Journal of Research and Development
Open Access

Study of the Clogging Phenomenon of Submerged Nozzles in Continuous Casting

Florian Huber^*, Bruno Schachner, Jessica Zofia Gottsberger, Peter Benedikt and Andreas Florian Zierer ^*Correspondence: Florian Huber, Department of Metallurgy, Badji Mokhtar University, Annaba, Algeria, Email:

Author info »

Introduction

This work deals with the quality of mild steels intended for the manufacture of thin sheets to subsequently undergo tinning. It particularly concerns the prevention of the formation of quantities of non-metallic inclusions which could damage the quality of the metal [1]. Above all, it is known that ultra-low carbon steel (with a carbon content lower than 0.07%) is exclusively calmed by aluminum due to the constraints imposed by its chemical composition [2]. In fact, the calming of this type of steel is conventionally carried out in a ladle by adding aluminum [3]. The aluminum reacts with the oxygen present in the metal bath, producing aluminates that primarily settle on the surface of the liquid metal, first in the ladle and then in the dispatcher.

Unfortunately, some of these non-metallic inclusions remain in suspension within the liquid metal [4]. In addition, it is known that the continuous casting of a semi-finished product with a large cross-section usually requires the use of a submerged nozzle to feed the mold.

Moreover, it is known that these nozzles are subject to fattening, leading in the more or less long term to their total clogging and consequently to the immediate stoppage of the casting in progress, which affects the slab in the form of a defect called returns of casting. It should be pointed out that practically 64% of the castings produced at the oxygen steelworks (case studied) suffered nozzle clogging following the erosion of the internal wall.

In addition, it is known that it is the particles of non-metallic inclusions, remained in suspension which, while crossing the nozzle, cling to the wall of the tube and by accretion phenomenon cause over time the total blockage of the passage metal. The phenomenon has been the subject of several individual or collective studies in order to make these inclusions less harmful.

One of the peculiarities of the production of steel in the oxygen steelworks concerns blowing back due to a high content of Manganese in the cast iron (Mn=2.36% on average) thus contributing strongly to the overoxidation of the metal. The production of steel always includes a refining phase in which the impurities contained in the cast iron are partially or totally burned using oxygen. The oxides formed pass into the slag, with the exception of carbon monoxide which escapes with the gas [5].

It has been demonstrated that among the factors which favor the solubility of oxygen in the liquid metal, the temperature is a very important element. Indeed, the increase in the temperature of the bath considerably increases the concentration of oxygen which, for example from [O]=0.005% in the solid metal it goes to 0.23% in the liquid steel at the metallurgical temperature 1640°C.

The decrease in the carbon content in the liquid metal increases the activity of oxygen, hence the difficulty of deoxidizing ultra-low carbon steels (grade which is the subject of our research) which exclusively requires a large quantity of aluminum (this steel is intended for the manufacture of thin sheets) [6,7].

The factors, which favor the increase in the content of dissolved oxygen in the liquid metal, are high temperature, low carbon concentration and high Ferrous oxide (FeO) content [7]. Obviously, other factors can also have an influence such as hydrogen and nitrogen; but for the rest of our work, we will limit ourselves only to the most important in our opinion: Carbon [8].

Recall that the steel production process at the oxygen steelworks (case studied) is distinguished by a large number of reblown castings for Mn not reached at the end of refining, which results in a significant increase in temperature, the amount of FeO, and carbon over oxidation [9].

The blockages of the nozzles affect the quality of the slabs in the form of "belts" which can be defined as a discontinuity in cross section of the slab; Figure 1 shows a resumption of casting or belt in the slab.

Figure 1: Resumption of belt in the slab.

Materials and Methods

Location and morphology of the deposit

The localization of the deposit having blocked the nozzle was carried out by observation with the naked eye. The morphological identification of the deposit consists of observing a sample using a binocular magnifying glass. The sample was cut from the lower part and cut by an ordinary chainsaw, and then we detached the metal envelope, as well as the terminal deposit and leaving only the dense deposit, which adheres to the refractory body of the nozzle. Figure 2 shows clogged worn nozzle and cross section of a submerged nozzle.

Figure 2: View of a submerged nozzle. A): Clogged worn nozzle, Note: 1): Clogged nozzle body; 2): White deposit; 3): Metal casing. B): Cross section of lower part of submerged nozzle.

Chemical characterization of deposit

After the macro morphological identification of the deposit, we took a sample from each zone as shown in the Figure 3.

Figure 3: Layout diagram of the different layers.

The is carried out by X-ray fluorescence to determine the concentration of chemical elements and metal oxides, as it gives a more precise measurement.

The metal casing and the deposit of the lower part of the nozzle are made up of a mixture of slag from the converter and metal oxides resulting from deoxidation, Figure 3 indicates layout diagram of the different layers.

