Numerical Modeling of the Ladle Flow by a LES-Based Eulerian–Lagrange Approach: A Systematic Survey

Abstract

AbstractTo account for increasing economic and ecological pressure, the steel industry is obligated to continuously optimize all processes. An important optimization approach is numerical modeling although its potential is limited by the accuracy of the mathematical models. In a previous work, a validation database was created and a validation score was derived from this data which allows a comprehensive qualitative accuracy assessment for those models. Here, this system is employed for a systematic optimization of the isothermal flow in the casting ladle. For that, different submodels, namely the turbulence models, subgrid turbulence models, bubble-induced turbulence and interfacial closure models as well as influencing factors, such as the grid resolution or the initial bubble size, are analyzed. It is shown that the large eddy turbulence model is more accurate than the Reynolds-average approach because it is able to reproduce the anisotropy of turbulence in the bubble region. In accordance with the literature, a grid dependency of the lift force is found which can be reduced using an averaged shear field as an additional variable. For the interfacial closure models, the combination of the Tomiyama drag model for fully contaminated systems and the Tomiyama lift correlation showed the best agreement with the experimental data. The results of the survey are summarized to a best-practice guideline with which the validation score can be increased from 38.7 with the Reynolds-average approach to 85.1 on a coarse grid respectively, and 87.8 on a fine grid. However, some upscaling problems of the numerical system from the water model to the real ladle are revealed. There is a need to find accurate yet efficient grid resolutions which make the large eddy turbulence model affordable with the current computational resources. Furthermore, alloying elements or non-metallic inclusions might alter the interfacial forces considerably. However, no studies on their effect have been published yet.