[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [ESPResSo-users] Predefined E-field and irregular boundaries.
From: |
Stefan Kesselheim |
Subject: |
Re: [ESPResSo-users] Predefined E-field and irregular boundaries. |
Date: |
Sat, 3 May 2014 08:30:24 +0200 |
Hi Joe,
I have implemented stuff for that. It is not yet released but can go into the
master very soon (monday perhaps).
What is there, is an "external potential" for all particles that is weighted by
some factors depending on the particle type (=reflecting the charge). We need
the values on a cubic lattice (anisotropic lattice constants are possible) and
it calculated the E-Field from the potential by taking the numerical derivative
with a certain stencil. This way it is assured, that energy is conserved up to
integrator precision.
The complex shape thing could be more difficult to do. We will soon (in a few
weeks) have something which can calculate the distance to a wireframe model on
a cubic lattice. Konrad is on holiday at the moment but it seems to work
already. Then we can use the same interpolation for the distance function and
plug the result into a potential. This will be a "tabulated constraint". You
will also need to tabulate your distance function somehow. One possible way is
solving the corresponding eikonal equation abs(\grad d)=1 with BC d=0. Maybe
this is easy in Comsol, I have never tried…. Then this can be read into the
tabulated constraint (which is not written yet but is really straightforward to
do).
Let's keep in touch, I think we have more or less everything you need….
Stefan
On May 3, 2014, at 3:11 AM, Joe <address@hidden> wrote:
> Dear All,
>
> I have a microfluidic channel and I have the E-field distribution calculated
> using COMSOL. As you can imagine the E-field values are distributed at
> discrete points in the domain. I don't know whether there is an easy way to
> apply this E-field distribution in espressomd and let espressomd calculate
> the E-field at any point in the space when needed.
>
> Also, my channel has complex shape/boundary whose curve doesn't have an
> analytical expression. Is there an easy way to apply this kind of confinement?
>
> Thanks,
> Joe