Jeff,
so your rho will be an expression or do you need to read in data?
Lutz
-----Original Message-----
From: Jeffrey E. McAninch <jeff at emc2fusion.com>
To: Lutz Gross <l.gross at uq.edu.au>
Subject: Re: [Escript_users] Examples of escript for 3D electrostatic
(poisson) with boundary conditions?
Date: Mon, 19 Apr 2010 09:21:36 -0600
Lutz,
Thanks for the link. I got sidetracked for the last few days, but it
looks like your tools will work well for me.
I am working on a simplified, poor-man's plasma physics code. Though,
with all the packages that have python
interfaces available, I'm starting to think I might reach a higher level
of sophistication that I had originally anticipated.
So to answer your question, ultimately the
rho will the time-dependent
net charge distribution of the ions and electrons.
Thanks,
Jeff
On Apr 18, 2010, at 5:13 PM, Lutz Gross wrote:
> Take a look at
>
http://esys.esscc.uq.edu.au/esys13/nigh ... node4.html
> This is more or less what you want. How exaclty is rho defined?
>
> Lutz
>
>
> Lutz Gross
> Earth Systems Science Computational Center
> School of Earth Sciences
> The University of Queensland
>
http://www.esscc.uq.edu.au
>
>
>
> ______________________________________________________________________
> From: Jeffrey E. McAninch [mailto:jeff at emc2fusion.com]
> Sent: 17-Apr-10 [Sat] 0:56
> To: Lutz Gross
> Subject: Re: [Escript_users] Examples of escript for 3D electrostatic
> (poisson) with boundary conditions?
>
>
> Lutz,
>
>
>
> Thanks for the offer to help.
>
>
>
Essentially I need to do the following (expressed as python
> meta-code):
>
>
>
> # p is a regular 3D mesh
> # with shape (3,101,101,101)
> #
> p = numpy.meshgrid[-1:1:101j,-1:1:101j,-1:1:101j]
>
>
> # phi_bdy is the defined "boundary" value
> # is_bdy is a logical telling whether a given point
> # on the mesh has a defined Vbdy or not
> #
> # phi_bdy.shape is (101,101,101)
> # is_bdy.shape is (101,101,101)
> #
> phi_bdy,is_bdy = set_boundary_values(p)
>
>
> #
rho is some charge density function on p
> # (would be 0. whereever is_bdy is False)
> #
> # rho.shape is also (101,101,101)
> #
>
rho = charge_density(p)
>
>
> # so need to
solve the Poisson equation
> #
del_squared(phi) = rho
> #
on the mesh
> phi = poisson_solver(p,phi_bdy,is_bdy,rho)
>
>
> Does it look to you like the tools in escript/finley would
> be a good starting point for trying to build "poisson_solver"?
>
>
> This is a
relatively simple, time-independent PDE, but is the
> basis for what
will evolve into a more complicated
> electromagnetic solve.
>
>
> Also, the example above is for ~1e6 mesh points,
> but will
ultimately be shooting for ~1e9 mesh points, so MPI
> is a big plus, which is why your tools caught my attention.
>
>
> Thanks for any help, examples, pointers you could provide.
>
>
> Jeff McAninch
>
>
>
>
>
>
> On Apr 14, 2010, at 7:44 PM, Lutz Gross wrote:
>
>
>
> > Not really but I am happy to help here.
> >
> > Lutz
> >
> > -----Original Message-----
> > From: Jeffrey E. McAninch <jeff at emc2fusion.com>
> > To: escript_users at esscc.uq.edu.au
> > Subject: [Escript_users] Examples of escript for 3D electrostatic
> > (poisson) with boundary conditions?
> > Date: Wed, 14 Apr 2010 18:01:51 -0600
> >
> >
> > New to this list...
> >
> >
Are there examples of using escript for 3D electrostatics with boundary conditions?
> >
> > –––––––––––––––––––––––––––––
> > Jeffrey E. McAninch
> >
> > jeff at emc2fusion.com