Closed abarret closed 7 months ago
I've pushed changes to account for physical boundary conditions. Note that this branch now targets physical_bcs
and should be merged before #68.
convec_op
is 2nd order accurate.
channel
with uniform grid runs for fine creeping_flow=FALSE (and also for TRUE). channel
on AMR grid produces the following error:
P=00000:Error in hier::GridGeometry<DIM> object with name = CartesianGeometry: in computeMaxGhostWidth(): Cannot add variables and increase maximum ghost width after creating the GridGeometry!
for creeping_flow=FALSE. It runs fine with this flag set to TRUE.
time_stepping
.
For a uniform grid, it produces the error: P=00000:Failed assertion: un_bc_coefs.size() == NDIM
. Seems like the input file might be missing the boundary conditions?
For an AMR grid, it produces the error: P=00000:Error in hier::GridGeometry<DIM> object with name = CartesianGeometry: in computeMaxGhostWidth(): Cannot add variables and increase maximum ghost width after creating the GridGeometry!
Grrrr.... The base class requires that boundary conditions are registered even if we are using periodic boundaries...
I will fix this tomorrow or Friday.
channel
with uniform grid runs for fine creeping_flow=FALSE (and also for TRUE).channel
on AMR grid produces the following error:
P=00000:Error in hier::GridGeometry<DIM> object with name = CartesianGeometry: in computeMaxGhostWidth(): Cannot add variables and increase maximum ghost width after creating the GridGeometry!
for creeping_flow=FALSE. It runs fine with this flag set to TRUE.
I can not recreate this. channel
runs fine for me with a refined region in the domain. Can you send me the input file that you used?
I have fixed the failing assertions for time_stepping
.
Whoops....
Ok I forgot which branch came first, so I rebased physical_bcs
onto convec_op
. I will fix that last issue in #68 if I can recreate it.
This adds support for the nonlinear convective operator in the momentum equation. The input file is searched for the flag
creeping_flow
, which is true by default. If it is false, then the convective operator is added to the discretization. @bindi-nagda, when the testing suite is ready,convective_op
would make a great test.I've also moved everything into a separate namespace.