News

Beam Blank Casting Technology

• satyendra
• May 14, 2014
• 1 Comment
• beam blank, continuous casting, flange, mould, SEN, strand, Tundish, web,

Beam Blank Casting Technology

The recent progress in continuous casting of steel has been remarkable as demonstrated by a noticeable rise in the adoption of the continuous casting technology. Different kinds of developments have been made aimed at achieving higher efficiency of operation and higher quality of the products. One of such developments is the continuous casting of the beam blank whose cross-sectional shape is as close as possible to the shape of the final product.

Near-net shape continuous casting offers efficient alternatives to the traditional continuous casting of slabs, blooms, or billets. The conventional 250 mm thick slabs have been replaced by thinner sections with thickness in the range of 50 mm to 90 mm, starting with the introduction of thin-slab casting machines in the late 1980s. Similarly, blooms have been replaced by dog bone-shaped beam-blank sections, the casting of which has been developed in the late 1960s. Beam blank casting is now a mature process.

Beam blank is ideal for the production of the structural sections of I-beam and H-beam. I-beam and H-beam steel sections are economic type of sectional materials which are widely being used in several constructional areas.  For gaining the highest benefit from beam blank casting technology, a direct coupling of the casting machine to the rolling mill is necessary. The development of the direct casting of beam blanks is one of the most outstanding success stories in the evolution of the continuous casting of steel. The continuous casting of near net shape cross sections, called beam blanks or dog bones, has been an efficient commercial process for the production of I-beams and H-beams.  Casting of these sectional shape saves not only on rolling costs, but also offers higher productivity and improved energy efficiency.

The first beam-blank continuous casting machine was commissioned at Algoma Steel (now Essar Steel Algoma Inc., Sault Ste. Marie, Canada) in 1968. As with several other innovations, the relatively conservative steel industry needed some time to accept this revolutionary concept. Its successful application depended on the inter-disciplinary co-operation and on the optimizing of casting and rolling processes. This pioneering effort immediately attracted wide interest not only in the steel industry, but also by academic world, e.g., studying of the solidification pattern of this complex strand shape. However, it took another five years until the next beam blank casting machine got off the ground at Mizushima works of Kawasaki Steel Corporation (Now part of JFE holdings)

Beam blank casting offers similar advantages for the production of medium and heavy sections because of the near net shape casting which are offered by the thin slab casting for the production of flat products. For the manufacture of I-beam and H-beam in the upper weight range, the application of cast beam blanks instead of conventional blooms is a very good alternative.

Direct casting of small size beam blanks reduces considerably the number of rolling passes. Normally, starting from a standard bloom / billet, 6 numbers of break down passes plus 10 numbers of finishing passes are needed for the shaping of the IPE 100 section (100 mm x 50 mm x 5 mm). If the size of cast beam blanks is reduced to 110 mm x 70 mm x 12 mm (25 kilograms per metre) then only a total of 6 numbers of passes are needed to shape the IPE 100 section.

Near net shape beam production is also one of the interests of beam producers around the world because of the reduction in the costs induced by reheating and rolling of beam blanks. This is done by combining the casting of near net shape beam blanks (web thickness of 50 mm) and direct rolling. The grade of the near net shape is determined by the needed minimum rolling passes to get the desired metallurgical micro-structure.

The plant based on the near net shape beam blank concept is very compact. It primarily consists of steelmaking furnace (mostly an electric arc furnace), a casting machine for the casting of the beam blank, a heating furnace to heat the beam blank for reaching the temperature distribution needed for rolling, a conventional break-down stand, and a U1-E1-U2 universal stand group which includes a universal roughing stand, a 2-high edging stand, and a universal finishing stand. The economic advantages of the beam blank casting can mainly be attributed to the rolling process. Because of the near net shape of the beam blanks, lesser rolling work is needed in the break down mill for achieving the necessary cross section for further rolling operation. For example, in one of the plants, earlier the beam IPE 300 was produced from an 80 mm x 300 mm rectangular bloom cross section by the application of 11 rolling passes in the break down rolling mill stand. With the usage of the beam blank cross section only 5 passes are necessary to get the same result.

