EP3904544A1 - Method for adjusting a permeability of a sintered good - Google Patents

Abstract

The present invention relates to the field of sintering in ferrous metallurgy. The object of the present invention is to enable optimal operation of a sintering plant by allowing maximum air passage through the sintered material mixture. The object is achieved by a method, a signal processing device and a machine-readable code solved to adjust the permeability. A data-based online model for the permeability is used to determine a target value for the moisture content of the sintered material mixture at which a maximum permeability of the sintered material mixture occurs. The data-based online model has - for ongoing updating - data of at least one moisture content and the permeability of the sintered material mixture as input variables. The determined moisture content is transferred as a setpoint to a moisture addition controller.

Description

Field of technology

The present invention relates to the field of sintering in ferrous metallurgy.

On the one hand, the invention relates to a method for setting a permeability of a sintered material mixture for a sintering plant.

On the other hand, the invention relates to a signal processing device and a machine-readable code for carrying out the method for setting a permeability. The invention also relates to a storage medium on which the machine-readable code is stored.

State of the art

In a sintering plant, fine-grained raw materials are used to produce lumpy iron ore sinter for use in a blast furnace. Fine ores, limestone, recycling dusts are mixed with coke breeze - as fuel - in a mixing drum to form a sintered material mixture. There are also known methods for producing mixtures of sintered goods which, in addition to a mixer, also have a granulator and a rolling drum - as in FIG EP 2848299 B1 shown. The sintered material mixture is piled up as a sinter bed on a sinter belt consisting of so-called grate carriages. The coke breeze on the surface is ignited by means of a gas burner via an ignition hood. Under the grate, a vacuum is applied via wind boxes or suction boxes, whereby air and resulting exhaust gases are sucked from top to bottom through the sintered layer and the combustion zone migrates from the surface in the direction of the suction boxes. The iron ores begin to melt, stick together and form a sinter cake. A speed of the sintering belt is chosen so that the sintered layer is completely burned through shortly before the end of the belt. At the end of the belt, the sintered material is thrown off, broken and cooled on coolers, and then fed to the blast furnace via long conveyor belts.

The productivity of the sintering plant is closely related to the permeability of the sintering bed. The higher the permeability of the sintered bed, the higher the productivity. High permeability leads to good burning behavior and thus to a sinter cake that is evenly burned through. In addition to particle size distribution, moisture content is one of the most important parameters influencing permeability. When the moisture content is low, fine particles block the passage of air and lead to low permeability. When the humidity is very high, water blocks the passage of air between the grains of the sinter mix. In between there is an optimum where the fine particles are bound to the larger grains by the moisture and the permeability and thus the air passage is maximized.

For optimal operation of the sintering plant, it is therefore of great importance to know the position of the maximum permeability.

Summary of the invention

The object of the present invention is to enable optimal operation of a sintering plant by allowing maximum air passage through the sintered material mixture.

The object is achieved by a method for setting the permeability which, using a data-based online model for the permeability, determines a target value for the moisture content of the sintered material mixture at which a maximum permeability of the sintered material mixture occurs.

The data-based online model has - for ongoing updating - data on at least one moisture content and the permeability of the sintered material mixture as input variables. The determined moisture content is transferred as a setpoint to a moisture addition controller. The data can be, for example, direct measurement data, data calculated from other measurement data, or estimated data.