The taking of the samples of the deposit to carry out the chemical composition is carried out by a chainsaw, then to take off the dense deposit (C), from the impregnated refractory zone (B), we used a hammer and a chisel, then in the same way we separated the affected refractory zone (B) from the healthy refractory (A). Figure 4 explains materials obtained after separation.

Figure 4: Materials obtained after separation. A): Refractory zone; B): Affected zone (impregnation); C): Dense deposit; D): Terminal sintered deposit.

We crushed these materials to obtain fine powders with a particle size of approximately 5-15 μm to produce pellets after fusion in order to carry out chemical analysis using X-ray fluorescence.

Adhesion test (interfacial indentation)

To prepare our samples, we recovered a used nozzle, which we cut in the transverse direction as shown in Figure 3.

To study the adhesion of the dense deposit (C) on the body of the nozzle we cut samples from the lower part, we removed the terminal deposit (D) which has very weak and friable mechanical properties, finally we completed cutting with a manual chainsaw.

Subsequently, we cut these samples with the Mecatome TZ 10 type micro-cutting machine, until a flat and uniform sample was obtained and, because of its low thickness (3-5 mm), this sample was coated with epoxy as shown in Figure 5A, samples were prepared to measure interfacial indentation.

To perform the interfacial indentation, we used a Vickers microhardness tester, a Vickers hardness tester, and finally a Rockwell hardness tester until the part was destroyed. Figure 5B explains determination of the nature of the break.

Figure 5: Interfacial indentation of the sample. A): Samples prepared to measure interfacial indentation; B): Determination of the nature of the break.

After the interfacial indentation, we marked a point using a wire cross on the two internal faces of the sample to determine exactly the chemical composition by Energy Dispersive Spectroscopy (EDS) of the two opposite points.

Influence of the preheating of the nozzles on the roughness of the internal wall of the nozzle

Preheating of nozzles is an important process in manufacturing, affecting the surface quality and durability of nozzle components.

Laboratory experimental setup: We have studied the influence of preheating new nozzles at a fairly low temperature (750°C), for variable heating times in an uncontrolled atmosphere. We carried out an experimental assembly, which consists of a bracket to hold the nozzle, a natural gas burner and an ordinary pyrometer for temperature control. Figure 6A shows nozzle experimental setup.

Nozzle heating sequence at the factory: The heating time from the nozzle to the surface is 8 h and more in some cases, the burner is oriented towards the bottom of the nozzle which undergoes intense heating over 1000°C while the top remains relatively cold, knowing that refractories are poor conductors of heat, this temperature difference causes the surface to burst, the consequence of which is spalling. Figure 6B shows sample nozzle view after heating.

Figure 6: View of a sample of a nozzle. A): Experimental setup for heating; B): Sample nozzle view after heating.

Roughness test: For the measurement of the roughness, we used a 2D roughness meter of the Sypertechnologie type.

We took three samples that we cut and heated at a constant temperature of 750°C for periods of 45, 75 and 120 min, to determine the influence of temperature and heating time on roughness.

Results and Discussion

Observation of the deposit with the naked eye of the clogged nozzle

The lower part is almost completely clogged by a deposit, while the deposit in the upper part only measures a few millimeters;

The deposit formed on the inner wall of the lower part can be divided into three layers:

• A gray metallic envelope.

• Layer of slag having a thickness comparable to that of the deposit formed on the upper part.

• An extremely thick deposit, several centimeters long with a crumbly appearance, in the form of a powder stuck to another white deposit, denser and less thick.

Binocular observation

The observation of the deposit was carried out using an Optika digital binocular loupe, the photos of the different samples with an enlargement of 200 µm, Figure 7 shows binocular observation of the deposit.

Figure 7: Binocular observation of the deposit.

The results obtained are contained in Table 1. From these results, we can say that in addition to the layer separated before the observation and the dense deposit adhered to our samples, we detected a third layer of refractory origin according to its macroscopic morphology and its color. After these observations, the new configuration of the deposit layers then becomes included inside the healthy zone according to the following order:

Table 1: Number, color and thickness of the different deposit layers of the 3 samples.

First zone: It is the decarburized zone called the impregnation zone; its thickness is between 300-3000 µm.

Second zone: It is the dense deposit zone, white color with dark brown spots and light brown spots. Its thickness is between 200- 6250 µm.

Third zone friable terminal deposit: It is a very thick and very friable layer.

Chemical analysis of deposit and refractory part

The comparison of the results obtained with analyses of the chemical compositions of the slag, the repair cement, and the gunning of the tundish.

To study the evolution of the composition of the sound refractory (sample A), as well as the composition of the impregnation layer (sample B), we stuck to the chemical composition of new nozzle refractory.