The output of the rolling mill is increased of around 1 %, because of the improved shape formation as a result of the near net shape beam blank, particularly at the beginning of the rolling of beam. There is an additional potential for reheating energy cost savings of around 8 % caused by the better surface / volume ratio for beam blanks in case of cold charging.

The advantages are available because of the casting and rolling of conventional beam blanks are (i) fewer rolling passes at the break down rolling mill, (ii) increase of productivity of an existing rolling mill of around 15 %, (iii) reduced energy consumption at the break down stands of rolling mill of around 55 %, (iii) lesser roll costs because of the reduced number of rolling stands, and (iv) lesser maintenance costs at the break down rolling stand (savings of around 55 %).

The economic advantages of beam-blank casting for the production of beams and sections can be mainly attributed to the reduced (or eliminated) rolling costs at the roughing stand of the hot-rolling mill. These advantages are summarized as (i) around 30 % lower investment costs, (ii) around 15 % of increase in the productivity, (iii) elimination of the rolling passes at the roughing stand, (iv) around 1.5 % higher yield, (v) lower operating costs, (vi) around 55 % reduction in the energy consumption because of the elimination of the roughing stand, (vii) lower CO2 (carbon di-oxide) and NOx (oxides of nitrogen) emissions, (viii) around 55 % lower maintenance costs at the roughing stand, (ix) reduction in man-hours per ton of steel produced, and (x) elimination of intermediate storage for blooms. These benefits have contributed substantially to the rapid increase of share of beam blank casting.

A near-net-shape casting of beam blank is ideal for the production of the I-beam and H-beam sections since it reduces the number of rolling stands in the beam mill. However, these beams have complex cross-section for casting and are associated with several difficult engineering and operating issues.

Continuous casting and rolling of beam blank have become a normal practice for the production of the steel beams. There is concentrated developments in recent years for the continuous casting of near net shape beam blanks. Fig 1 shows the difference between the conventional beam blanks and near net shape beam blank. Conventional beam blank has a relatively thicker flange, normally over 100 mm, while the near net shape beam blank has a flange thickness less than 100 mm, normally with a lower limit 50 mm in the practice. Several sizes of beams can be produced through rolling beam blank of one dimension.

Fig 1 Beam blanks

Tundish operation – For smaller beam blank sizes, open stream pouring and oil lubrication is applied. As in billet casting of commercial quality steels, two metering nozzles per mould are used for uniform steel feeding. The casting with metering nozzles needs a careful balance of steel oxygen activity to hit the ‘operation window’ between nozzle clogging and pinhole formation. Normally plain manganese (Mn) / silicon (Si) deoxidation is preferred, with pinhole control provided by aluminum (Al) wire feeding into the mould.

For larger sizes, mould powder application is preferred to minimize uneven solidification, strand surface depressions, cracks, and bleeders. In such cases open stream pouring is combined with sub-merged refractory funnels for preventing the powder entrainments. In case of aluminum addition for meeting the fine grain steels for high tensile requirements, active flow control by stopper rod and stream shrouding with a sub-merged entry nozzle (SEN) is used, normally using a single nozzle arrangement. This is advantageous with respect to mould level control and operating cost.

Casting of beam blanks – Casting of beam blanks can be carried out with open casting, semi sub-merged casting, and sub-merged casting. Open casting tends to concentrate on the smaller sizes. Normally two metering nozzles are used for a given strand, although the use of one has also been practiced. An automatic nozzle changer has been developed for open beam blank casting. Deoxidation is carried out with silicon and manganese, sometimes completed with aluminum injection in the mould.

Semi sub-merged casting is the most used process, since it combines the advantages of the other two casting methods. Here two funnels conduct the liquid steel flow to the mould, allowing for the addition of the casting powder addition. Straight, vertical entrance to the mould is normally carried out, although lateral ports have been tested. Sub-merged casting is used when the beam blank dimension is large and aluminum killed steel grades are to be cast. During sub-merged casting, stopper rod or slide-gate and sub-merged entry nozzle are used. Because of the complexities associated with the particular transverse section of the beam blanks, fluid flow and thermo-mechanical modeling has extensively been used, with much more intensity than for billet casting. Fig 2 shows the three casting modes.

