pair of sensors overhanging are collected. said meniscus, said signals being a function of the respective distances (h ₁ , h ₂ ) between said sensors and said meniscus, these two signals are combined so as to obtain a single signal representative of a fictitious level of said meniscus, and said signal is sent control means of a device for regulating the flow rate of metal entering the ingot mold, so that said control means actuate said device so as to bring said fictitious level of said meniscus to a predetermined set value (h), characterized in what we condition each signal from said sensors, eliminating oscillations having both a frequency greater than a threshold (F) and an amplitude below a threshold (D).

• on calcule la grandeur $$\text{(}\overline{\text{M}}\text{=}\frac{{\text{h}}_{\text{1}}{\text{+h}}_{\text{2}}\text{-2h}}{\text{2}}\text{)}$$
  
  et sa valeur absolue (| $\overline{\text{M}}$
  
  |) ;
• on compare (| $\overline{\text{M}}$
  
  |) à deux valeurs prédéterminées (diff_min) et (diff_max) avec (diff_min < diff_max) ;
• si | $\overline{\text{M}}$
  
  | ≤ diff_min, on prend ledit niveau fictif égal à $\overline{\text{M}}$
  
  ;
• si | $\overline{\text{M}}$
  
  | ≥ diff_max, on prend ledit niveau fictif égal à une valeur (Δh_max) qui est la plus élevée en valeur absolue, des grandeurs [(h₁ - h), (h₂ - h)]
• si diff_min < | $\overline{\text{M}}$
  
  | < diff_max, on prend ledit niveau fictif égal à ${\text{α Δh}}_{\text{max}}\text{+ (1-α)}\overline{\text{M}}$
  
  , α étant égal $$\frac{\text{(|}\overline{\text{M}}{\text{| - diff}}_{\text{min}}\text{)}}{{\text{(diff}}_{\text{max}}{\text{- diff}}_{\text{min}}\text{)}}\text{.}$$
  

Preferably, said signals are combined in the following manner:

• we calculate the size $$\text{((}\overline{\text{M}}\text{=}\frac{{\text{h}}_{\text{1}}{\text{+ h}}_{\text{2}}\text{-2h}}{\text{2}}\text{)}$$
  
  and its absolute value (| $\overline{\text{M}}$
  
  |);
• we compare (| $\overline{\text{M}}$
  
  |) with two predetermined values (diff _min ) and (diff _max ) with (diff _min <diff _max );
• if | $\overline{\text{M}}$
  
  | ≤ diff _min , we take said fictitious level equal to $\overline{\text{M}}$
  
  ;
• if | $\overline{\text{M}}$
  
  | ≥ diff _max , we take said fictitious level equal to a value (Δh _max ) which is the highest in absolute value, of the quantities [(h ₁ - h), (h ₂ - h)]
• if diff _min <| $\overline{\text{M}}$
  
  | <diff _max , we take said fictitious level equal to ${\text{α Δh}}_{\text{max}}\text{+ (1-α)}\overline{\text{M}}$
  
  , α being equal $$\frac{\text{(|}\overline{\text{M}}{\text{| - Diff}}_{\text{min}}\text{)}}{{\text{(Diff}}_{\text{max}}{\text{- Diff}}_{\text{min}}\text{)}}\text{.}$$
  

The invention also relates to a device for implementing this method.

As will be understood, the invention consists in conditioning the signals from these sensors prior to their combination, by eliminating from these signals the oscillations at high frequency and low amplitude, and by combining these signals into a single signal of a appropriately.

The invention will be better understood on reading the description which follows and refers to the attached single figure. This schematizes a section of a distributor and a mold for continuous casting of slabs equipped with a device according to the invention.

