Applied Sciences,
Journal Year:
2024,
Volume and Issue:
14(18), P. 8157 - 8157
Published: Sept. 11, 2024
A
comprehensive
framework
for
the
layout
optimisation
of
rigid-jointed
frame
structures
is
proposed,
addressing
multiple
mechanical
constraints
while
effectively
managing
geometric
complexity.
The
considered
include
displacement,
stress,
and
both
local
global
stability.
Geometric
complexity
controlled
by
minimising
low-stiffness
elements
reducing
number
in
resulting
layouts.
Numerical
examples
demonstrate
effectiveness
proposed
method,
showcasing
its
ability
to
generate
optimal
structural
layouts
with
desirable
performance
varying
levels
member
connectivity.
This
innovative
offers
significant
advantages
over
conventional
approaches
ensuring
optimality
manufacturability
structures,
thereby
facilitating
their
practical
application.
Research Square (Research Square),
Journal Year:
2024,
Volume and Issue:
unknown
Published: Nov. 12, 2024
Abstract
Existing
reinforced
concrete
(RC)
buildings
risk
seismic
damage
because
they
were
not
constructed
in
compliance
with
design
standards
and
may
have
irregular
mass
distribution
construction
defects.
Typically,
columns
these
are
designed
to
withstand
only
gravity
loads,
making
them
vulnerable
or
collapse
during
earthquakes.
Retrofitting
using
an
RC
jacket
system
is
a
standard
way
enhance
resilience.
However,
conventional
parametric
modeling
for
jacketed
structures
physics-based
(finite
element)
can
be
time-consuming
non-intuitive.
To
address
this
challenge,
the
present
study
proposes
novel
data-driven
machine-learning
approach
predict
columns'
demand-to-capacity
ratio
(DCR),
aiming
reasonably
accurate
reduced
computational
time.
Various
parameters
related
column
jacketing
considered
when
predicting
DCR.
The
datasets
generated
post-processing
used
train
Graphical
Neural
Network
(GNN)
Gaussian
Mixture
Model
(GMM).
dataset
encompasses
parameterization
of
variables,
including
retrofit
location,
compressive
strength,
cross-sectional
dimensions,
thickness,
longitudinal
transverse
reinforcement
areas,
yielding
slenderness
ratio.
Subsequently,
both
models
fitted
evaluated
against
test
identify
optimal
performer,
multiple
scorer
performance
index
as
model
evaluation
metric.
analysis
indicates
that
GMM
emerges
most
suitable
regressor
DCR
estimation,
exhibiting
lower
residual
error
than
GNN
model.
