• Home
  • Technical
  • Fluid Flow, Flow Control and Modifying Devices in a Tundish

Fluid Flow, Flow Control and Modifying Devices in a Tundish

Fluid Flow, Flow Control and Modifying Devices in a Tundish

For the transfer of liquid steel from a steel teeming ladle to the moulds of a continuous casting machine (CCM), an intermediate vessel, called a tundish, is used. The tundish is intended to deliver the liquid steel to the moulds evenly and at a designed throughput rate and temperature without causing its contamination by inclusions. The tundish acts as a reservoir during the ladle change periods and continues to supply the liquid steel to the moulds when incoming liquid steel is stopped, making possible sequential casting with a number of teeming ladles. The main causes for inclusion formation and contamination of the liquid steel include reoxidation of the liquid steel by air and carried over oxidizing ladle slag, entrainment of tundish and ladle slag, and emulsification of these slags into the liquid steel. These inclusions are required to be floated out of the liquid steel during its flow through the tundish before being teemed into the moulds.

The tundish is a big-end-up, refractory-lined vessel, which can have a refractory-lined lid on the top. The tundish bottom has one or more nozzle port(s) with stopper rod(s) or slide gate(s) for controlling the flow of liquid steel. The tundish is frequently divided into two sections namely (i) an inlet section, which normally has a pour box and where the liquid steel is fed from the ladle, and (ii) an outlet section from which liquid steel is fed into the mould(s). Different flow modifying devices (FMDs), such as dams, weirs, and baffles with holes, etc. can be arranged along the length of the tundish. The liquid steel flow path from inlet to outlet of the tundish is important. Longer path is preferred for prolonging the residence time of the liquid steel so as to promote the flotation of macro-inclusions.

The tundish is a refractory-lined vessel with a variety of possible geometries, but most commonly roughly rectangular in shape. It is designed to deliver liquid steel at a designed output rate without major fluctuations in the flow. The flow rate is mainly controlled by the depth of the liquid steel. Further control of the outlet flow can be performed by either stopper rods or slide gates. The number of outlets depends on the type of the cast product. Normally there are 1 to 2 moulds for a slab casting machine, 2 to 6 moulds for a bloom casting machine, and 3 to 8 moulds for a billet casting machine. The liquid steel flow transfer is a key operation of continuous casting which affects strongly solidified shell profile, floatation of the non-metallic inclusion, and mould flux entrainment by large meniscus level fluctuations.

Since the tundish is the final stage in the steelmaking process before continuous casting, it also presents the last opportunity for the control of the liquid steel composition. There is a constant demand for steel with improved properties, such as increased strength, ductility, durability, and corrosion resistance, which is needed for a large variety of applications. There is also the desire to make the steelmaking process more energy and cost efficient and to address the environmental concerns. These issues have promoted the evolution of the tundish into a metallurgical reactor, with the function of performing final control over the properties of the liquid steel before continuous casting for getting a final steel product with the desired mechanical properties.

Tundish has several functions in continuous casting. The primary role is to be a vessel where the liquid steel is distributed to different strands. In addition, tundish is necessary to provide a constant casting rate by controlling the liquid steel level. Tundish can also serve homogenization of steel composition and temperature if the liquid steel has proper flow and sufficient time to stay in the tundish so that the heat and concentration can be distributed homogeneously to the whole tundish. However, over the recent past, the tundish has a new important role which influences the quality of steel. Since it is the last vessel before mould in continuous casting, the inclusion removal in the tundish is necessary for clean steel production. Fig 1 gives the cross-section of a continuous casting machine where the tundish location is shown.

Fig 1 Cross-section of continuous casting machine

Tundish has become a fundamental component in clean steel production since it acts as a metallurgical reactor to remove non-metallic inclusions by allowing floatation of non-metallic inclusion towards the slag on the surface. Besides, tundish can also homogenize the steel composition and temperature in such a way by providing a proper flow which promotes floatation of inclusion and sufficient residence time so that the heat and concentration can be distributed homogeneously to the whole tundish. Hence, the liquid steel flow has an interesting parameter since it is an influencing factor for improving the tundish performance.

Presently, several tools for adjusting the liquid steel flow within the tundish has been developed as standard practices in casting foundries such as baffles, weirs, dam, or turbo-stopper. Those techniques, which normally are known as flow modifying devices, aim to optimize the liquid steel flow by creating optimum turbulence and mixing within the liquid steel. However, the consideration of flow control and flow modifying devices installation becomes more complicated in a multi-strand tundish. The large number of strands lead to increase the tendency of non-homogeneous liquid steel. The liquid steel flows to the strand closer to the ladle shroud only needs a shorter time to reach the strand so that it tends to have a higher temperature and contains more inclusions compared to the liquid steel which flows to the further strand. The uneven temperature of each strand can also lead to the different micro-structure of the cast steel.

In the recent past, the role of the tundish in the continuous casting process has evolved from that of a reservoir between the ladle and mould to being a grade separator, an inclusion removal device, and a metallurgical reactor. For both grade separation and inclusion removal, the flow patterns inside the tundish play an important role. For the former, the flow patterns determine the quantity of mixing which occurs, while for the latter, the paths of inclusion particles are influenced by the flow field. Several tundish designs and tundish flow control and the flow modifying devices exist which have as their aim the modification of tundish flow patterns, to either minimize mixed-grade length during transition, and / or to increase inclusion particle removal by the slag layer.

The designs of the flow modifying devices in a tundish vary, depending on the size, height, the internal design, and operating conditions. A flow modifying device has to be designed as per these parameters. Hence, the designs as well as the location of the flow modifying device is to change accordingly. With the proper design and placement of the flow modifying devices, liquid steels can be distributed in a uniform pattern as per the dimensions of a tundish. Hence, the single or twin strand tundish of continuous casting machine is used to make sure the liquid steel fluid is being transported in an even fashion to one, or both, of the strands respectively.