The liquid steel 1 contained in a distributor 2 flows through an outlet orifice 3, formed in the bottom 4 of the distributor 2, in an oscillating ingot mold 5 without bottom. The side walls 6, 7 of the mold 2 are vigorously cooled by an internal circulation of water. Against these walls 6, 7 begins the formation of a solidified crust 8. This progressively gains the entire cross-section of the cast slab as it is extracted from the machine, as symbolized by the arrow 9. On its path between the distributor 2 and the mold 5, the liquid steel 1 is protected by a tubular nozzle 10 made of a refractory material such as graphite alumina. The upper part of the nozzle 10 is fixed against the bottom 4 of the distributor 1, in the extension of the outlet orifice 3. The lower part of the nozzle 10 is provided with two lateral openings 11, 12 through which flows the liquid steel 1, each oriented towards one of the walls 7. The nozzle 10 crosses the meniscus 13 so as to bring the liquid metal 1 to the heart of the ingot mold 5 (for reasons of clarity of the drawing, we have not represented the layer of slag which usually covers the meniscus 13). The orifice 3 is partially closed (or completely closed when the casting is stopped) by a stopper 14 with a roughly conoid end, whose height position is adjusted by a device 15. The height position of the stopper 14, combined with the value of the speed of extraction of the slab from the ingot mold 5, determines the average level at which the meniscus 13 is in the ingot mold 5. The setpoint level 16 is thus marked with dotted lines. one wishes to maintain permanently during the casting of the slab.

This maintenance is ensured by means of a device which will now be described. It firstly comprises two level sensors 17, 18 of a type known in itself, for example eddy current sensors. They are located on either side of the nozzle 10, preferably at equal distances from the nozzle 10 and above the major median axis of the cross section of the ingot mold 5. In the general case, their lower ends are located at equal altitudes. The sensor 17 delivers an electrical signal representative of the distance h ₁ between its lower end and the meniscus 13, and the sensor 18 delivers an electrical signal representative of the distance h ₂ between its lower end and the meniscus 13. In the ideal case, these distances h ₁ , h ₂ should be equal to the distance h between the lower ends of the sensors 17, 18 and the reference level 16. In practice, this is very rarely the case, since the meniscus 13 always has undulations more or less significant amplitudes, depending on the variations in the flow of liquid metal 1 leaving the nozzle 10, the oscillation of the ingot mold 5, variations in the speed of extraction of the product, etc. These undulations practically never being completely symmetrical (in particular because the wear or blockage of the gills 11, 12 can be significantly different), h ₁ and h ₂ are generally uneven. This explains the previously reported impossibility of effecting reliable regulation of the level of the meniscus 13 on the basis only of the information delivered by a single sensor.

The analog signals delivered by the sensors 17, 18 are sent to analog / digital converters 19, 20, from which they emerge digitized. Each of these digitized signals is sent to a digital filtering device 21, 22 which operates in the following manner. The signals emitted by the sensors 17, 18 and representative of the variations in the level of the meniscus 13 which they detect, are the superposition of multiple undulations of frequencies and of various amplitudes. There are ripples of low frequencies, below a threshold which is arbitrarily set at 0.02 Hz, and ripples at higher frequencies, greater than 0.02 Hz and which can reach a few Hz.

It is considered that, for good regulation of the level of the meniscus 13, it is preferable not to take account of disturbances which have both a high frequency (greater than 0.02 Hz) and a low amplitude. Indeed, it is the disturbances at low frequency (less than 0.02 Hz) and the disturbances at high frequency but at high amplitude which are considered as harmful for the surface quality of the slabs. Disregarding disturbances at high frequency and low amplitude makes it possible not to exaggerate excessively and unnecessarily the device for adjusting the flow rate of the liquid metal and to limit its wear. In order to eliminate these disturbances from the processed signals, each of these is sent to a conditioning device 21, 22. These conditioning devices 21, 22 are identical, and operate in the following manner. The signal from each sensor 17, 18, after having been digitized by one of the converters 19, 20, is processed by a low-pass filter which suppresses or at least strongly attenuates the signals of frequency above a threshold F, that l for example, 0.02 Hz is fixed. The remaining low frequencies are then subtracted from the original, unfiltered signal, in order to obtain a new signal comprising no more, significantly, than the highest frequencies of the original signal. This new signal then crosses a dead band which strongly attenuates or eliminates from it the components whose amplitude does not exceed a predetermined threshold D, taken for example equal to 3 mm. Finally, the low frequency sampled at the output of the low-pass filter is added to the signal thus processed. We thus reconstitute a signal conforming to the original signal delivered by the sensor 17, 18, except that eliminated the components having both a high frequency (greater than F = 0.02 Hz) and a low amplitude (less than D = 3 mm).