Fig 2 Three casting modes for beam blank casting

Each of the three casting methods has advantages and drawbacks as given in Tab 1. The quality issues mentioned in Tab 1 suggest against open casting, despite its inherent simplicity and high productivity.

Moulds for beam blank continuous casting machine – Moulds for the beam blanks can be tubular, as is normal for the casting of square billets, or made of four plates, as in the case of the casting of slabs. Normally, tubular moulds tend to be preferred for the small and medium sections. Plate moulds have two narrow and two wide faces. An alternative has been devised for using plate mould in a mould jacket designed for tubular moulds. This is used for intermediate sections. Typically, plate moulds are cooled with holes (and spacers within the holes) for the wide faces, and slots for the narrow faces. Plate moulds have several features namely (i) more alternatives to manage water cooling such as slots, holes with spacer, full hole, distance between holes, and distance to hot face, (ii) more rigidity of the assembly, (iii) stability of transverse geometry of cooling channels, (iv) easy achievement of different taper modes, and (v) higher cost.

Mould design and operation are key factors in beam blank casting. Mould design is distinguished by three generation of design variants as given below. In case of first-generation mould, it has been a block mould with gun drilled water bores, consisting of two halves. The opening, closing and locking is done by a pneumatic motor to ease blank removal in case of an incident. This mould has been later modified by addition of two stages of foot rollers. In case of second-generation mould, the mould involves a hybrid block / plate design which means that the side walls are of the cold rolled copper plates featuring grooves for higher water velocity. In case of third-generation mould, the mould is similar to the above mould, but with narrow faces clamped between wide faces for achieving higher adjustability.

The third-generation mould design has been found especially useful for the production of an extensive range of wide flange beams sections. For smaller sections tubular moulds, the wall thickness of 6 mm to 32 mm are used. For larger beam blank sections, a plate mould is more suitable. Here individual copper plates are fixed on support plates and connected through screws to form the cross section. Primary cooling water is guided through cooling slots and holes. With this design, a negative taper in the shoulder area to compensate for web shrinkage and an improved arrangement of the cooling holes for homogenization of the copper surface temperature is possible. For ensuring long mould life, normally Cu-Cr-Zn (copper-chromium-zinc) is used as mould material for achieving high wear resistance. This is further improved by chromium plating, in some cases, a multiple coating (with three layers) being used. Fig 3 shows moulds for beam blank casting.

Fig 3 Moulds for beam blank casting

Presently there are two basic designs for beam blank moulds which are being used. The first is the tube mould, which is mainly used for beam blank formats up to 300 mm x 400 mm outer cross sectional dimensions. Depending on the beam blank size, the copper tube wall has a thickness of up to 25 mm and the primary cooling water is guided between the outer surface of the copper tube and a special baffle tube. For manufacturing reasons, it is not possible to design the mould with a negative taper on the shoulder area or with variations of the copper wall thickness for the temperature homogenization over the beam blank strand circumference.

The second is the plate mould which is more suitable for larger beam-blank sections. Here individual copper plates are fixed on support plates and connected through screws to form the cross section. Primary cooling water is guided through cooling slots and holes. With this design a negative taper in the shoulder area to compensate for web shrinkage and an improved arrangement of the cooling holes for homogenization of the copper surface temperature is possible.

A particular challenge in the mould design is the choice of adequate tapers for the intricate beam blank shapes. While a positive taper is applied for the outer side flanges, a zero or negative taper is needed for the web fillet. The inside angle of the flanges as well as radii between flange and web is also of importance. In recent years, taper design is supported by finite analysis method (FEM) analysis of mould heat transfer, shell growth, and shell contraction.

Continuous casting machines track the total energy removed from the solidifying steel by the mould, measured indirectly as the temperature change of the cooling water. Some moulds include a thin coating layer of nickel (Ni) or chromium (Cr) for reducing the wearing of the hot face, i.e., the face of the mould in contact with the strand. Several bolt holes are machined into the back side of the mould for mounting the mould into its support structure and water-delivery system, collectively called the water-box. Moulds are instrumented with thermocouples, either between the water channels or coaxially with the bolt holes, for on-line monitoring of the casting process. The cooling water temperature change and mould thermo-couple temperatures are the key validation points for the different models used for the mould heat transfer.

