EP0572848A2 - Method of determination of the end-point during oxygen steelmaking in a converter - Google Patents

Abstract

In a method of determination of the end point during oxygen steelmaking in a converter, it is proposed according to the invention that, during the conversion, waste gas constituents continuously escaping from the converter are analysed, preferably by mass spectrometry, and parameters are generated, from the measured values, whose significant change in time gives a signal for ending the oxygen feed.

Description

The invention relates to a method for determining the end point for the fresh process in oxygen converters in steel production.

Carbon is essentially oxidized during the freshening, that is to say when oxygen is blown onto or into the molten steel. At the end of oxygen blowing, the carbon content is relatively small. Therefore, the reaction of iron with oxygen comes to the fore. In addition to the fact that this reaction creates what is known as brown smoke, which is a heavy burden on the environment, the liquid iron oxide reacts with the expensive refractory lining of the converter and thus increases the fire test costs. At the same time, iron spreading is reduced. That is why it is particularly important to switch off the oxygen supply in good time. In addition, an increased FeO content of the slag is disadvantageous for the production of steel with a high degree of purity.

There are many ways to find the cheapest time to switch off. There are general mathematical models in which the point in time is determined on the basis of predetermined values, including those relating to the melt itself.

However, since these values change from batch to batch, these static models are unable to determine the switch-off time that applies to each batch.

In contrast, dynamic models that use a signal are more suitable for satisfactorily controlling the fresh process. One of these dynamic methods is the so-called sublance technique. Shortly before the end of blowing, one or two samples are taken by inserting a lance into the blowing position.

The carbon content is determined by a quick determination using its dependence on the solidification temperature. The disadvantages of sublance technology are the very high investment and maintenance costs as well as an unsatisfactory hit rate in terms of carbon content.

Other dynamic models use gas analysis. Such models would in themselves be particularly suitable for determining the switch-off time if there were a satisfactory method for using the signals to control the fresh process. In the conventional exhaust gas analysis, the decarburization rate is determined from the measured CO and CO₂ volume fractions, taking into account the exhaust gas rate G _A. A lower decarburization rate signals the end point. This method is not recommended due to the insufficient reproducibility of the final carbon content in the metal and the FeO content of the slag.

The invention is therefore based on the object of carrying out a method of the type mentioned at the outset in such a way that it is possible to decide individually for each melt as early as possible and at a specific bath carbon content to switch off the oxygen supply when freshening.

The invention solves this problem with the aid of the features of the characterizing part of claim 1.

Further advantageous embodiments of the invention are the subject of the dependent claims. In the method according to the invention, the course over time of a number of exhaust gas components is preferably determined by mass spectrometry. Towards the end of the freshening process, the proportion of individual exhaust gas components decreases, while the proportion of other exhaust gas components increases.

If e.g. B. with the help of the mass spectrometer, the exhaust gas components CO, CO₂ and N₂ are measured, it turns out that towards the end of the fresh process, the proportion of CO falls, while the proportions of CO₂ and N₂ increase. The increase in CO₂ is relatively small, while the increase in N₂ is very significant. This is due to the fact that at the end of the freshening process, a lot of false air gets into the exhaust gas stream. For environmental reasons, the exhaust gases have to be extracted. There is a gap between the converter opening and the exhaust hood through which more and more false air is drawn in when the suction pumps can no longer be offered enough converter gas.

, wobei N₂ bzw. O₂ Volumenanteile in % im Abgas sind.Since the decrease in the CO content or the increase in the N₂ content is initially very moderate, it is not immediately possible to decide whether the time to switch off the oxygen supply has already been reached. One can only be certain when the decrease in the CO content or the increase in the N₂ content becomes significant. Before this point in time, however, the reaction of oxygen with iron may have gained the upper hand over the reaction of oxygen with carbon. For this reason, parameters are derived from the measured exhaust gas components, as set out in claim 4 by way of example. Here means ${\text{G}}_{\text{K}}{\text{= G}}_{\text{A}}{\text{/ 100 · (100 - N₂ - O₂) ≈ G.}}_{\text{A}}\text{/ 100 · (100 - N₂)}$

, where N₂ or O₂ volume percentages are in the exhaust gas.

The parameters are selected so that they show a significant change even if the exhaust gas components only slowly rise or fall.

In this way it is possible to make a decision about switching off the oxygen supply about 2 minutes earlier than with the method just described, in which only the exhaust gas components themselves are used.

