> My MATLAB code isn't available at the moment. It shouldn't be a
> problem translating it to Python. The problem is translating the Maple
> code that derives the element matrices to Sage. I'm not exactly thrilled
> with the design of the MATLAB code, but it works for my problems. I'd
> probably make it object-oriented when converting it to Python.

And which exact symbolic features are still missing in Sage, if any? I
mean, was the problem just that python is
different than Matlab, or just because something in Sage should be improved?

> What I mean by the derivation is the steps to get the finite element
> approximation. There are two basic approaches to this.
>
> 1. Galerkin method - Convert the PDE to the weak form.
>
> 2. Ritz method - Requires a variational statement for the underlying
> physical problem (e.g., Hamilton's Principle). The PDE is never
> explicitly
> formed since the FE approximation is used directly in the variational
> statement.
>
> The Ritz method is less general, since you need an underlying
> variational
> principle, but it is easier to handle coupled equations. Getting the
> weak form of the equations is difficult if the equations are coupled.
> None of the introductory FE textbooks that I've seen covers it. Examples
> always seem to use only one variable or multiple decoupled variables.
>
> For my problems, the Ritz method is applicable so I use it.
> Unfortunately,
> most packages assume you use Galerkin. In addition, they assume you
> already
> have the weak equations and like I've stated, getting them can be
> challenging.
> So, given that this is part of a symbolic package, there should be some
> support for that.

I got it now. Yes, I only need the Galerkin method. I have the
equation, I have the weak form. Now I want to solve it, on a general
geometry with general boundary conditions. I use libmesh for that:

http://libmesh.sourceforge.net/

it's a very good FEM library. But still maybe it could be done
simpler, using the SyFi approach or similar.

Ondrej

--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/
-~----------~----~----~----~------~----~------~--~---

Reply via email to