University of Nova Gorica

c1011ef61ed9937904f4938c63d014ea_L

The Laboratory for Multiphase Processes at the University of Nova Gorica conducts fundamental research towards the development of advanced numerical methods for multiphase systems and the development of physical models for solid-liquid processes. The focus is on meshless methods for simulation of solid and fluid mechanics problems in the presence of moving boundaries. The models are developed within the continuum mechanics, cellular automata and phase field concepts. They involve systems without phase change, like dispersions and porous media as well as systems with melting or solidification and laminar or turbulent behaviour of liquid phase. The description of the processes is coupled on different scales, from the microstructure evolution to the macroscopic transfer of mass, energy, momentum and species. The laboratory is involved in the development of international test cases for Stefan problems and in the comparisons between numerical models and experiments. The research is incorporated into the numerous European, US and Chinese research projects. The members of Laboratory received numerous domestic and international awards and recognitions.

The applied research in the Laboratory is directed towards numerical modelling of a wide variety of processes with metallic, polymer or ceramic materials and their composites. The principal interest is in simulation and optimisation of relations between the process parameters and the product properties. Modelling is based on physical models, coupled with artificial intelligence. We focused on development of through process models for aluminium and steel industries. The models are validated on the basis of infrared thermography and other in-situ industrial measurements. The models encompass casting, rolling, extrusion and heat treatment. The collaboration reflects in improved regulation algorithms and substantial enhancement of product quality and process yield. All related industrial process modifications and upgrades are performed in close cooperation with industrial research teams. We started to numerically model also multiphase processes in nature like growth of the stalagmites and stalactites in nature.

We collaborate with numerous companies in simulation supported development of new products with higher knowledge, reduced environmental impact, higher quality and more efficient production.

The Fluent software is used in the Laboratory for benchmarking of our own developments and as a tool for CFD analysis of the described problem spectra.

Prepared by:
prof. dr. Božidar Šarler
dr. Robert Vertnik

1. Backward facing step

The solution of the incompressible turbulent flow over a backward facing step is presented. The geometry of the physical domain is taken from the experiment by [1] (JD), characterized by the step height H=1.0 m, the channel length after the step L=30 m and by the Reynolds number ReH=5000 based on the step height, which determines the inlet conditions at the step. In the experiment, the flow conditions in the channel upstream was achieved at the distance px/H=-3.05 before the step with the Reynolds number based on the momentum thickness Reθ=610. The computational domain in the Fluent considers only the physical domain after the step, where the inlet conditions are set above the step and pre-calculated from the boundary layer simulation over a flat plate. The inlet profiles of the velocity, turbulent kinetic energy and dissipation rate were defined by using the UDF. The low-Re turbulence model was used, with closure coefficients and damping functions by Abe-Kondoh-Nagano model. Image 1 represents the stream function contours, obtained on 78800 quadrilateral cells with second order upwind spatial discretization. The results were compared by the LRBFCM [2,3], JD experiment, and direct numerical simulation (DNS) by [4]. The comparison is shown in Image 2, with the horizontal velocity profile at position x/H=6. In Image 3, the skin friction at the bottom wall is presented. Excellent agreement between Fluent and LRBFCM is achieved.

References
[1] Jović, S. and Driver, D. M. (1994). Backward-Facing Step Measurements at Low Reynolds Number, Reh=5000. NASA Technical Report TM-108807.
[2] Vertnik, R. and Šarler, B. (2009). Solution of incompressible turbulent flow by a mesh-free method, CMES: Computer Modeling in Engineering & Sciences, 44:66-95.
[3] Vertnik, R. and Šarler, B. (2011). Local collocation approach for solving turbulent combined forced and natural convection problems, Adv. appl. math. mech., 3:259-279.
[4] Le, H., Moin, P. and Kim, J. (1997). Direct numerical simulation of turbulent flow over a backward- facing step. Journal of Fluid Mechanics, 330:349-374.

2. Continuous Casting

Numerical example represents the simulation of the continuous casting of steel. The process involves turbulent fluid flow with heat transfer and solidification of a carbon steel. Two-dimensional (2D) real curved geometry of length 1.8 m is generated, with billet dimension 0.140 m. The process parameters are taken from the real continuous casting machine. The submerged entry nozzle (SEN), which supplies the molten steel into the mould, is not considered in the numerical model. However, the velocity, turbulent kinetic energy and dissipation rate profiles at the SEN outlet are pre-calculated by the SEN flow simulation. The resulting profiles are then inserted into the inlet boundary conditions through the UDF functions. The pulling velocity of the solidified steel is defined by the custom field functions. The numerical solution is obtained on the 139200 quadrilateral cells with second order upwind spatial discretization. The low-Re turbulence model is used, with closure coefficients by Abe-Kondoh-Nagono model. The results are shown in Image 1 and Image 2, with velocity magnitude and temperature field in the billet, respectively. The comparison with the meshless LRBFCM [1÷3] was performed also, see Image 3, where very good agreement was achieved. The anomaly was observed at the outlet of the computational domain, see Image 3. By using larger computational domain, i.e. 2.8 m, the anomalies at the outlet disappear, but no difference was found in other regions.


References
[1] Vertnik, R. and Šarler, B. (2009). Simulation of continuous casting of steel by a meshless technique. International Journal of Cast Metals Research, 22:311-313.
[2] Vertnik, R. and Šarler, B. (2009). Solution of incompressible turbulent flow by a mesh-free method, CMES: Computer Modeling in Engineering & Sciences, 44:66-95.
[3] Vertnik, R. and Šarler, B. (2011). Local collocation approach for solving turbulent combined forced and natural convection problems. Advances in Applied Mathematics and Mechanics, 3:259-279.