> clearly. What I am interested in is the discretization of Stokes equation
> (here in weak form)
> a(y,v)+b(p,v)=(u,v) plus incompressibility and boundary conditions.
Just to be clear: y and v are vector valued, and p is a scalar? How about your
control function u? It must be a vector too, right?
> I now want to assemble N. The ingredients I have are the Stokes FE system
> from which I have assembled K, B, M_y,M_p and am now
> looking for a way to get the rectangular N.
What variables are in your FESystem, and consequently in your DoFHandler? If
your DoFHandler discretizes y,p,u, then you could build all the matrices that
correspond to your problem in one BlockSparseMatrix, where the (0,2) block
would then be exactly N.
> I guess if the numbering of the nodes is such that K=blkdiag(K_1,K_2) for
> two components I could assemble a matrix [M_y,N_1^{T};N_1 M_p]
> and then construct my N from the N_1 block. Would this be a sensible way to
> approach this?
Yes.
W.
-------------------------------------------------------------------------
Wolfgang Bangerth email: [email protected]
www: http://www.math.tamu.edu/~bangerth/
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
