Issue
Matériaux & Techniques
Volume 107, Number 5, 2019
Materials and Society: The Circular Economy (SAM13)
Article Number 510
Number of page(s) 12
DOI https://doi.org/10.1051/mattech/2020006
Published online 23 March 2020

© SCF, 2020

1 Foreword

Steel intermixing in slabs occurs during ladle change when pouring different steel grades (Fig. 1). The amount of “mixed” steel whose composition does not match any customer request can be downgraded or even scrapped. To enhance yield, it is therefore important to evaluate as more precisely as possible the mixed region volume to be skipped.

The length of the slab with “two-grade” steel depends on the tundish fluid flow pattern and operating conditions, namely liquid steel volume in the tundish at ladle change; tundish volume; casting speed during grade transition; slab format; tundish filling rate and chemistry of “new” and “old” steel grade.

Typical transition curves for the varying element content are shown in Figure 2, C-curve (response to “pulse” input) and F-curve (response to “step” input). In particular, the C-curve is relevant for deriving mixing time and volumes values, the F-curve allows to evaluate when a well-defined value for the “new” alloying element concentration is reached and is more immediate for the intermixing phenomena description. In the example of F-curve shown, the rising concentration value indicates a ladle change from steel with lower to higher element content.

thumbnail Fig. 1

Steel intermixing schematic (1).

Schéma de mélange de nuances differentes d’acier (1).

thumbnail Fig. 2

C- and F-curves features to characterize mixing in tundish.

C- et F-courbes pour la description du mélange de nuances en répartiteur.

2 Approach

The investigation here described was focused on the CC2 tundish of the Colakoglu Metallurji (CM) plant in Dilovasi (Turkey). The steel plant includes:

  • a 240-MWA (+ % 20) powered AC Electric arc furnace with 340 tons/h capacity;

  • two 48-MWA powered ladle furnaces with 300 tons/h capacity;

  • a twin tank vacuum degasser equipped with wire feeding system;

  • a 8-strands billet caster (130 × 130 mm, 150 × 150 mm, 6–16 m length);

  • a 2-strands slab caster (800–1650 mm width, 220 mm thickness, 6–16 m length).

Different approaches are used and detailed in literature to minimize intermixing effects [120]. They have been targeted at improving operating conditions and/or layout, even with suitable tools (e.g., particular ladle shroud geometry).

The first typical approach is based on simplified or analytical modelling. Here, analytical laws are used linking the relevant species concentration of the “new” steels to all the process and layout parameters previously mentioned. On one hand, such tools can be easily handled and roughly adapted to several similar conditions. On the other hand, for its semi-empirical nature, they should be validated by significant steel sampling vs time on plant, this making this technique often prone to margin of errors with significant underestimation or overestimation of the real intermixed volume.

The second approach is based on computational fluid modelling (CFD). In this case, the use of commercial codes allows a detailed description of the tundish flow pattern. As a main advantage, once designed and meshed the tundish, different operating conditions can be simulated and after flow validation, sampling on plant can support the correctness of the description. As main disadvantage, this kind of modelling is somewhat more complicated to describe in detail ladle change operations for different operating practices and should be validated.

The third is represented by physical (water) modelling. Tracing technique can allow to monitor in a water model the features of the transition among “old” and “new” steels. The flow features are easily evidenced, and different ladle changeover flow rates can be tested simply by varying tundish inlet water flow rate and drainage, but the technique should be accurate in order to achieve the proper new concentration curves.

For the study performed, CSM made use of the following means:

  • computational fluid modelling (CFD), via a validated Ansys-Fluent code already used for similar topics [21,22]. Meshing is provided via body-fitted coordinates with about 1,000,000 hybrid cells for the tundish investigated. Unsteady Reynolds Averaged-Navier-Stokes equations are solved, with standard wall functions treatment. As solutor, a SIMPLE scheme, spatial discretisation with momentum-2nd order upwind. Finally, turbulence was described via the κ-ε model;

  • physical modelling, via a flexible full scale tundish water model set to reproduce the CC2 55 ton tundish of CM under steady-state operating conditions and at ladle change. It was used to validate the CFD model and to define some tundish mixing parameters that, according to the chemical reactors’ theory [3] are needed for the subsequent curve fitting as shown next.