A data-based model is created in an offline phase. In order to ensure a suitable selection of a robust model and the influencing variables to be included, a validation of different models with different combinations of influencing variables is carried out in the offline phase on the basis of independent test data not used in the model creation. The model performance is assessed with the help of a root mean square error, which provides a measure of the deviation of the model prediction for the permeability and the measured permeability. Models are preferably used which provide a comprehensible formula for the relationship, for example a multiple linear regression. This data-based model has at least the moisture content as an input variable. The data-based model is continuously updated with data, preferably measurement data, the permeability and the moisture content when the sintering plant is in operation - i.e. in online mode. The permeability can also be determined indirectly by productivity. Productivity results from the weight of the sintered material produced over time. By recording the productivity, the permeability can also be determined to a good approximation. This data-based online model makes it possible to determine the moisture content of the sintered material mixture at which a maximum permeability occurs. This moisture content is then transmitted as a setpoint to a moisture addition controller. This transmission should take place in reasonable time segments, which depend on the measurement intervals for the moisture content and / or the permeability. An update of the setpoint should always take place after a time interval which is at least greater than or equal to the measurement interval. The measurement interval for the moisture content, which is advantageously determined with a mix tester, is, for example, in a range from 5 to 30 minutes. However, it is also conceivable to use other regression models from machine learning - such as lasso regression, ridge regression, symbolic regression, decision tree, random forest, gradient boosting or neural networks. Depending on which model is selected, it may be necessary to numerically determine a maximum permeability if an analytical determination is not possible. The moisture content is set during the preparation of the sintered mix in a unit provided for this purpose - for example in a mixer, intensive mixer, granulator and / or a roller-burnishing drum. If the maximum permeability cannot be clearly detected, it could also be regulated with a slope at the point of the last specified permeability value. Depending on the gradient, the new predetermined permeability value would be increased, decreased or left unchanged. In this case, a linear model could also be used from which the slope is calculated. The use of the slope can be used, for example, if the operating point is on the edge of the curve and is well away from the optimum. Alternatively, the maximum or optimum can be determined by the slope. A combination of the two methods - slope and determination of the maximum - is possible. If a maximum can be determined, its position is used for regulation. If this cannot be determined, the control can be switched to by means of a gradient.

• Schüttgutdichte der Sintergutmischung
• Zusammensetzung der Sintergutmischung
• Korngrößenverteilung
• Prozessparameter beim Aufbereiten des Sintermischgutes
• Umgebungseinflüsse, bevorzugt eine Umgebungstemperatur.

Durch die Schüttgutdichte kann die Materialzusammensetzung gut beschrieben werden. Die zusätzliche Eingangsgröße der Schüttgutdichte der Sintergutmischung für das datenbasierte online Modell verbessert die Bestimmung der Permeabilität, da dadurch Materialeigenschaften in das datenbasierte online Modell einfließen. Die Zusammensetzung der Sintergutmischung beinhaltet alle Materialen, Zuschlagsstoffe und/oder Brennstoffe. Die Korngrößenverteilung kann beispielsweise für alle Materialen der Sintergutmischung und/oder für Eisenerz als Eingangsgröße an das datenbasierte online Modell übergeben werden.
Als Eingangsgrößen der Prozessparameter beim Aufbereiten des Sintermischgutes können jene für einzelne oder für alle Aufbereitungsmaschinen verwendete werden. Unter den Aufbereitungsmaschinen werden beispielsweise Mischer, Intensivmischer, Granulatoren und/oder Rolliertrommeln verstanden. Bei den Umgebungseinflüssen sind neben der Umgebungstemperatur auch Luftfeuchtigkeit, Wetter, Niederschlagsmengen und andere Umgebungsbedingungen relevant. Je nachdem wo die Materialien für die Sintergutmischung gelagert werden, sind solche Einflüsse in Bezug auf den Feuchtegehalt von Bedeutung.A preferred embodiment provides that at least one of the following parameters of the sintering material mixture is an additional input variable of the data-based online model:

• Bulk density of the sinter mix
• Composition of the sinter mix
• Grain size distribution
• Process parameters when preparing the sintered mix
• Ambient influences, preferably an ambient temperature.

The material composition can be described well through the bulk density. The additional input variable of the bulk material density of the sintered material mixture for the data-based online model improves the determination of the permeability, since material properties flow into the data-based online model. The composition of the sinter mix includes all materials, additives and / or fuels. The grain size distribution can, for example, be transferred to the data-based online model as an input variable for all materials of the sintered material mixture and / or for iron ore.
The input variables of the process parameters when preparing the sintered mix can be used for individual or for all preparation machines. The processing machines are understood to mean, for example, mixers, intensive mixers, granulators and / or roller drums. When it comes to environmental influences, not only the ambient temperature but also humidity, weather, rainfall and other ambient conditions are relevant. Depending on where the materials for the sintered mixture are stored, such influences are important with regard to the moisture content.

