problem with convergence using densitiy coupled flow

hi everybody,

i am testing a 2D vertical box model in order to get used to the modelling of density coupled flow.
i am applying constant head boundary conditions for flow and mass transport. boundaries for mass transport are 0 and 35000 mg/l simulating seawater conditions on one end and fresh water conditions on the other. after taking into account all the required settings (density ratio, reference and maximum mass, seawater head instead of hydraulic head) i can sea a well shaped seawater wedge in my model but it does not converge to the usual default error of 1e-03. i don't have much experience with convergence criteria and accepable errors at all. can anybody help me with that? are errors in the range of 1e-02 a comonly acceptable value which is just due to the complex coupled equations? if not what might be wrong in my model?

thanks in advance

Density-coupling leads to strong nonlinearities in the set of equations. For this, steady-state convergence may be difficult to achieve. Typically such kinds of models (also unsaturated models, for example) are then run in transient mode for a long time to obtain steady-state conditions. For density-dependent models, in this case the maximum time step lenth should be limited to avoid osciallations in the concentration.

thanks a lot for the useful advice!
it seems to work well although it takes a bit more time to run the simulation.
now i am working with the default value of 1e+01. is that a value one should work with or are there any further recommendations according to the purpose of the simulation?

You can try to use a maximum time step longer than 10 days. If you get oscillations in the concentration, reduce the maximum time step again.