Several quality issues which originate during continuous casting can be directly attributed to the poor control of fluid flow conditions. The task of the flow system is to transport the liquid steel at a desired flow rate from the ladle into the mould cavity through the tundish and to deliver the liquid steel to the meniscus area which is neither too cold nor too turbulent. In addition, the flow conditions are required to (i) minimize exposure to air, (ii) avoid the entrainment of slag or other foreign material, (iii) aid in the removal of inclusions into the slag layer, and (iv) encourage uniform solidification. Achieving these somewhat contradictory tasks need careful optimization of the flow.

Flow objective in the ladle, tundish, and the moulds is to encourage uniform mixing of the liquid steel so as (i) to have a uniform composition, (ii) to agglomerate inclusions, and (iii) to encourage their removal into the slag layer at the top surface. Flow is driven by the injection of gas bubbles or by natural thermal convection. In the absence of gas injection, the differences in density of liquid at different temperatures creates considerable flow. This tends to stratify the liquid into distinct layers, where convection cells circulate. The flow challenge then is to optimize the accompanying gas flow to achieve the flow objectives while minimizing problems such as disrupting the surface slag layer.

The existence of non-metallic inclusion in steel leads to a detrimental effect on the mechanical properties of steel. Besides, non-metallic inclusion can also cause clogging problems in the nozzle of continuous casting. Also, the non-metallic inclusion in clogging can also be carried by the stream during the process, which results in non-homogeneous mechanical properties in steel product. Hence, the quantity and size of non-metallic inclusion are necessary to be limited.

One of the fundamental mechanisms of inclusion removal in the tundish is by promoting the floatation of non-metallic inclusion to the tundish slag. The inclusion can be then trapped on the surface so that it can be removed together with the slag. The floatation of inclusion spontaneously happens, since the density of inclusion is lighter than hot liquid steel. However, it is more problematic for small inclusions as the buoyancy force is highly dependent on the diameter of inclusion as can be seen in equation 1 which is ‘Vr = 2(Rinc)square x g(Df − Dinc)9Lvf, where  ‘Vr’ is the Stoke rise velocity, ‘Rinc’ is the radius of inclusion, ‘g’ is the gravity, ‘Lvf’ is the laminar viscosity of the steel, ‘Df’, and ‘Dinc’ are the density of liquid steel and inclusion respectively. Hence, it is evident that inclusions with the larger size can float to the surface more quickly than the smaller inclusions.

There are several mechanisms of inclusion removal in the tundish and one of the dominant mechanisms is by the Stokes floatation. Hence, the inclusion removal in the tundish is highly dependent on how long the liquid steel can stay in the tundish so that the inclusion has enough time to float towards the slag at the surface before it goes to the outlet of tundish. The time, which is called as residence time, can be affected by different parameters of the tundish, whether it is related with the tundish design or the process parameter. One of the parameters related to the design is the size of the tundish. A larger tundish results in a longer residence time since the liquid steel has to travel with a longer distance to the outlet. However, this option is not preferable because of the limitation of cost and space in reality. There is, however another way to promote the flotation of inclusion by adjusting the liquid steel flow in the tundish using flow control and flow modifying devices. The addition of these flow control and flow modifying devices can change the liquid steel pathway towards the surface as well as prevent any short-circuiting flow. However, it is to be noted that the low turbulence at the wall and the surface is to be maintained otherwise the exogenous inclusion because of the refractory wear or slag entrainment appears.

However, the addition of flow control and flow modifying devices is limited by the available space inside the tundish. In addition, it is also highly dependent on tundish design. The parameter, such as the location or the height of turbo-stopper, influences the mixing phenomena considerably. The different design of baffles can also create different phenomena which are very specific for every tundish. Also, the position of strand and casting parameter such as shroud immersion depth can also affect the performance of flow control and flow modifying devices. Hence, the consideration when selecting a type and design of flow control and flow modifying devices involve both tundish and casting parameters.

Another disadvantage is that the flow control and flow modifying devices are susceptible to wear for long time usage. This can affect the productivity, quality and as well as the total cost of its implementation. The last, the flow control and flow modifying devices cannot provide the adjustment of flow for the whole process time. The other method to adjust the liquid steel flow inside tundish is by performing a stirring by argon gas or electro-magnetic stirring. Theoretically, the stirring improves the probability of collision and agglomeration between inclusions resulting in a larger size and floatation velocity. In addition, it also helps the temperature and composition homogenization because of increase of mixing intensity. However, the gas stirring is limited by injection rate because of the possibility of slag entrainment which can occur. The limitation of using flow control and flow modifying devices and argon stirring can be partly solved by using electro-magnetic stirring.

The tundish provides a continuous flow of liquid steel from the batch ladle operation to the continuous casting machine. It also serves as an important metallurgical reaction vessel, where quality can be improved, maintained, or lost. The flow objective in the tundish is to encourage uniformity and inclusion removal, while avoiding flow-related issues. Tundish flow issues include surface turbulence, short-circuiting, dead zones, and vortexing. Excessive flow directed across the top surface can produce turbulence and lead to reoxidation and slag entrainment. Short-circuiting allows incoming liquid steel from the ladle to exit prematurely into the mould with insufficient time for inclusion flotation. Dead zones are stagnant, colder regions which prevent inclusion removal and can slowly mix and contaminate the new liquid flowing through the tundish. If the liquid level is too shallow, high-speed, asymmetric can produce vortexing, which can entrain surface slag down into the mould.