With the support of the mentioned tools the adopted work approach was the following:

  • a CFD model was set up to derive the remaining parameters also after tuned with the information from water modelling;

  • then, a suitable and generalizable analytical algorithm was developed to foresee the steel allying element concentration after ladle change in function of the operating conditions;

  • finally, parameters tuning of the general analytical expression was made by using plant data (slab sampling and verification of the intermixed length).

The validated intermix model was therefore implemented, at first as off-line tool, and then on-line connected to levels 2 and 3, by developing a user-friendly interface.

The mentioned approach made use of the following subsequent technical items:

  • thermo-fluid-dynamic CFD simulations of the cases under representative steady-state conditions and supporting and validating water modelling tests;

  • use of the flow results to have information on the parameters relevant for the set-up of the typical mixing curves: minimum residence time for the “new” fluid, average residence time for the “new” fluid, dead volume in the reactor, that can be derived from the C-curve; minimum residence time, that can be derived from both curves. The need of having both the group of curves is the following. The F-curves are the most relevant for describing the concentration change which is the intermixing feature. But to derive general analytical expressions, where the parameters have a well-defined physical meaning as shown next, come information pertinent to the behaviour of tundish as chemical reactors are needed as “dead volume”, average time and the first appearance time [23];

  • definition of the typical grade change curve (form of a F-curve) fitted with analytical expressions based on the parameters found. As a matter of fact, such curve has the typical shape of an exponential-asymptotic curve with suitable fitting constants at the exponent [24]. These fitting constants were just found to have physical meaning (just average time, mixing time…) and once found, they allow a generalization of the curve.

2.1 Activity performed

The considered cases are listed in Table 1. Since the fluid mixing is strongly affected by the use of flow modifiers currently used in tundishes, the possible representative cases where considered with different modifiers types: pads (like “boxes”, braking the inlet stream after pouring in the tundish), multiple-hole baffles (internal walls with inclined holes through which steel is guided towards the surface to favour a longer residence time beneficial for steel cleanliness), and with no modifiers.

Table 1

Conditions simulated.

Conditions simulées.

2.1.1 Thermo-fluid-dynamic simulations of the cases under steady-state conditions.

The velocity fields in relevant planes (longitudinal and transverse mid-planes, surface) are represented in Figures 35.

For symmetry, ¼ of the tundish is represented. It is evident the flow pattern in tundish for the different configurations. Qualitatively, in case of no flow deviators, flow short-circuiting occurs towards the exits. When a pad is used, enhanced flow appears towards the surface. Finally, when using a baffle, the overall flow behaviour is similar to that with a pad, and the flow deviation occurs by the holes. Some residual flow short-circuiting also appears through the baffle lower holes.

thumbnail Fig. 3

Velocity map (m/s) at surface and symmetry planes for the cases examined with no flow modifiers. For symmetry, ¼ of the tundish is represented.

Vitesse (m/s) à la surface et plans de symétrie pour les cas  sans modificateurs de flux. Pour la symétrie, ¼ des le répartiteur est représenté.

thumbnail Fig. 4

Velocity map (m/s) at surface and symmetry planes for the cases examined with pad. For symmetry, ¼ of the tundish is represented.

Vitesse (m/s) à la surface et plans de symétrie pour les cas avec un tampon. Pour la symétrie, ¼ du répartiteur est représenté.

thumbnail Fig. 5

Velocity map (m/s) at surface and symmetry planes for the cases examined with multiple hole baffles. For symmetry, ¼ of the tundish is represented.

Vitesse (m/s) à la surface et plans de symétrie pour les cas avec des chicanes à trous multiples. Pour la symétrie, ¼ du répartiteur est représenté.

2.1.2 Use of the flow results to have mixing information

From the CFD calculations, information was also achieved for each case on the distribution of the dead zones (taken as zones where the local steel velocity is below 50% of the average value in all tundish volume). This information is achieving by deriving a C-curve from the reactor operating conditions and it was needed to allow a generalised analytical form of the intermixing curves as previously explained. The qualitative “dead volume” distribution is shown in Figures 68. As expected, the dead zones are localised at the tundish corners, and the worst cases are achieved with no flow deviators able to favour stream spreading inside the tundish.

thumbnail Fig. 6

Distribution of the dead zone (in black) after CFD calculations. No flow modifiers. For symmetry, ¼ of the tundish is represented.

Distribution du volume ’mort’ (en noir) après calculs CFD. Pas de modificateurs de flux. Pour la symétrie, ¼ du répartiteur est représenté.

thumbnail Fig. 7

Distribution of the dead zone (in black) after CFD calculations. Pad as flow modifier. For symmetry, ¼ of the tundish is represented.

Distribution du volume « mort » (en noir) après calculs CFD. Pad comme modificateur de flux. Pour la symétrie, ¼ du répartiteur est représenté.

thumbnail Fig. 8

Distribution of the dead zone (in black) after CFD calculations. Multiple hole baffles as flow modifiers. For symmetry, ¼ of the tundish is represented.

Distribution du volume « mort » (en noir) après calculs CFD. Déflecteurs à trous multiples comme modificateurs de débit. Pour la symétrie, ¼ du répartiteur est représenté.

2.1.3 Determination of the mixing curves

To validate the CFD model, water model results were used. As comparison, another parameter was accounted for, relevant in the chemical reactors’ theory, i.e., the first appearance time of the fluid at the reactor outlet, indicative of the minimum fluid residence time in the reactor.

To do this, a tracing technique was used and the first appearance time for the colour tracer the exit was measured after direct visualization of the test. In Figure 9, representative images of the trials are shown for the case of “high” flow rate.

The comparison among first appearance times is shown in Table 2. A good agreement was found among CFD and water modelling results.

At this stage, the information gained allowed to derive the C-and F- curves for all cases of Table 1 with the CFD modelling. They are represented in Figure 10.

The normalised results for dead volumes are shown in Table 3. The values resulted to be independent from the flow rate and, based on the previous validation, were considered valid also for the water modelling.

It is interesting to note that, using flow modifiers, the dead volumes can be reduced up to about 30%. This is due to the fact that the flow modifiers allow steel to be directed toward the surface enabling a broad eddy before joining the strand.

Conversely, with no flow modifiers, flow short-circuiting occurs towards the strands, causing in turn a dead volume zone localized at the tundish corner.

thumbnail Fig. 9

Images of water modelling trials with colour tracer at the CSM labs.

Images d’essais de modélisation de l’eau avec un traceur de couleur dans les laboratoires du CSM.

Table 2

Comparison among first appearance time (normalised) of steel at tundish exit.

Comparaison entre temps (normalisé) de première apparition d’acier à la sortie du répartiteur.

thumbnail Fig. 10

C- and F-curves for all the cases examined after CFD calculations.

Courbes C et F pour tous les cas examinés après calculs CFD.

Table 3

Dead volume calculated from C-curves derived from CFD calculations. Results independent from the flow rate.

Volume mort calculé à partir des courbes C obtenues après les calculs CFD. Résultats indépendants du debit.

2.1.4 Fitting of mixing curves with analytical expressions

The F-curves achieved were fitted with suitable analytical curves for representing the variation of the alloying element concentration in the steel in a compact and physically meaningful way, and generalisable for all cases examined. Since the F-curve is typically a time-dependent exponential, with asymptotic value reached at complete mixing, the idea was to start from a general analytical expression, and then to fit the constant determining the curve shape, on the modelling results considering a well-defined physical meaning for the constants. The function is so generalised, for the element content C in function of time t during intermixing occurrence: for increasing element concentration (transition from C1 to C2 content with C2 > C1, and for decreasing element concentration (transition from C1 to C2 content with C2 < C1.

The fitting curves parameters are reported in Table 4.

It was found that the constant α could have been related to perfect mixing volume Vp, dead volume Vd and total volume V and is expressed by (V – Vd – Vp)0.5.

The previous relationships for C(t) can be also rewritten in function of slab length L (in m, with v = casting speed in m/s): for increasing element concentration (transition from C1 to C2 content with C2>C1), and for decreasing element concentration (transition from C1 to C2 content with C2 < C1).

Figure 11 shows all the fit curves to give an idea of the reasonably low differences among CFD and analytically derived curves after fitting. The close-up in Figure 12 evidences the maximum time difference among curves vs time, which anyway is present in an initial phase of the intermixing step and does not affect the final result.

Finally, as interesting technical information, the different attainment of intermixing is evidenced in Figure 13. As expected, the use of flow modifiers shortens the grade transition (less time is needed to achieve the 80% value of the “new” steel concentration).

A validation of this formulation is shown in Figure 14. Comparison is shown between fitting curve and points representing values achieved from test samples taken on CM CC2 tundish.