Die dabei verwendeten Koeffizienten c0, c1, c2 werden aufgrund von vorhanden Datensätzen aus Messdaten angepasst. Aus dieser Formel kann nun jener Punkt bestimmt werden, an welchem die Permeabilität einen Maximalwert bei variierender Feuchtigkeit annimmt. Durch Berechnung der ersten Ableitung nach der Feuchtigkeit und Nullsetzen dieser Ableitung, erhält man den Wert für die Feuchtigkeit, an dem die Permeabilität maximal/minimal wird.
Aus dem Koeffizienten des quadratischen Terms der Feuchtigkeit c2 kann abgelesen werden, ob es sich um ein Minimum (positiver Koeffizient) oder ein Maximum (negativer Koeffizient) handelt. Es ist auch denkbar, dass zusätzlich noch ein Term mit einem Koeffizienten c3 multipliziert mit der Schüttgutdichte der Sintergutmischung zur Gleichung 1 addiert wird.Another preferred embodiment provides that the online model of the permeability has a functional relationship between the moisture content, a square of the moisture content and preferably the bulk material density. It has been shown that the model can be further improved by taking into account the square of the moisture content and the determination of the maximum permeability is further improved by such a relationship. The bulk density as a further input parameter enables a further improvement of the model and the determination of the permeability. The permeability p can therefore be approximated, for example, by the following functional relationship using moisture f and the square of moisture f ² : $$\mathrm{p}=\mathrm{c}\mathrm{0}+\mathrm{<figure-callout\; id="1"\; label="c"\; filenames="imgf0001.png,imgf0002.png"\; state="\{\{state\}\}">c</figure-callout>}\mathrm{1}*\mathrm{f}+\mathrm{<figure-callout\; id="2"\; label="c"\; filenames="imgf0002.png"\; state="\{\{state\}\}">c</figure-callout>}\mathrm{2}*{\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="imgf0002.png"\; state="\{\{state\}\}">f</figure-callout>}}^{\mathrm{2}}$$

The coefficients c0, c1, c2 used here are adapted on the basis of existing data records from measurement data. From this formula, the point at which the permeability assumes a maximum value with varying humidity can now be determined. By calculating the first derivative according to the humidity and setting this derivative to zero, the value for the humidity at which the permeability becomes maximum / minimum is obtained.
From the coefficient of the quadratic term of the humidity c2 it can be read off whether it is a minimum (positive coefficient) or a maximum (negative coefficient). It is also conceivable that a term with a coefficient c3 multiplied by the bulk material density of the sintered material mixture is added to equation 1.

An advantageous embodiment provides that the moisture content, as a setpoint value, is transferred to the moisture addition controller only after at least thirty minutes, preferably after one hour. With this measure, the moisture addition controller has enough time to adjust to the new default value, and fluctuations are avoided. The goal is the most stable operation possible with few changes.

An expedient embodiment provides that the desired value of the moisture content is in the range between 3.5% and 10%, preferably 3.8% -8%, particularly preferably 3.8% -4.3%. For the sinter mix, this is the area in which the moisture content should move. The maximum permeability normally occurs in this area.

Another preferred embodiment provides that the setpoint of the moisture content, which is transferred to the moisture addition controller, has a maximum deviation of 0.2% from the previous setpoint. This measure prevents the controller from overshooting and ensures stable operation.

