Learning Forum

[Simon Tempelaars]

Hi, I want to simulate the deaeration of air out of water depending on pressure using a venturi. The goal is to see how bubbles form from the fluid and how they would move relative to the water flow. I have the issue that when I enable Henry's law within the species mass transfer interaction my simulation diverges. My mesh quality is fine and a 1-phase flow model converges correctly. I simulate a transient model.

I am using a k-omega SST turbulence model with the following parameters



I created a multiphase mixture model with:
- Phase 1 being a mixture template consisting water (l), nitrogen and oxygen
- Phase 2 being a mixture template of nitrogen and oxygen. 
The model parameters are as follows



Boundary conditions: 
Inlet: Velocity based where both species enter with the same velocity. Mass fraction of water in phase 1 is 0.962, phase 2 has a nitrogen mass fraction of 0.77. 
Outlet: Pressure based
Wall: Standard wall 
Symmetry: The 3D model of the venturi is split in half so this boundary is selected for the symmetry axis. 

For the methods I'm using the following parameters which were chosen to make the solution less complex: 


Currently I have left the Controls to their default setting:

I have tried to lower the Flow Courant Number to around 10 to stabilise the simulation however this did not work. 

I am using standard initialisation computed from the inlet. 

What can be the reason that my simulation diverges? 

Thanks

[SRP]

Hi,

Can you please share your residual plot?

Thank you.

[Simon Tempelaars]

Hey, 

This is the residual plot. I failed to mention that in this simulation I am currently only simulating the species transfer of nitrogen from water to air. 



For this simulation I lowered the under-relaxation factors for pressure, momentum, turbulent kinetic energy and spec. dissipation rate: 


Thanks for the reply