Please wait...
×

Error



Posted Wed, 17 Feb 2010 21:44:00 GMT by Denim Umeshkumar Anajwala
Hi,

In a model that I am working on, I am estimating the flow across a series of small line segments. However, changing the darcy flux calculation to elemental rather than nodal gives very different values in places. Can anyone explain why this might be?

My model has a series of different K values e.g from west to east, with the node locations indicated by the yellow asterisks:

                                                                               [color=grey][b]NORTH[/b][/color]


[color=grey][b]WEST [/b] [/color]    |------------------------------[color=red]K1[/color]---------------------------|--------[color=green]K2[/color]-------|--------[color=blue]K3[/color]------|      [color=grey][b]EAST[/b][/color]
[color=yellow]            *           *          *          *          *          *           *          *           *          *           *          *     [/color]

                                                                                   |
                                                                                   |  
                                                                                   |   (flow direction)
                                                                                   V
                                                                               [color=grey][b]SOUTH[/b][/color]

...the model extends above and below the diagram (with nodes in the same pattern) and all flow going from north to south.

[b]With the nodal setting I get the following amounts:[/b]
Q in [color=red]K1[/color] = 52 m3/d
Q in [color=green]K2[/color] = 34 m3/d
Q in [color=blue]K3[/color] = 210 m3/d

[b]With the elemental setting I get the following amounts:[/b]
Q in [color=red]K1[/color] = 50 m3/d
Q in [color=green]K2[/color] = 8 m3/d
Q in [color=blue]K3[/color] = 240 m3/d

For some reason the flow in the [color=green]K2[/color] region seems to change significantly, where as the other two seem fairly similar, what might be the reason for this? Is the low number of nodes a problem? It's a fairly unrefined part of the model..

Thanks,

SB
Posted Wed, 17 Feb 2010 22:43:08 GMT by Denim Umeshkumar Anajwala
I tried refining the mesh significantly (~100 nodes in each K region) and the values appeared to converge a bit towards the discontinuous element values, but not completely - there was still a discrepancy. I'm still not sure which method I should use.
Posted Thu, 18 Feb 2010 07:01:56 GMT by Denim Umeshkumar Anajwala
The difference between the nodal and elemental modes is the way the Darcy velocity is calculated: In nodal mode, it is calculated at each node. Thus, when you calculate the flux across a line, the nodal velocities of all the nodes surrounding the cut elements are used. Hereby you may get an influence of conductivities in elements that are not cut directly by the line. In elemental mode, one velocity for each element is calculated, and only the Darcy velocities of the elements cut by the line are used.
It is hard to tell the reason for the difference in your case - it seems a relatively simple situation with not much room for getting a difference. However, maybe my explanations still help in detection.
Posted Fri, 19 Feb 2010 23:12:08 GMT by Denim Umeshkumar Anajwala
Thanks Peter, I think I'll go with the elemental one in this case.

You must be signed in to post in this forum.