Pumping well in an axis symmetric model (Discrete Feature Element)

Hi FEFLOW forum,

I am quite new to FEFLOW so I plea for your forgiveness if these questions are silly.

I need to incorporate a pumping well, at the center of a small-scale, variably saturated, axis symmetric FEFLOW model. So far I have tried to use a Hagen-Poiseuille Discrete Feature Element (DFE) with either a Sink term or a 4th order well BC situated within the DFE. If I use the 4th order well BC I get a huge negative pressure (very low saturation) just around the pumping well which leads to instability in the simulation. If I use a Sink term there are no convergence problems.

My question is this: What exactly is the Sink term in the DFE options? It looks as if it is normalized by some volume hence the unit 1/d. I have tried normalizing my pumping rates with the volume of the bore casing but I feel that I am fumbling in the dark. I have tried to use the Help function but I dont see that it holds a solid explanation of the Sink term. In fact the conceptual understanding of the DFE besides the "different domain - different rules" seems a bit hard to grasp. How should I understand the DFE's extent in an axis symmetric model (e.g. selecting Rubberbox and then marking an area after pressing Discrete element). Is it in fact the extent of the vertical borehole and how is this related to the Thickness parameter (internal diamater of the casing?)?

I have tried searching this forum for information but I may have missed a thread/answer that applies to this post. If this is the case I apologize in advance!

Thank you,
Søren

Dear Søren,

the DFE source term refers to the unit (complete written) m^3/(m^3 d). It is a volumetric source.
However, in case of an axis symmetric FEFLOW model, it is a good approach to use a DFE which has to be located where the x-coordinate is zero. Ideally at the outermost left boundary. This DFE should be connected to a 4th kind boundary condition at the lowest point of the 'borehole'.

Any areal conditions in an axis symmetric model lead to cylindrical configurations. Which are usually unwanted.

Kind regards
Wolfram

Thank you very much Wolfram. I will use the approach that you suggest here.

The unit m^3/(m^3 d) still puzzles me as I am wondering what the different volumes refer to.

Anyways - this is sufficient for me - thanks again :)

Cheers,
Søren

As I am now succesfully interpreting drawdown data from a pumping test performed at my fieldsite I would like to share my experience on how to set up a well in an axis-symmetric model.

I abandoned the DFE combined with the 4th order well condition as it is unclear to me what the geometrical and hydrological interpretation of this setup is. Instead I moved the origin of the model one borewell radius (r_w) away from x = 0 so that the boundary related to the pumping well is now in x = r_w. Then I calculate the flux along the screen by normalizing the pumping rate by the area of the screen. The flux is then applied as a standard 2nd order flux condition along the screen. This boundary is equivalent to the BC applied in the analytical solution of flow to a well in a water-table aquifer by e.g Moench et al. (2001).

I did parameter estimation (of the aquifer parameters) using PEST and obtained a very good agreement between model and observation - and more importantly, realistic aquifer parameters in agreement with previous findings.