2021 IEEE International Conference on Big Data (Big Data),
Journal Year:
2023,
Volume and Issue:
unknown, P. 2221 - 2228
Published: Dec. 15, 2023
The
usage
of
Lithium-ion
(Li-ion)
batteries
has
gained
widespread
popularity
across
various
industries,
from
powering
portable
electronic
devices
to
propelling
electric
vehicles
and
supporting
energy
storage
systems.
A
central
challenge
in
Li-ion
battery
reliability
lies
accurately
predicting
their
Remaining
Useful
Life
(RUL),
which
is
a
critical
measure
for
proactive
maintenance
predictive
analytics.
This
study
presents
novel
approach
that
harnesses
the
power
multiple
denoising
modules,
each
trained
address
specific
types
noise
commonly
encountered
data.
Specifically,
auto-encoder
wavelet
denoiser
are
used
generate
encoded/decomposed
representations,
subsequently
processed
through
dedicated
self-attention
transformer
encoders.
After
extensive
experimentation
on
NASA
CALCE
data,
broad
spectrum
health
indicator
values
estimated
under
set
diverse
patterns.
reported
error
metrics
these
data
par
with
or
better
than
state-of-the-art
recent
literature.
Energies,
Journal Year:
2025,
Volume and Issue:
18(3), P. 746 - 746
Published: Feb. 6, 2025
The
sustainable
reuse
of
batteries
after
their
first
life
in
electric
vehicles
requires
accurate
state-of-health
(SoH)
estimation
to
ensure
safe
and
efficient
repurposing.
This
study
applies
the
systematic
ProKnow-C
methodology
analyze
state
art
SoH
using
machine
learning
(ML).
A
bibliographic
portfolio
534
papers
(from
2018
onward)
was
constructed,
revealing
key
research
trends.
Public
datasets
are
increasingly
favored,
appearing
60%
studies
reaching
76%
2023.
Among
12
identified
sources
covering
20
from
different
lithium
battery
technologies,
NASA’s
Prognostics
Center
Excellence
contributes
51%
them.
Deep
(DL)
dominates
field,
comprising
57.5%
implementations,
with
LSTM
networks
used
22%
cases.
also
explores
hybrid
models
emerging
role
transfer
(TL)
improving
prediction
accuracy.
highlights
potential
applications
predictions
energy
informatics
smart
systems,
such
as
grids
Internet-of-Things
(IoT)
devices.
By
integrating
estimates
into
real-time
monitoring
systems
wireless
sensor
networks,
it
is
possible
enhance
efficiency,
optimize
management,
promote
practices.
These
reinforce
relevance
machine-learning-based
resilience
sustainability
systems.
Finally,
an
assessment
implemented
algorithms
performances
provides
a
structured
overview
identifying
opportunities
for
future
advancements.
Processes,
Journal Year:
2023,
Volume and Issue:
11(8), P. 2333 - 2333
Published: Aug. 3, 2023
Lithium-ion
batteries
are
widely
utilized
in
various
fields,
including
aerospace,
new
energy
vehicles,
storage
systems,
medical
equipment,
and
security
due
to
their
high
density,
extended
lifespan,
lightweight
design.
Precisely
predicting
the
remaining
useful
life
(RUL)
of
lithium
is
crucial
for
ensuring
safe
use
a
device.
In
order
solve
problems
unstable
prediction
accuracy
difficultly
modeling
lithium-ion
battery
RUL
with
previous
methods,
this
paper
combines
channel
attention
(CA)
mechanism
long
short-term
memory
networks
(LSTM)
propose
hybrid
CA-LSTM
model.
By
incorporating
CA
mechanism,
utilization
local
features
situations
where
data
limited
can
be
improved.
Additionally,
effectively
mitigate
impact
capacity
rebound
on
model
during
charging
discharging
cycles.
ensure
full
validity
experiments,
National
Aeronautics
Space
Administration
(NASA)
University
Maryland
Center
Advanced
Life
Cycle
Engineering
(CALCE)
datasets
different
starting
points
validation.
The
experimental
results
demonstrated
that
proposed
exhibited
strong
predictive
performance
was
minimally
influenced
by
point.
Batteries,
Journal Year:
2024,
Volume and Issue:
10(5), P. 152 - 152
Published: April 30, 2024
Lithium-ion
batteries
are
currently
widely
employed
in
a
variety
of
applications.
Precise
estimation
the
remaining
useful
life
(RUL)
lithium-ion
holds
significant
function
intelligent
battery
management
systems
(BMS).
Therefore,
order
to
increase
fidelity
and
stabilization
predicting
RUL
batteries,
this
paper,
an
innovative
strategy
for
prediction
is
proposed
by
integrating
one-dimensional
convolutional
neural
network
(1D
CNN)
bilayer
long
short-term
memory
(BLSTM)
network.
Feature
extraction
carried
out
through
input
capacity
data
model
using
1D
CNN,
these
deep
features
used
as
BLSTM.
The
BLSTM
applied
retain
key
information
database
better
understand
coupling
relationship
among
consecutive
time
series
along
axis,
thereby
effectively
trends
batteries.
Two
different
types
datasets
from
NASA
CALCE
were
verify
effectiveness
method.
results
show
that
method
achieves
higher
accuracy,
demonstrates
stronger
generalization
capabilities,
reduces
errors
compared
other
methods.