An advantageous embodiment provides that the data-based online model takes into account at least two, preferably at least five, particularly preferably at least ten data sets with data - preferably measurement data. The data used are, among other things, the moisture content and the permeability of the sintered material mixture. The period of data acquisition has a different length of time for each of these data records. Using multiple data sets makes the data-based model more stable. For example, data records are used which cover a short period of time but are very current. These data sets allow a quick adjustment to changed material properties. Other data records that can be used are those with a long period of time, which are usually more stable. The data records can therefore overlap or be completely contained in another data record. For example, records with a short period of time can be completely contained in one of the records - with a longer period of time. It is conceivable that all data records have the same end time and differ only in a different start time. The combination of several data sets, each with different time periods, ensures that the model can react stably on the one hand and react quickly to changed conditions on the other. In connection with this invention, the periods of time range from a few hours to a few weeks and months.

Another useful embodiment provides that at least one data record is weighted differently. The data sets with the more recent periods are preferred a higher weighting factor than those data sets with a longer period in the past. The data-based online model can be additionally improved by the weighting factors, since more recent data sets - i.e. data sets with a current and short period of time - have a stronger influence than those that are further back in time. This leads to a stable model, but also to a model that reacts very quickly to changed material conditions. It is therefore conceivable to combine data sets without weighting factors and data sets with weighting factors with one another or to weight all data sets. It is conceivable that individual or a number of data points in the respective data records have different weighting factors.

An embodiment that has proven to be advantageous provides that only data records are used for the data-based online model which have a clearly determinable maximum. If this is not the case, this data set is not used for the data-based online model. This measure is intended to ensure that only data records are used which each have a clear maximum. This ensures an unambiguous determination of the maximum permeability.

A further expedient embodiment provides that the online model has the functional relationship of the third degree. For example, the third-degree moisture content can also flow into the data-based online model.

Another useful embodiment provides that the online model is determined by means of symbolic regression. With the help of symbolic regression, a number of operations and functions can be permitted, such as powers, logarithmic functions, exponential functions, trigonometric functions and / or types of calculation - such as plus, minus, division and / or multiplication. A symbolic regression algorithm then searches for the best formula using heuristic optimization methods matches the given training data. The relationship in the data records can possibly be described better because one also has more degrees of freedom.

The object is further achieved by a signal processing device for a sintering plant, with a machine-readable program code which has control commands which cause the signal processing device to carry out a method according to the method described above.

The object is also achieved by a machine-readable program code for a signal processing device for a sintering plant. The program code has control commands which cause the signal processing device to carry out the method described above.

The object is also achieved by a storage medium with a machine-readable program code stored on it, as described above.

• Fig. 1 und Fig. 2 zeigen eine schematische Darstellung einer Sinteranlage und eine Aufbereitung des Sintermischgutes.
• Fig. 3 zeigt ein Diagramm der gemessenen und der durch das datenbasierte online Modell bestimmte Permeabilität.
• Fig. 4 zeigt ein Diagramm des durch das datenbasierte online Modell ermittelten Feuchtegehaltes.

Beschreibung der AusführungsformenBrief description of the drawings

• Fig. 1 and Fig. 2 show a schematic representation of a sintering plant and a preparation of the sintered mix.
• Fig. 3 shows a diagram of the measured permeability and that determined by the data-based online model.
• Fig. 4 shows a diagram of the moisture content determined by the data-based online model.