It is important for the method that at the beginning of the end-point determination measurement the carbon monoxide content is above 40% by volume and the nitrogen content is below 40% by volume in the exhaust gas stream. Because only when these initial conditions are met will there be an evaluable drop in the carbon monoxide content or increase in the nitrogen content in the course of freshening.

Since it is important in the method according to the invention that the decisive significant change in the time course of the exhaust gas components or the parameters derived therefrom is recognized as early as possible, this time course must be subjected to a pattern recognition. According to the invention, this is accomplished with the aid of a computer program. The measured exhaust gas values are analog values that are digitized in an analog-digital converter. The derived parameters are then formed from the digital values obtained using the computer program. Measuring point for measuring point, which are approximately 3 seconds apart, is scanned and the respective time change is compared with a target state.

This desired state means, for example, "fall" with CO, while it means "rise" with nitrogen. Returns the scan, i.e. the comparison of the temporal change of the measuring points with the respective target state in several runs in succession the same tendency (CO falls constantly, N₂ rises continuously) a signal is generated which is used to switch off the oxygen supply.

The number of comparison runs (loops) is determined in advance and is, for example, 7.

However, since situations in which the target states are met several times in succession can occur in the course of the measured exhaust gas values, pattern recognition is only started when a point has been reached in the course of time from which the final drop occurs of the CO portion or the final increase in the N₂ content can be expected. This is roughly determined, for example, using one of the static models given above. However, there are other criteria, such as B. the so-called coal stop. In some steelmaking processes, for example, carbon is blown into the batch through the bottom of the converter. If there is enough carbon in the melt, the carbon supply is stopped while freshening continues. The pattern recognition process is then 5 minutes, for example. started after the coal stop, d. that is, when it is approximately certain that only a little carbon can react with oxygen, the carbon monoxide content in the exhaust gas stream thus begins to decrease steadily.

If the above-mentioned conditions are then fulfilled over a predetermined time or a predetermined number of comparison runs, the computer generates a signal that either stops the oxygen supply automatically in DDC mode or is sent to the converter control station, where the oxygen supply from the Operating personnel can be ended. This second alternative has the advantage that, based on experience, the operating personnel can still maintain the oxygen supply a little despite the signal.

Fig. 1

den prinzipiellen Aufbau der Meß- und Auswerteapparatur am Konverter,

Fig. 2

den zeitlichen Verlauf der Volumenanteile von CO, CO₂ und N₂,

Fig. 3

den zeitlichen Verlauf des Verhältnisses von CO und N₂ multipliziert mit dem Konvertergasanteil,

Fig. 4

den zeitlichen Verlauf des Verhältnisses von CO und N₂ und

Fig. 5

ein Flußdiagramm des erfindungsgemäßen Auswerteverfahrens.

The invention is illustrated below with reference to drawings and explained in more detail. Show it:

Fig. 1

the basic structure of the measuring and evaluation apparatus on the converter,

Fig. 2

the time course of the volume fractions of CO, CO₂ and N₂,

Fig. 3

the time course of the ratio of CO and N₂ multiplied by the proportion of converter gas,

Fig. 4

the time course of the ratio of CO and N₂ and

Fig. 5

a flowchart of the evaluation method according to the invention.

A converter 1 is shown schematically in FIG. 1. The oxygen blowing device is omitted for reasons of clarity. Above the converter opening 2 there is a suction hood 4 at a short distance 3, via which the exhaust gases from the converter 1 enter the chimney 5. The suction pumps are also not shown for reasons of clarity. A schematically illustrated branch 6 from the chimney 5 supplies a mass spectrometer 7 with the exhaust gas components to be analyzed. The measurement signals are plotted over time, for example with the aid of a recorder 8. The analog values are sampled and fed to an analog-digital converter 9. The digitized values arrive in the computer 10 and are further processed there. After it has been determined that the right time to switch off the oxygen supply has been reached, the computer 10 generates a signal which in the present example is transmitted via the digital output 11 to the converter control station (not shown).

2 shows the time course of the measured volume fractions in% of carbon monoxide, carbon dioxide and nitrogen. The measurements have been started during the freshening after a predetermined time. It can be seen that at the freshness end, i.e. H. in this case in the range of 10 to 12 minutes, the CO curve drops and the CO₂ and N₂ curves rise. The courses of the measurement signals are after approx. 10.5 min. clearly.

