Hello guys, I’ve been trying to solve an electro-Chemo-elasticity coupling problem, and I’ve been faced with some problems in the mesh, it’s a simple geometry, but it requires a great refinement.
I’ve been trying to test my mesh through a Poisson model, and apparently there’s an error in it or the way I’m using it. If anyone has any comment about the code, please, don’t hesitate to tell me
# Scaled variable
carregamento= -50000
E = 210e6
poisson = 0.3
lambda_ = E*poisson / ((1+poisson)*(1-2*poisson))
G= E / 2*(1+poisson)
import numpy as np
import ufl
from mpi4py import MPI
from petsc4py.PETSc import ScalarType
from dolfinx.io import gmshio
from dolfinx import mesh, fem, plot, io
#Importação da geometria e das condições de contorno.
domain, cell_tags, facet_tags = gmshio.read_from_msh("malha_estruturada.msh", MPI.COMM_SELF,0, gdim=2)
"""Facets_tags numbers options:
1 9 "face superior"
1 10 "face_esquerda"
1 11 "face_inferior"
1 12 "face_direita"
2 13 "dominio"
"""
# Define function space
V = fem.VectorFunctionSpace(domain, ("CG", 1))
# Dirichlet boundary
u_D = np.array([0,0], dtype=ScalarType)
dofs_2 = fem.locate_dofs_topological(V, facet_tags.dim, facet_tags.find(10))
bc=fem.dirichletbc(u_D,dofs=dofs_2,V=V)
#especificar a medida de integração, que deve ser a integral sobre a fronteira do nosso domínio
ds = ufl.Measure('ds', domain=domain, subdomain_data=facet_tags)
# Define test functions in weak form
u = ufl.TrialFunction(V)
v = ufl.TestFunction(V)
def epsilon(x): #tensor deform small
return ((1/2)*((ufl.nabla_grad(x)) + (ufl.nabla_grad(x).T) ))
def sigma(y):
d = len(y)
I = ufl.variable(ufl.Identity(d))
return 2.0 * G * epsilon(y) + lambda_ * ufl.tr(epsilon(y)) * I
#funcao carregamento
f = fem.Constant(domain, ScalarType((0,carregamento )))
#Formulação variacional
a = ufl.inner(sigma(u), ufl.grad(v)) * ufl.dx
L = ufl.dot(f, v) *ds(9)
problem = fem.petsc.LinearProblem(a, L, bcs=[bc], petsc_options={"ksp_type": "gmres", "pc_type": "lu","ksp_max_it": 100}, )
uh = problem.solve()
from dolfinx.io import XDMFFile
with XDMFFile(domain.comm, "resultados/malha.xdmf", "w") as xdmf:
xdmf.write_mesh(domain)
xdmf.write_meshtags(facet_tags)
xdmf.write_meshtags(cell_tags)
xdmf.write_function(uh)
No exit from error is presented, contundo, the solution does not make sense. Thus, I would like to know if this mode of implementation, the contour faces is the most appropriate or not, so that in this way, I can reflect on the construction of this mesh .
Here is a model of malha built na gmsh platform completely manually, versao used é 4.11: https://drive.google.com/file/d/1nmo2PnBmQHSJ2pz09rHNGNdfhMbATfbn/view?usp=drive_link
This result obtained in paraview:
