Denim Umeshkumar Anajwala

Hi,

I'm having problems with the mesh generator. After adding in points from a .pnt file and manually attemping to move/delete one point, the Gridbuilder mesh generator no longer finds a solution and gives me the following message:

"A node with incidence zero arose" or sometimes "Node outside outer boundary, area    1"

There are no visible nodes outside the area, and I am pretty certain that this is not really the case, but perhaps a bug? The TMesh generator also fails to converge.

The strange thing about this problem is that it appears to corrupt all version of the supermesh that I was working on. I can load older versions of the file that I know for sure worked before, but I now get the same message. The only way I can get the mesh builder to converge is to delete all line add-ins, but that is the absolute last resort as the mesh design is quite intricate...and adding in lines from the map does not yield the element geometries that I require.

Any ideas?

Thanks.

Laurie,
you do not have to set a no-flow condition in FEFLOW. If no boundary condition is set, FEFLOW assumes no-flow. If you set a flux BC with a value of 0, however, FEFLOW will calculate the budget. Depending on mesh geometry etc. this might lead to fluxes that are not exactly 0, and summing over a huge number of nodes you might see a flux in the budget analyzer. The fluid flux analyzer is based on an integration over the velocity field. Nodal velocities at no-flow boundaries will seldom be exactly 0 depending on discretization, therefore the fluid flux analyzer typically will show some boundary flow even if there is none. You should always prefer the budget analyzer at model boundaries for this reason.

Peter

Thanks, that sounds like the right kind of thing. The only problem is that the model has a few layers due to the need to model a significant downthrow across a fault. The confined aquifer is the bottom layer...hence there will be a couple of layers above representing aquitards (in the unconfined section the parameters are changed so that the top layer represents the aquifer).

Does the confining layer need to be the top layer of the model? If not will there be a problem with layers going dry?

Cheers.

In 2D confined models, a depth-integrated Darcy velocity is given as FEFLOW does not have any information about aquifer thickness (transmissivity is used as the material property).

Thanks !

The reason for the error message is that by default FEFLOW tries to load the entire results to the memory for transient particle tracking - which is not possible in case of large files. In later FEFLOW versions, after a warning like the one mentioned FEFLOW should continue, but with obtaining the information from the hard disk. This is much slower, but at least allows to do transient particle tracking also for very large files.

Arnauld,

all the slices have the same extent. Please also see forum topic http://feflow.info/forum/index.php/topic,16.0.html, answer number 7:
"How can I model areally limited layers like sand lenses in FEFLOW? In FEFLOW for 3D modeling all layers have to extend
over the whole model area. For modeling areally limited layers, e.g. a sand lense, you can choose between two options: A)
Modeling of the sand lense by changing material parameters inside a model layer, B) Modeling the shape of the sand lense by
two extra slices which in the area where the sand lense does not exist form a very thin model layer with material
parameters of the surrounding stratigraphic layers."

Best regards,
Peter

Hi !

For my PhD, I want to use FeFlow to model a spring and its catchment area.

The model is a simple three layered aquifer whom layers are cut by the topography. But when I built it in FeFlow, despite the adaptation of the layer to the topography, it appears that all slices remain the same (i.e. they do not be clip by the intersection of topography, as pointed out by grey arrows in the below illustration) :

[url=http://i.imagehost.org/download/0445/FeflowGrid][img width=200 height=150]http://i.imagehost.org/0445/FeflowGrid.jpg[/img][/url]

So the questions are: Is it possible to clip slices (so to have slices with various geographical extents) ? And if is it possible, how to do that ?

Seepage face boundary conditions can also be applied around an excavated section within the mesh. Then you would simply not look at what happens inside the excavation.

Thanks Giovanni, I'll see if I can get that to work. 8)