, in Abhängigkeit von der Frischzeit aufgetragen, so ist ein eindeutiger Verlauf noch früher als nach 10, 5 min. zu erkennen.Are certain derived quantities, such as. B. CO / N₂ or $\text{CO² (1 / N₂ - 1/100)}$

, depending on the freshness time, a clear course is even earlier than after 10, 5 min. to recognize.

, was $\text{CO² (1/N₂ - 1/100)}$

entspricht, als Funktion der Frischzeit durchgeführt worden. Hier ist schon ein eindeutiger Trend nach ca. 8,5 min. erkennbar. Der Verlauf von CO/N₂ in Abhängigkeit von der Frischzeit wird in Fig. 4 dargestellt. Auch in diesem Fall ist ein deutlicher Abfall nach ca. 8,5 min. zu sehen.In FIG. 3, for the same batch as in FIG. 2, an application of $\text{(CO / N₂) · (CO [100 - N₂] / 100)}$

, What $\text{CO² (1 / N₂ - 1/100)}$

corresponds as a function of the fresh time. Here is a clear trend after about 8.5 minutes. recognizable. The course of CO / N₂ depending on the fresh time is shown in Fig. 4. In this case too there is a clear drop after approx. 8.5 min. to see.

The method is based on the palpation of the time-varying variables specified above. To do this, it is necessary to create a computer program, as shown schematically in FIG. 5, and to convert the analog values into digital data. The computer program first of all generates the derived quantities and constantly compares the change over time with the target state.

The program is only started at a later point in time so that this comparison does not take place at the beginning of the fresh process or at the beginning of the measuring time, i.e. in an area in which no meaningful decisions can yet be made. In the present example only 5 minutes after the coal stop.

At the start of the measurement, the mass spectrometer is completely exposed to nitrogen. Now that the measurement begins, ie if the exhaust gases to be examined are fed to the mass spectrometer, the proportion of calibration gas nitrogen in the mass spectrometer drops. If it drops below 95%, the program is deactivated by a reset. Are the 5 min. Once the carbon supply has been reached, the system asks whether the proportion of calibration gas in the mass spectrometer has dropped to below 95%. If this question is answered in the affirmative, the counter in the program is set to 0 and the first condition is queried. In this case the condition is "the nitrogen content increases". If the answer is no, the run is restarted. If the condition is met, the second condition is queried, in this case "the carbon dioxide content increases". If the answer is no, the run is restarted, if the answer is yes, the third condition is queried, in this case "the carbon monoxide content drops". If the answer is no, the system restarts. Otherwise, other conditions are taken into account, e.g. B. falls over time the ratio of CO to N₂. Another condition could be whether the gradient of the falling edge of this ratio (CO / N₂) exceeds a certain value. If all conditions are met, the counter is increased by one and the conditions are queried for the next measuring point. If the number of loops reaches a previously set value, the process is ended and the signal to switch off the oxygen supply is generated.

5 is to be understood purely as an example. So it is of course possible that only two conditions are queried and that the number of runs between n = 2 and n = 10, but advantageously can be set to 7 beforehand.

Claims (7)

Method for determining the end point for the fresh process in oxygen converters during steel production, in which exhaust gas components escaping from the converter are continuously analyzed during the fresh work,
characterized,
that the changes over time of the exhaust gas components analyzed by mass spectrometry and / or parameters derived therefrom are continuously compared with a target curve with the aid of a computer program, and after the specified conditions have been recognized as fulfilled by the computer over a certain number of measured values in succession, the oxygen supply is terminated . Method according to claim 1,
characterized,
that as exhaust gas components C0 and / or CO₂ and N₂ are analyzed, from the measured value the parameters are derived. Method according to claim 2,
characterized,
that as a parameter the exhaust gas rate G _A and / or converter gas G _K and / or dC / dO₂ and / or $\text{CO (100 - N₂) / 100}$

and / or CO / N₂ and / or $\text{CO² (1 / N₂ - 1/100)}$

can be determined from the measured values of the exhaust gas components. Method according to one of claims 1 to 3,
characterized,
that at the beginning of the end point determination measurement the carbon monoxide content should be over 40% by volume and the nitrogen content under 40% by volume in the exhaust gas stream. Method according to one of claims 1 to 4,
characterized,
that the measured values are analog values that are digitized and then fed to the computer. Method according to claim 5,
characterized,
that the computer automatically stops the oxygen supply in DDC mode. Method according to claim 5,
characterized,
that the computer sends the signal to terminate the oxygen supply via a digital output to the converter control station, where the oxygen supply is then switched off.

Images