For beam-blank moulds, the regions of the hot face furthest from the water channel become very hot, especially at the meniscus. These hot spots are found at the shoulder, and are alleviated with smaller and / or more water channels in the region. The mould normally bows outward, away from the steel, with a slight twisting motion.

The relatively heavy moulds need a robust oscillation. A short lever design with motor driven eccentric and push rod is virtually maintenance free. It also assures high guidance accuracy of a pass line deviation less than 0.02 mm. For shallow oscillation marks, the short stroke / high frequency mode is the most suitable for the low C range of structural steel.

Strand support length – For the design of the strand-support length, a transient heat-transfer analysis of the beam-blank section is performed. This type of analysis provides the necessary information about the shell growth within the strand support and the exact metallurgical length. A web-strand support which is too short can cause bulging or even an opening of the web centre. This can lead to steel segregation and web-thickness variations. A flange strand support which is too short can also cause interface cracks.

Because of the unique shape of the beam blank section, four different areas on the surface of the beam-blank section have to be individually supported as given below.

Web – In order to prevent bulging of the web, and hence more pronounced centre segregation, the web of the beam-blank section needs to be supported until sufficient solidification over its width is achieved. Two dimension (2-D) thermal analysis provides the information for the necessary supporting length.

Flange – The flange is to be supported in order to prevent bulging and internal cracking. A 2-D thermal analysis yields the temperature field and the corresponding shell thickness. A subsequent stress analysis displays the stress / strain and displacement fields, which results from the internal ferro-static pressure from the liquid steel core. The criterion for the support length in this area is the generated interface strain because of the ferro-static pressure at the liquid / solid transition of the flange inner surface.

Flange tip – Similar criteria apply to the flange tip as for the whole flange and in general the supporting length depends on the casting size and on the casting speed. In several cases, particularly for the lower casting speeds and small beam blank cross sections, no additional support other than the mould foot rollers is necessary.

Shoulder – Because of its physical shape, the shoulder area acts like an arch, and hence no support is normally necessary. A 2-D finite-element analysis shows the stress and displacement field.

Mathematical modeling – The efficiency and quality of continuous-cast steel is constantly improving because of the increased automation and advances gained mainly through plant experimentation. However, empirical solutions are now very expensive without the aid of tools such as computational modeling. Practical applications of mathematical models include the design of mould geometry to control the mould temperature and gap formation in order to avoid crack formation in both the mould and solidifying steel shell. In particular, the development of mould tapers to match the steel shrinkage is an on-going challenge which are to be met for each new cross-section, mould design, and steel grade.

Because of the complexities associated with the particular transverse section of the beam blanks, fluid flow and thermo mechanical modeling has been used extensively, with much more intensity than for billet casting. Clearly there exists a strong incentive to develop quantitative computational models which can predict temperature, deformation, and stress in the solidifying steel shell in the mould during continuous casting of near-net-shape sections with sufficient accuracy to solve practical issues such as the design of mould taper. Several computational models have been developed of thermal and mechanical behaviour during continuous casting of steel. Only a few models of the beam-blank casting process have been attempted, owing to the computational difficulties associated with complex geometry and behaviour.

Lait and Brimacombe pioneered the application of solidification models to beam blank casters. The predictions of their one-dimensional (1-D) moving-slice model compared well with plant measurements, but variations around the perimeter have not been studied. Thomas and coworkers applied a two-dimensional (2-D) model of mould heat transfer to study water channel design in a beam-blank mould, and identified how the concave internal-corner region where the web intersects the flange is susceptible to overheating and hot face cracks in the meniscus region of the mould. Some specialists at Voest-Alpine Industrial Services have used a 2-D transient finite-element model of heat transfer and stress in a horizontal slice through both the shell and mould to gain insights into beam-blank mould design and shell behaviour.

Lee and coworkers show-cased multi-physics modeling by coupling a three-dimensional (3-D) finite-difference model of fluid flow with a similar 2-D transient thermal-stress model for predicting solidification, gap formation, stress, and crack formation in a beam-blank casting machine. Impingement of steel flow from the nozzle inlet has been found to retard shell solidification in the central regions of the flange (narrow-face) and the web. Air gaps because of the solidification shrinkage have been predicted near the flange-tip corner. Surface cracks in the web and fillet regions and internal cracks in the flange-tip region have been predicted.

