User demand should return concentrations instead of removing it

[evetion]

Because the UserDemand in and outflows are on separate edges, #1849 sees them as different boundaries altogether, ignoring that the return flow is actually a part of the inflow.

edit: This should work now in Ribasim, leaving it open to investigate how much Delwaq supports this.

[visr]

In time, we may want to support some process or mass injection or removal at the UserDemand node. Until then, indeed it is best to assume the outflow concentrations are equal to the inflow concentrations.

[evetion]

Injection is already support in the current PR, there's the UserDemandConcentration table.

[visr]

That setting the concentration, not injection, right? Injection adds mass to the inflow.

[evetion]

Each outflow from the UserDemand is multiplied times the set concentration of the UserDemandConcentration table, so it's a mass injection?

[gijsber]

The result is a mass injection and I assume that the User demand return flow has only one concentration, either as specified in this table or (when this table is not populated by the user) the same concentration as the abstraction.

[evetion]

Note that the offline Delwaq coupling correctly returns concentrations for cyclic UserDemands (from/to the same Basin, as the flows are merged), but like Ribasim not for the other User Demands (seperate boundaries). I'll investigate with Erwin how to support this.

update: It seems Ribasim does now support this, but a change/workaround is needed for Delwaq. Best is to setup a small toy model to test this (just a userdemand and a basin or two).