PHYSICAL MODELING OF THERMAL EFFECTS ON STEEL FLOW AND MIXING IN TUNDISH

In recent years there has been considerable interest in the effect of temperature gradients on liquid steel flow in the tundish. Temperature gradients occur as a result of heat losses in the ladle-tundish syste m and are most pronounced during the start of heat and end of heat per iods. The density difference due to temperature gradient!; could cause changes in liquid steel flow that could, in turn, affect steel slab q uality and may have an influence on the intermix region experienced du ring casting of dissimilar grades. In the present study, extensive wat er modelling work was conducted on a 0.6 scale model of the US Steel G ary No. 2 tundish to determine the nature of the steel flow and how it is affected by a step change in incoming steel temperature into the t undish, The influence of the presence and absence of flow control devi ces (FCDs) at two casting rates - high and low - on flow patterns was investigated. A pulse tracer injection technique was used to evaluate the behaviour of liquid steel flow and mixing in the tundish. Temperat ure probes were placed at multiple locations in the tundish to establi sh more precisely the distribution of local time for temperature homog enisation. Analysis of the tracer results was performed on the basis o f a mixed model method. It was found that temperature gradients genera l ly influence flow patterns in the tundish and cause deviations in tr acer response from the isothermal case. In the presence of FCDs and at the higher casting rate of 4.3 t min(-1), as the temperature of the i ncoming fluid is increased from 0 to 36 degrees C the minimum residenc e time or plug flow volume of fluid in the tundish and the eddy diffus ivity of the sweeping tracer front increased continuously, while the a verage residence time remained approximately constant. At a lower cast ing rate of 3.7 t min(-1), minimum and average residence times and edd y diffusivity of the tracer front increased with temperature different ial but reached a plateau at a lower temperature differential than the high casting case, and thereafter remained approximately constant abo ve temperature differential values of 20 K. In the absence of FCDs, th e minimum residence time increased while the average residence time de creased with temperature differential. Non-isothermal tracer response results from the model have been compared with plant tracer data and b etter agreement has been found than that reported earlier for isotherm al models. IS/1339. (C) 1998 The Institute of Materials.