Modeling heat transfer in the continuous casting process needs accurate incorporation of the mould, the solidifying strand, and the interface between them. The behaviour of the material in the interface, a ceramic slag, governs the heat extraction from the strand. Continuous casting of steel or any other metal is a complicated process with several coupled and non-linear phenomena, and needs advanced modeling techniques to understand what is important for the process. Majority of the process phenomena are dependent upon the mould heat transfer, e.g., the rate-dependent solidification shrinkage of the solid shell, the time dependent crystallization and flow of the interfacial slag, or the multi-phase turbulent flow of the liquid steel with a free surface and particle transport.

Mathematical modeling is very convenient for investigating the suitable structure of sub-merged entry nozzle and has a wide application as a powerful tool in the analysis of the transport phenomena in the mould. The majority of the mathematical models about beam-blank continuous casting are based on the mould thermal-mechanical analysis, mould water channel design, and secondary cooling strategy. These models are mainly on the straight sub-merged entry nozzle. With such sub-merged entry nozzles, two major problems, which the engineers are encountering in stee plants, are the inactive meniscus status and the non-uniform shell thickness distribution at the mould exit. This can seriously affect the function of the mould flux and the uniform distribution of the stress and ultimately lead to a bad quality.

The geometric and thermal mould conditions for the initial solidification of the strand are extremely important in order to get a strand with outstanding surface and internal quality. A properly designed primary cooling system and mould taper are, hence, necessary pre-conditions to meet these needs. A 2-D, fully coupled thermo mechanical finite element model is used to calculate the temperature and displacement fields of the strand during initial solidification in the mould. This type of simulation provides a better understanding of the complex shrinkage behaviour of a particular beam blank section, enabling the shape and taper of the mould inner contour to be accurately determined. This 2-D finite element model has been successful with respect to shell growth, internal and surface beam blank quality, and mould wear.

A transient analysis, neglecting heat flux in the longitudinal direction, provides the temperature and displacement fields. The influence of different mould tapers on the shell growth, temperature fields and contact pressures because of the shell shrinkage can easily be studied. The internal ferro-static pressure is increased as the strand shell moves through the mould.

Defects arise if the mould has either very little or very large taper. If any part of the internal mould shape has insufficient taper, gaps can open up between the solidifying steel shell and the mould wall, which drop the local heat transfer and lead to locally hot and thin sections. Combined with the ferro-static pressure from the internal liquid pool, this can cause transverse strains on the newly solidified shell, which concentrate at the thin spots, resulting in longitudinal cracks or break-outs. Alternatively, if the mould taper is excessive anywhere, the mould pushes on the shell, leading to severe mould wear and / or buckling of the shell, and again leading to longitudinal cracks or break-outs. In addition, axial stresses from the mould friction can cause excessive local cooling and binding of the shell in the mould, generating transverse cracks and other issues.

Thermal and mechanical models are powerful tools for providing a complete understanding of the product quality. An analysis of stress and distortion is necessary to interpret the deformations suffered by the solidified skin and the mechanisms of crack formation. Majority of the studies in this area are focused on the straightening region, where the formation of cracks can occur depending on the temperature, level of applied force, and additional stresses because of the roller mis-alignment or the soft reduction region, which is used to improve the internal quality of the material by the application of compression force. The use of inappropriate process parameter values can cause break-outs, shape defects, and the formation of cracks. Several plants for continuous casting aim to optimize the operational parameters to minimize the occurrence of such defects.

Defects in cast beam blanks – Beam blank casting is an established process with a long history. However, it is not free of surface and inner defects. Some of them share features with billet defects while other have more to do with slab defects. The complex shape of beam blanks induces specific solidification defects. The occurrence of defects needs carrying out improvement plans for ensuring their minimal occurrence. Defect characterization is important. Simulation can help to explain the formation mechanism and to suggest corrective measures. Fig 4 shows some defects in cast beam blank.

