Abstract
The
modeling
of
inelastic
phenomena
with
tensor‐valued
internal
variables
requires
a
regularization
to
counteract
mesh
dependence.
Here,
we
consider
the
anisotropic
damage
at
finite
strains
and
seek
for
an
efficient
formulation
based
on
reduced
number
nonlocal
degrees
freedom.
We,
thus,
equip
brittle
version
model
different
gradient‐extensions
in
micromorphic
framework
using
full
tensor.
models
are
compared
structural
simulation
asymmetrically
notched
specimen
perforated
shear
band
section.
High
agreement
is
observed
between
two
Furthermore,
present
novel
study
evolution
field
variables.
International Journal of Damage Mechanics,
Год журнала:
2025,
Номер
unknown
Опубликована: Май 14, 2025
This
paper
establishes
a
generic
framework
for
the
nonlocal
modeling
of
anisotropic
damage
at
finite
strains.
By
combination
two
recent
works,
new
allows
flexible
incorporation
different
established
hyperelastic
strain
material
formulations
into
whilst
ensuring
mesh-independent
results
by
employing
set
micromorphic
gradient-extensions.
First,
model,
generally
satisfying
growth
criterion,
is
investigated
specific
choice
neo-Hookean
on
single
element.
Next,
model
applied
with
gradient-extensions
in
structural
simulations
an
asymmetrically
notched
specimen
to
identify
efficient
form
volumetric–deviatoric
regularization.
Thereafter,
framework,
which
without
loss
generality
here
specified
gradient-extension,
successfully
serves
complex
simulation
pressure-loaded
rotor
blade.
The
codes
subroutines
are
accessible
public
https://doi.org/10.5281/zenodo.11171630
.
International Journal for Numerical Methods in Engineering,
Год журнала:
2024,
Номер
125(24)
Опубликована: Авг. 11, 2024
Abstract
Modern
inelastic
material
model
formulations
rely
on
the
use
of
tensor‐valued
internal
variables.
When
phenomena
include
softening,
simulations
former
are
prone
to
localization.
Thus,
an
accurate
regularization
variables
is
essential
obtain
physically
correct
results.
Here,
we
focus
anisotropic
damage
at
finite
strains.
a
flexible
with
isotropic,
kinematic,
and
distortional
hardening
equipped
three
gradient‐extensions
using
full
two
reduced
regularizations
tensor.
Theoretical
numerical
comparisons
yield
excellent
agreement
between
based
volumetric‐deviatoric
only
nonlocal
degrees
freedom.
Abstract
The
characterization
of
the
material
behavior
inelastic
materials
requires
a
high
degree
expert
knowledge
to
identify
and
constitutively
describe
response.
In
addition,
specific
models
are
usually
pre‐selected
in
course
only
best
parameters
for
these
determined,
but
therefore
not
necessarily
models.
Unfortunately,
more
general
description
results
an
increased
effort
during
characterization,
which
is
barely
practicable
by
hand.
This
where
machine
learning
algorithms
may
help
us.
To
get
both
worlds,
powerful
sound
thermodynamic
considerations,
Constitutive
Artificial
Neural
Networks
(iCANNs)
discover
generic
formulations
Helmholtz
free
energy
pseudo
potential.
relations
guide
us
towards
thermodynamically
consistent
descriptions
stresses
strains;
concept
that
applicable
wide
range
phenomena
from
viscoelasticity,
elastoplasticity,
phase
transformations
growth
remodeling
living
tissues.
Here,
we
equip
original
iCANN
framework
guarantee
polyconvexity
priori,
ensures
at
least
one
minimizing
deformation.
We
investigate
ability
our
viscoelastic
polymer
different
stretch
levels
strain
rates.
made
source
code,
data,
example
accessible
public
https://doi.org/10.5281/zenodo.11084354
.
Results in Engineering,
Год журнала:
2024,
Номер
23, С. 102685 - 102685
Опубликована: Авг. 8, 2024
Virtual
testing
of
helmets
using
finite
element
(FE)
analysis
can
be
a
valuable
tool
during
product
development.
Still,
its
usefulness
is
limited
by
the
quality
constitutive
model
energy-absorbing
material,
usually
foam.
Built-in
models
in
commercial
FE
software
are
developed
for
traditional
linear
compression
loading.
However,
modern
oblique
test
methods
load
foam
combined
and
shear.
Therefore,
we
aim
to
evaluate
what
extent
built-in
represent
Expanded
Polystyrene
(EPS)
shear
loading
(CCSL).
EPS
tested
experimentally
newly
rig
CCSL
(V-test).
The
response
compared
against
simulation
three
different
available
LS-DYNA
(M83,
M126,
M181).
assessed
their
ability
capture
correct
response,
focusing
on
how
well
continuum
phenomenological
events
seen
experiments.
results
show
that
perform
compression,
as
expected.
point
out
limitations
significant
unloading
both
important
helmet
testing.
Due
these
limitations,
conclude
existing
inadequate
accurately
simulating
impacts.
There
clear
need
develop
implement
new
focused
capturing
including
unloading.
Additionally,
frictional
sliding
was
found
substantially
influence
measured
V-test
method.
Minimizing
interface
therefore
critical
isolating
material
behavior.