Flow behaviour in the tundish is governed mainly by the size and shape of the vessel and the location of flow-control and flow modifying devices. The flow pattern is also affected by the liquid steel flow rate and its temperature distribution. Thermal buoyancy tends to lift up the hotter, lower-density flowing steel, while colder steel tends to flow down the walls and along the bottom. A temperature difference of only a few degrees is enough to lift the jet flowing beneath the weir and completely reverse the flow direction in the second chamber of the tundish. Flow in the tundish is also highly affected by the ladle-tundish nozzle geometry and gas in the ladle stream. Issues related to surface turbulence can be reduced by avoiding excessive argon levels in the ladle stream and by using fully-shrouded and immersed nozzles. For producing an upward velocity component over a large portion of the liquid steel within the tundish is very important to the designs of liquid steel flows, since it is the essential mechanism which improves and has a large influence on the removal of macro-inclusions.

It has been shown that the optimum design of a tundish can eliminate dead zones and improves fluid flow control. The effect of different parameters such as inclination of tundish walls, the geometrical size of tundishes, and the double wall inclination tundish on fluid flow have been studied. The most important parameter is the length / height ratio. It has been found that the ratio is to be larger than 3.5 in order to reduce the quantity of fluid flow which is unable to ascend towards the surface and to proceed with the removal of the macro-inclusions. A wall inclination of 10-degree to 15-degree to the vertical improves the fluid flow, eliminating stagnant zones inside the tundish, while the double inclined wall largely dissipates the fluid motions, decreasing the intensity of the recirculation and has an effect in making the fluid less turbulent in the downstream half of the tundish.

In one of the studies, where the advantages of increasing the volume of the tundish have been discussed, it has been proposed that the longer is the residence time, the higher is the numbers of inclusions removed. In another study, the fluid flow performance inside a six-strand billet casting machine tundish in which one side is curved is compared against a conventional delta shaped tundish using a three-dimensional mathematical model. The curved area, along the pouring chamber of the tundish has improved inclusion flotation. It also has increased the plug flow as well as residence time distribution and the dead volume has decreased.

Transfer systems – Liquid steel flows from the ladle, through the tundish into the mould where it solidifies. Flow between vessels is driven by gravity. Between the tundish and mould, this driving force is proportional to the head of liquid steel between the top surface of the liquid steel in the tundish and the liquid steel level in the mould. Control of the flow rate into the mould is achieved by metering nozzles, slide gates / stopper rods.

The simplest means to control the flow rate is to choose the appropriate size of the opening in the bottom of the tundish which restricts the flow to the desired rate. This inexpensive method is used in conjunction with open-stream pouring of lower-quality steel, where some air entrainment can be tolerated. Air entrainment can be reduced, but not avoided, by surrounding the stream with argon gas through one of the shrouding methods as shown in Fig 2. The bellows appears to be the most effective of the choices shown in the figure.

Fig 2 Free-stream pouring practices between tundish and mould

Metering nozzles have the disadvantage that variations of the liquid level in the tundish produce variations in flow rate. These flow-rate variations are to be compensated for by changes in withdrawal rate of the strand in order to maintain a constant average liquid steel level in the mould. A second concern is that the condition of the stream is very sensitive to the flow pattern in the tundish and to any imperfections in the nozzle shape, such as oxide build-up or notches in the outside of the nozzle lip. The condition of the stream is important in open pouring since it directly affects the quality of the final steel product.

There are different stream conditions which need comparison. A rough, turbulent stream is to be avoided since (i) its irregular shape entrains much more air, leading to more and larger oxide inclusions in the product, and (ii) it creates fluctuations in the liquid level in the mould when it impacts the surface, leading to non-uniform solidification, surface defects, and even break-outs. A tight, smooth stream entrains less air and creates less disturbance of the meniscus region while it penetrates deeper below the liquid surface. Tight streams are hence very desirable. The funnel in Fig 2 helps to improve surface quality by substantially lowering meniscus turbulence, although it has little effect on air entrainment.

For avoiding reoxidation and produce higher quality steel, a protective ceramic nozzle can be used to shroud the flow of steel into the mould. The flow rate is controlled by restricting the opening with either a stopper rod or a slide gate.In case of slide gate flow control, liquid steel flows into the upper tundish nozzle, through the slide gate opening, down the long submerged-entry-nozzle (SEN) tube, and through the submerged entry nozzle ports into the mould cavity, as shown schematically in Fig 3a. In the three-plate slide gate, shown in the figure, the central plate is moved hydraulically for adjusting the opening between the upper and lower stationary plates by misaligning the hole in the sliding plate relative to the nozzle bore. Alternatively, the two-plate slide gate is missing the lowest plate, so the submerged entry nozzle is attached to the moving plate and travels as the opening is adjusted. This has the disadvantage of continuous variation in the alignment of the nozzle relative to the strand centre-line. In both systems, the joints are all flooded with low-pressure inert gas (argon) to protect against air entrainment in the case of leaks.

Fig 3 Schematic of continuous casting process showing tundish, SEN, and mould

Flow through the slide gate is governed by the size of the overlapped openings of the plates, as shown in Fig 4a. This opening size can be quantified in several different ways. Two popular measures are ‘area opening fraction’ ‘fA’, defined by the ratios of the shaded area to the total bore area, and ‘linear opening fraction’, ‘fL’, which is defined as the ratio of the distance ‘S’ to distance ‘T’.

