Instability with transient modelling of discrete feature elements

Does anyone have experience/success using discrete feature elements in transient models? I am using an existing model which is stable in transient mode with complex arrays of multilevel wells and varying  boundary conditions. I am trying to add into the model vertical open boreholes which have been drilled to drain an upper perched aquifer through to a lower partially saturated one.

I have tried using discrete feature elements as single nodes connected vertically across several layers. I am using Darcy's law with a high conductivity (1 m/s) and a 1m2 cross section (they are BIG boreholes). My nodal spacing is a minimum of 1m and is tightly discretized around the discrete elements and I'm using extremely small timesteps to start, but the model is unstable. Any suggestions why it might be unstable? is there an alternate way I can model drainholes that will provide a stable transient solution?

(The instability is particularly apparent at the free and moveable upper surface which is also a seepage face boundary condition)


Thanks
Martin

Hi Martin,

Descrete element makes the model unstable in general, I would recommend you to make high K zone using small FEFLOW elements directly.

Cheers,
Robin

try removing the free and movable option...run confined to simplify the problem.  its likely that your model set up is ill posed.

Pete