Fig 4 Some defects in cast beam blank

Surface defects – Surface defects in beam blanks are much like in billets, but have some specific characteristics. Surface defects are pinholes, bleeding, scum / casting powder entrapment, and longitudinal facial cracks.

Pinholes defect occurs mainly for casting with metering nozzle and with oil lubrication. They can be harmful for the final product if they are concentrated in a particular zone (nest). They are deep enough as to not disappear in the reheating furnace or if in the first rolling steps the material has free spreading (it is not contained) somewhere. They are almost scale-free in the beam blank, but then in the reheating furnace they become filled with scale. Some of possible causes for pinholes are moisture in the oil (or moisture pick-up in the oil circuit), very high oil rate, in-homogeneous distribution, very thick oil slot gap (more than 0.5 mm), partial obstruction of oil slot gap by splashing, sudden variations in steel mould level, use of pulsing bomb, and lack of deoxidation. The use of electro-magnetic stirring can help in the elimination of pinholes.

Bleeding defect occurs when small strand break-out takes place, and healing immediately, without metal loss. It can be attributed to the effect of annular strain in hot zones, or sticking. Scum forms during open casting because of the thorough reoxidation of the liquid steel in contact with air and oxidizing slag. This scum is normally a liquid manganese silicate. If the silicon content is very high (because of a low manganese-silicon ratio), silica precipitation occurs, bringing about higher viscosity and the risk of scum entrapment in the surface of the beam blank, and in extreme cases, strand break-out. Another way to promote higher viscosity is through aluminum wire injection in the mould, whenever it is excessive or it is not in the right point.

Somewhat similar phenomena can occur with casting powder (with funnel or sub-merged entry nozzle). Higher viscosity can occur in this case through alumina pick-up, or reduction reactions between elements in the steel and oxides in the casting powder, e.g., dissolved titanium (Ti) reacting with silica (SiO2) in the slag. Casting powder entrapment is enhanced through turbulence which can be because of excessive electro-magnetic stirring, and small sub-merged entry nozzle / funnel submersion.

Network cracks, related with a high copper content in the steel, are common to other semi-products produced in the steel melting shops operating with high scrap charge. In zones, where the gap between strand and mould becomes large, the grain size increases and if copper content is high, gives place to such cracks.

The defect of longitudinal facial cracks is fairly common for beam blanks. Formed in the mould, it has certain similitudes with longitudinal cracks in slabs and blooms. When observed in the rolled product, its metallographic features are the presence of internal oxidation (as polished, no etching), decarburization (etching with Nital 2 %), and oxygen penetration (hot etching with alkaline sodic chromate). Influencing factors are the chemistry of the liquid steel, the properties of the casting powder, deviations of casting machine radius caused by mould oscillation, as well as flow rate and temperature of primary cooling.

Steel chemistry can be cause of defects. From early times, the influence of sulphur content on longitudinal cracks is well known. Another steel chemistry-related factor is the carbon content. Peritectic transformation need to be avoided. Several experiences point to their influence on longitudinal cracking as is well known from slab casting. For example, specialists of a plant have studied these cracks for scarfed beam blanks corresponding to 2,000 heats. They found that the more sensible chemistry range has been 0.12 % carbon (C) to 0.13 % carbon, corresponding to steel suffering peritectic transformation. Specialists of another plant have recommended 0.08 % maximum carbon, with 0.6 % minimum manganese for getting the needed mechanical strength. Recently, the criteria to calculate the range of carbon content for which peritectic reaction occurs has been revisited, using thermo-dynamic commercial codes. Regarding residuals, their presence seems not to affect very much the presence of longitudinal facial cracks. One of the steel plants has reported no quality problems with 0.35 % copper (Cu) and 0.2 % chromium (Cr) or nickel (Ni), the same with 0.03 % zinc (Zn) and 0.04 % tin (Sn).

The influence of casting powder for funnel or sub-merged entry nozzle casting on longitudinal crack formation is well established. For example, high basicity, low viscosity mould fluxes have given good results in casting of small beam blanks at low speed (less than 1 meter per minute) which includes achievement of ‘soft’ cooling at meniscus level with that flux. The lower capacity for infiltration and lubrication has been partially compensated by the low viscosity. For another set of conditions, one of the plants has found the opposite situation, i.e., low viscosity giving place to longitudinal cracks (among other reasons). In cold zones of the meniscus (e.g., close to the nozzle), casting powder can reach the limit of its performance and give place to surface cracks.