The signals thus reconstituted are then sent to a combination device 23, to be combined into a single signal which synthesizes them, so as to provide the information necessary for the control of the stopper 14. This signal constitutes in a way a medium level fictitious for metal in an ingot mold. It is sent to a digital regulator 24 which in turn provides the device 15 with a signal which enables it to adjust the position of the stopper nose 14 in the outlet orifice 3 adequately, and therefore the flow of liquid metal entering in the ingot mold 5. The aim is thus to reduce the imaginary level of the liquid metal in the ingot mold to the set value, if a difference is detected between them.

Advantageously, the converters 19, 20, the conditioning devices 21, 22, the combination device 23 and the regulator 24 can be arranged inside the same housing 25. The devices downstream of the converters 19, 20 can even be made up of a single digital processing card designed and programmed to perform each of their functions.

The choice of the way in which the signals are combined in the device 23 is of great importance for the quality of the final result, that is to say an adequate regulation of the level of the meniscus 13. One could be satisfied to take as stopper control signal 14 the simple average of the signals collected by each sensor, and reflecting the deviations of the level from the setpoint. But there is then a risk of minimizing the importance of a strong disturbance as long as it is limited to one side of the mold. It is therefore advantageous to combine these two signals in a more complex manner. However, avoid going overboard by giving undue importance to a disturbance of average amplitude limited to one side. We would then fall back into the trap of the single sensor regulation systems described above.

To this end, the inventors have proposed the following method which gives satisfactory results. As explained above, the difference to be ideally maintained between the meniscus 13 and the sensors 17, 18 is called h, this difference corresponding to the setpoint level 16. Similarly, the differences measured between the sensors are called h ₁ and h ₂ respectively. 17 and 18 and the meniscus 13. The differences (h ₁ - h) and (h ₂ - h) reflect the differences in the levels of the metal in the ingot mold in line with the sensors 17, 18 compared to the setpoint level 16. If these differences are positive, the metal level at the place of measurement is lower than the setpoint level 16. If they are negative, the metal level at the place of measurement is higher than the setpoint level.

de (h₁ - h) et (h₂ - h) soit $$\overline{\text{M}}\text{=}\frac{{\text{h}}_{\text{1}}{\text{+h}}_{\text{2}}\text{-2h}}{\text{2}}\text{.}$$

La valeur absolue de $\overline{\text{M}}$

appelée | $\overline{\text{M}}$

| est ensuite comparée à deux valeurs prédéterminées qu'elle peut prendre, dont la plus petite est appelée diff_min et la plus grande est appelée diff_max. Trois cas peuvent alors se présenter.

• 1) Si | $\overline{\text{M}}$
  
  | ≤ diff_min, le signal qui est envoyé au régulateur 24 correspond à $\overline{\text{M}}$
  
  . On considère donc que l'écart par rapport au niveau de consigne 16 est convenablement traduit par la simple moyenne arithmétique des écarts mesurés par chacun des capteurs 17, 18.
• 2) Si | $\overline{\text{M}}$
  
  | ≥ diff_max, le signal qui est envoyé au régulateur 24 correspond à la plus élevée, en valeur absolue des différences (h₁ -h) et (h₂ - h), qu'on appelle Δh_max. On tient alors compte exclusivement de celle qui traduit l'écart le plus important par rapport à la consigne.
• 3) Si diff_min < | $\overline{\text{M}}$
  
  | <diff_max, le signal qui est envoyé au régulateur 24 correspond à un compromis entre $\overline{\text{M}}$
  
  et Δh_max, calculé de manière à assurer une transition progressive entre les deux modes de régulation précédents. A cet effet, ce signal est pris égal à ${\text{α Δh}}_{\text{max}}\text{+ (1-α)}\overline{\text{M}}$
  
