Manifold Learning for Aerodynamic Shape Design Optimization
Aerospace,
Год журнала:
2025,
Номер
12(3), С. 258 - 258
Опубликована: Март 19, 2025
The
significant
computational
cost
incurred
due
to
the
iterative
nature
of
Computational
Fluid
Dynamics
(CFD)
in
traditional
aerodynamic
shape
design
frameworks
poses
a
major
challenge,
especially
context
modern
integrated
requirements
and
increasingly
complex
conditions.
To
address
demands
design,
we
developed
an
efficient
framework
based
on
our
previous
work
involving
locally
linear
embedding
plus
constrained
optimization
genetic
algorithm
(LLE+COGA)
high-fidelity
reduced-order
model
(ROM).
An
active
manifold
(AM)
auto-en/decoder
was
employed
dimensionality
curse
arising
from
excessively
large
space.
fast
mesh
deformation
method
utilized
for
high-precision,
rapid
deformation,
significantly
reducing
associated
with
transferring
geometric
deformations
CFD
fine
mesh.
This
addressed
transonic
problem
undeflected
Common
Research
Model
(uCRM)
three-dimensional
wing
(with
aspect
ratio
9),
241
variables.
results
demonstrate
that
optimized
achieved
reduction
drag
coefficient
by
38.9%
54.5%
compared
baseline
Case
1
2,
respectively.
Additionally,
total
time
shortened
62.6%
57.7%
two
cases.
Moreover,
outcomes
aligned
well
those
obtained
FOM-based
framework,
further
validating
effectiveness
practical
applicability
proposed
approach.
Язык: Английский
Multi-Objective Topology Optimization of Thin-Plate Structures Based on the Stiffener Size and Layout
Electronics,
Год журнала:
2024,
Номер
13(24), С. 4968 - 4968
Опубликована: Дек. 17, 2024
To
address
the
limitations
of
existing
optimization
methods
that
focus
on
single
objectives
or
neglect
stiffener
features,
a
multi-objective
topology
(MOTO)
method
is
proposed
based
size
and
layout.
By
constraining
initial
structural
performance
parameters,
optimal
height
determined
through
optimization.
Based
height,
single-objective
used
to
achieve
best
material
distribution.
The
width
treated
as
design
variable,
while
MOTO
performed
load
point
displacement,
first
natural
frequency,
mass,
thereby
yielding
an
performance.
Finally,
multi-dimensional
analysis
width,
dynamic
static
characteristics
stiffened
thin-plate
structure
conducted.
results
indicate
optimized
layout
considerably
improved.
Compared
structure,
maximum
average
displacements
are
reduced
by
23.26%
8.62%,
respectively.
frequency
increases
3.81%,
resonance
amplitude
overall
mass
decrease
39.97%
1.99%,
achieves
lightweight
maintaining
better
stiffness
low-frequency
vibration
resistance.
feasibility
effectiveness
validated.
Язык: Английский