Björn Kaiser

Your question is difficult to answer here in the forum. Could you please contact the FEFLOW support?

https://www.mikepoweredbydhi.com/support

http://www.feflow.info/html/help71/feflow/mainpage.htm#t=09_Parameters%2FBoundary_Conditions%2FFlow%2Fhydraulic_head_bc_saltwaterhead.html

Thank you for your note. We will fix the link as soon as possible.

As you correctly pointed out, you have the possibility to represent the wells by means of fixed heads (Dirichlet). Use a maximum flow rate constraint with 0 m³/d to avoid any artificial infiltration. You may additionally impose a minimum flow rate constraint with an exfiltration rate corresponding to the pumping rate. The latter case enables you to avoid that more water is leaving the system as the pumping well is technically able to handle (pump capacity).

In 2D axisymmetric projection the flow is not integrated along the length and unit depth of one meter. Instead, the flow is integrated over the area spanned over the rotation axis.

In which range (numbers) do you encounter the difference and for which variable / BC / material property? Is it outside the numercial noise?

Theoretically, a debug and a release can be behave different with regard to e.g. numerical noise, uninitialized memory read. However, a different behavior is quite exceptional. Which API functions and callbacks are you using? Please let me also know the Operating System, FEFLOW version and compiler.

You could try to use the Fluid-Flux BC (Gradient):
http://www.feflow.info/html/help71/feflow/mainpage.htm#t=09_Parameters%2FBoundary_Conditions%2FFlow%2Ffluid_flux_bc_gradient.html

It's difficult to say anything about the reason for the high imbalance. In general, a groundwater flow model with a small imbalance for the fluid does not necesserily imply a small imbalance for the temperature. In most of the cases the large imbalance can be attributed to the spatial descrizitations. For more details, I suggest to contat the FEFLOW Support: https://www.mikepoweredbydhi.com/support

Thank you for further clarification. I understand the problem now. Yes, you are right, FEFLOW does not have the option to assign a time-series to a SaltwaterBC. This is a missing feature. I will add this missing feature on the wish list for possible future developments. In the meanwhile, I suggest to assign the time-series to the equivalent fresh-water head (Hydraulic-head BC). Of course, you need to recalculate the time-series before you assign. Thank you for the note on the missing feature.