Journal of The Electrochemical Society,
Journal Year:
2025,
Volume and Issue:
172(4), P. 040503 - 040503
Published: April 1, 2025
A
hybrid
model
based
on
black-winged
kite
algorithm
and
dual-attention
mechanism
optimized
temporal
convolutional
network
(TCN)
with
simple
recurrent
unit
(SRU)
is
proposed
to
improve
the
accuracy
of
online
remaining-useful-life
(RUL)
prediction
for
Li-ion
batteries
(LIBs).
Health
indicators
(HIs)
correlated
battery
capacity
are
extracted
from
calculated
variables
verified
Spearman
correlation
coefficient
constructed,
applying
TCN
multi-head
self-attention
capture
in
spatial
dimension
decay
pattern
HIs,
introducing
attention
ability
SRU
timing
patterns
input
sequences
as
well
BKA
further
optimize
hyper-parameters,
enhancing
performance.
Experimental
data
used
validate
model’s
predictive
performance
LIBs
at
different
usage
levels
under
complex
conditions
such
regeneration,
sharp
fluctuations,
plunges.
The
results
achieve
MAE
less
than
3.66%,
MAPE
below
2.02%,
RMSE
not
exceeding
5.03%,
R
2
greater
0.96,
absolute
error
RUL
5.
experimental
demonstrate
that
can
accurate
perform
good
robustness.
Energy Science & Engineering,
Journal Year:
2025,
Volume and Issue:
unknown
Published: Jan. 8, 2025
ABSTRACT
Rolling
bearings
are
fundamental
components
of
contemporary
machinery,
yet
their
prolonged
usage
frequently
leads
to
wear,
performance
deterioration,
and
potential
faults.
In
scenarios
characterized
by
limited
sample
sizes
complex,
noisy
environments,
traditional
diagnostic
methods
often
encounter
difficulties
achieving
satisfactory
fault
identification
results.
To
address
these
challenges,
this
study
introduces
an
innovative
approach
for
rolling
bearing
diagnosis.
Initially,
the
black‐winged
kite
algorithm
(BKA)
is
enhanced
through
integration
a
differential
evolution
strategy
iterative
search
method,
enabling
precise
determination
optimal
parameters
variational
mode
decomposition
(VMD).
Subsequently,
comprehensive
index
evaluation
criterion
established
identify
signal
components,
which
then
subjected
detailed
analysis
extract
diverse
sensitive
features,
ultimately
forming
hybrid
feature
set.
further
improve
accuracy
efficiency
diagnosis,
proposes
extreme
learning
machine
model,
termed
twin
(TELM).
Moreover,
TELM
model
seamlessly
integrated
into
architecture
convolutional
neural
network
(CNN),
specifically
as
component
its
output
layer,
resulting
in
novel
diagnosis
model.
Rigorous
data
validation
performed
on
testbed
underscores
that
proposed
significantly
surpasses
conventional
approaches,
including
SVM,
KELM,
ELM,
LSTM,
softmax,
terms
accuracy,
recall,
F1
score.
Notably,
maintains
robust
capabilities
even
environments
with
varying
degrees
noise
interference.
International Journal of Intelligent Systems,
Journal Year:
2025,
Volume and Issue:
2025(1)
Published: Jan. 1, 2025
The
rapid
advancement
in
unmanned
aerial
vehicle
(UAV)
technology
has
marked
a
transformative
shift
various
industries,
with
logistics
distribution
service
being
one
of
the
prime
sectors
reaping
benefits.
UAVs
offer
substantial
benefits
speed,
cost,
and
reach,
promising
to
revolutionize
logistics,
especially
remote
areas.
On
hand,
they
are
poised
meet
demands
for
quick
versatile
delivery
options.
other
their
deployment
comes
challenges.
Weather
variabilities
such
as
rainfall,
wind
need
safe
take‐off
intervals
can
compromise
UAV
safety
operation.
Conventional
route
optimization
often
overlooks
these
dynamic
factors,
resulting
inefficient
or
unworkable
routes.
repeated
time‐consuming
calculations
caused
by
trials
when
making
group
plans.
Recognizing
gaps,
this
study
proposes
data
representation
effectively
transform
flight
flyable
area
into
time‐varying
network
that
maintains
spatiotemporal
connectivity
establishes
mathematical
model
represents
complexities
distribution.
Then,
multistage
algorithm
specifically
tailored
large‐scale
search
is
designed
obtain
stable
optimal
solution.
Subsequent
experimental
validations
on
actual
case
datasets
have
confirmed
correctness,
effectiveness,
adaptability
algorithm.
Benchmarking
against
traditional
CPLEX
methods
demonstrated
not
only
rivals
best
solutions
but
does
so
38.8
times
increase
computational
speed.
When
pitted
shortest
path
Dijkstra
A
∗
algorithms,
method
consistently
outperformed,
delivering
up
3.5
faster
applications.
Moreover,
parameter
sensitivity
analysis
performed
adjusting
thresholds
rainfall
speed
parameters
revealed
performance
strong
positive
correlation
size
network.
Computational Urban Science,
Journal Year:
2025,
Volume and Issue:
5(1)
Published: March 26, 2025
Abstract
Digital
twins
are
enjoying
widespread
and
growing
success
in
both
theoretical
practical
applications.
A
recent
development
that
is
gaining
increasing
traction
the
application
of
digital
to
cities.
The
aim
this
article
discuss
whether
there
inherent
limitations
case.
At
present,
scientific
literature
on
urban
dominated
by
“technical”
approaches.
Critical
investigation
–
especially
from
a
philosophical
perspective
still
at
its
beginnings.
This
aims
contribute
line
inquiry.
It
mainly
analytical.
On
basis
specific
conceptual
framework,
it
examines
their
applications
contexts.
starts
distinguishing
among
simple,
complicated
complex
systems,
reaches
conclusion
that,
while
using
generally
appropriate
(and
often
helpful)
first
two
these
some
structural
use
case
systems.
In
latter
case,
depend
certain
distinctive
aspects
such
as
emergent
unpredictable
nature,
role
played
regard
“dispersed
knowledge”
(that
is,
form
diffused
knowledge
crucial
for
functioning
large
systems
but
cannot
be
collected
re-unified
because,
coherent
integrated
whole,
does
not
exist
anywhere).
