Journal of engineering system.,
Journal Year:
2024,
Volume and Issue:
2(4), P. 38 - 45
Published: Dec. 1, 2024
Hydraulic
systems
are
widely
used
in
wind
turbines.
However,
with
the
extension
of
operation
time,
failure
rate
hydraulic
system
gradually
increases,
which
seriously
drags
down
efficiency
turbine.
Therefore,
it
is
urgent
to
find
a
fast
and
accurate
fault
identification
method
identify
abnormality
turbine
system.
The
analysis
documents
collected
over
past
15
years
constitute
data
set
that
has
been
processed
by
BIO
annotation
technique
make
suitable
for
analysis.
Based
on
classic
Bert-BILSTM-CRF
entity
recognition
model,
an
optimized
version
developed.
Firstly,
Bert
model
collect
extract
features
scattered
light;
then,
these
fused
output
vector
BILSTM
model;
finally,
CRF
complete
classification
labels.
By
embedding
strategy
adversarial
learning
BERT
architecture,
robustness
successfully
enhanced.
Subsequently,
we
will
analyze
overall
architecture
obtained
triad
information
save
Neo4j
graph
database
promote
its
adaptability.
Finally,
help
Python,
have
created
knowledge
mapping.
study
ultimately
revealed
achieved
superior
performance
93
F1
score.
Up
1.8
percentage
points.
proposed
exhibits
enhancement
roughly
56
points
model.
Energy Informatics,
Journal Year:
2024,
Volume and Issue:
7(1)
Published: Aug. 8, 2024
Abstract
In
recent
years,
there
has
been
growing
interest
in
digital
twin
technology
both
industry
and
academia.
This
versatile
found
applications
across
various
industries.
Wind
energy
systems
are
particularly
suitable
for
platforms
due
to
the
integration
of
multiple
subsystems.
study
aims
explore
current
state
predictive
wind
by
surveying
literature
from
past
five
identifying
challenges
limitations,
addressing
future
research
opportunities.
review
is
structured
around
four
main
questions.
It
examines
commonly
employed
methodologies,
including
physics-based
modeling,
data-driven
approaches,
hybrid
modeling.
Additionally,
it
explores
data
sources
such
as
IoT
sensors,
historical
databases,
external
application
programming
interfaces.
The
also
delves
into
key
features
technologies
behind
real-time
systems,
communication
networks,
edge
computing,
cloud
computing.
Finally,
addresses
platforms.
Addressing
these
questions
enables
development
modeling
strategies
with
fusion
algorithms,
which
allow
interpretable
real
time.
Filter
methods
dimensionality
reduction
algorithms
minimize
computational
resource
demand
operating
algorithms.
Moreover,
advancements
high-bandwidth
networks
facilitate
efficient
transmission
between
physical
assets
twins
reduced
latency.