Fig 4 Quantifying and effect of slide gate opening size

For equal sized openings, these different measures of opening fraction are related by equation 2, which is ‘fA = 2/pi(cos to the power -1)[1-L/D] – 2/pi[1-L/D]root x [1-{1-L/D}square], where L/D = (1 + R/D)fL – R/D (equation 3), ‘fA’ is gate opening (area fraction), ‘fL’ is gate opening (linear fraction), ‘L’ is length of opening in metres, ‘D’ is nozzle diameter in metres, and ‘R’ is the offset distance from nozzle bore to ‘reference line used to measure ‘S’ and ‘T’ in metres. It is to be noted that ‘fA’ is always slightly less than ‘L/D’, while ‘fL’ is always more than ‘L/D’. The linear opening fraction, ‘fL’ in equation 2, simplifies to exactly ‘L/D’ if ‘R’ is zero.

The stopper rod system is shown schematically in Fig 3b. Flow control with a stopper rod is slightly more difficult than with slide gates since the stopper rod is to be manipulated through the entire depth of the liquid steel in the tundish, and the area of the annular opening which controls the flow and is more sensitive to displacement. In addition, a continuous nozzle does not allow fast exchange of submerged entry nozzle tubes and needs some other means for emergency flow stoppage.

However, the stopper rod offers several significant advantages over slide gates consisting of (i) natural prevention of liquid steel from entering the upper tundish well and freezing prior to startup without the need for special flow control devices, (ii) natural prevention of vortex formation above the tundish well and possible slag entrainment into the nozzle when the liquid level is low, (iii) easier sealing to avoid air entrainment because of the reduced number of moving surfaces, and (iv) more uniform distribution of flow to both ports, so that flow entering the mould cavity is more symmetrical.

The steel flow rate depends mainly on the height of liquid steel in the tundish driving the flow and the pressure-drop across the slide gate. Flow rate naturally increases with increasing slide gate opening position and with increasing tundish depth, as quantified in Fig 4b for typical conditions. The flow rate is also influenced by the nozzle bore size, the quantity of gas injection, the constriction of the ports, and the extent of clogging and wear. The flow rate produced under ideal conditions, such as presented in Fig 4b, can be compared with the measured flow rate in order to identify the extent of nozzle clogging.

Flow modifying devices (FMDs) – While linings are an important part of the tundish refractory system, the major work for removing inclusions is being carried out by the refractory shapes in the tundish known as flow modifying devices. If the turbulent energy from the ladle shroud can be harnessed, the steelmaker can then focus on metallurgical measures to improve inclusion flotation in the tundish. Principal among these flow modifying devices are dams, weirs, and baffles etc., which serve to assist in directing inclusions upward to a captive tundish slag layer.

The contour shape along the pouring chamber can be useful in reducing the recirculation of the liquid steel fluid, rather than using sharp edges. In recent years, efforts have been oriented into the internal design of the tundish. Different designs of flow modifiers within the tundish have been studied by mathematical and water modeling, with the main objective of having better fluid flow characteristics. The flow modifying devices can be pour-pads, weirs, dams, baffles with holes and turbulence inhibitors. The turbulence inhibitor controls the degree of splashing of liquid steel during ladle changes or at the beginning of the casting sequences. It is also useful in increasing the plug flow, the residence time, and decreasing the dead volume.

The turbulence inhibitor reorients the impingement jet from the ladle shroud upwards. This is useful for inclusion flotation purposes and in the prevention of high liquid steel velocities in the flowing liquid, which can cause lining wear of the walls. Weirs are used to control the incoming flow which disturbs the surface and are placed in the higher part of the tundish. The dams are placed on the floor and do not cover completely the transversal area of the tundish. The baffles cover completely the bath surface height, and are structured with internal angled holes, to distribute the inclusions towards the bath surface. The flow modifying devices are used alone or with different combinations. Fig 5 shows schematic diagram of a tundish showing flow control and flow modifying devices.

Fig 5 Schematic diagram of a tundish showing flow control and flow modifying devices

Flow modifying devices are typically manufactured by casting shapes with 70 % alumina refractories, through occasionally areas subject to strong erosive flow can necessitate a switch to magnesite-based refractories. Less costly refractory grades, such as 60 % alumina, can be used to manufacture flow modifying device shapes but have shown a propensity for premature failures in field trials when tundishes are used for multiple heat sequences.

For the purpose of definition, dams are barrier devices placed on the bottom of the tundish which force steel to flow over their top. Dams force inclusions upward. Weirs, on the other hand, are refractory plates hinged on the top of the tundish lining which leave a gap between their bottom and the tundish floor, hence forcing the fluid to flow downward and underneath the shape. Weirs are normally effective in containing carry-over ladle slag from reaching to the well nozzles. Baffles, which are the third major category of the flow modifying devices, are straight-through walls which span the height of the tundish from its floor to its top and which incorporate a series of holes to allow passage and directional flow of the liquid steel. It is typical in baffles that the upper rows of holes are horizontal, while the lower holes are angled upward and even outward at angles up to 60-degree. This is done for preventing slag metal mixing during ladle exchanges when the tundish level drops. Examples of dam, weir, and / or baffle combinations for a single strand tundish are shown in Fig 6.

Fig 6 Schematics of complex configurations of flow modifying devices

Advances made in flow modifying devices include the recognition that increasing the surface area on the shape is beneficial for the inclusion capture. This can be achieved simply with a rough surface finish or through more intricate designs such as dimpled surfaces. Similarly, openings in baffles, as well as drainage holes in dams, have evolved from their traditional round or elliptical shape to include fluted and castellated openings with low flow areas designed to capture inclusions.

Besides dam, weir, and baffles, the fourth type of flow modifying devices is a group collectively termed as the impact pad. These pads, which range from simple rectangular plates to complex shapes with lips for reducing the energy of the impact stream, are inserted on the tundish bottom immediately below the ladle shroud region. Impact pads have been developed to cushion the tundish bottom from the ladle stream impact. Without such a pad, the impact energy can eventually work through the spray lining and result in steel penetration into the back-up shapes. Like dams, weirs and baffles, these high-alumina refractory plates are installed on the back-up lining and are anchored using the spray lining. Within the recent past, large quantity of work has been done to turn the simple impact pad into a turbulence-inhibiting device and a true flow modifying device. The working principle of these high-tech impact pads is to revert the impact stream onto itself, hence using own energy of the stream for dampening it.