In the Figure 3, we have divided our nozzle profile into two parts according to the nature of the base materials (refractory or steel), the results obtained after chemical analyses by X-ray fluorescence are collated in Table 2, namely:

Table 2: Results of chemical analysis by X-ray fluorescence.

Refractory: It comprises the zone of sound refractory (A), as well as the zone of the layer of refractory affected by impregnation (B).

Deposit: It comprises the dense deposit (C) as well as the terminal sintered deposit (D).

For more, we took by convention the names of the samples by following the direction that starts from the healthy refractory towards the terminal deposit and the flow channel of liquid steel, the results obtained after chemical analysis by fluorescence X-ray are gathered in the Table 2.

According to the chemical analyses, we found a very high quantity of iron of the order of 34%, with total decarburization; Table 3 represents chemical composition of nozzle and chemical composition of slag, this result shows us that the decarburized or impregnation zone is more spread out and thicker.

Table 3: Chemical composition of the new nozzle and slag.

To facilitate the comparison, we have mentioned in Table 4, the chemical analysis of gunitage, cement and chamotte bauxite. The results tell us about the nature of the deposit in general. Al₂O₃ mainly forms the terminal deposit with a percentage of 90%. The dense deposit layer remains predominant with an Al₂O₃ percentage of 53.348%, followed by total iron with a percentage of 40.262%.

Table 4: Chemical analysis of gunning products, tundish repair cement and bauxite chamotte from the nozzle refractory.

The affected refractory zone is a zone which has undergone a variation in the chemical composition. Indeed, the content of Al₂O₃ and that of SiO₂ have dropped respectively, from 66% to 42.258% and from 9% to 2.214%, we believe that this is due to the corrosion of our refractory material. We have also noticed that the iron content of 40.262% is very high; knowing that the initial iron content in the composition of the new nozzle is zero, so there is a great impregnation (diffusion of iron) in this layer. The impregnation in this zone is accompanied by a very intense effect, where we observed an absence of material.

We observe a significant increase in the iron content, in fact, it went from 15.5% to 54.98% in the affected area, as well as the significant decrease in the SiO2 and carbon content and that the impregnation layer is much higher than the calculated thickness, this change in morphology as well as the variation in color, in fact its color changed from metallic gray to light gray due to the increase in the alumina content (42% to 61%). The tundish repair refractory materials (guniting, cement, and bauxite chamotte) do not enter into the chemical composition of the plugging deposit.

Oxygen activity measurement

At the end of the refining operation and just before casting, it is useful to measure the oxygen activity in order to be able to determine the quantity of deoxidizer to add to the bath at a temperature of 1680°C. The results of the measurements of the activity of oxygen in our case enabled us to construct Figure 8 where we note that the equilibrium concentration of oxygen at the end of blowing is clearly higher than that of the theoretical equilibrium curve.

Figure 8: Relationship between carbon and oxygen.

All the conditions favorable to the increase of oxygen dissolved in the liquid metal are therefore united. To deoxidize steel, a large amount of aluminum is used. The activity of the oxygen in the liquid steel must be reduced as much as possible to avoid the formation of blisters, but too much quenching with aluminum causes the clogging of the nozzles by the inclusions of alumina.

Adhesion result

We have started the interfacial indentation tests by the Vickers microhardness with loads varying from 300 kgf^-1 kgf. These rather weak loads did not give visible results. This is why we increased these loads from 1 kgf-30 kgf, this time using a Vickers durometer, but we could not get results. This pushed us to change equipment and the Rockwell durometer seemed to us the most adequate.

We operated in the same way as the previous tests except that this time we started with a load of 60 kgf, but only at 187.5 kg. Force that the damage has taken place as shown in Figure 8. We consider that this load represents the level of adhesion of the deposit on the internal wall of the refractory, which is quite high. The observation by the Scanning Electron Microscopy (SEM) of the two faces after the interfacial indentation and the chemical analysis by EDS of the two points selected by the crosses in wire showed that there was a brittle rupture and not a detachment, indeed, one observed in Figures 9A and 9B, the same chemical composition, (the iron and aluminum (percentage (%) peaks are superimposable).

Figure 9: SEM image and chemical composition after EDS. A): Left side; B): Right side.

