Scientific Reports,
Journal Year:
2024,
Volume and Issue:
14(1)
Published: Oct. 28, 2024
Abstract
This
research
proposes
a
day-ahead
scheduling
utilizing
both
demand
side
management
(DSM),
and
Energy
Management
(EM)
in
grid-tied
nanogrid
comprises
of
photovoltaic,
battery,
diesel
generator
for
optimizing
the
generation
cost
energy
not
supplied
(at
grid-outage).
Wider
terminology
is
introduced
to
combine
load
controllability
(considered
traditional
DSM),
interval
capability
accommodate
additional
loads
defined
as
flexible,
non-flexible,
semi-flexible
intervals.
Moreover,
user
selection
EM
or
combined
operation
with
DSM
at
different
degrees
flexibility
preference.
In
addition,
three
utility’s
operations
are
considered
denoted
fixed
rate
pricing
(FRP),
time-of-use
(ToU)
pricing,
FRP
grid-outage.
Hence,
suggested
framework
utilizes
opportunities
diversity,
electricity
strategy,
flexibility.
The
obtained
result
show
that,
flexible
intervals
reduces
by
21.02%,
25.23%,
18.15%
FRP,
ToU,
grid-outage
scenarios
respectively.
And
reduction
20.41%,
22.42%,
17.81%
16.24%,
21.15%,
13.8%
non-flexible
associated
full
utilization
renewable
from/to
battery
which
enhances
its
lifetime
required
size
during
design
stage
provisions
saving
A
hybrid
optimization
technique
Moth-flame
algorithm,
Lagrange’s
multiplier
proposed
confirms
effectiveness
detailed
comparison
other
techniques.
Fluctuations
in
real
time
and
vulnerability
to
cyber
threats
make
smart
nano
grids
less
stable
efficient.
In
addition,
most
control
algorithms
need
help
adapting
such
fluctuations,
leading
inefficiency
instability.
To
improve
the
dependability
scalability
of
grids,
a
novel
"Federated
Reinforced
LSTM-Crayfish
Whale
Optimization
Detection
(FRLC-WOD)"
is
introduced,
which
integration
Reinforcement-LSTM-Crayfish
Technique
(RL-LSTM-CAO)
Federated
Graph
Intrusion
(FG-WOA-ID).
This
system
consists
RL-LSTM-CAO,
uses
Bidirectional
Long
Short-Term
Memory
for
accurate
prediction,
Reinforcement
Learning-based
Power
Distribution
real-time
adaptability,
Crayfish
Algorithm
energy
management
optimal
terms.
It
enhances
stability
efficiency
grid
as
demand
adjustments
are
possible
ensures
improved
distribution.
Additionally,
FG-WOA-ID
formed
by
Learning
with
Anomaly
decentralized
data
processing,
Neural
Networks
intrusion
detection,
adaptive
security
optimization
concerning
addressing
issues.
Results
show
enhanced
efficiency,
cybersecurity,
renewable
utilization,
ensuring
reliable
optimized
performance.
Renewable energy focus,
Journal Year:
2024,
Volume and Issue:
50, P. 100596 - 100596
Published: July 3, 2024
The
electricity
distribution
network
is
experiencing
notable
transformations
influenced
by
a
cluster
of
dynamic
forces
for
electrical
grid
remodelling.
One
the
technologies
supporting
this
transformation
transactive
energy
system
(TES).
TES
solution
can
dynamically
balance
demand
and
supply
using
economic
control
techniques.
paves
way
new
approach
to
independent
retail
markets.
This
article
comprehensively
examines
analyzes
relevant
recent
literature
on
technological
advancements
in
with
special
focus
concepts,
models,
metrics,
technologies,
policies,
drawbacks,
future.
review
demonstrates
viability
as
future
offer
between
growth
terms
provisioning
at
affordable
cost,
accessibility
energy,
utilization.
In
addition,
sustainable
environments
promote
implementation
utilization
green
energy.
Furthermore,
work
presents
balanced
but
critical
state-of-the-art
advance
knowledge
research
area.
recommendations
include
adopting
advanced
techniques
delivering
effective
management
cost
reduction
solutions.
Others
are
experimental
validation;
combining
blockchain
Artificial
Intelligence
solve
identified
challenges.
Smart Grids and Sustainable Energy,
Journal Year:
2024,
Volume and Issue:
9(2)
Published: Oct. 8, 2024
Accurate
solar
irradiation
prediction
is
directly
linked
to
estimating
the
power
output
of
photovoltaic
systems,
making
it
essential
for
optimal
operation
and
management
facilities.
Over
past
few
years,
there
has
been
an
increasing
interest
in
domain
irradiance
prediction,
where
numerous
long-
short-term
memory
(LSTM)
convolutional
neural
network
(CNN)-based
models
have
emerged
as
promising
approaches
forecasting.
This
paper
aims
at
developing
a
simple
accurate
LSTM
CNN-based
global
horizontal
(GHI)
forecasting
predictor.
More
specifically,
present
study
examines
impact
several
parameters
on
accuracy,
including
timestep
length,
dataset
size,
number
units,
size
CNN
filters,
input
configuration
model.
Using
Los
Angeles
Irradiance
datasets,
yearly
seasonal
are
explored
predicting
one-hour
ahead
irradiance.
It
observed
through
simulation
results
that
using
enhances
accuracy
quality
increases
stable
datasets
free
contradictory
examples.
In
models,
clearly
influenced
by
kernel
max
pooling,
where,
after
tests,
lag
$$N\_steps=72\;hours$$
pooling
=
1
found
be
parameter
values.
While
stacked
layers
improve
stacking
does
not
necessarily
lead
improvements.
obtained
show
annual
LSTM-based
model
with
three
descending
(32,
16,
8
units)
outperforms
all
other
examined
structures,
achieving
root
mean
square
error
(RMSE)
absolute
(MAE)
values
approximately
37.08
W/m2
15.23
W/m2,
respectively.
However,
GHI
without
might
surpass
schemes
terms
efficiency.