A good fitting between experimental points and model curve appears. The maximum difference is of about 10% of the alloying concentration value − a very low amount and occurs at the early stage of the curve when steel is far from being completely “changed” in composition, so it is not affected the prediction of the “optimum intermixing length”.

In the same figure, the chart on the right shows that the requested correspondence among calculated and “real” data can be found after about 12’ casting, in line with the developed curve on the left. As a matter of fact, a change in heat transfer behaviour appears from plant data after the same time lapse (zones evidenced in the chart). This can be ascribed to the changed properties of steel in mould that during the change of composition passes through the peritectic range. The consequent strong shrinkage of the forming shell causes the drop-in mould heat exchange. Therefore, it represents an indication of a “changed” steel behaviour and in turn of the presence of the “new” steel grade.

Summarising, from both model and samples/chart information, it is foreseen a change grade for which it can be assumed the presence of almost all (80%) “new” steel.

The model was checked also for different refilling policies (different tundish level at ladle opening), also giving good fittings.

The model was subsequently fitted as well to account for is able to take into considerations all possible casting (e.g., only one-strand casting available in case of blocked strands) and it was successfully implemented in the tool being currently used at CM steelworks.

Table 4

Parameters for the F-curve analytical relationship in (3).

Paramètres de la relation analytique de la courbe F dans (3).

thumbnail Fig. 11

Overview of C-Curves from CFD and fittings for each case. Low flow rate.

Vue d’ensemble des courbes C de CFD et des raccords pour chaque condition. Débit faible.

thumbnail Fig. 12

Close-up of zones with maximum fit error. Use of multiple-hole-baffle, low flow rate.

Gros plan des zones avec une erreur d’ajustement maximale. Utilisation de chicane à trous multiples, faible débit.

thumbnail Fig. 13

Comparison among fitted C-curves. Low flow rate.

Comparaison entre les courbes C ajustées. Faible débit.

thumbnail Fig. 14

Model data validation by chemical analysis on slab samples (points on the curve) and comparison with plant process parameters (screenshot of the CC2 process control system on the right).

Validation  du modèle par analyse chimique sur des échantillons de brame (points sur la courbe) et comparaison avec les paramètres du processus de l’usine (capture d’écran du système de contrôle du processus CC2 à droite).

2.2 Off-line and on-line tools development

The algorithms and rules described before were the base to develop the off-line and on-line intermix tools. The main input data are: steel flow rate, slab size, refilling conditions, presence of flow modifiers. The model gives as output the length of intermixed slab and steel chemical composition at that stage.

CSM developed two tools which uses Intermix Model. The first one (Offline intermix tool) allows technologists to perform offline simulation based on a manual input (the user can insert chemical compositions, cut plan, casting speed, etc.). The second (On-line Intermix tool) performs online simulations based on real time data acquired from Levels 2 and 3 automation.

2.2.1 Offline intermix tool

The tool was developed with Microsoft. NET Framework 3.5 technology for Windows XP or higher version. Examples of screenshots of the user interface set up and installed are shown in Figure 15. After the implementation of input data (top), the simulation is made and as output (bottom), the application displays the evolution of the steel concentration for the strands involved.

For each slab falling within the intermix area, the average composition is also given.

An example of optimization of slab cutting after intermixing, using the offline tool, is shown in Figure 16. The first output chart (Fig. 16, top) shows the slabs with different colors with fixed slabs length, based on input data (cutting plane). The first bold line on the left indicates start intermixing occurrence, the second indicates the end. In this case, three slabs with 10 m can be identified, two of them falling within the intermix range.

The second output chart (Fig. 16, down) shows the suggested cutting plane after the model optimization to minimize steel downgrading. In this case, the tool suggests introducing a further slab (S2) with a sub-length of 6 m (chosen within the allowable lengths in production plan) where intermix is “concentrated” and achieve two slabs 10 m long out from the intermix range (S1, S3).

thumbnail Fig. 15

Screenshot of the user interface (top, input; bottom, output) of the off-line tool developed and used at Colakoglu plant.

Capture d’écran de l’interface utilisateur (haut, entrée; bas, sortie) hors ligne développé et utilisé à l’usine de Colakoglu.

thumbnail Fig. 16

Output chart for the offline tool developed.

Graphique de sortie pour l’outil hors ligne développé.

