I would like to run a simulation where species concentration from Advection-Diffusion influences the density, and drives buoyancy flow. Unfortunately, there is only "Heat Expansion Coefficient" and "Reference Temperature", but no solute coefficient or reference concentration.
Is this possible in Elmer now, but not documented? If not, how can I add it?
Solutal buoyancy
-
raback
- Site Admin
- Posts: 4841
- Joined: 22 Aug 2009, 11:57
- Antispam: Yes
- Location: Espoo, Finland
- Contact:
Re: Solutal buoyancy
Hi
In the Navier-Stokes equation you might try to set the "Compressibility Model" to "User Defined" and then give "Density" as a function of the consentration. All this in the "Material" section. This should take care of the correct treatment of density in the N-S equations. However, for this material law the bodyforce is not accounted for automatically. Hence, you need to give the components of the gravity vector in "Body Force" section as "Flow Bodyforce i", i=1,2,3. This will be multiplied with density in the code. Did not try it but it might still work...
As you have a loose coupling between N-S and the bubble consentration you might use most of the bullets in the coupled level i.e. just take one nonlinear iterations within the specific solvers. To boost convergence a relaxation of ~0.5 is often a good guess.
-Peter
In the Navier-Stokes equation you might try to set the "Compressibility Model" to "User Defined" and then give "Density" as a function of the consentration. All this in the "Material" section. This should take care of the correct treatment of density in the N-S equations. However, for this material law the bodyforce is not accounted for automatically. Hence, you need to give the components of the gravity vector in "Body Force" section as "Flow Bodyforce i", i=1,2,3. This will be multiplied with density in the code. Did not try it but it might still work...
As you have a loose coupling between N-S and the bubble consentration you might use most of the bullets in the coupled level i.e. just take one nonlinear iterations within the specific solvers. To boost convergence a relaxation of ~0.5 is often a good guess.
-Peter
Re: Solutal buoyancy
Thanks very much, this works nicely! And it couples with the thermal buoyancy driving force as well -- though I haven't benchmarked it just yet.
On the other hand, it seems like a large hammer to use for this -- to have a body force of rho*g. How much work would it be to include this in the more elegant Boussinesq framework?
On the other hand, it seems like a large hammer to use for this -- to have a body force of rho*g. How much work would it be to include this in the more elegant Boussinesq framework?