Multilayer Wells in 2D Cross-sections

Hi all,

Is there a method (or a tutorial) on setting up a multilayer well for a 2D cross-section?

Cheers,

Technically, a MLW is composed of a 1D Discrete Feature (DF) + a WellBC at the lowest elevated node.  Accordingly, you may use a 1D DF and a WellBC.

Great - thanks Bjorn.

Is there anything I can do to prevent 'upwards' flow from the well?

e.g. simulate the water injection through screened intervals?

When I tried to implement a well with a Discrete fracture condition as described, many of the streamlines were going straight up, as opposed to leaving the discrete fracture first.

Sorry if that doesn't make sense - I'll draw it out to try and describe better.

If you represent the well by a DF plus a WellBC at the lowest node with an injection (negative algebraic sign for the rate), a forced upward flow direction will inevitably occur. In contrast, if you extract water (positive algebraic sign for the rate), a downward directed flow direction will be triggered. DF act as preferential pathways for the groundwater flow induced by highly conductive material along the trace. In the standard implementation, DF's are perfectly hydraulically connected with the surrounding void space of the porous medium. If you want to have multiple screens along the trace of DF's I suggest to use either arbitrary node paths (http://www.feflow.info/html/help71/feflow/mainpage.htm#rhhlterm=arbitrary&rhsyns=%20&t=09_Parameters%2FDFE%2Fdfe.htm) or Dual-Node-Connectors (http://www.feflow.info/html/help71/feflow/mainpage.htm#rhhlterm=arbitrary%20node%20path&rhsyns=%20&t=09_Parameters%2FDFE%2Fnodesplit.htm).

Thanks Bjorn.

(Sorry if this next question is a little abstract.)

Could you briefly touch on how the dimensionality affects the well boundary condition for a 2D cross section?

Obviously the Well BC relies on a volume of water at a constant rate, and the 2D cross-section assumes a thickness of 1 m[sup]2[/sup] (if I understand the online help well enough).

Does the volume of water only apply to the 2D cross-section, and then - how can that accurately represent a 3D volume of water affectecting the system? i.e. what are the in-and-out constraints applied to the 2D model? Does the volume of injected or abstracted water apply only in the 2D (x/y) plane, and then surely that will be overcompensating the effect of the injection on the groundwater system?

A cross-sectional model has a unit depth of 1 m. accordingly, the volume of water only apply to the 2D cross-section (x/y-plane). If the lateral flow dimension in the direction of "unit depth" is large, you need either re-locate the cross-section (if the other lateral flow component can be neglected) or model in 3D (f the other lateral flow component cannot be neglected)