Energies,
Journal Year:
2024,
Volume and Issue:
17(17), P. 4423 - 4423
Published: Sept. 3, 2024
Trends
in
the
use
of
renewable
energy
sources
to
power
buildings
do
not
bypass
objects
for
which
maintaining
a
supply
is
critical.
This
also
applies
railway
signal
boxes.
The
aim
research
work
was
test
multisource
system
box
with
electronic
converter
systems
and
DC
bus,
built
as
part
project.
assumption
powering
building
sources,
storage
devices,
3
kV
traction
network
second
required
grid.
Both
grids
were
connected
by
converters,
values
converters
set
based
on
calculated
balance
using
measured
at
nodes
constraints.
tests
primarily
tested
response
changes
load
generated
photovoltaic
source,
well
charge
level
devices.
correctness
control
algorithm’s
operation
assessed
recorded
nodes.
carried
out
60
scenarios
that
covered
all
normal
emergency
operating
conditions.
During
tests,
delays
supplied
circular
flow
between
grid
connections
recorded.
ranged
from
2
about
50
s
flows
did
exceed
1500
W.
Based
results
it
found
necessary
improve
measurement
quality
communication
transmission
time
data
time.
Energy Science & Engineering,
Journal Year:
2024,
Volume and Issue:
12(7), P. 3142 - 3156
Published: July 1, 2024
Abstract
The
accurate
prediction
of
photovoltaic
(PV)
power
is
crucial
for
planning,
constructing,
and
scheduling
high‐penetration
distributed
PV
systems.
Traditional
point
methods
suffer
from
instability
lack
reliability,
which
can
be
effectively
addressed
through
interval
prediction.
This
study
proposes
a
short‐term
method
based
on
the
framework
sparrow
search
algorithm
(SSA)‐variational
mode
decomposition
(VMD)‐convolutional
neural
network
(CNN)‐gate
recurrent
unit
(GRU).
First,
data
undergo
similar
day
clustering
permutation
entropy
VMD
applied
to
solar
radiation
signals
with
high
correlation.
Then,
hyperparameters
GRU
are
optimized
by
SSA
according
comprehensive
evaluation
indicator
proposed
in
this
study.
Subsequently,
quantile
results
obtained
CNN‐GRU
using
optimal
parameters
optimization.
Finally,
composed
multiple
results.
A
MATLAB
R2022b
program
developed
compare
different
methods.
demonstrate
that
compared
single
methods,
improves
coverage
width‐based
criterion.
In
sunny
rainy
days,
indicators
only
54.3%
37.4%
GRU,
respectively,
indicating
significantly
improved
accuracy.
