2022 IEEE 10th Joint International Information Technology and Artificial Intelligence Conference (ITAIC),
Год журнала:
2023,
Номер
unknown, С. 165 - 169
Опубликована: Дек. 8, 2023
To
improve
the
utilization
efficiency
of
wind
energy,
this
research
proposes
a
hybrid
model
based
on
Temporal
Convolutional
Network
(TCN)
and
two-level
speed
decomposition.
Firstly,
original
data
is
decomposed
into
main
residual
signals
through
Singular
Spectrum
Analysis
(SSA).
Then,
usage
Variational
mode
decomposition
(VMD)
decomposes
several
sub-sequences.
The
next
step
involves
predicting
signal
all
sub-sequences
using
TCN.
Eventually,
Grey
Wolf
Optimizer
(GWO)
employed
to
perform
optimization
stack
prediction
results,
resulting
in
outcomes.
results
demonstrate
that
proposed
SSA-VMD-TCN-GWO
outperforms
reference
models.
Thus,
provides
new
solution
Results in Engineering,
Год журнала:
2024,
Номер
21, С. 101899 - 101899
Опубликована: Фев. 12, 2024
The
imperative
of
accurate
forecasting
spans
diverse
industrial
sectors,
notably
impacting
the
tent
manufacturing
industry.
This
study
embarks
on
a
rigorous
examination
and
development
novel
models,
specifically
tailored
for
this
sector.
We
introduce
juxtapose
two
distinct
approaches:
Holt-Winters
method
Artificial
Neural
Networks
(ANN).
Our
analysis
is
grounded
in
case
company,
delving
into
dynamics
demand
variation,
particularly
under
seasonal
influences.
Through
meticulous
comparison,
we
demonstrate
efficacy
ANN
model,
highlighting
its
superior
accuracy
forecasting,
especially
Elite
Party
Canopy
albeit
with
noted
prediction
error
15%
Vista
tents.
paper
also
explores
broader
supply
chain
context
industry,
examining
influential
factors
affecting
commercial
sales
identifying
key
players.
findings
underscore
nuanced
capabilities
capturing
intricate
patterns,
offering
promising
direction
refining
practices
Energies,
Год журнала:
2024,
Номер
17(6), С. 1270 - 1270
Опубликована: Март 7, 2024
Wind
prediction
has
consistently
been
in
the
spotlight
as
a
crucial
element
achieving
efficient
wind
power
generation
and
reducing
operational
costs.
In
recent
years,
with
rapid
advancement
of
artificial
intelligence
(AI)
technology,
its
application
field
made
significant
strides.
Focusing
on
process
AI-based
modeling,
this
paper
provides
comprehensive
summary
discussion
key
techniques
models
data
preprocessing,
feature
extraction,
relationship
learning,
parameter
optimization.
Building
upon
this,
three
major
challenges
are
identified
prediction:
uncertainty
data,
incompleteness
complexity
learning.
response
to
these
challenges,
targeted
suggestions
proposed
for
future
research
directions,
aiming
promote
effective
AI
technology
address
issues
therein.
e-Prime - Advances in Electrical Engineering Electronics and Energy,
Год журнала:
2023,
Номер
6, С. 100293 - 100293
Опубликована: Сен. 21, 2023
As
the
global
population
is
growing
at
a
high
rate,
so
electricity
demand
also
increasing
faster
rate.
This
exerts
pressure
on
electricity-generating
plants
and
maintenance
engineers
because
of
variability
in
demand.
Avoiding
disruption
supply
to
meet
requires
forecasting
what
future
will
look
like
be
able
plan
adequately
towards
it.
study,
therefore,
develops
new
model
using
feature
extraction
(FE)
where
statistical
information
hourly
data
extracted
which
serves
as
input
variables
for
Backpropagation
neural
network
(BPNN)
optimized
by
particle
swarm
optimization
(PSO)
Ghana.
The
known
FE-PSO-BPNN
compared
other
seven
models
such
Radial
Basis
Function
(RBFNN),
Random
Forest
(RF),
Gradient
Boosting
Machine
(GBM),
Multivariate
Adaptive
Regression
Splines
(MARS),
BPNN,
PSO-RBFNN
FE
selects
all
models.
Electricity
from
Ghana
Grid
Company
period
including
1st
September
2018
30th
November
2019
used
testing
model's
performance.
Evaluation
criteria
Root
Mean
Square
Error
(RMSE),
Absolute
(MAE),
Percentage
(MAPE),
Scatter
Index
(SI)
were
used.
proposed
more
powerful
than
others
it
has
RMSE
(0.5344),
MAE
(3.3845),
MAPE
(0.1773),
SI
(0.0003).
expected
better
option
sector
managers
when
considering
forecasting.
