Engineering Applications of Artificial Intelligence,
Journal Year:
2024,
Volume and Issue:
133, P. 108573 - 108573
Published: May 11, 2024
Understanding
the
effects
of
extreme
floods
on
critical
infrastructures
such
as
bridges
is
paramount
for
ensuring
safety
and
resilient
design
in
face
climate
change
events.
This
study
develops
robust
computational
predictive
modeling
tools
assessing
impacts
hydraulic
response
structural
resilience
bridges.
A
fluid
dynamic
(CFD)
model
utilizing
RANS
equations
k-ω
Shear
Stress
Transport
(SST)
simulating
supercritical
flows
adopted
to
compute
hydrodynamic
pressures
water
levels
bridge
piers
cylindrical
rectangular
shapes
during
a
flood
event.
The
CFD
validated
based
case
data
obtained
from
Haj
Omran
Bridge,
built
Khorramabad
River
Iran.
numerical
simulations
consider
hydrological
conditions
exclude
geotechnical
parameters
abutment
damages.
results
are
evaluated
well-established
guidelines.
Machine
learning
techniques,
including
Extreme
Gradient
Boosting
(XGBoost),
Random
Forest
(RF),
Support
Vector
Regression
(SVR),
optimized
with
Grid
Search
Cross-Validation
(GSCV),
enhance
accuracy
pressure
forecasting
at
piers.
XGBoost
exhibits
superior
performance
(R2
=
0.908,
RMSE
0.0279,
E
3.41%)
compared
RF
SVR
models.
All
estimated
by
falls
within
±6
percent
error
lines,
highlighting
model's
robustness
out-of-range
prediction.
Additionally,
an
Long
Short-Term
Memory
(LSTM)
effectively
predict
free
surface
flow
profiles
(i.e.
depth)
over
0.937
0.083),
demonstrating
its
potential
practical
applications
depth
predictions
infrastructures.
proposed
methodological
framework
outlined
this
can
facilitate
bridges,
enabling
assessment
Frontiers in Built Environment,
Journal Year:
2024,
Volume and Issue:
10
Published: Feb. 9, 2024
Seawalls
are
critical
defence
infrastructures
in
coastal
zones
that
protect
hinterland
areas
from
storm
surges,
wave
overtopping
and
soil
erosion
hazards.
Scouring
at
the
toe
of
sea
defences,
caused
by
wave-induced
accretion
bed
material
imposes
a
significant
threat
to
structural
integrity
infrastructures.
Accurate
prediction
scour
depths
is
essential
for
appropriate
efficient
design
maintenance
structures,
which
serve
mitigate
risks
failure
through
scouring.
However,
limited
guidance
predictive
tools
available
estimating
scouring
sloping
structures.
In
recent
years,
Artificial
Intelligence
Machine
Learning
(ML)
algorithms
have
gained
interest,
although
they
underpin
robust
models
many
engineering
applications,
such
yet
be
applied
prediction.
Here
we
develop
present
ML-based
predicting
seawall.
Four
ML
algorithms,
namely,
Random
Forest
(RF),
Gradient
Boosted
Decision
Trees
(GBDT),
Neural
Networks
(ANNs),
Support
Vector
Regression
(SVMR)
utilised.
Comprehensive
physical
modelling
measurement
data
utilised
validate
models.
A
Novel
framework
feature
selection,
importance,
hyperparameter
tuning
adopted
pre-
post-processing
steps
In-depth
statistical
analyses
proposed
evaluate
performance
The
results
indicate
minimum
80%
accuracy
across
all
tested
this
study
overall,
SVMR
produced
most
accurate
predictions
with
Coefficient
Determination
(
r
2
)
0.74
Mean
Absolute
Error
(MAE)
value
0.17.
algorithm
also
offered
computationally
among
tested.
methodological
can
datasets
rapid
assessment
facilitating
model-informed
decision-making.
Scientific Reports,
Journal Year:
2024,
Volume and Issue:
14(1)
Published: March 29, 2024
Abstract
The
consumption
of
water
constitutes
the
physical
health
most
living
species
and
hence
management
its
purity
quality
is
extremely
essential
as
contaminated
has
to
potential
create
adverse
environmental
consequences.
This
creates
dire
necessity
measure,
control
monitor
water.
primary
contaminant
present
in
Total
Dissolved
Solids
(TDS),
which
hard
filter
out.
There
are
various
substances
apart
from
mere
solids
such
potassium,
sodium,
chlorides,
lead,
nitrate,
cadmium,
arsenic
other
pollutants.
proposed
work
aims
provide
automation
estimation
through
Artificial
Intelligence
uses
Explainable
(XAI)
for
explanation
significant
parameters
contributing
towards
potability
impurities.
XAI
transparency
justifiability
a
white-box
model
since
Machine
Learning
(ML)
black-box
unable
describe
reasoning
behind
ML
classification.
models
Logistic
Regression,
Support
Vector
(SVM),
Gaussian
Naive
Bayes,
Decision
Tree
(DT)
Random
Forest
(RF)
classify
whether
drinkable.
representations
force
plot,
test
patch,
summary
dependency
plot
decision
generated
SHAPELY
explainer
explain
features,
prediction
score,
feature
importance
justification
estimation.
RF
classifier
selected
yields
optimum
Accuracy
F1-Score
0.9999,
with
Precision
Re-call
0.9997
0.998
respectively.
Thus,
an
exploratory
analysis
indicators
associated
their
significance.
emerging
research
at
vision
addressing
future
well.
Engineering Applications of Artificial Intelligence,
Journal Year:
2024,
Volume and Issue:
133, P. 108156 - 108156
Published: March 6, 2024
Ground
settlement
prediction
during
mechanized
tunneling
is
of
paramount
importance
and
remains
a
challenging
research
topic.
Typically,
two
paradigms
are
existing:
physics-driven
approach
utilizing
numerical
simulation
models
for
prediction,
data-driven
employing
machine
learning
techniques
to
learn
mappings
between
influencing
factors
the
settlement.
To
integrate
advantages
both
approaches
assimilate
data
from
different
sources,
we
propose
multi-fidelity
deep
operator
network
(DeepONet)
framework,
leveraging
recently
developed
methods.
The
presented
framework
comprises
components:
low-fidelity
subnet
that
captures
fundamental
ground
patterns
obtained
finite
element
simulations,
high-fidelity
learns
nonlinear
correlation
real
engineering
monitoring
data.
A
pre-processing
strategy
causality
adopted
consider
spatio-temporal
characteristics
tunnel
excavation.
results
show
proposed
method
can
effectively
capture
physical
information
provided
by
simulations
accurately
fit
measured
(R2
around
0.9)
as
well.
Notably,
even
when
dealing
with
very
limited
noisy
(with
50%
error),
model
robust,
achieving
satisfactory
R2>0.8.
In
comparison,
R2
score
pure
simulation-based
only
0.2.
utilization
transfer
significantly
reduces
training
time
20
min
within
30
s,
showcasing
potential
our
real-time
construction.
