Flicker reduction in electric arc furnaces by means of flicker prediction from the state determination in the initial phase of the smelting process

Abstract

Flicker values to be expected may be determined and achieve a high probability from suitable state and operating variables which are acquired during the first minutes in the initial smelting phase. In this way, flicker can effectively be reduced and kept below predefined limiting values. This is in particular suitable during steel production using electric arc furnaces.

Claims

1. A method for reducing flicker in steel production using an electric arc furnace, which comprises the steps of: providing a flicker database storing a plurality of curves of flicker; measuring a curve of flicker during an initial melting phase of a steel production; comparing, via a computer, the curve of flicker during the initial melting phase of the steel production with the plurality of curves of flicker in the flicker database; selecting, via the computer, one of the plurality of curves of flicker in the flicker database by performing a similarity search comparing the plurality of curves of flicker in the flicker database with the curve of flicker during the initial melting phase of the steel production; in the computer, using the selected one of the plurality of curves of flicker as a predicted curve of flicker; and changing an electrode regulation based on a comparison of the predicted curve of flicker with the curve of flicker during the initial melting phase of the steel production.

2. The method according to claim 1, which further comprises measuring and storing in each case a dimension and a slope of the instantaneous flicker.

3. The method according to claim 1, wherein the initial melting phase has in each case a drilling phase and a collapse phase for scrap metal that is introduced in each case by means of a basket, and a measurement and a determination take place in each case during a first 100 to 200 seconds after melting of the introduced scrap metal.

4. The method as claimed in claim 1, wherein if predicted flicker values are less than predefined limiting values, a controller controls the steel production process by performing an electrode regulation.

5. The method according to claim 1, wherein if predicted flicker values are greater than predefined limiting values, a controller controls the steel production process in a manner adapted for flicker reduction during times flicker values are greater than predicted.

6. The method according to claim 5, which further comprises taking into consideration, via the controller, auxiliary information that flicker values which are greater than predicted occur for each basket at an end of a drilling phase and/or during a collapse phase.

7. The method according to claim 5, wherein the control unit increases an inductance of the electric arc furnace by means of higher throttle steps or switching in a throttle during a time that the flicker values are greater than predicted.

8. The method according to claim 5, wherein the controller adjusts electric arcs and currents by an electrode regulation during the times the flicker values are greater than predicted.

9. The method according to claim 5, wherein the control unit applies targeted periodic movements to electric arcs during the times the flicker values are greater than predicted.

10. The method according to claim 5, wherein the control unit changes parameters during the times the flicker values are greater than predicted.

11. The method according to claim 1, wherein when predicted flicker values are adjacent predefined limiting values, a controller controls the steel production process to set a compromise is between a melting performance and flicker reduction.

12. The method according to claim 11, wherein the control unit controls the steel production process in a manner adapted for flicker reduction during a time at an end of a drilling phase for each basket and during a collapse phase.

13. The method according to claim 1, which further comprises storing the plurality of curves of flicker in dependence on variables.

Description

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWING

(1) The invention will be described in greater detail on the basis of exemplary embodiments in conjunction with the figures. In the figures:

(2) FIG. 1 shows a first exemplary embodiment of an instantaneous flicker curve;

(3) FIG. 2 shows a second exemplary embodiment of an instantaneous flicker curve;

(4) FIG. 3 shows a third exemplary embodiment of an instantaneous flicker curve;

(5) FIG. 4 shows an exemplary embodiment of a method according to the invention.

DESCRIPTION OF THE INVENTION

(6) FIG. 1 shows an exemplary embodiment of an instantaneous flicker curve. FIG. 1 shows a typical curve of an instantaneous flicker of a melt. The time t in seconds s is plotted on the abscissa. The ordinate specifies numeric values for the instantaneous flicker. The scrap metal of a first basket is melted between 3200 seconds and 4200 seconds. This initial melting phase is divided into a drilling phase B and a collapse phase E. Within the time window from 4500 seconds to 5200 seconds, scrap metal located in a second basket is melted in the electric arc furnace. The melting phase is also divided into the drilling phase B and the collapse phase E for this additional step. For example, a third basket having scrap metal can influence the instantaneous flicker curve. According to FIG. 1, a liquid phase F follows the second collapse phase E. MF indicates instantaneous flicker. FIG. 1 shows a typical curve of the instantaneous flicker during a complete batch. The instantaneous flicker was determined for each of the three phases on the high-voltage side using a standard flicker meter and describes the short-term occurrence of the network flicker. In the case of melting of the first basket K1 during the drilling phase B, a strong increase of the instantaneous flicker is observed, then a characteristic drop and a second increase during the so-called collapse phase E, during which scrap metal which is not yet melted slips down or also collapses from the furnace walls into the molten zones under the electrodes. This process results in high flicker values. In the case of the second basket K2, this behavior is repeated, wherein the so-called flat bath phase F then follows, during which the scrap metal is substantially melted and the electric arcs burn stably on the melt. This results in very low flicker values. The curve and the height of the flicker can have very different profiles depending on the scrap metal used and the melting mode of operation, however. This is shown in FIGS. 2 and 3.

