Has anyone experience with computing the flow of heat and/or contaminants in large bodies of water? In our application, there are significnat thermal gradients, but velocites are very small (<1 inch/s) and relevant time scales are on the order of days. What are the important issues? Can commercial packages handle such flows?

(1). The problem is you need to model it first. (2). The geometry, and the boundary conditions need to be specified first by you. (3). Some commercial codes should be able to handle the natural convection problem as well as the diffusion problem.

I think that you must model the problem first and the geometry, and the boundary conditions need to be specified by you like says John.

For this kind of problem you can contact David Glynn at cfd@flowsolve.com in order to discuss a bout some difficulties of the model.

I totally agree with Mr. Chien's recommendation. Once you model your problem properly, you will have some ideas if you need to include the time term or the convective term in N-S equation. You, then, can narrow your choices of commercial codes.

Jared

Start from the original equations. Non-dimensionalize them and look at order of magnitude of the various terms to decide which terms can be (safely) eliminated. Time scale in the order of days does not mean much; it all depends on what you are looking for - whether you want to see what happens during the first hour, what happens in days, or in a year ... This will determine whether you have a quasi-steady problem at hand or not. That may also determine whether you have a purely diffusive problem or whether natural convection should also be included. That's all one can say given the amount of info you've provided

Adrin Gharakhani