Hi everyone, I am trying to convert a code written in Fenics to FenicsX. Here we use a fully incompressible formulation, solving for the displacement and the hydrostatic pressure simultaneously. However, it doesn’t work (aka, doesn’t converge), and I’m not able to see why. I’ve tried to adapt this…

I think the error is in your Dirichlet bc. You should not use the collapsed subspace, i.e. [image] aashildte: bcs = [fem.dirichletbc(u_bc, left_dofs, V)] is wrong. Instead, you should use the non-collapsed subspace for Dirichlet bc creation, as described in my previous answer , e.g.: left_do…

Hi @conpierce8 , and thanks a lot for answering! That makes a lot of sense, thanks for the clarification. I completely get why it won’t work. However, if I try replacing the two lines with yours, I get an error message I don’t understand … $ python mwe_pressure.py Traceback (most recent call last)…

My mistake. I should have tried running my proposed solution before recommending it. Yes, the error message makes sense. When you call locate_dofs_topological with a list of two function spaces, it returns a list of two ndarrays holding the dof indices for each function space. fem.dirichletbc is the…

Thank you so much @conpierce8 - worked like a charm! I’m totally ok with using ufl.split without understanding why :slight_smile: Also, the link to your other post will be very useful – it is actually my next step to do this component-wise!

Hello, I tried add one more displacement load on the right face in the follow, and make the traction and body force as 0. ### Add displacement on u_x right U_right_x = 1 facet_right = mesh.locate_entities_boundary(domain, fdim, right) dofs_right_x = fem.locate_dofs_topological(V.sub(1), domain.to…

Could you post your complete code? Without it, I would be completely unable to say why it doesn’t converge.

Thanks for your help in advance, here it is: import numpy as np import ufl from petsc4py import PETSc from mpi4py import MPI from dolfinx import fem, mesh, plot, io, nls, log, cpp from petsc4py.PETSc import ScalarType domain = mesh.create_box(MPI.COMM_WORLD,[[0.0,0.0,0.0], [1, 1, 1]], [1, 1, 1], …

[image] Q: dofs_right_x = fem.locate_dofs_topological((V.sub(0),V), domain.topology.dim-1, facet_right) [image] Q: bc_right = fem.dirichletbc(ux, dofs_right_x, V.sub(0)) You should not use V.sub(0), but rather state_space.sub(0).sub(0) as I explained above . [image] Q: The last par…

Thanks a lot for your explanation. Could I ask one more basic question here? I am confused with the way of setting dofs, is there any difference between fem.locate_dofs_topological() and fem.locate_dofs_geometrical() as I thought they are the same to locate dofs.

Fully incompressible formulation for a non-linear hyperelasticity problem in FenicsX

dolfinx

dokken September 28, 2023, 7:09pm 11

See Difference between locate_dofs_topological and locate_dofs_geometrically - #2 by dokken

1 Like

Topic		Replies	Views
Please help me to check whether boundary condition are right set dolfinx	13	1819	August 28, 2022
Nonlinear hyperelasticity in dolfin-x: Minimal example, and a few issues dolfinx	15	5122	May 29, 2022
Switching from Fenics to FenicsX for Incompressible Mixed Element Code General	2	257	July 19, 2024
Non linear solver in FEniCSx. Comparison to SNES in legacy FEniCS	11	2297	April 24, 2023
How to applying BC using V..sub().collapse() in dolfinx dolfinx	6	676	September 13, 2023

Fully incompressible formulation for a non-linear hyperelasticity problem in FenicsX

Related topics