Posted Thu, 30 Jul 2015 23:54:27 GMT by LK81
Hi all,

I am applying some Dirichlet conditions (with max constraint) to the locations of some river nodes in a steady state Richard´s model. The fixed head conditions are equal to the respective node elevation. Although the calculated hydraulic head in the steady state solution for most of these nodes is higher than the corresponding elevation, the BCs are deactivated. Has anyone had a similar experience or knows what can be the reason for this? This causes my model to be saturated and "pressurized" at the fixed BC nodes.

I have already checked that there is a gradient from the neighboring nodes towards the boundary condition, so this could not be the problem. Also, I do not want to use fluid-transfer in this case which could be an alternative.

Many thanks,
Posted Fri, 31 Jul 2015 10:48:10 GMT by Björn Kaiser
Seepage face Boundary Conditions (h=z and max flow-rate constraint) combined with a steady-state representation of the system needs to be considered with caution. Configurations where the water level h swings around z may trigger in ambiguous constrain activation or constrain deactivation. This situation intensifies if (even small) numerical oscillations are also present.

Accordingly, I recommend switching to a quasi-steady approach by adopting time-constant boundary conditions and time-constant material properties. In this manner, you enable the system more degree of freedom to evolve “through time”. In a steady-state mode you only have one step which may result in a wrong decision on whether or not the constraints are activated. If you switch to a quasi-steady approach FEFLOW has more (time)-steps for the decision and fluctuations may disappear.

To switch the setting please go to the Problem Settings and switch to Problem Class – Transient.
Run the model over a certain period and until a steady-state a solution is reached. If the solution between two different time steps do not change and if the capture and release from the storage is close to zero you may consider the solution as steady-state.

You must be signed in to post in this forum.