(7) FIGS. 2 and 3 clearly show that the curve and the height of flicker can have very different profiles depending on the scrap metal used and the melting mode of operation. FIGS. 2 and 3 show two different scenarios on an enlarged scale for two melts. While in the first scenario according to FIG. 2, a curve as described in FIG. 1 is shown, a completely different curve of the instantaneous flicker takes place in FIG. 3 as a second scenario. In the second scenario, the scrap metal composition and possibly also the operating parameters were different in comparison to the first scenario in FIG. 2, so that the flicker does not substantially increase either in the drilling phase or in the collapse phase and is very low as a whole.

(8) Scenario 1 shows a steep increase and high flicker values in the drilling phase B. Scenario 2 according to FIG. 3 shows a gentle increase in the drilling phase B with low flicker values.

(9) FIG. 4 shows an exemplary embodiment of a method according to the invention.

(10) It is particularly advantageous to use the items of information from the curve of the instantaneous flicker and the associated state variables and operating variables in particular in the starting phase from approximately 100 to 200 seconds during the melting of each basket K, to predict the future flicker.

(11) According to a first step S1, a knowledge database about the curve of the flicker is prepared for each furnace in dependence on the melting process and steel qualities, in which a sufficient number of typical cases is stored. The following data are stored in this so-called flicker database:

(12) Dimension and slope of the instantaneous flicker in the starting phase in the time range of approximately 100 to 200 seconds and in the further curve, as well as basket number, steel quality, electrical parameters, which can be, for example, voltage, current, transformer step, throttle step, effective power and apparent power, etc., and scrap metal parameters, which can be, for example, the scrap metal quality, the scrap metal volume, the scrap metal weight, etc. With a second step S2, this flicker database is transferred into a classifier, using which a similarity search can be carried out over a suitable feature space, for example, with a so-called closest neighbor classifier. For this purpose, the flicker curve most similar to the measured instantaneous flicker values, which can be the dimension and slope of the instantaneous flicker, for example, and the associated state variables and operating variables, is found. This classifier can also operate on a dynamically growing knowledge database and can be implemented as a learning system. The knowledge database can also be stored in a decentralized manner in a so-called enterprise cloud. Since the data of many different melting furnaces are stored here, the initial learning effort would be reduced. After the most probable flicker curve has been found from the starting phase, the further mode of operation can be dynamically optimized in a third step S3. For this purpose, three cases F1, F2, and F3 can be roughly differentiated. Case F1: the predicted flicker values are significantly less than predefined limiting values. In this case, the mode of operation is trimmed to optimum energy introduction and highest performance. This means, for example, for the electrode regulation, a setting for longer electric arcs and lower currents. Case F2: the predicted flicker values are greater than predefined limiting values. In this case, the mode of operation is adapted for the periods in which high flicker values are expected, as outlined hereafter. As was apparent from FIGS. 1 to 3, the periods having high flicker values occur in particular at the end of the drilling phase, i.e., approximately 100 to 250 seconds after the beginning of melting, for each basket K and in the collapse phase E. The inductance can be increased in these periods by higher throttle steps or by switching in a throttle or coil, which results in more stable electric arcs. The electrode regulation is set for shorter electric arcs having higher currents. Furthermore, targeted periodic movements in the electric arcs could result in flicker reduction especially in the collapse phase E. In addition, parameters can be dynamically changed in a possibly provided SVC. Case 3: the predicted flicker values lie in a border region here. A compromise can now be set between a high melting performance and the flicker reduction in the periods having high flicker values. In particular, the flicker reduction can be executed at the end of a drilling phase B and during a collapse phase E.

(13) The methods proposed here ensure optimum melting performance and can generally keep the flicker values below predefined limiting values. However, there is not absolute certainty in this regard, since the method is based on a prediction and flicker values occurring in future may deviate therefrom. For example, scrap metal collapses are subject to a certain level of randomness.

(14) The present invention uses the finding that the flicker values to be expected may be determined with high probability from suitable state variables and operating variables, which are captured during the first minutes in the melting phase. In this manner, flicker can be effectively reduced and kept under predefined limiting values. The invention is suitable in particular in the case of steel production using electric arc furnaces.