International Journal of Turbo and Jet Engines,
Journal Year:
2024,
Volume and Issue:
unknown
Published: Aug. 26, 2024
Abstract
Capturing
degradation
trends
from
the
Condition
monitored
signals
is
a
proven
technique
for
predicting
Remining
Useful
Life
(RUL)
of
equipment,
which
has
gained
more
prominence
in
Prognostics
and
Health
Management
(PHM)
Industry
4.0.
However,
this
process
tiresome
comprehending
all
physical
parameters
system
to
construct
Index
that
characterize
health
state
complex
process,
especially
if
multiple
sensors
are
involved.
This
work
proposes
Deep
residual
ensemble
model
constructs
Fused
(FHI)
by
harnessing
temporal
property
signals.
The
proposed
Residual
network
integrates
Bi-directional
Long
Short
Term
Memory
(Bi-LSTM)
Neural
Network
(DNN)
absorbs
individual
residuals
both
forward
reverse
LSTMs
acts
as
an
important
feature
improve
overall
prediction
process.
validated
using
CMAPPS
dataset
various
unique
performance
metrics
portray
effectiveness
model.
Fundamental Research,
Journal Year:
2024,
Volume and Issue:
unknown
Published: Jan. 1, 2024
As
one
of
the
key
technologies
to
maintain
safety
and
reliability
stochastic
degrading
systems,
remaining
useful
life
(RUL)
prediction,
also
known
as
prognostics,
has
been
attached
great
importance
in
recent
years.
Particularly,
with
rapid
development
industrial
4.0
internet-of-things
(IoT),
prognostics
for
systems
under
big
data
have
paid
much
attention
years
various
prognosis
methods
reported.
However,
there
not
a
critical
review
particularly
focused
on
strengths
weaknesses
these
provoke
new
ideas
research.
To
fill
this
gap,
facing
realistic
demand
background
data,
paper
profoundly
analyzes
basic
research
ideas,
trends,
common
problems
data-driven
methods,
mainly
including
statistical
machine
learning
(ML)
based
hybrid
ML
methods.
discusses
emerging
topic
incomplete
possible
opportunities
future
are
highlighted.
Through
discussing
pros
cons
existing
we
provide
discussions
challenges
steer
data.
While
an
exhaustive
remains
elusive,
hope
that
perspectives
can
serve
stimulus
era
Proceedings of the Institution of Mechanical Engineers Part B Journal of Engineering Manufacture,
Journal Year:
2025,
Volume and Issue:
unknown
Published: Jan. 15, 2025
Remaining
useful
life
(RUL)
prediction
is
vital
to
formulate
a
suitable
maintenance
strategy
in
manufacturing
systems
health
management.
Multisensor
data
fusion
of
complex
engineering
has
attracted
substantial
attention
due
the
fact
that
single
sensor
can
only
collect
partial
information.
Health
indicator
(HI)
construction
plays
crucial
role
multisensor
and
machinery
prognostic,
mainly
because
it
attempts
quantify
history
ongoing
degradation
process
by
fusing
advantages
multiple
sensors.
However,
large
numbers
coefficients
are
involved
for
most
existing
HIs.
Additionally,
simplifications
during
modeling
may
inhibit
wide
application
constructed
HI.
To
address
these
two
challenges,
new
method
proposed
this
paper
constructing
HI
characterization
process.
Firstly,
sensors
invalid
or
conflicting
removed
through
correlation
coefficient
operation.
Then,
principal
component
analysis
(PCA)
adopted
reduce
number
before
Furthermore,
objective
function
under
comprehensive
consideration
three
factors
HI,
is,
monotonicity,
trendability,
fitting
errors.
The
effectiveness
verified
using
C-MAPSS
dataset.
Multiple
comparison
results
show
possesses
excellent
performance
both
remaining
prediction.