Description of the embodiments

In the Fig. 1 is a schematic representation of a sintering plant 1 and an exemplary preparation of a Sinter mix 5 shown. In an intensive mixer 10, iron ore 11, which is in a grain size range from 200 μm to 10 mm, and raw materials 12 such as coke breeze, limestone, sintered debris, binders, dust and steel mill residues are introduced. A moisture content is set by adding 13 water. The addition of water 13 is controlled by a valve 13a which is connected to a water supply line 13b.
The moisture content of the sintered mix 5 is set by a moisture control 15. The humidity control 15 receives a setpoint value for the humidity content 16a from a data-based online model 16. This nominal value of the moisture content 16a is compared with the measured moisture content 17a. The humidity control 15 adjusts the valve 13a with a control variable 14 in order to ensure the optimal moisture content of the sintered material mixture 5 - as specified by the data-based online model 16.
The data-based online model 16 has as input variables at least the moisture content 17a and the permeability 17b of the sintered material mixture 5. It is also conceivable that the productivity is used instead of the permeability 17b. Further data 17 can also be transmitted to the data-based online model 16 as input parameters - for example grain size distribution, bulk material density and others. The moisture, permeability and bulk material density are preferably determined by a mixed material tester. The sintered material mixture is applied to the sintering plant 1 at a material addition point 2. A gas burner ignites coke breeze on the surface of the sintered material and a vacuum is applied through wind boxes 4, whereby air and resulting exhaust gases are sucked from top to bottom through the sintered layer. The iron ores begin to melt, stick together and form a sinter cake. At the end of the sintering plant there can also be a pin crusher (not shown) in order to crush the sinter cake. The sintered material 6 is dropped at a material delivery point 3. Subsequently, the sintered material 6 is cooled on coolers and one Blast furnace fed (not shown). The adjustment of the moisture content of the sintered material mixture 5 is essential in order to adjust the permeability 17b to an optimum. The optimum of the permeability 17b makes it possible to ensure a maximum passage of air so that at the end of the sintering plant 1 a sintered material 6 that is uniformly well burned through is produced.

In Fig. 2 is analogous to Fig. 1 an intensive mixer 10 is available. In this embodiment, the intensive mixer 10 is connected to a granulator 18 into which the mixed material 10a is transferred. The mixed material 10a is fed to a further process step in the granulator 18 and then the granulated material 18a can be transferred to a roller-burnishing drum 19. Instead of the rolling drum 19, another unit such as a disk pelletizer is also possible. In this embodiment, the moisture content 17a is set not only in the intensive mixer 10 but also in the granulator 18 and the roller-burnishing drum 19. The moisture addition controller 15 transmits the control signals 14 to the valves 13a for setting the water addition 13. In this embodiment, the moisture content can be specifically adjusted in each unit - intensive mixer 10, granulator 18 and roller drum 19. However, it is also conceivable that the moisture content 17a is set in only two or one unit.

In Fig. 3 a diagram of the permeability over time is shown. The permeability is recorded as the Japanese Permeability Unit (JPU). Measurement data from the permeability measurement 17b and a permeability prediction 20 of the data-based online model are entered in the diagram. This diagram shows the good agreement between the measurement data and the prediction of the permeability 20. The data-based online model is based for the permeability prediction on the functional relationship according to equation 1 (p = c0 + c1 ^∗ f + c2 ^∗ f ² ). The coefficients c0, c1, c2 were determined on the basis of existing data sets.

In Fig. 4 On the one hand, the humidity value 23 determined by the data-based online model is plotted in a diagram and the set humidity value 22. The set humidity value 22 was only changed in this example if the determined humidity value changes by an amount of 0.05%. This amount can also be higher or lower. It would also be conceivable to make the settings almost continuously. A number of individual model results 21 used are also shown in the diagram. The number of individual model results used 21 are those data sets that are taken into account for the data-based online model. These data sets are individual model results from which a maximum can be clearly determined. The widely varying number of data sets used shows that it can be important to use several data sets, as only a few data sets have a well-determinable maximum at certain points in time. This can be due to the fact that the material composition has changed or one is already very close to the optimum, which means that many data sets do not have a determinable maximum. If several data sets are used, the probability is very high that at least some data sets are included which have a determinable maximum.

Although the invention has been illustrated and described in more detail by the preferred exemplary embodiments, the invention is not restricted by the disclosed examples and other variations can be derived therefrom by the person skilled in the art without departing from the scope of protection of the invention.