2.2.2 On-line intermix tool

The tool developed with Microsoft. NET Framework 4.5 technology for Windows XP or higher version, communicates with levels 2 and 3 by using database interfaces tables. The scheme in Figure 17 represents the main information flow and its interaction with the levels 2 and 3 systems.

In Figure 17, level 3 is the master database for the management of target steel grade, target chemical composition and cutting plan. Level 2 is the master database for the management of process data, current chemical composition and steel grade, tracking information.

The interface tables contain all information acquired by level 3, leve 2 and PLC. The system reads the process data and the leve 3 data from this table.

The server stores the data acquired by interface tables on CSM database and simulates cutting for the heat being cast based on process data acquired by the plant.

The on-line tool reads the data from the server DB and allows user to track online heats or simulate the old heats.

The tool is then able to:

  • track the continuous slab casting production;

  • display the planned slabs affected by intermix;

  • calculate a new cutting plan based on selected rules;

  • suggest compatible steel grades for intermix slabs;

  • allow user to make “reverse intermix simulation” of heats already produced.

A screenshot of the on-line user interface set up and installed is shown in Figure 18.

thumbnail Fig. 17

Information flow for the on-line tool developed.

Flux d’informations pour l’outil en ligne développé.

thumbnail Fig. 18

Screenshot of the on-line tool developed.

Capture d’écran de l’outil en ligne.

3 Conclusions

As outcome of a modelling work performed by CSM and focused on Colakoglu CC2 tundish casting conditions:

  • a model was set up describing steel intermixing (element concentration variation during ladle change)

  • a general analytical form for the curves describing the grade change history in tundish was achieved with Computational Fluid Dynamics & water modelling;

  • the curves were fitted on technical basis (function of mixing quantities);

  • the results were validated on available plant data and applied to other cases;

  • the model, with all the possible grade change trends in tundish when varying flow rate (format, casting speed) and layout (with or without flow modifiers), at constant casting speed, and also accounting for anomalous process conditions (e.g., asymmetrical feeding of the strand) was implemented in a tool currently and successfully used on plant.

The tool has been integrated into the CM CC2 control system and is currently used offline for optimizing a priori mixed grade production scheduling, and online for slab cutting optimization in order to reduce the downgraded product amount.

For its nature, the model and the related off- and on-line tools can be generally applied to any other casting plant, after a preliminary tuning on well-defined plant operations.

