CFD SIMULATION OF A GAS-LIQUID MULTIPHASE SYSTEM

CFD simulations: a valid support for real high risk systems analysis

In manufacturing it is quite common to face severe conditions and environments. Such a situation can be encountered in foundries, where the risk connected to daily activities and employed tools is very high.

In this type of scenario, we tried to predict the dynamics characterizing a molten steel bath moved by oxygen jets flowing through special nozzles in a supersonic regime, by means of a CFD analysis.

The objective was to understand which problems can arise as regard numerical modelling and to verify if the blending and therefore its product can be more homogeneous and less flawed, showing a good quality.

Would have been possible to model and simulate such a physical phenomenon using Computational Fluid Dynamics (CFD)?

The only way to get an answer was to take on this challenge and implement a simulation.

Two-phase system involving water and air are nowadays commonly studied through CFD, but what would happen if we consider a different liquid material, such as molten metal, much more dense and affected by gravity, interacting with high speed supersonic gas? Would the calculus be able to hold on? If it this happens, will the obtained results be realistic and thus useful for further analyses?

Before proceeding in the analysis of a real test case, we decided to implement the simulation on an artificial simple geometry, however having velocity and temperature conditions similar to the real ones.

Typically, a factory should build a series of furnaces to be used only for experimental purposes, employing a specific measurement equipment (i.e it should resist to high temperature).

TecnoHit is able to offer the possibility of taking advantage of numerical analysis in order to implement and simulate the process.

In this specific case, an unsteady (time dependent variables) and multiphase (the materials employed appears in different physical state) was developed.

Considering the different characteristics of the fluids and the particular operating conditions (velocity, temperature are really prohibitive), one can understand the difficulties this analysis can exhibit.

In this scenario, for example, oxygen has to be considered as an ideal gas (i. e. with variable density), increasing the possibilities to incur in numerical instabilities.

As it almost happens in every simulation, it is possible to simplify the system in order to reduce CPU time, without giving away the possibility of equally reaching realistic and physically acceptable results.

Firstly, the domain was simplified considering just a small slice of an ideal furnace, thus reducing the geometry from a 3D system to a 2D one.

Furthermore, using a second symmetry property, the domain was cut in a half, using a second symmetry plan, allowing to considering just one nozzle.

These simplifications show their importance during the mesh creation, that need a much smaller number of cells than the 3D one, thus reducing CPU time.

With this concept in mind and taking advantage of TecnoHit analysts experience, it was possible to create a mesh more appropriate for the physics of the problem, i.e. more dense in those zones where it is supposed that the variables can experience great variations, such as around the interface between gas and liquid which defines the profile of the wave formed by molten steel. This technique allows to further reduce CPU time while maintaining the same results accuracy.

 Despite these expedients lead to a very good quality mesh, the modelling of oxygen as compressible gas due to high velocities can determine difficulties during the convergence of the solution.

This circumstance is clearer in the initial moments of the simulation, since the system changes its status from rest to one characterized by extremely high values of velocity, temperature and volume fractions, causing numerical instabilities.

The habit of TecnoHit team in facing this type of situations permitted to overcome these obstacles, assigning to oxygen an initial low value for velocity and increasing it linearly until the reach of the capacity value.

Simulations results can be analyzed through the investigation of several physical quantities (volume fraction, pressure, velocity profile, etc.). The simulated wave formed by molten steel is quite realistic, confirming that CFD simulation can accurately predict the motions inside the furnace.

It’s clear that numerical simulations permit to produce very good indications on the process development even for complex phenomena, thus eliminating the need for real tests, which would require longer times and more elevated costs.

TecnoHit value lies not only on its capability in numerically reproducing elaborate physical phenomena in prohibitive conditions, but also in supporting its partners in results evaluation and in operating the right choice for improve manufacturing processes and product quality based both on numerical results and also on experience.