River influence in dewatering project

[Vera-geo]

I'm working on a project with some rivers and would like some advice if possible.
Started with steady, to deal with less parameters first.

Farmsum_114 (1).ipynb
Some general information:
pumped aquifer, horizontal k of 20 m/day.

"Zee" to the north, has a depth of around-13 mNAP, waterlevel 0.0 mNAP (not true tidal, but assumptions...)
"Oost" to south has a depth of around -5 mNAP, waterlevel 0.53 mNAP.
Assumed both rivers to be straight (not true) and at a distance of 330 and 600 meters (on average). Also assumed contact of both rivers into the pumped aquifer (starting at -5 mNAP). Could off course up the resistance of Oost to 100 days, to add for the probable prescence of left over resistance in the aquitard. Both 'rivers' are for shipping

Adding a resistance of 1 day between layers, I have noticed a large dependency on this number. f.e. end of sheet piles 1 day resistance etc.

1. I have the feeling I'm off with regards to anisotropy, using ModelMaq, thoughts?
2. Run with both sheet piling and wells untill -15 mNAP, gives a total discharge of 7236 m3/day.
   BUT run with sheet piling untill -13 and wells untill -15 gives lower total discharge of 3586 m3/day. I assumed discharge would be slightly higher not lower.

[mbakker7]

@Vera-geo , thanks for your question.

You are right, starting off with ModelMaq is not what you want. For your case, use Model3D. There you can specify a semi-confining layer at the top, and a vertical anisotropy in the aquifer. The resistance between aquifer layers is then automatically computed by timflow.

Crank up the order of the ImpermeableWallString, as order 3 is certainly too low with those wells close by.

The two rivers don't have to be straight, as you are certainly aware, but you also need to crank up that order, or use a few extra segments, to make sure the head along the rivers is indeed equal to the value you specify.

Try these things out and let us know if you have any further questions.

[Vera-geo]

Way better already. Also I added a more realistic width to 'wh' in riverstring. But I'm guessing it sees the line as the centerline, so when adding width I should move the line to the center of the river instead of the edge.
My canals are quite wide, >100 m.

[MattBrst]

If the width of the canals is important to you, maybe you should consider modeling those as PolygonInhoms.

[Vera-geo]

@MattBrst Good idea! Thanks.