  , α étant défini par : $$\text{α =}\frac{\text{(|}\overline{\text{M}}{\text{| - diff}}_{\text{min}}\text{)}}{{\text{(diff}}_{\text{max}}{\text{- diff}}_{\text{min}}\text{)}}$$
  

The combination device first calculates at an instant t, the arithmetic mean $\overline{\text{M}}$

from (h ₁ - h) and (h ₂ - h) either $$\overline{\text{M}}\text{=}\frac{{\text{h}}_{\text{1}}{\text{+ h}}_{\text{2}}\text{-<figure-callout id="2" label="2h" filenames="imgf0001.png" state="{{state}}">2h</figure-callout>}}{\text{2}}\text{.}$$

The absolute value of $\overline{\text{M}}$

called | $\overline{\text{M}}$

| is then compared to two predetermined values it can take, the smallest of which is called diff _min and the largest of which is called diff _max . Three cases can then arise.

• 1) If | $\overline{\text{M}}$
  
  | ≤ diff _min , the signal which is sent to regulator 24 corresponds to $\overline{\text{M}}$
  
  . It is therefore considered that the deviation from the setpoint level 16 is suitably translated by the simple arithmetic mean of the deviations measured by each of the sensors 17, 18.
• 2) If | $\overline{\text{M}}$
  
  | ≥ diff _max , the signal which is sent to regulator 24 corresponds to the highest, in absolute value of the differences (h ₁ -h) and (h ₂ - h), which is called Δh _max . Only the one which reflects the largest deviation from the set point is then taken into account.
• 3) If _min diff <| $\overline{\text{M}}$
  
  | <diff _max , the signal which is sent to regulator 24 corresponds to a compromise between $\overline{\text{M}}$
  
  and Δh _max , calculated so as to ensure a gradual transition between the two previous regulation modes. For this purpose, this signal is taken equal to ${\text{α Δh}}_{\text{max}}\text{+ (1-α)}\overline{\text{M}}$
  
  , α being defined by: $$\text{α =}\frac{\text{(|}\overline{\text{M}}{\text{| - Diff}}_{\text{min}}\text{)}}{{\text{(Diff}}_{\text{max}}{\text{- Diff}}_{\text{min}}\text{)}}$$
  

Following these calculations, the regulator 24 and the control means 15 impose on the stopper 14 a movement such that it aims to correct the difference between the set value 16 and the fictitious level represented by the signal leaving the device. combination, established as just exposed. The operation is then repeated at an instant t + Δt, Δt being, for example, equal to 0.1 sec, and an almost continuous regulation of the level of liquid metal in the mold is thus obtained.

• a) Si le capteur 17 mesure h₁ = 70 mm et le capteur 18 mesure h₂ = 79 mm, on a (h₁ - h) = - 5 mm et (h₂ - h) = + 4 mm. On a alors $\overline{\text{M}}$
  
  = - 0,5 mm. Comme | $\overline{\text{M}}$
  
  | = 0,5 mm est inférieure à diff_min, le régulateur 24 envoie au dispositif de commande 15 un signal lui imposant d'actionner la quenouille 14 de manière à compenser un écart de M = - 0,5 mm par rapport au niveau de consigne 16. On n'a pas à tenir compte de la valeur de Δh_max (qui est égale à - 5 mm).
• b) Si le capteur 17 mesure h₁ = 70 mm et le capteur 18 mesure h₂ = 91 mm, on (h₁ - h) = - 5 mm et (h₂ - h) = + 16 mm. On a alors Δh_max = + 16 mm et $\overline{\text{M}}$
  
  = + 5,5 mm. Comme | $\overline{\text{M}}$
  
  | = 5,5 mm est supérieure à diff_max, le régulateur 24 envoie alors au dispositif de commande 15 un signal lui imposant d'actionner la quenouille 14 de manière à compenser un écart de Δh_max = + 16 mm par rapport au niveau de consigne 16.
• c) Si le capteur 17 mesure h₁ = 70 mm et le capteur 18 mesure h₂ = 85 mm, on (h₁ - h) = - 5 mm et (h₂ - h) = + 10 mm. On a alors Δh_max = + 10 mm, et $\overline{\text{M}}$
  