References

  1. The making, shaping and treating of steel, 2003, The AISE Steel Foundation, Pittsburgh, PA − Chapter 13 [Google Scholar]
  2. F. Oeters, Metallurgy of steelmaking, Ed. Springer-Verlag, Berlin, 1995 [Google Scholar]
  3. Y. Sahai, T. Emi, Tundish technology for clean steel production, World Scientific Publishing Co., Singapore, 2008 [Google Scholar]
  4. M.J. Cho, I.C. Kim, Simple Tundish mixing model of continuous casting during a grade transition, ISIJ Int. 46(10), 1416–1420 (2006) [CrossRef] [Google Scholar]
  5. M. Alizadeh, H. Edris, A.R. Pishevar, Behavior of mixed grade during the grade transition for different conditions in the slab continuous casting, ISIJ Int. 48(1), 28–37 (2008) [CrossRef] [Google Scholar]
  6. A.K. Sinha, Y. Sahai, Mathematical modeling of inclusion transport and removal in continuous casting tundishes, ISIJ Int. 33(5), 588–594 (1993) [CrossRef] [Google Scholar]
  7. European Steel Commission − Science Research Development, Contract no. 7210-CA/841-431-336-189, Report EUR 19851EN, 2001 [Google Scholar]
  8. X. Huang, B.G. Thomas, Intermixing model of continuous casting during a grade transition, Metall. Met. Trans. B 27, 617–628 (1996) [CrossRef] [Google Scholar]
  9. R. Kumar, A. Maurya, I.H. Siddiqui, P. Jha, Some studies in different shapes of tundish intermixing and flow behavior, IV International Conference on Production and industrial engineering, CPIE-2016, India, Vol. 4., https://www.researchgate.net/publication/311806506, last consulted: 11.4. 2019 [Google Scholar]
  10. B.G. Thomas, Modeling study of intermixing in tundish and strand during a continuous-casting grade transition, Iron Steelmaker (ISS Transaction), Vol. 24, 12, Iron and Steel Society, Warrendale, PA, 1997, pp. 83–96. [Google Scholar]
  11. K. Michalek, K. Gryc, M. Tkadlečková, D. Bocek, Model study of Tundish steel intermixing and operational verification, Arch. Metall. Mater. 57(1), (2012) [CrossRef] [Google Scholar]
  12. I.H. Siddiqui, P.K. Jha, Effect of inflow rate variation on intermixing in a steelmaking Tundish during ladle change-over, Steel Res. Int. 87(6), (2015). DOI: 10.1002/srin.201500210 [Google Scholar]
  13. Krashnavtar, D. Mazumdar, Transient, multiphase simulation of grade intermixing in a Tundish under constant casting rate and validation against physical modeling, JOM 70(10), 2139–2147 (2018) [CrossRef] [Google Scholar]
  14. J.-H. Bi Jing-han, T. Ping, W.N. Guang-hua, D. Ling, S. Gang, Prediction model of intermixing slab length and location in continuous grade transition casting process, Chinese J. Process Eng. 12(2), 271–276 (2012) [Google Scholar]
  15. Zhang, Y. Yang, S. Li, J. Tang, H. Jang, The effect of a dissipative ladle shroud on mixing in Tundish: mathematical and experimental modelling, https://scholars.uow.edu.au/display/publication118430, last consulted: 11.4. 2019 [Google Scholar]
  16. M. Bartosiewicz, A. Cwudziński, Influence of immersion depth of ladle shroud in liquid steel on range of the transition zone for one-strand tundish during continuous casting of steel, https://www.researchgate.net/publication/320515089, last consulted: 11.04. 2019 [Google Scholar]
  17. K. Chattopadhyay, M. Isac, R.I.L. Guthrie, Physical and mathematical modelling to study the effect of ladle shroud misalignment on liquid metal quality in a tundish, ISIJ Int 51(5), 759–768 (2011) [CrossRef] [Google Scholar]
  18. G. Wang, M. Yun, C. Zhang, G. Xiao, Flow mechanism of molten steel in a single-strand slab caster tundish based on the residence time distribution curve and data, ISIJ Int 55(5), 984–992 (2015) [CrossRef] [Google Scholar]
  19. H. Zhang, Q. Fang, S. Deng, C. Liu, H. Ni, Multiphase flow in a five-strand tundish using trumpet ladle shroud during steady-state casting and ladle change-over, Steel Res Int 90(3), (2018). [Google Scholar]
  20. A. Maurya, P.K. Jha, Two-phase analysis of interface level fluctuation in continuous casting mold with electromagnetic stirring, Int. J. Num. Methods Heat Fluid Flow, (2018). DOI: 10.1108/HFF-08-2017-0310 [Google Scholar]
  21. I.H. Siddiqui, M.-H. Kim, Two-phase numerical modeling of grade intermixing in a steelmaking tundish, Metals 1(40), 9 (2019). DOI: 10.3390/met9010040 [Google Scholar]
  22. V. Battaglia, M. De Santis, V. Volponi, M. Zanforlin, Steel quality and transient flow in tundish casting: features and countermeasures through validated modelling, Proceedings of the 4th International Conference on Modelling and Simulation of Metallurgical Processes in Steelmaking (STEELSIM) − METEC InSteelCon® 2011 − Düsseldorf. [Google Scholar]
  23. www.ansys.com, User manual, last consulted 11.4.2019. [Google Scholar]
  24. O. Levenspiel, Chemical reaction engineering, chap. 11, J. Wiley & Sons, New York, 1999 [Google Scholar]

Cite this article as: M. De Santis, N. De Santis, D. Fera, R. Tonelli, S. Oktay, A. Oran, Online tool based on Tundish steel CFD model to monitor and minimise steel intermix in CC slabs, Matériaux & Techniques 107, 510 (2019)

All Tables

Table 1

Conditions simulated.

Conditions simulées.

Table 2

Comparison among first appearance time (normalised) of steel at tundish exit.

Comparaison entre temps (normalisé) de première apparition d’acier à la sortie du répartiteur.

Table 3

Dead volume calculated from C-curves derived from CFD calculations. Results independent from the flow rate.

Volume mort calculé à partir des courbes C obtenues après les calculs CFD. Résultats indépendants du debit.

Table 4

Parameters for the F-curve analytical relationship in (3).

Paramètres de la relation analytique de la courbe F dans (3).

