ff3d-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [ff3d-users] How to define a slope domain?


From: Hao
Subject: Re: [ff3d-users] How to define a slope domain?
Date: Tue, 6 Mar 2007 16:49:21 -0500

Hi Stephane,

Thank you so much for your answer.

I still want to confirm
Is it possible to define the following domain(see cross section) in FEM, not FDM?
If yes, how to define the mesh M here?

solve(H) in  M krylov(precond=diagonal,type=bicg), bicg(epsilon=1e-5, maxiter=50000)
 {
...
}

        /|
      /  |
    /    |
  /      |
/        |
|        |
|        |
|____ |

Thanks,

Hao


On 3/6/07, Stephane Del Pino <address@hidden> wrote:
Le mardi 6 mars 2007 15:13, Hao a écrit:
> Hi,
Hello.

> > I am trying to solve a slope stability problem with ff3d.
> > The surface of space is not flat, but with a slope, such as 10 degree.
> > How to define the problem space in FEM? and how to identify the upper
> > surface?
The upper surface of a box is given by  'zmax' or the reference number 5 if
you build the mesh with ff3d.

> > I know there are two methods implemented in ff3d, one is FEM, another one
> > is FDM.
> > FEM can only run on regular hexahedral mesh defined  by vector a, b, and
> > n, is it right(here, FEM used only mesh)?
No. You can use also tetrahedral meshes build with external meshes in
format 'gmsh' or 'mesh'. Or, if the geometry is simple, you can generate it
with ff3d:
-------8<--------------
// build a cartesian grid of hexahedra
mesh m = structured(n,a,b);
// tetrahedrize it: each hexa is divided into 5 tetra
m = tetrahedrize(m);
// transform your mesh with a field
m = transform([f,g,h],m);
-------8<--------------
f,g and h are functions defining the transformation field

> > However, FDM can run on other irregular domain with the help of pov ray,
> > but the domain must be contained in the Mesh completely, is it true(here,
> > FDM used doamin and mesh)?
Yes.

> What is the  difference between FEM and FDM?  Does they use same algorithms
> to solve the problem?
FDM is fictitious domain method. In ff3d, it relies on a Q1 FEM method.

Best regards,
Stéphane.



_______________________________________________
ff3d-users mailing list
address@hidden
http://lists.nongnu.org/mailman/listinfo/ff3d-users


reply via email to

[Prev in Thread] Current Thread [Next in Thread]