The effect of casting speed has been studied in one of the plants. It has been found that there is a lineal relationship. As the casting speed increases, the solidifying shell is thinner, heat flow increases, and strain is larger, giving rise to more cracks as a result.

In case of secondary cooling, it has been found that an increase in secondary cooling intensity brings about a proneness to the longitudinal crack formation. An experience at a plant shows this relationship taking into account the already discussed effect of sulphur also. Extensive use of mathematical modeling has been carried out to solve this defect. For example, one plant modeled an optimization of secondary cooling to avoid these cracks, using ANSYS for the thermo-mechanical model and MATLAB for parameter optimization. One of the plants has done a very thorough modeling of secondary cooling with the same purpose, taking into account all the mechanisms involved in heat transfer.

The corrective actions can be divided as (i) metallurgy- low sulphur for avoiding peritectic transformation, (ii) mould flux – high basicity and even heat transfer, (iii) mould design – for avoiding longitudinal cracks in the shoulder, and (iv) secondary cooling – less water, mostly for the first segments, and better transverse distribution.

Internal defects – The internal defects in beam blanks have similitude to some extent with defects in billets. The internal defects are (i) blow-holes, (ii) web central crack, and (iii) inner crack in wing end.

The blow-holes are localized close to the surface, with a direction perpendicular to the surface. They can be seen in the oxy-cutting, if their occurrence is important. Depending on the root cause, they can be concentrated in the first heat of the sequence or in some given heat, or all along the sequence. The blow-holes depart close to the beam blank surface, when there is enough gas segregation to the inter-dendritic spaces. They come to an end when somewhere below the meniscus, the ferro-static pressure is higher than the gas pressure.

Blow-holes are attributed to an excess of gases dissolved in the steel (oxygen, nitrogen, and hydrogen), enough to produce a bubble. This phenomenon has been modeled since the early times of continuous casting. From the point of view of deoxidation, for manganese-silicon killed steels there is a compromise between the risk of clogging (if deoxidation is very strong) and the risk of blow-holes.

Some typical industrial cases of heats with blow-hole occurrence, mostly for open casting, are (i) high oxygen – heats sent to the casting machine with deoxidation not finished, because of the coordination issues of complications derived from furnace slag carry-over, (ii) high oxygen and nitrogen – sequence start, (iii) high hydrogen – moisture in new lining of ladle or tundish, and (iv) high nitrogen – ladle with long treatment, when nitrogen is used for stirring.

One steel plant has reported a case of blow-holes in beam blanks cast with metering nozzle. The deoxidation practice included an 80 kg aluminum addition during tapping, and 40 kg Ca-Fe (calcium-iron) to get oxygen less than 10 ppm (parts per million). Oxygen is injected in the tundish if steel temperature has been too low. The beam blanks are scarfed for inspection. After thorough study of ladle furnace and continuous casting variables, it has been concluded that high moisture in tundish repair refractory material is responsible for the issue.

The defect of web central crack is equivalent to the centre-line segregation in slabs. Not enough support length and / or insufficient secondary cooling have a consequence of the bulging of the beam blank and in severe cases, an internal opening in the web.  High central segregation and crack formation can appear in the rolled product. Tools for avoiding central cracking are the roll checker and the equipment for segment alignment.

One of the plants has reported a case of web central cracks for its casting machine of 12.5 m radius, casting with funnel beam blanks of 400 mm x 460 mm x 120 mm and 287 mm x 560 mm x 120 mm sizes. There has been an influence of sulphur content and casting speed. The issue has been solved with intensive spray cooling on the web portion, and strict maintenance of roll gap.

The defect of inner crack in wing end has been studied since it has given place to strand break-outs. An improvement plan has been carried out at one plant to get rid of the problem. The approach included several studies. The aspect of cracks in normal heats has certain features in common with off-corner cracks in billets and slabs. The problem has been solved through an optimization of mould taper in the ends of the wing.