Search

The success of the continuous casting process owes much to the performance of the mould powder. Mould powders have been used for more than 50 years and our knowledge of how they work and perform has increased steadily over this time. The performance of the mould powder should always be judged by how it affects the steel shell. Since this is frequently due to the nature of the slag film formed between the shell and the mould, the performance of the mould powder can also be judged by the nature of the slag film formed. However, other factors (such as the depth of the molten pool) must also be taken into account since they affect the formation of defects. Our current state of knowledge on how mould powders work is summarised.

The success of the continuous casting process owes much to the performance of the mould powder. Mould powders have been used for more than 50 years and our knowledge of how they work and perform has increased steadily over this time. The performance of the mould powder should always be judged by how it affects the steel shell. Since this is frequently due to the nature of the slag film formed between the shell and the mould, the performance of the mould powder can also be judged by the nature of the slag film formed. However, other factors (such as the depth of the molten pool) must also be taken into account since they affect the formation of defects. Our current state of knowledge on how mould powders work is summarised.

When you look into the continuous casting mould you can see very little. Consequently, steelmakers have had to rely on plant trials, simulation experiments and physical property measurements on fluxes and steels to gain an understanding of the mechanisms responsible for process problems and product defects. However, in recent years, mathematical modelling has advanced to the stage where they can provide us with great insight into these mechanisms. As a non-mathematical modeller, I was initially sceptical of some of the predictions of the mathematical models. However, I have been completely won over by the ability of these models to simulate accurately the mechanisms responsible for various defects, such as slag entrapment, oscillation mark formation, etc. Mathematical modelling literally allows us to ‘see’ what is happening in the mould. It is a remarkable tool.

The continuous casting (CC) mould may appear very peaceful when viewed from above, but the powder bed hides relentless fluctuations in the following phenomena: metal flow, thermal gradients, chemical reactions and multiple phase transformations. When observed separately, some of these phenomena seem to have a ‘simple behaviour’, which may appear easy to control through the main casting parameters (e.g. casting speed, pouring temperature and powder type) and associated control systems (e.g. mould level control, automatic powder feeding and mould oscillation). However, when combined, these phenomena exhibit periodic fluctuations in behaviour, which is both difficult to predict and control. For instance, the combination of casting speed, submerged entry nozzle design and slab size can cause the metal flow pattern to shift from double roll to single roll and back, which can cause unstable fluctuations in metal level, standing waves, etc. In this respect, the CC process closely resembles a meteorological system where both variations and local fluctuations in temperature, humidity, pressure, etc., can result in effects that are difficult to predict in the long term. This is equivalent to the famous Lorenz premise: ‘Does the flap of a butterfly’s wings in Brazil set off a tornado in Texas?’ In this paper, we give some examples of the ‘butterfly effect’ in CC discussed below by using a mathematical model able to predict the slab solidification inside the mould in which various factors affecting the process stability are analysed and the probable sources of fluctuation are identified.

Fluoride containing mould fluxes may damage both the environment and casters, so the development of fluorine free mould fluxes is of great interest. The key issue for the development of fluorine free or low fluoride mould fluxes is to find a suitable crystal phase as a substitute for cuspidine. Some researchers have suggested that additions of TiO2to fluxes can help to form perovskite (CaTiO3) as a suitable substitute for cuspidine. However, the rate of sticker breakouts was found to increase greatly after the use of the TiO2containing, fluorine free mould fluxes in the continuous casting of peritectic steel grades. One of the possible reasons for this phenomenon is the formation of titanium nitride and carbonitride during the melting process of mould flux. This possible explanation has been proved in this study using both thermodynamic calculation and laboratory experiments on TiO2containing, fluoride free mould fluxes.

Plant data from steelworks throughout the world have been analysed to determine: the effects of mould dimensions on powder consumption, the factors affecting the required melting rates of the fluxes, the reliability of empirical rules to predict both powder consumption and the optimum casting conditions, and the relation between break temperature and the type of steel being cast. The results are discussed with special reference to thin slab casting.

Ladle glazes are a thin coating of slag formed when the top slag comes in contact with the refractory during draining of the ladle. It is believed that ladle glazes are a major source of inclusions in the next cast of steel held in that ladle. Pilot plant trials, simulation tests, and thermophysical property measurements have been carried out on the glazes to: (i) obtain more information about ladle glazes and the factors affecting their formation and subsequent removal from the refractory; (ii) provide direct evidence that glazes are a major source of inclusions; (iii) study the factors affecting the formation, adhesion, and thickness of the glaze; (iv) study the formation of the slag penetrated layer and the factors affecting penetration. Doping trials have shown that ladle glazes are a major source of inclusions. The calculated thickness of the adhered slag layer was appreciably lower than measured experimental values. Calculated horizontalslag infiltration depths were in reasonable agreement with values measured on plant, but measured verticalinfiltration depths were much lower than calculated values. It has been shown that slag penetration can be significantly reduced by the application of an electrical potential.

The thermal conductivities of alkali silicate melts containing fluorides have been determined using the non-stationary hot wire method. Measurements were carried out over the temperature range 1050–1550 K on the following six samples: Li2O–2SiO2, Na2O–2SiO2, 0·7Li2O–0·6LiF–2SiO2, 0·9Na2O–0·2NaF–2SiO2, 0·7Na2O–0·6NaF–2SiO2, and 0·5Na2O–NaF–2SiO2. In the temperature range investigated, the thermal conductivities of these silicates decreased with increasing temperature. Additions of fluorides were found to cause decreases in thermal conductivity. This supports the view that fluorides depolymerise the network structure of silicates to form bonds terminating in non-bridging oxygen and fluorine ions. It was concluded on the basis of this structural model that, from the viewpoint of the thermal resistance, the alkali fluorosilicate melts contain three types of chemical bonds: (i) primary covalent bonds (denoted by Si–O) between silicon and oxygen in the network structure, including Si–O-and Si–F bonds because both are covalent, (ii) ionic bonds (denoted by O–O) between non-bridging oxygens via sodium or lithium ions, and (iii) F–F bonds which are actually dissociated by the formation of Si–F bonds. The magnitudes of the contributions to the thermal resistivities of the silicates r from these bonds were in the hierarchy: ?F–F> ?O–Oin Na2O–SiO2> ?O–Oin Li2O–SiO2> ?Si–O. The contribution from the F–F bonds was the greatest.

A slag film is formed in the channel between the mould and shell during the continuous casting of steel, and the amount of crystalline material in the slag film has a significant effect on the heat transfer between the steel shell and the mould. Techniques to measure the degree of crystallisation of these slag films, such as X-ray diffraction, calorimetry, microscopy, and thermal analysis, have been examined in this investigation. The ratio of (% crystallised/% glass) phases has been determined for various slag films taken from the mould when casting different steel grades; these showed that the glassy fraction varied between 0 and 100%. These results are discussed.