The study of the images of the SEM and EDS of the two faces of the damaged sample shows a similarity from the point of view of chemical composition and structure so we can affirm that the sample rather suffered a brittle fracture at the maximum load of 187.5 kg and not a detachment for the separation, of the dense layer and the refractory wall of the body of the nozzle thus showing a fairly high level of adhesion, which is explained by:

The impregnation zone of the internal walls of the nozzle leaves cavities where steel droplets, containing non-metallic inclusions, freeze, and cause the growth of the initial deposit by accretion phenomenon, we know that these are particles of non-metallic inclusions, remained in suspension which while crossing the nozzle cling to the wall of the tube and cause in time, the reduction of the passage of the metal until the complete reduction of the internal diameter of the nozzle. After nozzle internal fattening, indicates the level of growth of refractory nozzle flow channel clogging, the components of the clogging which react easily with the walls of the nozzle during the passage of the molten steel are oxides.

Roughness test result

The roughness test gives us the graphs of Figure 10.

Figure 10: Roughness profile of the sample for A): 45 minutes of holding 750°C; B): 75 minutes of holding 750°C; C): 120 minutes of holding 750°C.

The histogram in Figure 11 shows the roughness of the sample.

The results of the graphs (Figure 10) enabled us to construct the histogram of Figure 11. Indeed, the study of this histogram shows that the roughness decreases with the increase of the time of maintenance of the temperature. The roughness, after 120 min of maintenance at a temperature of 750° C., increased from 8.3 μm before heating to 2.81 μm, we can explain this decrease by the crumbling of the spikes during the roughness tests. In the industrial case, the preheating of the nozzles is rather anarchic and unbalanced, which has caused spalling at the level of the internal wall of the nozzle.

Figure 11: Histogram indicates the roughness.

Conclusions

Analysis by X-ray fluorescence of the deposit, which closes off the casting nozzles, revealed that it consists of practically 90% Al2O3 which comes from the over-consumption of aluminum for the calming of the liquid steel. The dissolved oxygen content in the steel at the end of refining measured using electrochemical cells is high compared to that of theory (approximately 30 ppm) this can be explained by the fact that the refining cycle is longer long in our case because of the high manganese content in the cast iron, which requires additional blowing in almost 64% of cases.

The deposit adheres very strongly to the lower internal walls of the nozzle. Indeed, the interfacial indentation practiced on a deposit sample showed that it was only at more than 187.5 kgf that the sample broke, which represents a fairly high level of adhesion, the examination by eye nu first followed by Material and Energy Balances (MEB) and EDS of both sides of the sample showed us that the fracture is of a brittle nature since we have the same chemical composition on both sides of the sample.

Among the factors that increase the adhesion of non-metallic inclusions (Al₂O₃) on the internal walls of the casting nozzle is the presence of flaking which is due to the overheating of the lower part of the nozzles (more than 1000°C) for a fairly long time (more than 8 h) in the conditions of steelworks. Indeed, the experiment carried out in our laboratory on three samples heated to 750°C with different holding times has shown that the internal surface condition of the nozzles is smoother, and that there is rather has an improvement in the state of the interior surface as evidenced by the roughness tests indeed.

To minimize the phenomenon of clogging of the nozzles, it is proposed to preheat the nozzles immersed in a furnace, which ensures homogeneous preheating under a controlled atmosphere for an optimal lifespan.

References

1. Bai H, Thomas BG. Effects of clogging, argon injection and casting conditions on flow rate and air aspiration in submerged entry nozzles. Steel Conf Proceed. 2000; 83:183-200.

   [Google Scholar]

2. Harcsik B, Tardy P, Karoly G. Examination of nozzle clogging in continuous casting. Metall Res Technol. 2012;109(3):177-186.

   [Crossref] [Google Scholar]

3. Hauke F, Petske J. Wear of submersible nozzles and formation of alumina deposits in the continuous casting of steel. Refract Contin Cast Mach: Conf Proceed. 1986;62-75.

   [Google Scholar]

4. Smirnov AN, Efimova VG, Verzilov AP, Maksaev EN. Clogging of submersible nozzles in continuous slab-casting machines. Steel Transl. 2014;44:833-837.

   [Crossref] [Google Scholar]

5. Spink J, Carberry M, Cassar P, Mizobe A, Yasuda T, Oki K. Performance of a new tundish nozzle at operation in a steel plant. 2011.

   [Google Scholar]

6. Mizobe A, Yasuda T, Tsuduki T. New tundish nozzle with the optimum bore profile. Refract Mater. 2011.

   [Google Scholar]

7. Mizobe A, Ueki M. Inner bore profile of the tundish nozzle minimising alumina adhesion. Refract Appl News. 2011;16(2):6-13.

   [Google Scholar]

8. Barati H, Wu M, Kharicha A, Ludwig A. A transient model for nozzle clogging. Powder Technol. 2018;329:181-198.

   [Google Scholar]

9. Barati H, Wu M, Kharicha A, Ludwig A. Calculation accuracy and efficiency of a transient model for submerged entry nozzle clogging. Metall Mater Trans B. 2019;50:1428-1443.

   [Crossref] [Google Scholar]