I have been working during a lot of time on this task and it is a very complicate one.
The geometry is a 3D horizontal tube of 2m length and 6.35 mm. inner diameter with a wall thickness of 3.2 mm. (I’m taking into account the wall thickness in the simulation. Only the half of the tube is being drawing to reduce the computational time).
The refrigerant reaches the tube inlet with a determined mass fraction and the mass fraction increases along the tube due to a constant heat flux imposed at the external wall.
The simulation is steady. To resolve it I’m using the eulerian model with the Non-equilibrium boiling model. For the turbulence simulation the k-ω model is being used as can be modified as it is mentioned later.
The boundary conditions are mass flow at the inlet and pressure outlet at the outlet. At the outer wall a constant heat flux is imposed and at the inner wall a coupled boundary condition is imposed (as it is imposed also at the inner wall shadow automatically created).
The main problem with this simulation is that the results are not comparable with the experimental results. After doing a deeper research I found the MsC of Hazari 2016 at which the author imposes by an udf a modification of the turbulence ω value. What he did is defined the ω value at the interface between the liquid and the gas incorporating the effect of surface roughness on it. In this way the ω value is defined as function of the liquid thickness on the tube.
Because the liquid thickness is not a variable which Fluent controls I calculate it analytically and I defined the ω value as function of the liquid velocity in the tube.
For the smallest mass fraction (x=0.10 at 0.46m from the inlet) I achieve to resolve the simulation although the result is very oscillating (the values of mass flow rate at the outlet, wall temperatures and volume fraction are oscillating around its final value). But when the mass fraction increase I need to change the boiling model to the CHF. And for this case the simulation quickly diverges or if I reduce the under relaxation factors to 0.1 it doesn’t diverge but a constant result is not achieved.
So I want to kindly ask you for any advice about this case. Thanks in advance.
try to have a look to the OpenFoam solver interPhaseExchangeFoam.
It handle phase changes.
Franco
- Badges
- Science topic