dolfin team mailing list archive

["Garth N. Wells" <gnw20@xxxxxxxxx>]

On Aug 23 2007, Jake Ostien wrote:

> Hi,
>
> I am trying to compute some reaction forces, currently I am trying to 
> assemble the stiffness matrix and compute the matrix vector product with 
> the converged solution vector to get a vector of forces.  Here is what I 
> am trying.
>
> <u is the converged solution>
> Matrix A;
> assemble(A, *a, mesh);
> Vector f;
> A.mat().mult(u.vector().vec(), f.vec());
>
> First an issue.  I am getting a seg fault coming in 
> DiscreteFunction::interpolate as part of the assemble step.  I haven't 
> figured out why yet, but I'm looking.
>
> Next, a few questions.
> Is there an easier (or cleaner) way to accomplish this?

Yes, you should be able to assemble a linear form which will contain the 
reaction forces. For elasticity, plasticity, etc

 L = -grad(v)*sigma*dx - v*f*dx  (f is the body force)

Assemble this into a vector, and the vector will contain the reaction 
forces.

Garth

> I have a mixed formulation.  Is is possible to only assemble the 
> stiffness matrix for a SubSystem? 
>
> Thanks,
> Jake
>
> _______________________________________________
> DOLFIN-dev mailing list
> DOLFIN-dev@xxxxxxxxxx
> http://www.fenics.org/mailman/listinfo/dolfin-dev