List of reference symbols

1

Sintering plant

2

Material addition point

3

Material delivery point

4th

Wind boxes

5

Sinter mix

6th

Sintered material

7th

Gas burner

10

mixer

10a

Mixed material

11

Iron ore

12th

raw materials

13th

Addition of water

13a

Valve

13b

Water supply line

14th

Control variable

15th

Moisture addition regulator

16

Data-based online model

16a

Setpoint of the moisture content

17th

data

17a

Moisture content

17b

Permeability measurement

18th

Granulator

18a

granulated material

19th

Rolling drum

20th

Permeability prediction

21

Number of single model results used

22nd

set humidity value

23

determined humidity value

JPU

Japanese Permeability Unit

n

number

%

Moisture content in%

Claims (15)

Method for setting a permeability of a sintered material mixture (5) for a sintering plant (1), characterized in that a target value of the moisture content (16a) of the sintered material mixture (5) is determined using a data-based online model (16) of the permeability of the sintered material mixture (5) , at which a maximum of the permeability of the sintered material mixture (5) occurs, the setpoint value of the moisture content (16a) is transferred to a moisture addition controller (15), the data-based online model (16) being preferred as the input variable (17) for continuous updating Has measurement data of at least a moisture content (17a) and a permeability (17b) of the sintered material mixture (5). Method for setting a permeability (17b) of a sintered material mixture (5) for a sintering plant (1) according to Claim 1, characterized in that at least one of the following parameters is an additional input variable (17) of the data-based online model (16): - Bulk material density of the sintered material mixture - Composition of the sinter mix - grain size distribution - Process parameters when mixing the sintered mix - Environmental influences, preferably an ambient temperature. Method for setting a permeability of a sintered material mixture (5) for a sintering plant (1) according to Claim 1 or 2, characterized in that the data-based online model (16) of the permeability shows a functional relationship between the moisture content of the sintered material mixture, a square of the moisture content of the sintered material mixture and preferably has a bulk density of the sintered material mixture. Method for setting a permeability of a sintered material mixture for a sintering plant according to Claims 1 - 3, characterized in that the moisture content is only transferred to the moisture addition controller as a target value when there is a change of greater than 0.05% to the previous target value. Method for setting a permeability (17b) of a sintered material mixture (5) for a sintering plant (1) according to claim 1 - 4, characterized in that the moisture content (17a) as a setpoint value only after at least thirty minutes, preferably after one hour, at the addition of moisture Controller (15) is passed. Method for setting a permeability (17b) of a sintered material mixture (5) for a sintering plant according to Claims 1 - 5, characterized in that the desired value of the moisture content (17a) is in the range between 3.5% and 10%, preferably 3.8% - 8%, particularly preferably 3.8% - 4.3%. Method for setting a permeability (17b) of a sintered material mixture (5) for a sintering plant (1) according to claims 1-6, characterized in that the desired value of the moisture content which is transferred to the moisture addition controller (15) has a maximum deviation of 0.2 % of the previous setpoint. Method for setting a permeability (17b) of a sintered material mixture (5) for a sintering plant (1) according to claims 1-7, characterized in that the data-based online model (16) has at least two, preferably at least five, particularly preferably at least ten data sets with data , preferably measurement data, are taken into account, each of which has a different period of data acquisition. Method for setting a permeability (17b) of a sintered material mixture (5) for a sintering plant (1) according to claim 8, characterized in that at least one data record is weighted differently, data records with the more recent time periods preferably having a higher weighting factor than those data records with an earlier period. Method for setting a permeability (17b) of a sintered material mixture (5) for a sintering plant (1) according to claim 8 or 9, characterized in that only data sets are used for the data-based online model (16) which have a clearly determinable maximum. Method for setting a permeability (17b) of a sintered material mixture (5) for a sintering plant (1) according to Claims 3 - 10, characterized in that the data-based online model (16) has the functional relationship of the third degree. Method for setting a permeability (17b) of a sintered material mixture (5) for a sintering plant (1) according to Claim 1 or 2, characterized in that the data-based online model (16) is determined by means of symbolic regression. Signal processing device for a sintering plant (1), with a machine-readable program code which has control commands which cause the signal processing device to carry out a method according to one of Claims 1 to 12. Machine-readable program code for a signal processing device for a sintering plant (1), the Has program code control commands which cause the signal processing device to carry out the method according to one of Claims 1 to 12. Storage medium with a machine-readable program code according to claim 14 stored thereon.

Images