The primary function of the flow modifying devices in the tundish is to provide a more twisting path for the liquid steel during its residence time in this vessel. These flow modifying devices prevent short-circuiting, help to promote inclusion separation, and can impart chemical and thermal homogeneity prior to casting. The efficiency of these devices is best assessed by means of physical and / or mathematical modeling prior to their installation. The characteristic information such as plug, mixed and dead flow fractions can be derived from residence time distribution (RTD) curves for a particular flow modifying device set-up.

In multi-strand tundishes, the placement of flow modifying devices is normally symmetric around the centre-line of the tundish, or more specifically around the centre-line of the ladle shroud. The role of flow modifying devices here is expanded for including minimization of discrepancies between the different strands. This is particularly important in billet and bloom casting machine tundishes, where the natural flow phenomena (i.e., no flow modifying devices) produce extreme short-circuiting, with order of magnitude discrepancies in minimum residence time not being unexpected.

A good example of how flow modifying devices can be designed to minimize such discrepancies can be found in a study which has shown that asymmetric baffles can be used to provide metallurgical quality benefits in four-strands and six-strand caster tundishes (Fig 7a). These baffles prevented short circuiting while allowing more calm passage of the liquid steel from the turbulent pour region to the remaining tundish volume. Benefits have been observed in terms of decreased inclusion levels, a lower incidence of entrapped surface slag on billets and blooms, and in improved thermal homogeneity between the strands.

Fig 7 Flow patterns and cleanliness rating

It is to be understood that there exists no universal solution in terms of flow modifying devices. A configuration which apparently works excellently in one plant, can be an inferior solution for another steel melting shop. Normally speaking, a weir / dam combination is neither superior nor inferior to baffles, though in specific cases, one flow modifying device can produce considerably better results than the another. An example of this is shown in Fig 7b, where a ‘slotted dam’ (which really has to be termed a slotted baffle to remain consistent with the terminology normally used) produced better cleanliness rating results than a conventional baffle, which has previously provided a more homogeneous flow distribution and improved cleanliness over a weir / dam combination. By contrast, a similar slotted baffle has produced worse fluid flow results than a weir / dam combination for the single-strand casting machine shown in Fig 6. Similar examples abound, and the lack of an apparent cure-all solution is not to be taken too lightly by the casting machine operators. Continuous improvement is to drive the process, and this can involve a radical but controlled change in the flow modifying device refractories.

In short, the flow modifying device refractories in the tundish are necessary for uninterrupted, safe, and, of course, profitable quality production in the continuous casting process. Sacrificial in nature, the flow modifying devices have a measurable cost and have quality impact on the steel production process. Hence, these devices are to be controlled as much and as well as the liquid steel itself. The devices also need expertize. In harnessing the flow and energy of the liquid steel in the tundish, the flow modifying device refractories help to turn this vessel into a continuous refining vessel, both through their physical presence and, as shown, by controlled chemical reactions.

A wrong selection or application of th flow modifying device materials can have disastrous results. The safety of the operating personnel needs to be first and foremost on the mind of the designer and operator in the selection process of the flow modifying devices. Supreme to protecting against the potential for steel penetration and eventual tundish break-outs, there is necessity not only to select materials with the appropriate insulation, hot strength, and erosion-resistance properties, but also to avoid straight through joints. Steps in the refractories provide opportunities for steel to freeze, if it penetrates, with the skulled steel providing some measure of retention of the overlying liquid.

Filters and flow modifiers – benefits to the casting operation – Ceramic foam filters (CFFs) are an example of the potential offered by cross-pollination of technology. Already routinely used in the commercial production of foundry irons, specialty steel castings, and aluminum manufacturing, ceramic foam filters have been increasingly studied for tundish applications. As per Hascher, there are five basic issues which are to be addressed in considering filtration process. These are (i) mechanical and thermo-mechanical stability, including priming, (ii) chemical stability, especially at long contact times, (iii) high filtration efficiency, (iv) capacity to feed the entire casting operation, and (v) cost effectiveness.

With requirements for high thermal shock resistance, quick priming, and the need to handle large quantities of liquid steel while maintaining the full casting rate in the continuous casting machines and not interrupting operations, the material of choice for filters is frequently magnesia-stabilized zirconia. Magnesia and alumina ceramics have also been used. Providing that the materials can withstand the high temperature environment for their designed exposure period, benefits being  highlighted for ceramic foam filters include lower operating costs because of a lesser number of submerged entry nozzle exchanges, fewer down-grades because of clogging, and a reduction in the number of sliver surface defects. Sliver defects are frequently caused by the entrapment of fine nonmetallic inclusions at the meniscus. These defects normally do not manifest themselves until after the product has reached a certain aspect reduction, e.g., cold rolled sheet in flat rolled products.

There are several types of ceramic foam filters, ranging from simple sponge-like foam filters through multi-hole reticulated cell filters and loop filters reminiscent of chain-mail armour to the more modern filter / flow modifiers (FFMs). The latter is a linear cell filter in which the perforations are designed not only to capture inclusions, but also to streamline the incoming eddies into an aligned laminar-type flow down-stream of the filter. Fig 8a shows types of these filters.

