Why do I obtain ~185 near-zero eigenvalues in a clamped 3D elasticity eigenproblem (DOLFINx + SLEPc)?

[dokken]

As you haven’t given us the file

@nadine

from function import import_CAD_to_FEniCSx

we can’t reproduce your example.

I would guess the issue is similar to:

Structural Modal Analysis

one should remove the strongly enforced dofs from the system, i.e. something along the lines of: I would do something along the lines of: from dolfinx.fem import Function, form, functionspace, dirichletbc, locate_dofs_topological from dolfinx.fem.petsc import assemble_matrix from slepc4py import SLEPc from ufl import dx, grad, inner, TrialFunction, TestFunction from dolfinx.io import XDMFFile, gmsh as gmshio from mpi4py import MPI from petsc4py import PETSc import numpy as np from ufl import …

where the solution is to remove the rows and columns that have the clamped conditions by using

K_sub, (idx_set_row, idx_set_row) = extract_sub_matrix(K, V, [bc])
M_sub, _ = extract_sub_matrix(M, V, [bc])

as described in the aforementioned post.
Please let me know if this helps.

All the best,
Jørgen