[CIG-SHORT] embedded fault plane

[Brad Aagaard]

Rebecca-

Looking at Carl's example a little further. The figure with the aspect 
ratio shows some very distorted cells (dark blue) adjacent to the fault 
surface. I suspect those cells are causing the problem. For reasonable 
convergence of elastodynamics problems, you need a tetrahedral mesh such 
that the worst aspect ratio in LaGriT is greater than about 0.25 
(preferably about 0.3 or 0.4). The LaGriT command "quality" will tell 
you the aspect ratio.

Note to CUBIT users:

CUBIT uses a different definition of aspect ratio. See the CUBIT 
documentation for more details; generally it is the inverse of the 
aspect ratio in LaGriT, so you a maximum aspect ratio of about 4. The 
precise number depends somewhat on which quality measure you use and the 
CUBIT documentation gives some general guidelines for the desired values 
for each quality measure.

Brad


Brad Aagaard wrote:
> Rebecca-
> 
> I think the error message indicates that PyLith has detected a 
> tetrahedral cell has multiple faces on a fault. This would indicate that 
> the LaGriT mesh does not conform to what PyLith is expecting. The most 
> likely explanation is that either you have some very distorted cells so 
> that multiple faces lie on the fault surface or the group of fault 
> vertices is incorrect.
> 
> If you have not gone through generating the mesh in examples/tet4, I 
> would start there. It uses a different approach that uses a surface for 
> the fault so that the initial mesh honors the fault. For an embedded 
> surface, you could simply extend your surface to the edges of the 
> domain. This approach is less likely to cause difficulties for 
> relatively simple geometries. Carl's embedding approach can be useful 
> for very complex geometries but it is easier to stray from what PyLith 
> expects.
> 
> Brad
> 
> 
> Rebecca Bendick wrote:
>> Hi, everyone.  This may be a very rookie question; I'm just trying to  
>> learn how to use PyLith.  Are there any examples using a dislocation  
>> on a fault plane embedded in a larger volume?  I've had no trouble   
>> reproducing results for a throughgoing fault (cutting though the whole  
>> model volume) using the examples provided in the distribution and in  
>> the latest manual.  However, now I've used the demos from LaGRiT, in  
>> particular variations on the example at http://meshing.lanl.gov/proj/crustal_dynamics_CFEM/insert_fault/gallery.html 
>>   to create a fully embedded dislocation, and I can't get through the  
>> PyLith run.  Here's the screen dump, but (given my newbie-ness) maybe  
>> an already-worked example would be more helpful than debugging because  
>> I might have made many nested mistakes.
>>
>>   -- meshiolagrit(info)
>>   -- Reading finite-element mesh
>>   >> meshio/PsetFileBinary.cc:70:<unknown>
>>   -- psetfile(info)
>>   -- Reading binary Pset file 'tet4_1000m_binary.pset'.
>>   >> meshio/PsetFileBinary.cc:86:<unknown>
>>   -- psetfile(info)
>>   -- Reading 7 point sets from file.
>>   >> meshio/PsetFileBinary.cc:203:<unknown>
>>   -- psetfile(info)
>>   -- Reading point set 'fault' with 0 points.
>>   >> meshio/PsetFileBinary.cc:215:<unknown>
>>   -- psetfile(info)
>>   -- Done.
>>   >> meshio/PsetFileBinary.cc:203:<unknown>
>>   -- psetfile(info)
>>   -- Reading point set 'boundary_xm' with 287 points.
>>   >> meshio/PsetFileBinary.cc:215:<unknown>
>>   -- psetfile(info)
>>   -- Done.
>>   >> meshio/PsetFileBinary.cc:203:<unknown>
>>   -- psetfile(info)
>>   -- Reading point set 'boundary_xp' with 287 points.
>>   >> meshio/PsetFileBinary.cc:215:<unknown>
>>   -- psetfile(info)
>>   -- Done.
>>   >> meshio/PsetFileBinary.cc:203:<unknown>
>>   -- psetfile(info)
>>   -- Reading point set 'boundary_ym' with 287 points.
>>   >> meshio/PsetFileBinary.cc:215:<unknown>
>>   -- psetfile(info)
>>   -- Done.
>>   >> meshio/PsetFileBinary.cc:203:<unknown>
>>   -- psetfile(info)
>>   -- Reading point set 'boundary_yp' with 287 points.
>>   >> meshio/PsetFileBinary.cc:215:<unknown>
>>   -- psetfile(info)
>>   -- Done.
>>   >> meshio/PsetFileBinary.cc:203:<unknown>
>>   -- psetfile(info)
>>   -- Reading point set 'boundary_zm' with 49 points.
>>   >> meshio/PsetFileBinary.cc:215:<unknown>
>>   -- psetfile(info)
>>   -- Done.
>>   >> meshio/PsetFileBinary.cc:203:<unknown>
>>   -- psetfile(info)
>>   -- Reading point set 'boundary_zp' with 49 points.
>>   >> meshio/PsetFileBinary.cc:215:<unknown>
>>   -- psetfile(info)
>>   -- Done.
>> faults/TopologyOps.cc:330: failed assertion `numFaultCorners ==  
>> faceSize'
>> mpirun noticed that job rank 0 with PID 7750 on node kham.gs.umt.edu  
>> exited on signal 6 (Abort trap).
>> /Users/rbk/pylith/pylith-1.4.2-darwin-powerpc/bin/nemesis: mpirun:  
>> exit 134
>> /Users/rbk/pylith/pylith-1.4.2-darwin-powerpc/bin/pylith: /Users/rbk/ 
>> pylith/pylith-1.4.2-darwin-powerpc/bin/nemesis: exit 1
>>
>>
>>
>> Thanks in advance,
>> Becks
>> ---------------------------------------------------------------
>> Rebecca Bendick
>> SC 331
>> Department of Geosciences
>> University of Montana
>> Missoula, MT 59812
>> (406) 243-5774
>>
>> bendick at mso.umt.edu
>>
>>
>>
>> _______________________________________________
>> CIG-SHORT mailing list
>> CIG-SHORT at geodynamics.org
>> http://geodynamics.org/cgi-bin/mailman/listinfo/cig-short
>>
> 
> _______________________________________________
> CIG-SHORT mailing list
> CIG-SHORT at geodynamics.org
> http://geodynamics.org/cgi-bin/mailman/listinfo/cig-short
>