[col16]

Hi there.

I am working towards a PhD, doing some research on (particular configurations/types of) underwater discharges of waste-water. When wastewater is discharged to the ocean, it is invariably of a different density to the surrounding fluid - unless of course the wastewater has exactly the same amount of salt in it than the ocean does (at that point). In other words, we are dealing with buoyancy created by the presence of salt.

My question, most simply put, is: Does OpenFOAM have any solvers which would be appropriate for salt-generated buoyancy, or more directly, has anyone written any solvers which would be appropriate?

From what I can see, the answer to the first part of my question is no. As far as I can see, 'buoyancy' in OpenFOAM means temperature is involved. This is not appropriate for my problem: one of the most basic assumptions we are making (or which our research is targetted at) is that both the inlet and the surrounding fluid are at the same temperature. In fact, we know that laboratory experiments where the buoyancy (to an equivalent level) is produced by temperature not salinity exhibit quite different behaviours. So I cannot just play with temperature differentials to get the effect I'm wanting.

In addition, 'incompressible' in OpenFOAM seems to automatically mean that the density (rho) is constant. This is a mistake that many textbooks make, so you're not the only ones. But the point is this: Incompressible means that density cannot change because of pressure, but it doesn't have to mean that density can't change due to anything else. In particular, salt can change densities in flows where the Mach number is much less than 0.3, and so which the incompressible assumption is still a very good one. Underwater discharges are a great case in point - velocities are much smaller than 450m/s. To use a compressible solver seems a bad use of CPU time.

I have had a go writing up the governing equations which I am interested in solving, and have put them on this page. This document is a draft... there are probably mistakes in it and bad logic. I was just trying to flesh out what I was actually wanting/needing to solve. In the short term I want to be able to do a RANS simulation of the problem; however with that done I wish to move to LES.

Maybe I have understood the description of the solvers incorrectly, and what I am describing is already possible. To be honest, I hope that's the case! Maybe on the other hand, someone else has written a solver themselves which can do this? Failing that I suspect that I will need to create a customised solver myself... a task which may take some time, considering I haven't used C++ before - and so something which I might need some pointers on!

Thank you very much for your help. It's definitely appreciated.
Cheers,
--cam--
Cameron Oliver

[eugene]

Hmm, what you want is a salt or lack of salt transport equation. Then you derive your fluid density from this.

Someone has written a buoyancy solver based on the Bousinesq approach. I think it might be available on the wiki. Take this, rip out the temperature transport and dependencies and replace with salt.

[dhebert]

Hello all,

My post is somewhat similar to Cameron's so I thought here would be a good spot.

I would like to use OpenFOAM with some small scale ocean circulation modeling that is fully 3 dimensional (that is, not using hydrostatic equations of motion). For simplicity density can be computed in a linear fashion, typically rho = alpha * T + beta * S, where T and S are temperature and salinity fields, alpha and beta are expansion/contraction coefficients. My questions regarding this are:

1) where to incorporate an equation of state in OpenFOAM? Could I simply copy the perfectGas code and modify the rho equation? Beta would need to be added somewhere wouldn't it?
2) Does it make more sense to add a thermophysicaModels\liquid called seawater that has the simple equation of state? There seems to be alot of extra stuff in the H20 that I might not need.
3) Since I am going to be treating the flow as incompressible, is turbFoam ok to base my solver on (and add transport equation for T and S), or do I need to start from rhoTurbFoam?

Thanks in advance for any suggestions/help.

David