Fig 8 Ceramic foam filtration of liquid steel

Several mechanisms have been identified which describe how ceramic foam filters provide inclusion capture. These are shown schematically in Fig 8b. Mechanical sieving, also referred to as screening filtration, prevents passage of particles which exceed the cell dimension of the filter medium. Adherence separation is the principle which also underlies deep-bed filtration, where particles are captured along the walls of the filter medium as a result of surface forces. Cake filtration refers to straight sieving by a surface build-up over the filter medium induced by either or both of the prior mechanisms. In this sieving mechanism, particles blocking the passages and surface sites of the filter medium result in the build-up of a cake layer which is capable of filtering both small and large inclusions. It is normally the cake filtration phenomenon which dooms the filters to failure, since a caked filter interferes with normal casting operations.

In filtration, it is normal to classify ceramic foam filters and filter / flow modifiers by ‘pores per centimeter’ number. The higher the pores per centimeter number, the smaller the pore dimensions in the filter / flow modifier, and the more likely it is to entrap inclusions by mechanical or cake sieving. One concern frequently cited with filters is that it needs relatively few inclusions to clog them up. As shown in Fig 8c, single inclusion is to be extremely large to completely clog a 10 pores per centimeter number filter / flow modifiers, even a 200 micrometers inclusion, which is enormous by clean steelmaking standards, occupies only a fraction of the pore opening. This can be different in foam and loop ceramic foam filters, since the overlapping nature of successive layers of pores can result in premature cake filtration phenomena. This also serves to show the advantage of rectangular cell filter / flow modifiers, which provide a sieving performance equivalent to square and circular arrays but allow a higher percent open area.

Priming is an important consideration in the applications of ceramic foam filters and filter / flow modifiers. Priming refers to the period in which liquid steel first touches the filter. There are two components to the priming action. The first is thermal priming which is necessary to heat the ceramic to steelmaking temperatures such that liquid steel can flow through the ceramic foam filters and filter / flow modifiers without solidifying. The second is capillary priming which is dictated by the need to provide sufficient ferro-static pressure to overcome capillary effects and force the liquid steel through the ceramic foam filters and filter / flow modifiers. Capillary priming head is a function of orifice opening, liquid steel density, and interfacial tension, and it increases directly with the pours per centimeters number.

In tundish applications, ceramic foam filters and filter / flow modifiers are to be designed to meet steady-state and non-steady-state filtration requirements, and are not to interfere with normal casting throughput requirements. Filtering applications in multi-strand billet casting machines have shown to stabilize mould level considerably, which is visible by a considerable reduction in the number of stopper rod ‘bumps’ needed during casting.

Other applications have shown considerable reductions in total oxygen levels between filtered and unfiltered applications in liquid steel throughout an entire ladle casting period, including ladle change (Fig 8d). It is worth mentioning the work conducted at Kobe Steel, where a laboratory- versus-plant comparison of filtering in the tundish has been performed. A comparison of the inclusion size distributions between the laboratory and plant (Fig 9a) shows that, although filtering efficiency is overstated in the laboratory, the plant results still provide for considerable reduction of oxide loading to the mould. Also, the data suggest that inclusions higher than 20 micrometers does not reach the mould when a filter is used.

Fig 9 Inclusion size distribution and effect of tundish filters

An alternative to ceramic foam filters and filter / flow modifiers has been to design flow modifying devices with mechanical sieving and adherence separation features which incorporate lime-bearing materials (Fig 10). The lime works to assimilate filtered inclusions and ideally transforms them into liquid-phase calcium aluminates which reduce the clogging tendency in the system. Results from industrial interstitial-free (IF) quality liquid steel comparing a dam / weir / dam flow modifying device set-up with a dam / filter-baffle / dam set-up have been mixed, with high titanium ultra-low-carbon liquid steels show a 40 % reduction in overall inclusion population, while conventional ultra-low-carbon (ULC) liquid steels have shown little change. Here, the filter-baffle is a ‘filter dam’, and the triple dam set-up is a dam / weir / dam set-up. It is to be noted that the results of this industrial trial have shown impressive results where the number of fine inclusions (size 5 micrometer maximum) have decreased along with the larger inclusions when the filter baffle has been used.

Fig 10 Industrial trials with a lime bearing filter baffle

Similar data has been reported in another study. A comparison of sub-surface inclusion concentration under three conditions have been performed. The single-strand tundish contained a single baffle and has contained 33 tons at full operating depth. In all cases, the casting rate has been 5 m/minute. Only the flow modifying device material has been changed, ranging from no filter to alumina and lime filters. The data in Fig 9b show that a lime-bearing filter flow modifying device can produce considerable inclusion modification in a tundish supporting high casting speeds on an intermediate thickness continuous casting machine. In a true case of ‘not all filters are created equal’, the use of an identical filter composed of alumina in the same application actually intensified the population of inclusions measuring less than 100 micrometers. It is also to be noted that the study has also determined a considerable increase in dead flow fraction in the tundish when either filter baffle rather than a conventional baffle has been used, though no clear explanation can be provided for this observation.

It can be surmised that the use of ceramic foam filters and filter / flow modifiers in tundishes is an application full of promise but with little to show in terms of a proven track record. This field can yet evolve into a significant component of tundish metallurgy, especially as flow modification, turbulence suppression, and shrouding systems reach their capability limits in terms of removal of small inclusions. It is here that filtering can make considerable inroads. The emerging role of ceramic foam filters and filter / flow modifiers suggests that filtration systems are likely to be used on specific grades for eliminating the identifiable issues, rather than across the board where the operating improvements can be more difficult to justify the cost.

During the recent past, several different techniques, including physical and mathematical simulations, have been used in order to study the transport processes in tundishes. These include tracer dispersion, agglomeration and float out of inclusions and the transport of momentum and energy. The physical phenomena, in nature, are convective and turbulent diffusion processes. As such, the correct prediction and analysis of the turbulent flow fields is fundamental in understanding and analyzing the other important variables. Special emphasis in this area is given to the prediction, reliability, and visualization of the velocity fields within the liquid steel in the tundish.

