Negative concentrations

Hello everybody,

I'm working on a saltwater intrusion problem. The results indicate a zone of negative mass concentration values near the coastline. Mass concentration becomes zero within the inner parts of the aquifer. Is this a numerical method problem or a parameter problem? Should I refine the mesh? The problem is that the aquifer is quite big and a dense aquifer would result in a really time intensive model.

Any ideas would be great help

Thanks

George

You can to refine the mesh in the coastline. For mass transport problems feflow suggest to not accept more than 5% of obtuse angled nodes (triangular mesh).
And to review the boundary conditions of flow and mass for the sea.

Paola

The reason for the negative concentrations are probably numerical oscillations (I know this from a lot of FEFLOW computations). Mesh refinements could help but probably they will not solve the whole problem.

You could try different Upwind techniques (temporal and controll data menu).

Or you have to adjust your dispersivity. Your mesh Peclet number (edge length of your elements divided by dispersivity) should be less than 1. So a high dispersivity willhelp, but of course it will also smooth your results.