salazardetroya / bridge_post Goto Github PK

Hi,

Thank you for this great post!!
I have a couple of questions about this as follows.

Q1) You mentioned above for a reason to use the Firedrake that “Firedrake” is more integrated with the PETSc library,which is very helpful for setting solver options required for large scale problems. I am just wondering if you had a chance to solve the same or similar problem using “FEniCS”. If so, what’s the main differences? If someone wants to use the FEniCS rather than Firedrake, can I guess that is not only a general implementation including solver setting more tricky but also computational speed slower?

Q2) In this bridge example, are you using a parallel solver based on MPI?

Q3) How the function”DirichletBC()” and the UFLs “dx()” and “ds()” recognize the boundary and domain markers? It seems that you just define the specified numbers for the boundary and domain, but not the specified locations. Are the information defined in the Gmsh file, “bridge.geo”?

Q4) Is it possible to define the boundary marker and domain marker separately? I mean that can we define the boundary markers from 0 to N_bd and the domain markers from 0 to N_domain, respectively, like below;

“For domains”
DECK = 0
DOMAIN = 1

“For boundaries”
SUPPORT = 0
LOAD = 1
SYMMETRY_X = 2
SYMMETRY_Y = 3

Sorry, I am new to this Python-based TO things, so my questions could be a bit silly.

Thank you again for this helpful post!!

bridge.py (line 49) loads the geometry from a file called bridge.msh
The repo appears to have that info stored in a file called bridge.geo

bridge.geo does not appear to be called by any of the scripts in the repo