Key engineering materials,
Год журнала:
2024,
Номер
999, С. 109 - 115
Опубликована: Дек. 16, 2024
Platform
jacket
structures
are
offshore
buildings
that
support
oil
and
gas
drilling
activities.
The
jacket's
structure
is
generally
tubular
steel
connected
between
the
legs
brace.
selection
of
brace
topology
essential
to
prevent
from
failure,
such
as
fractures.
research
was
carried
out
by
modelling
analyzed
using
SACS
finite
element
method
software.
Analysis
reliability
Monte
Carlo
simulation
method.
purpose
this
study
analyze
determine
optimal
topological
clamp
model
for
use
on
it.
results
show
X
has
better
than
K
N,
where
can
withstand
a
maximum
force
44000
kN
with
index
value
3.35.
In
comparison,
working
35200
2.94,
N
32000
3.02.
Recent
advancements
in
sensor
technology
and
data
processing
algorithms
have
revolutionized
Structural
Health
Monitoring
(SHM),
enabling
real-time
monitoring
analysis
of
structural
responses
to
dynamic
loads.
As
a
result,
many
buildings
are
permanently
instrumented
with
sensors,
typically
accelerometers,
continuously
record
vibrational
over
time,
hence
generating
huge
amounts
data.
However,
the
extraction
meaningful
insights
from
recorded
assist
engineers
building
managers
assessing
conditions
would
be
challenge.
systems
can
programmed
ground
motion-induced
vibrations
that
surpass
specific
trigger
threshold
levels.
Nonetheless,
there
challenges
long-term
damage
detection
including
automated
previously
data,
limited
number
available
nonlinear
under
severe
earthquakes,
name
but
few.
In
this
paper,
new
methodology
based
on
adaptive
time-series
(TS)
models
for
SHM
subjected
earthquakes
is
introduced
overcome
these
challenges.
Using
proposed
technique,
large
set
establishing
reliable
baseline
structure
using
even
as
few
two
accelerometers
(one
one
ground)
achievable.
The
efficiency
method
large-scale
structures
sensors
verified
3-D
Finite
Element
(FE)
model
5-story
reinforced
concrete
(RC)
SAP2000
platform.
simulation
results
demonstrated
accurate
identification
potential
provided
clear
indication
progression
severity
induced
increases.
Proceedings of the Institution of Mechanical Engineers Part M Journal of Engineering for the Maritime Environment,
Год журнала:
2025,
Номер
unknown
Опубликована: Янв. 30, 2025
Vibration-based
damage
detection
techniques
are
widely
applied
in
the
structural
health
monitoring
of
offshore
platforms.
This
study
employs
a
parallel
multi-scale
convolutional
neural
network
(PMSCNN)
to
analyze
acceleration
response
signals
collected
from
platforms,
achieving
localization
fatigue
cracks.
The
effectiveness
proposed
method
is
validated
through
numerical
simulations
jacket-type
platforms
subjected
random
wave
excitations
different
directions.
focuses
on
identifying
crack
platform
components,
addressing
both
single
and
multiple
scenarios
involving
small
To
assess
robustness
against
noise,
Gaussian
white
noise
varying
intensities
was
added
signals.
results
demonstrate
that
approach
effectively
identifies
locates
cracks
exhibiting
strong
resistance.
Journal of Marine Science and Engineering,
Год журнала:
2025,
Номер
13(2), С. 327 - 327
Опубликована: Фев. 11, 2025
Precise
prediction
of
mooring
tension
is
essential
for
the
safety
and
operational
efficiency
semi-submersible
aquaculture
platforms.
Traditional
numerical
methods
struggle
with
real-time
performance
due
to
nonlinear
dynamic
characteristics
environmental
loads.
This
study
proposes
a
novel
neural
network
approach
enhance
forecasting
line
responses,
combining
Ensemble
Empirical
Mode
Decomposition
(EEMD),
Temporal
Convolutional
Networks
(TCNs),
Self-Attention
(SA)
mechanism.
The
training
dataset
encompasses
time-domain
analysis
results,
including
tensions,
motion
total
structural
forces.
Firstly,
Pearson
Correlation
Analysis
(PCA)
utilized
assess
linear
relationships
among
hydrodynamic
variables.
Subsequently,
EEMD
applied
decompose
data,
which
then
combined
highly
correlated
variables
form
input
dataset.
Finally,
TCN
model
trained
predict
time
series,
while
an
SA
mechanism
integrated
weigh
significance
different
moments
within
sequence,
thereby
further
enhancing
accuracy.
results
demonstrate
that
evaluation
metrics
EEMD-TCN-SA
outperform
those
other
models,
effectively
predicting
platforms
significantly
reducing
errors.
Metals,
Год журнала:
2025,
Номер
15(4), С. 408 - 408
Опубликована: Апрель 4, 2025
In
steel
structural
engineering,
artificial
intelligence
(AI)
and
machine
learning
(ML)
are
improving
accuracy,
efficiency,
automation.
This
review
explores
AI-driven
approaches,
emphasizing
how
AI
models
improve
predictive
capabilities,
optimize
performance,
reduce
computational
costs
compared
to
traditional
methods.
Inverse
Machine
Learning
(IML)
is
a
major
focus
since
it
helps
engineers
minimize
reliance
on
iterative
trial-and-error
by
allowing
them
identify
ideal
material
properties
geometric
configurations
depending
predefined
performance
targets.
Unlike
conventional
ML
that
mostly
forward
predictions,
IML
data-driven
design
generation,
enabling
more
adaptive
engineering
solutions.
Furthermore,
underlined
Explainable
Artificial
Intelligence
(XAI),
which
enhances
model
transparency,
interpretability,
trust
of
AI.
The
paper
categorizes
applications
in
construction
based
their
impact
automation,
health
monitoring,
failure
prediction
evaluation
throughout
research
from
1990
2025.
challenges
such
as
data
limitations,
generalization,
reliability,
the
need
for
physics-informed
while
examining
AI’s
role
bridging
real-world
applications.
By
integrating
into
this
work
supports
adoption
ML,
IML,
XAI
analysis
design,
paving
way
reliable
interpretable
practices.
