[aspect-devel] Cylindrical coordinates

[Jonathan Perry-Houts]

Ah, cool. Thanks for clarifying. That does seem like it would be useful,
and presumably trivial to adapt from a cylindrical solver.

Cylindrical coordinates has been on my radar for a while, but I'm
probably not going to pursue it right now. I'm trying to wrap up this
whole dissertation thing, and need to weigh the time commitment of
adding this feature vs. time to just run the models in 3D. Seems like 3D
Cartesian wins again. As always, XKCD sums up my predicament well:
https://xkcd.com/974/

Anyway, if someone at the Hackathon is looking for something to do,
maybe this could be a fun project. Just throwing the idea out there :)

-JPH


On 05/10/2018 11:10 AM, Max Rudolph wrote:
> My understanding is that in 2D, the spherical shell model is equivalent
> to a 2D annulus.
> I was suggesting the spherical annulus, which is actually a three
> dimensional equatorial slice with a very small latitudinal opening
> angle. This is like taking a (thin) slice of pizza, tipping it sideways,
> and making a volume of revolution :). In this geometry the area ratios
> of the surface and CMB are preserved. See Hernlund and Tackley 2008:
> https://www.sciencedirect.com/science/article/pii/S0031920108001921
> 
> On Thu, May 10, 2018 at 11:02 AM, Jonathan Perry-Houts
> <jperryh2 at uoregon.edu <mailto:jperryh2 at uoregon.edu>> wrote:
> 
>     Do you mean the axisymmetric case with a non-zero inner radius (i.e.
>     still in (r, z) coordinates)? The perpendicular plane (r,theta)
>     already exists as the 2D 'spherical shell', correct?
> 
> 
>     On 05/09/2018 11:07 PM, Maxwell Rudolph wrote:
> 
>         If you decide to pursue this it might make sense to also
>         implement the 2D spherical annulus simultaneously.
> 
>             On May 9, 2018, at 10:16 PM, Jonathan Perry-Houts
>             <jperryh2 at uoregon.edu <mailto:jperryh2 at uoregon.edu>> wrote:
> 
>             Thanks Wolfgang and Felipe. Sounds like there's no library
>             magic that saves me in this case. Thanks for the advice!
> 
>             -Jonathan
> 
>                 On 05/09/2018 08:11 PM, Wolfgang Bangerth wrote:
> 
>                     On 05/10/2018 10:16 AM, Jonathan Perry-Houts wrote:
> 
>                     Does anybody have a sense of how difficult it would
>                     be to implement 2D
>                     cylindrical coordinates in Aspect? It doesn't look
>                     like there's support
>                     for non-Cartesian derivatives in Deal.II, but maybe
>                     somebody has a
>                     suggestion for hacking something together.
> 
>                 You'd have to rewrite the assemblers to implement the
>                 bilinear form that describes the equation in cylindrical
>                 coordinates. You'd also have to hack all of the other
>                 places where we do any integrals (nullspace removal,
>                 postprocessors).
>                 It's likely going to be a substantial amount of work --
>                 not difficult, just going through it all.
>                 Best
>                   W.
> 
> 
>             -- 
>             Jonathan Perry-Houts
>             Ph.D. Candidate
>             Department of Earth Sciences
>             1272 University of Oregon
>             Eugene, OR 97403-1272
>             _______________________________________________
>             Aspect-devel mailing list
>             Aspect-devel at geodynamics.org
>             <mailto:Aspect-devel at geodynamics.org>
>             http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel
>             <http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel>
> 
>         _______________________________________________
>         Aspect-devel mailing list
>         Aspect-devel at geodynamics.org <mailto:Aspect-devel at geodynamics.org>
>         http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel
>         <http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel>
> 
> 
>     -- 
>     Jonathan Perry-Houts
>     Ph.D. Candidate
>     Department of Earth Sciences
>     1272 University of Oregon
>     Eugene, OR 97403-1272
>     _______________________________________________
>     Aspect-devel mailing list
>     Aspect-devel at geodynamics.org <mailto:Aspect-devel at geodynamics.org>
>     http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel
>     <http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel>
> 
> 
> 
> 
> _______________________________________________
> Aspect-devel mailing list
> Aspect-devel at geodynamics.org
> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel
>