Hi,
I am trying to solve the laplace equation on an adapted grid. When I call
constrain_element_matrix_and_vector() with asymmetric_constraint_rows set to
false, The constraint on the hanging nodes does not seem to be satisfied.
Instead the solver seems to set all hanging nodes to zero, as shown in the
picture attached. If I set asymmetric_constraint_rows to be true, on the other
hand, I need to use gmres, which is something I'd like to avoid. Is there a
reason for this? I see that nothing is done to ensure the application of the
constraint with asymmetric_constraint_rows set to false, beyond multiplying the
stiffness matrix with the constraint matrix. However, after this, the diagonal
values at the constrained point are set to 1.0, which will lead to zeros in the
solution as the corresponding term in the rhs is set to 0.0. Are solvers that
assume symmetric matrices bound to fail for Adaptive Mesh Refinement?
Thanks,
Subramanya
------------------------------------------------------------------------------
How ServiceNow helps IT people transform IT departments:
1. Consolidate legacy IT systems to a single system of record for IT
2. Standardize and globalize service processes across IT
3. Implement zero-touch automation to replace manual, redundant tasks
http://pubads.g.doubleclick.net/gampad/clk?id=51271111&iu=/4140/ostg.clktrk
_______________________________________________
Libmesh-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-users
