Re: cylindrical coordinates
Dear Daniel, Thanks for the reply. It is helpful to know about this. We have started working with the 3D Cartesian grid, first testing with a simple spherical diffusion problem for which an analytic solution exists. (We still make some use of symmetry, since we need to calculate only one octant of the solution...). The FiPy results agree really well with the analytic solution, and the calculation time is very acceptable for now. Best wishes, Martin On 10/10/2018 17:46, Daniel Wheeler wrote: > On Thu, Sep 20, 2018 at 4:15 AM Martinus WERTS > wrote: >> Now I would like to go 3D, and the symmetry of our system would allow to >> use a 2D cylindrical grid (r,z) - with zero flux at z=0 and z=L (and >> r=0), and either Dirichlet/zero flux at r=R. Looking at the mailing list >> archive and GitHub, it appears that cylindrical coordinates are at the >> moment not working properly (missing factor) > Yes, that's correct. Sorry about that. > >> Is this still the case? >> >> If so, I will start with a simple 3D Cartesian mesh, and then perhaps >> move to more adapted meshes, depending on the calculation speed and the >> precision required. Perhaps the cylindrical coordinates will be fixed in >> a near future? > Maybe, the developers of FiPy aren't doing an awful lot of work on it > so there is no timeline for fixing it. > >> Many thanks for any advice that you may have > Sorry that I can't be more helpful. > ___ fipy mailing list fipy@nist.gov http://www.ctcms.nist.gov/fipy [ NIST internal ONLY: https://email.nist.gov/mailman/listinfo/fipy ]
Re: cylindrical coordinates
On Thu, Sep 20, 2018 at 4:15 AM Martinus WERTS wrote: > > Now I would like to go 3D, and the symmetry of our system would allow to > use a 2D cylindrical grid (r,z) - with zero flux at z=0 and z=L (and > r=0), and either Dirichlet/zero flux at r=R. Looking at the mailing list > archive and GitHub, it appears that cylindrical coordinates are at the > moment not working properly (missing factor) Yes, that's correct. Sorry about that. > Is this still the case? > > If so, I will start with a simple 3D Cartesian mesh, and then perhaps > move to more adapted meshes, depending on the calculation speed and the > precision required. Perhaps the cylindrical coordinates will be fixed in > a near future? Maybe, the developers of FiPy aren't doing an awful lot of work on it so there is no timeline for fixing it. > Many thanks for any advice that you may have Sorry that I can't be more helpful. -- Daniel Wheeler ___ fipy mailing list fipy@nist.gov http://www.ctcms.nist.gov/fipy [ NIST internal ONLY: https://email.nist.gov/mailman/listinfo/fipy ]
Re: Boussinesq Equations
Don't you still have a $\nabla . \vec{u} = 0$ equation though? It
doesn't go away. That equation becomes like a constraint.
https://www.comsol.com/multiphysics/boussinesq-approximation
On Wed, Oct 10, 2018 at 5:58 AM fgendr01 wrote:
>
> Hi Daniel,
> Thank you for your answer.
> I thank you for trying to solve my problem.
> About my set of the equation here is my reasoning.
>
--
Daniel Wheeler
___
fipy mailing list
fipy@nist.gov
http://www.ctcms.nist.gov/fipy
[ NIST internal ONLY: https://email.nist.gov/mailman/listinfo/fipy ]
