[CIG-SHORT] pseudo plane-strain in Pylith

Fri Apr 5 16:32:39 PDT 2013

Hi Lucas,

Since I was going from flow code to PyLith, I just created a spatial database by subtracting the pore fluid pressure from the fault normal stress that I had already computed.  I doubt that you can rely on a node numbering relationship, although it may work for some very specific cases.  Using the coordinates would definitely be more robust.  An alternative is to figure out the mapping from one mesh to the other (you just have to do this once using coordinates).  Then you would have the indices to use.

Good luck,
Charles

On 6/04/2013, at 11:42 AM, Lucas Abraham Willemsen wrote:

> Thanks for your suggestions,
> 
> I tried the pseudo-dirichlet you suggested Brad, but I did not have good results. I think the best way of dealing with this would be to probably be to do transfer the information from the 2D mesh in Pylith to the 3D mesh for the flow simulation. The 3D mesh would then be generated by extruding the 2D mesh so they have exactly the same in-plane geometry. 
> 
> I haven't written the flow code or the connecting piece myself and I am cooperating with a colleague for that part. I'm wondering if it would be hard to connect the information in the different meshes. Charles, do you have any suggestions on how you have done this? Is there some pattern in the node numbers you exploited when feeding your 3D pressures into the 2D Pylith mesh? Or did you look at the x-z coordinates for nodes in the 3D mesh and find the corresponding node with same x-z coordinates in the 2D mesh?
> 
> I will discuss this option with my colleague and if it turns out it is not viable I may reopen this thread to find out why the pseudo-dirichlet does not seem to work properly.
> 
> cheers,
> Lucas
> 
> 
> From: cig-short-bounces at geodynamics.org [cig-short-bounces at geodynamics.org] on behalf of Charles Williams [willic3 at gmail.com]
> Sent: Friday, April 05, 2013 18:14
> To: cig-short at geodynamics.org
> Subject: Re: [CIG-SHORT] pseudo plane-strain in Pylith
> 
> Hi Lucas,
> 
> I've done a similar model, but going the other way (feeding fluid pressures from a flow code into the fault-normal stress for PyLith.  To do this, I used a 2D mesh for PyLith and a 3D mesh (1 element thick) for the flow code.  If you are careful when designing the meshes, you can have elements that are identical, except that the quadrilaterals are extruded to form hexahedra.  The flow code provided identical fluid pressures on the front and back faces, so I was able to select one to use.  Hopefully, you can do something similar for your problem.
> 
> Cheers,
> Charles
> 
> 
> On 6/04/2013, at 9:07 AM, Lucas Abraham Willemsen wrote:
> 
>> Hello,
>> 
>> Currently I am trying to investigate the geomechanical response of a reservoir with a single fault when CO2 is injected. The fluid-flow part of the problem is performed by a code which is coupled to Pylith. The results from Pylith are fed into the flow code, which marches time forward and the  new fluid pressures are then fed back into Pylith, etc.  
>> 
>> The CO2 sequestration geometry I am considering is plane-strain, but the complication is that the flow-code (currently experimental) only works with hexahedrals. So I turned the 2D problem into a 3D-problem that is 1-element thick in the out-of-plane direction (y-direction). To enforce plane strain conditions I put zero displacement in the y-direction on the front and back plane. But the problem is that the fault intersects these front and back planes. This is a natural consequence of turning the 2D geometry with fault into a 1-element thick 3D geometry with fault. 
>> 
>> All of the input parameters are independent of 'y' since I am trying to model a plane strain problem. So there should be no reason for slip in the y-direction on any of the fault nodes (intuitively speaking). But Pylith does not allow me to explicitly constrain the y-displacements of the fault where they intersect the front and back plane. Therefore, the only way to run the model is by excluding the nodes on the fault from the front- and back- plane nodesets. But this way the boundary condition on these fault nodes is effectively zero normal stress, and elastic extrusion of material takes places. (Applying Neumann B.C. with normal stress to these fault nodes on the front and back plane to somehow reduce the extrusion does not seem to affect the fault nodes either. It would not be a great solution anyway).
>> 
>> Is there a way to deal with faults in this 'pseudo plane-strain' geometry I described?
>> 
>> sincerely,
>> Lucas
>> _______________________________________________
>> CIG-SHORT mailing list
>> CIG-SHORT at geodynamics.org
>> http://geodynamics.org/cgi-bin/mailman/listinfo/cig-short
> 
> Charles A. Williams
> Scientist
> GNS Science
> 1 Fairway Drive, Avalon
> PO Box 30368
> Lower Hutt  5040
> New Zealand
> ph (office): 0064-4570-4566
> fax (office): 0064-4570-4600
> C.Williams at gns.cri.nz
> 
> _______________________________________________
> CIG-SHORT mailing list
> CIG-SHORT at geodynamics.org
> http://geodynamics.org/cgi-bin/mailman/listinfo/cig-short

Charles A. Williams
Scientist
GNS Science
1 Fairway Drive, Avalon
PO Box 30368
Lower Hutt  5040
New Zealand
ph (office): 0064-4570-4566
fax (office): 0064-4570-4600
C.Williams at gns.cri.nz

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://geodynamics.org/pipermail/cig-short/attachments/20130406/ffa0ff8f/attachment.htm 