Dear users,
First and foremost, sorry to bother, this is the second question in two 
days :/ I have been trying to parallelize a simple code that computes the 
solution to the torsion problem for an ideal 3D body, i.e. a warping 
function. The solution can be written in terms of an harmonic function that 
must satisfy an inhomogeneous Neumann condition at the boundaries. This 
arises when we try to enforce that the derived stress vector must be 
tangent to the boundary. The condition is as follows
du/dn = y*nx - x*ny 
where n = [nx, ny] is the normal vector. The solution u(x,y) must satisfy 
Laplace equation in the domain. 
The serial code works just fine and the results agree with a few known 
cases. I have been trying to parallelize it using step 40 as a reference. 
The code is similar to the tutorial since the problem is almost the same. 
When I try to run it in parallel with more than one processor the solver 
sometimes converges to a wrong solution. I have no idea why this is 
happening but I know it must have something to do with enforcing the 
boundary condition where two subdomains meet. 

[image: Screenshot from 2020-05-08 10-40-27.png] <about:invalid#zClosurez>














The green rectangle and the red one are a zoomed portion of two subdomains 
and the black arrow represents the stress vector. That vector should be 
tangent to the boundary but right at the intersection this does not happen. 
Sometimes the parallel solver converges to the right solution. I have tried 
changing the domain, the solver, the preconditioner but nothing worked so 
far. I suspect that I might be doing something very stupid but I haven't 
been able to correct the code.

-- 
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see 
https://groups.google.com/d/forum/dealii?hl=en
--- 
You received this message because you are subscribed to the Google Groups 
"deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/dealii/e99a5230-61ca-4736-b43a-0bcdb6c243c7%40googlegroups.com.

Reply via email to