Some of the physical techniques which have been used for these studies, are ‘particle image velocimetry’ (PIV), and ‘laser doppler velocimetry’ (LDV) for visualizing the velocity fields. The tracer addition technique is used to predict the ‘residence time distribution’ (RTD). This has been used extensively in different studies rather than the PIV technique, and is mainly applied to validate mathematical models by comparing the similarity in the two RTD curves got, i.e. by physical experiments and by mathematical simulations. One of the studies describes the procedures and important features for analyzing RTD curves in multi-strand tundishes. Here, a description for getting the measurements of different flow volumes fractions such as dead, plug, and well mixed etc. is reviewed.

For the mathematical simulations, commercial software packages are normally used nowadays mainly because of the powerful framework which they have for coupling different phenomena, the variety of options available to treat different turbulence models, and the reliability when comparing with physical experimentation and the visualization tools for the data post-processing. For the tundish fluid flow analysis, different software and turbulence models are used. The different variables which are taken into account when analyzing tundish systems, are boundary conditions, inlet turbulence parameters, turbulence models, and nodal configurations, for assessing the basics of mathematical modeling and to get good agreement with physical simulations.

Initial approaches in characterizing the fluid flow are by combining mathematical and physical modeling of a full-scale water model tundish. These approaches have been very useful in improving the operating conditions and arrangements of flow modifiers inside a tundish. Recently, some studies have been dedicated towards the efforts for maintaining the fluid for longer periods of time inside the tundish and improving the performance of the normally used flow modifying devices such as weirs, dams, and baffles etc.

Some approaches have been oriented towards giving a rotational motion to the fluid using different flow modifying devices during the stay of liquid steel inside the tundish. One of the studies has proposed a new design of a ladle shroud named the ‘swirling ladle shroud’ (SLS). This consisted of a pipe plus three intermediate chambers of larger diameter which works as buffers to the fluid velocity. At the top, the shroud is equipped with a blade whose function is to initiate the swirling motion. Fig 11a shows the swirling ladle shroud.

Fig 11 Swirling ladle shroud and tundish with swirling chamber

The swirling ladle shroud controls the jet entering inside the tundish, decreasing the impact velocity to a third at the bottom of the tundish, as compared to that of the conventional ladle shroud. Different physical and mathematical simulations, including the particle image velocimetry technique and a variety of turbulence models, such as ‘Reynolds stress model’ (RSM) are being used. The use of ‘Reynolds stress model’ is desired in the tundish when dealing with swirling motions, since the predicted velocity fields are in good agreement with the results got by the physical experimentation executed by the ‘particle image velocimetry’ technique.

There are experiments which have been carried out in the swirling ladle shroud involving non-isothermal conditions and the role of this flow control device for the removal of inclusions. Under thermal gradients, flows intensify the formation of vortices close to the outlets. This is harmful since when inclusions are entrained by these vortices, they are directed towards the outlets. As the swirling ladle shroud decreases the turbulence at the entry of the jet into the tundish, the entrainment of slag is decreased if compared with the conventional ladle shroud.

One of the studies has proposed a ‘swirling flow tundish’ (SFT), whose main objective is to give the entry jet a swirling motion through a cylindrical ‘swirling chamber’ (SC). To analyze the fluid motion inside the swirling flow tundish, physical and mathematical simulations have been executed. The trials have been done using an asymmetrical one-strand 1:25 and a symmetrical two-strand 1:3 scale model. The trajectory of the inclusion has been studied using the ‘discrete phase model’ (DPM). It has been shown that the swirling flow tundish is more efficient in the removal of small inclusions with sizes in the range of 20 micrometers than a tundish fitted with a turbulence inhibitor. This can be important for a high-quality product in the steel industry. One important characteristic is that the swirling chamber is in an asymmetric position, leading to asymmetrical flows along the length of the tundish with the possibility of vortexing flows and entrainment of upper slag phase. Fig 11b shows (a) single strand tundish with the swirling chamber arrangement, and (b) two-strand tundish fitted with the swirling chamber arrangement.

Another new proposal, also following along the lines of creating a centrifugal velocity inside a single-strand tundish has been developed, using physical tracer experiments and mathematical simulations. Here, the effects of the dam spacing and rotation speed of the fluid flow have been analyzed. The mechanism of inclusions flotation is based on the rotation of the fluid in the centrifugal chamber induced by electro-magnetic forces. This procedure has been demonstrated to have good results during plant trials. After the results got by the mathematical simulation using liquid steel as a fluid, it has been concluded that the rotational motion induced by the magnetic field in the centrifugal chamber reduces the dead volume related to the position of a dam. By reducing the space in the dam, the evolution of the rotational flow increases the flow path, resulting in a larger fraction of plug flow. The highest ratios of plug to dead volumes are reached when the dam spacing cause a stronger transverse circulation. A balanced magnetic field has been suggested for maintaining a balanced rotational motion.

One of the studies has simulated mathematically a centrifugal flow tundish by the ‘large eddy simulation’ (LES) technique. This technique model offers a deeper insight into transport phenomena regarding to the fluid flow taking place under transient conditions. Three cases have been studied namely (i) electro-magnetic force with the direct nozzle (EDN), (ii) bending nozzle (BN) with height potential energy of liquid steel, and (iii) a combination of electro-magnetic force and bending nozzle (EBN). It has been found that when the electro-magnetic force with the direct nozzle arrangement is applied, inside the rotation chamber, two strong recirculations are generated. In the case of bending nozzle, the horizontal swirling gets stronger than the vertical segment. In the last case, combination of electro-magnetic force and bending nozzle, the evolution of the swirling flow is highly improved. The study found the bending nozzle increases the velocity in a comparable range with that of the electro-magnetic force with the direct nozzle case, which makes the bending nozzle a simple and alternative option for centrifugal tundish operations. The large eddy simulation technique has been used less than other turbulence models.