  = + 2,5 mm. Comme | $\overline{\text{M}}$
  
  | = 2,5 mm est comprise entre diff_min et diff_max, il faut calculer $$\text{α =}\frac{\text{2,5 - 1}}{\text{5 - 1}}\text{= 0,375.}$$
  
  Le régulateur 24 envoie alors au dispositif de commande 15 un signal lui imposant d'actionner la quenouille 14 de manière à compenser un écart de $${\text{αΔh}}_{\text{max}}\text{+ (1 - α)M=0,375 x 10 + (1 - 0,375) x 2,5 = 5,3 mm}$$
  
  par rapport au niveau de consigne 16.

By way of example, it is assumed that the setpoint level 16 is located at a distance h = 75 mm from the two sensors 17, 18. It is also posed that diff _max = 1 mm and diff _min = 5 mm.

• a) If sensor 17 measures h ₁ = 70 mm and sensor 18 measures h ₂ = 79 mm, we have (h ₁ - h) = - 5 mm and (h ₂ - h) = + 4 mm. We then have $\overline{\text{M}}$
  
  = - 0.5 mm. Like | $\overline{\text{M}}$
  
  | = 0.5 mm is less than diff _min , the regulator 24 sends a signal to the control device 15 requiring it to actuate the stopper 14 so as to compensate for a deviation of M = - 0.5 mm from the set level 16. We do not have to take into account the value of Δh _max (which is equal to - 5 mm).
• b) If sensor 17 measures h ₁ = 70 mm and sensor 18 measures h ₂ = 91 mm, we have (h ₁ - h) = - 5 mm and (h ₂ - h) = + 16 mm. We then have Δh _max = + 16 mm and $\overline{\text{M}}$
  
  = + 5.5 mm. Like | $\overline{\text{M}}$
  
  | = 5.5 mm is greater than _max diff, the regulator 24 then sends to the control device 15 a signal requiring it to actuate the stopper 14 so as to compensate for a deviation of Δh _max = + 16 mm from the set level 16.
• c) If sensor 17 measures h ₁ = 70 mm and sensor 18 measures h ₂ = 85 mm, we have (h ₁ - h) = - 5 mm and (h ₂ - h) = + 10 mm. We then have Δh _max = + 10 mm, and $\overline{\text{M}}$
  
  = + 2.5 mm. Like | $\overline{\text{M}}$
  
  | = 2.5 mm is between diff _min and diff _max , calculate $$\text{α =}\frac{\text{2.5 - 1}}{\text{5 - 1}}\text{= 0.375.}$$
  
  The regulator 24 then sends to the control device 15 a signal requiring it to actuate the stopper 14 so as to compensate for a deviation of $${\text{αΔh}}_{\text{max}}\text{+ (1 - α) M = 0.375 x 10 + (1 - 0.375) x 2.5 = 5.3 mm}$$
  
  relative to setpoint level 16.

It will be recalled that the mode of combining the signals from the sensors 17, 18 which has just been described constitutes only one example, and other modes of combination can be envisaged. Likewise, the numerical values cited for the operating parameters of the conditioning and combination devices are only examples and must be adjusted according to the local conditions of each machine according to the quality of the results obtained.

As a variant, one could also do without the operation of digitizing the signals from the sensors 17, 18 before their processing, and carry out their conditioning and their combination by purely analog means. But it goes without saying that we could not adjust with the same precision, and especially quickly modify if necessary the different operating parameters of the installation, such as, for the conditioning device, the width of the dead band and the cutoff frequency of the filter, and, for the combination device, the parameters diff _min and diff _max .

Likewise, all types of sensors delivering an electrical signal as a function of their distance from the meniscus, and not only eddy current sensors, can be used.