All Figures

thumbnail Fig. 1

Steel intermixing schematic (1).

Schéma de mélange de nuances differentes d’acier (1).

In the text
thumbnail Fig. 2

C- and F-curves features to characterize mixing in tundish.

C- et F-courbes pour la description du mélange de nuances en répartiteur.

In the text
thumbnail Fig. 3

Velocity map (m/s) at surface and symmetry planes for the cases examined with no flow modifiers. For symmetry, ¼ of the tundish is represented.

Vitesse (m/s) à la surface et plans de symétrie pour les cas  sans modificateurs de flux. Pour la symétrie, ¼ des le répartiteur est représenté.

In the text
thumbnail Fig. 4

Velocity map (m/s) at surface and symmetry planes for the cases examined with pad. For symmetry, ¼ of the tundish is represented.

Vitesse (m/s) à la surface et plans de symétrie pour les cas avec un tampon. Pour la symétrie, ¼ du répartiteur est représenté.

In the text
thumbnail Fig. 5

Velocity map (m/s) at surface and symmetry planes for the cases examined with multiple hole baffles. For symmetry, ¼ of the tundish is represented.

Vitesse (m/s) à la surface et plans de symétrie pour les cas avec des chicanes à trous multiples. Pour la symétrie, ¼ du répartiteur est représenté.

In the text
thumbnail Fig. 6

Distribution of the dead zone (in black) after CFD calculations. No flow modifiers. For symmetry, ¼ of the tundish is represented.

Distribution du volume ’mort’ (en noir) après calculs CFD. Pas de modificateurs de flux. Pour la symétrie, ¼ du répartiteur est représenté.

In the text
thumbnail Fig. 7

Distribution of the dead zone (in black) after CFD calculations. Pad as flow modifier. For symmetry, ¼ of the tundish is represented.

Distribution du volume « mort » (en noir) après calculs CFD. Pad comme modificateur de flux. Pour la symétrie, ¼ du répartiteur est représenté.

In the text
thumbnail Fig. 8

Distribution of the dead zone (in black) after CFD calculations. Multiple hole baffles as flow modifiers. For symmetry, ¼ of the tundish is represented.

Distribution du volume « mort » (en noir) après calculs CFD. Déflecteurs à trous multiples comme modificateurs de débit. Pour la symétrie, ¼ du répartiteur est représenté.

In the text
thumbnail Fig. 9

Images of water modelling trials with colour tracer at the CSM labs.

Images d’essais de modélisation de l’eau avec un traceur de couleur dans les laboratoires du CSM.

In the text
thumbnail Fig. 10

C- and F-curves for all the cases examined after CFD calculations.

Courbes C et F pour tous les cas examinés après calculs CFD.

In the text
thumbnail Fig. 11

Overview of C-Curves from CFD and fittings for each case. Low flow rate.

Vue d’ensemble des courbes C de CFD et des raccords pour chaque condition. Débit faible.

In the text
thumbnail Fig. 12

Close-up of zones with maximum fit error. Use of multiple-hole-baffle, low flow rate.

Gros plan des zones avec une erreur d’ajustement maximale. Utilisation de chicane à trous multiples, faible débit.

In the text
thumbnail Fig. 13

Comparison among fitted C-curves. Low flow rate.

Comparaison entre les courbes C ajustées. Faible débit.

In the text
thumbnail Fig. 14

Model data validation by chemical analysis on slab samples (points on the curve) and comparison with plant process parameters (screenshot of the CC2 process control system on the right).

Validation  du modèle par analyse chimique sur des échantillons de brame (points sur la courbe) et comparaison avec les paramètres du processus de l’usine (capture d’écran du système de contrôle du processus CC2 à droite).

In the text
thumbnail Fig. 15

Screenshot of the user interface (top, input; bottom, output) of the off-line tool developed and used at Colakoglu plant.

Capture d’écran de l’interface utilisateur (haut, entrée; bas, sortie) hors ligne développé et utilisé à l’usine de Colakoglu.

In the text
thumbnail Fig. 16

Output chart for the offline tool developed.

Graphique de sortie pour l’outil hors ligne développé.

In the text
thumbnail Fig. 17

Information flow for the on-line tool developed.

Flux d’informations pour l’outil en ligne développé.

In the text
thumbnail Fig. 18

Screenshot of the on-line tool developed.

Capture d’écran de l’outil en ligne.

In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.