Another important factor which is nowadays taken into account in different studies is the coupling of different phases and phenomena in order to get a more realistic view of the actual transport processes occurring. The volume of fluid (VOF) has been considered to simulate the interaction between the phases, taking into account surface tension. Using different positions and conditions of argon injection inlets, some studies are able to predict a decrease of the slag layer opening, hence making the multi-phase modeling an important approach to increase the resemblance to the actual processes.

It can be seen that the present trends are to reduce as much as possible the usage of flow control and flow modifying devices inside the tundish, and to improve the mathematical models by coupling several phenomena involving different phases. Large computational investments into parallel processing are needed. In some parts of the world, the tundish has been increased in its length, and depth. Similarly, it is treated by plasma reheating in order to maintain isothermal conditions during casting. This can reduce the use of flow control and flow modifying devices consumption. The present efforts are now more oriented towards the development and optimization of mathematical models compared to physical experiments, leading to a high computational investment.

It has been shown that flow control and flow modifying devices are useful tools for controlling the fluid flow through a tundish. In general, the turbulence inhibitor (TI) is the most efficient tool, alone or together with other flow control and flow modifying devices. A reliable prediction of the turbulent flow fields is fundamental to the understanding and analysis of the passage of liquid steel through the tundish. Hence, techniques such as particle image velocimetry and residence time distribution are used extensively in this area. Overall, the residence time distribution technique is more widely used rather than lasers when analyzing fluid flow in tundish, mainly because of its simplicity, economic advantages, and the useful information obtained from them, such as, plug volume, and dead volume etc. However, if an in-depth observation of the fluid flow is desired, the laser techniques are more detailed and superior. Nowadays, a combination of residence time distribution technique with mathematical modeling is more common for analyzing fluid flows in tundishes. Increasing reliance is being made of fluid flow visualization through mathematical simulations, carried out using commercial packages, such as FLUENT-ANSYS, and COMSOL etc.

On the other hand, it can be seen that the present trends are to reduce the usage of flow control and flow modifying devices. Larger designs of the tundish, combined with external heating, can be useful towards achieving larger averaged residence times along the length of the tundish, as well as an efficient temperature distribution. The use of electro-magnetic forces in the tundish is an interesting approach which can be useful in increasing the flotation of micro-inclusions. However, these techniques have not been fully implemented yet. Hence, the control of the fluid flow through the ladle shroud, as seen in the swirling ladle shroud, can be a simple solution and a reliable option to achieve higher steel qualities. However, potential difficulties such as clogging by entrained inclusions need to be tested.

The particular combination of flow control and flow modifying devices can still be used to overcome the similarity problem in the multi-strand tundish. Several studies related to the use of flow control and flow modifying devices in multi-strand tundish have been conducted.  One of the studies has used a physical model of the ten-strand tundish and it has been stated that the combination of the specific design of turbulence inhibitor and baffles can reduce the inclusion around 42 % and lead an evener distribution of inclusion among the strands. This result has a good agreement with another similar study carried out for six-strand tundish by numerical modelling and industrial measurements. It is concluded that specific design of baffle can reduce the transient zone in the tundish.

However, there are some limitations when using flow control and flow modifying devices in the multi-strand tundish. Firstly, the multi-strand tundish normally has insufficient working spaces which make challenging to install several flow control and flow modifying devices. In addition, the shape and design of flow control and flow modifying devices are also very case sensitive, which means the selection of flow control and flow modifying devices are to involve both of tundish design and casting parameters. Moreover, the refractory material of the flow control and flow modifying devices is also susceptible to wear during the long practice so that it affects the productivity, and quality as well as the total cost of its implementation. Also, the flow control and flow modifying devices cannot provide the adjustment of flow for the whole process time.

One of the solutions which can overcome the above problems is to use electro-magnetic stirring (EMS) installed on the tundish to control the flow of the liquid steel. One or two electro-magnetic stirrers which is installed on the side wall of tundish can provide stirring during the whole casting process hence there is a possibility to replace or simplify the flow control and flow modifying devices using electro-magnetic stirring technology. Theoretically, a horizontal stirring generated in the liquid steel mixes the liquid steel and adjust the flow of the liquid steel so that the composition and temperature can be homogenized. From the flexibility point of view, it can be adjusted easily based on the tundish design and the casting process. Also, although the electro-magnetic stirring equipment is more expensive than flow control and flow modifying devices, the longer lifetime and lower running cost makes the total cost becomes comparable. All these advantages make this technology seems to be promising in the future.

However, a comprehensive analysis of the liquid steel flow generated from electro-magnetic stirring has not well-understood because of the lack of studies related to the application of this technology in the tundish application. Hence, the focus of some studies is to investigate the possibility to replace flow control and flow modifying devices with electro-magnetic stirring in multi-strand tundish application by comparing the flow pattern generated in the tundish for both cases. These studies have been conducted through physical modelling of eight strand tundish and numerical modelling of the experiment. The water model experiment has been focused on revealing the effect of stirring while the numerical modelling has been used to study the effect of such flow control and flow modifying devices. The analysis of flow movement, residence time distribution curve and dye-colour injection have been used to measure two parameters of tundish performance, i.e., flow characteristics and strand similarity in different tundish configurations. In addition, a simulation of inclusion injection has also been conducted to get more understanding regarding the effect of baffle wall and turbo-stopper on inclusion removal.

Leave a Comment