On the other hand, it is perfectly conceivable to use several pairs of sensors, distributed over the length of the mold, if one wishes to increase the precision of the detection of irregularities in the level of the meniscus. One can also use such a device on a square mold with blooms or billets.

Finally, it goes without saying that the regulating device described can also be used on a continuous casting machine on which the flow rate of the liquid steel leaving the distributor is regulated by a device other than a stopper, for example a nozzle with drawer. .

Claims (8)

1. Method for regulating the level of the meniscus of the liquid metal in a mould of a machine for the continuous casting of metals, according to which method the electrical signals supplied by at least one pair of sensors overhanging the said meniscus are picked up, the said signals being a function of the respective distances (h₁, h₂) between the said sensors and the said meniscus, these two signals are combined so as to obtain a single signal representing an imaginary level of the said meniscus and the said signal is sent to means for controlling a device for regulating the flow rate of metal penetrating the mould, so that the said control means actuate the said device so as to bring the said imaginary level of the said meniscus back to a predetermined set value (h), the method being characterized in that each signal coming from the said sensors is conditioned, eliminating therefrom the oscillations having both a frequency greater than a threshold (F) and an amplitude less than a threshold (D).
2. Method according to Claim 1, characterized in that, on combining the said signals emitted by the said sensors:

   - the quantity $$\text{(}\overline{\text{M}}\text{=}\frac{{\text{h}}_{\text{1}}{\text{+h}}_{\text{2}}\text{-2h}}{\text{2}}\text{)}$$

   

   and its absolute value (| $\overline{\text{M}}$

   

   |) are calculated;

   - (| $\overline{\text{M}}$

   

   |) is compared to two predetermined values (diff_min) and (diff_max), where diff_min < diff_max;

   - if | $\overline{\text{M}}$

   

   | ≤ diff_min, the said imaginary level is taken to be equal to $\overline{\text{M}}$

   

   ;

   - if | $\overline{\text{M}}$

   

   | ≥ diff_max, the said imaginary level is taken to be equal to a value (Δh_max) which is the higher in absolute value of the quantities [(h₁ - h), (h₂ - h)];

   - if diff_min < | $\overline{\text{M}}$

   

   | < diff_max, the said imaginary level is taken to be equal to ${\text{αΔh}}_{\text{max}}\text{+ (1-α)}\overline{\text{M}}$

   

   , α being equal to $$\frac{\text{(|}\overline{\text{M}}{\text{| - diff}}_{\text{min}}\text{)}}{{\text{(diff}}_{\text{max}}{\text{- diff}}_{\text{min}}\text{)}}\text{.}$$

   
3. Method according to Claim 1 or 2, characterized in that the signals coming from the said sensors are put into digital form and in that the said conditioning and combining operations are performed on the said signals thus digitized.
4. Method according to one of Claims 1 to 3, characterized in that the threshold (F) is taken to be equal to 0.02 Hz.
5. Method according to one of Claims 1 to 4, characterized in that the threshold (D) is taken to be equal to 3 mm.
6. Device for regulating the level of the meniscus (13) of liquid metal in a mould (5) of a machine for the continuous casting of metals, of the type comprising at least one pair of sensors overhanging the said meniscus (13), each of these sensors (17, 18) delivering a signal representing its distance (h₁, h₂) from the said meniscus (13), means (23) for combining the said signals and for delivering a single signal representing an imaginary level of the said meniscus to means (24, 15) for controlling a device (14) for regulating the flow of the liquid metal penetrating the mould, characterized in that it also comprises means (21, 22) for conditioning the said signals before combining them, so as to eliminate therefrom the undulations having both a frequency greater than a threshold (F) and an amplitude less than a threshold (D).
7. Device according to Claim 6, characterized in that it comprises means (19, 20) for digitizing the said signals emitted by the said sensors (17, 18) and in that the said means (21, 22, 23) for conditioning and combining the said signals are digital processing means.
8. Device according to Claim 6 or 7, characterized in that the said sensors (17, 18) are eddy-current sensors.

Images