Energies,
Journal Year:
2024,
Volume and Issue:
17(2), P. 352 - 352
Published: Jan. 10, 2024
This
study
focuses
on
using
machine
learning
techniques
to
accurately
predict
the
generated
power
in
a
two-stage
back-pressure
steam
turbine
used
paper
production
industry.
In
order
by
turbine,
it
is
crucial
consider
time
dependence
of
input
data.
For
this
purpose,
long-short-term
memory
(LSTM)
approach
employed.
Correlation
analysis
performed
select
parameters
with
correlation
coefficient
greater
than
0.8.
Initially,
nine
inputs
are
considered,
and
showcases
superior
performance
LSTM
method,
an
accuracy
rate
0.47.
Further
refinement
conducted
reducing
four
based
analysis,
resulting
improved
0.39.
The
comparison
between
method
Willans
line
model
evaluates
efficacy
former
predicting
power.
root
mean
square
error
(RMSE)
evaluation
parameter
assess
prediction
algorithm
for
generator’s
By
highlighting
importance
selecting
appropriate
techniques,
high-quality
data,
utilising
refinement,
work
demonstrates
valuable
estimating
energy
Building Services Engineering Research and Technology,
Journal Year:
2023,
Volume and Issue:
44(4), P. 459 - 480
Published: April 26, 2023
Smart
grid
technology
has
given
users
the
ability
to
regulate
their
home
energy
in
a
much
more
effective
manner.
In
such
scenarios,
Home
Energy
Management
(HEM)
potentially
becomes
an
arduous
task,
as
it
necessitates
optimal
scheduling
of
smart
appliances
order
reduce
usage.
this
research,
hybrid
Harris
Hawk
Optimization-Sine
Cosine
Algorithm
(hHHO-SCA)
been
proposed
develop
meta-heuristic-based
HEM
system.
The
hybridization
HHO
with
SCA
done
enhance
exploration
and
exploitation
stages
HHO,
hence
improving
its
global
search
phase
effectively
optimizing
usages.
addition
this,
several
knapsacks
are
utilized
guarantee
that
load
demand
for
power
does
not
surpass
certain
level
during
peak
hours.
terms
electricity
prices
Peak
Average
Ratio
(PAR)
reduction,
is
demonstrated
be
beneficial
achieving
objectives.
Simulations
performed
multi-family
housing
complex
range
equipment.
results
achieved
approach
suggest
hHHO-SCA
comparatively
efficient
cost
PAR,
when
compared
other
optimization
techniques.
Practical
Application.
This
management
system
can
applied
optimally
schedule
all
building
minimize
consumption
provide
consumer
potential
savings
costs.
Journal of Computational Design and Engineering,
Journal Year:
2023,
Volume and Issue:
10(4), P. 1315 - 1349
Published: June 1, 2023
Abstract
A
metaheuristic
algorithm
that
simulates
the
foraging
behavior
of
remora
has
been
proposed
in
recent
years,
called
ROA.
ROA
mainly
host
parasitism
and
switching
remora.
However,
experiment,
it
was
found
there
is
still
room
for
improvement
performance
When
dealing
with
complex
optimization
problems,
often
falls
into
local
optimal
solutions,
also
problem
too-slow
convergence.
Inspired
by
natural
rule
“Survival
fittest”,
this
paper
proposes
a
random
restart
strategy
to
improve
ability
jump
out
solution.
Secondly,
inspired
remora,
adds
an
information
entropy
evaluation
visual
perception
based
on
With
blessing
three
strategies,
multi-strategy
Remora
Optimization
Algorithm
(MSROA)
proposed.
Through
23
benchmark
functions
IEEE
CEC2017
test
functions,
MSROA
comprehensively
tested,
experimental
results
show
strong
capabilities.
In
order
further
verify
application
practice,
tests
through
five
practical
engineering
which
proves
competitiveness
solving
problems.
Sensors,
Journal Year:
2022,
Volume and Issue:
22(15), P. 5793 - 5793
Published: Aug. 3, 2022
The
massive
environmental
noise
interference
and
insufficient
effective
sample
degradation
data
of
the
intelligent
fault
diagnosis
performance
methods
pose
an
extremely
concerning
issue.
Realising
challenge
developing
a
facile
straightforward
model
that
resolves
these
problems,
this
study
proposed
One-Dimensional
Convolutional
Neural
Network
(1D-CNN)
based
on
frequency-domain
signal
processing.
Fast
Fourier
Transform
(FFT)
analysis
is
initially
utilised
to
transform
signals
from
time
domain
frequency
domain;
was
represented
using
phasor
notation,
which
separates
magnitude
phase
then
fed
1D-CNN.
Subsequently,
trained
with
White
Gaussian
Noise
(WGN)
improve
its
robustness
resilience
noise.
Based
findings,
successfully
achieved
100%
classification
accuracy
clean
simultaneously
considerable
exceptional
adaptation
ability.
retained
up
97.37%,
higher
than
CNN
without
training
under
noisy
conditions
at
only
43.75%.
Furthermore,
98.1%
different
working
conditions,
superior
other
reported
models.
In
addition,
outperformed
state-of-art
as
Signal-to-Noise
Ratio
(SNR)
lowered
-10
dB
achieving
97.37%
accuracy.
short,
1D-CNN
promising
rolling
bearing
diagnosis.
Water,
Journal Year:
2022,
Volume and Issue:
14(21), P. 3549 - 3549
Published: Nov. 4, 2022
Precise
estimation
of
pan
evaporation
is
necessary
to
manage
available
water
resources.
In
this
study,
the
capability
three
hybridized
models
for
modeling
monthly
(Epan)
at
stations
in
Dongting
lake
basin,
China,
were
investigated.
Each
model
consisted
an
adaptive
neuro-fuzzy
inference
system
(ANFIS)
integrated
with
a
metaheuristic
optimization
algorithm;
i.e.,
particle
swarm
(PSO),
whale
algorithm
(WOA),
and
Harris
hawks
(HHO).
The
data
acquired
period
between
1962
2001
(480
months)
grouped
into
several
combinations
incorporated
models.
performance
was
assessed
using
root
mean
square
error
(RMSE),
absolute
(MAE),
Nash–Sutcliffe
Efficiency
(NSE),
coefficient
determination
(R2),
Taylor
diagram,
Violin
plot.
results
showed
that
maximum
temperature
most
influential
variable
compared
other
input
variables.
effect
periodicity
investigated,
demonstrating
efficacy
improving
models’
predictive
accuracy.
Among
developed,
ANFIS-HHO
ANFIS-WOA
outperformed
models,
predicting
Epan
study
different
Between
these
two
performed
better
than
ANFIS-HHO.
also
proved
when
they
used
prediction
given
station
obtained
another
station.
Our
can
provide
insights
development
hybrid
analysis
conducted
data-scare
regions.
