[petsc-users] CFD code with immersed boundary method

Matthew Knepley knepley at gmail.com
Wed Nov 7 15:40:58 CST 2012

On Wed, Nov 7, 2012 at 10:59 AM, Arun Kamath <arun.kamath.m at gmail.com> wrote:
> Yes, I would like to know if it is possible to do what DA does for an
> unstructured grid.

There is a way to do this now. You can use DMComplex, from petsc-dev.
Here is the idea:

  1) You must input the topology of your mesh as a DMComplex object

  2) You make a PetscSection object encoding your data layout

  3) You call DMSetDefaultSection(dm, section)

  4) Now the regular DMGetGlobal/LocalVector() and DMCreateMatrix()
work as normal

I am almost finished with the manual section on this. I will push it
soon. First, how is your
mesh expressed, since we have helper functions for importing meshes.


> Thanks.
> On Wed, Nov 7, 2012 at 3:31 PM, Arun Kamath <arun.kamath.m at gmail.com> wrote:
>> Hi!
>> I am working with a CFD code  where I need to solve the Poisson Pressure
>> equation. We solve a two phase flow problem which results in a jump in the
>> coefficients and a non symmetric system.
>> The boundary conditions are fulfilled using immersed boundary method ghost
>> cells at the solid boundaries of the domain.
>> I read about the DAs to deal with the boundary conditions, but they are
>> applicable to only "IJK" matrices as per my understanding.
>> So when I am using the "Aij" (CSR) matrix, how would you suggest I deal
>> with the ghost cells of the physical domain while forming the matrix and the
>> vector?
>> Thanks a lot.
>> --
>> Arun Kamath
>> PhD Stipendiat
>> NTNU, Trondheim
> --
> Arun Kamath
> PhD Stipendiat
> NTNU, Trondheim

What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to which
their experiments lead.
-- Norbert Wiener

More information about the petsc-users mailing list