Frontiers in Ecology and Evolution,
Journal Year:
2023,
Volume and Issue:
11
Published: Oct. 17, 2023
Introduction
Large-scale
construction
projects
such
as
sports
stadiums
are
known
for
their
significant
energy
consumption
and
carbon
emissions,
raising
concerns
about
sustainability.
This
study
addresses
the
pressing
issue
of
developing
carbon-neutral
by
proposing
an
integrated
approach
that
leverages
advanced
convolutional
neural
networks
(CNN)
quasi-recurrent
long
short-term
memory
(QRLSTM)
models,
combined
with
dynamic
attention
mechanisms.
Methods
The
proposed
employs
CNN-QRLSTM
model,
which
combines
strengths
CNN
QRLSTM
to
handle
both
image
sequential
data.
Additionally,
mechanisms
adaptively
adjust
weights
based
on
varying
situations,
enhancing
model's
ability
capture
relevant
information
accurately.
Results
Experiments
were
conducted
using
four
datasets:
EnergyPlus,
ASHRAE,
CBECS,
UCl.
results
demonstrated
superiority
model
compared
other
achieving
highest
scores
97.79%
accuracy,
recall
rate,
F1
score,
AUC.
Discussion
integration
deep
learning
models
in
stadium
management
offers
a
more
scientific
decision
support
system
stakeholders.
facilitates
sustainable
choices
reduction
resource
utilization,
contributing
development
stadiums.
Energies,
Journal Year:
2024,
Volume and Issue:
17(8), P. 1926 - 1926
Published: April 18, 2024
Short-term
load
forecasting
(STLF)
plays
a
vital
role
in
ensuring
the
safe,
efficient,
and
economical
operation
of
power
systems.
Accurate
provides
numerous
benefits
for
suppliers,
such
as
cost
reduction,
increased
reliability,
informed
decision-making.
However,
STLF
is
complex
task
due
to
various
factors,
including
non-linear
trends,
multiple
seasonality,
variable
variance,
significant
random
interruptions
electricity
demand
time
series.
To
address
these
challenges,
advanced
techniques
models
are
required.
This
study
focuses
on
development
an
efficient
short-term
model
using
forest
(RF)
algorithm.
RF
combines
regression
trees
through
bagging
subspace
improve
prediction
accuracy
reduce
variability.
The
algorithm
constructs
bootstrap
samples
selects
feature
subsets
at
each
node
enhance
diversity.
Hyperparameters
number
trees,
minimum
sample
leaf
size,
maximum
features
split
tuned
optimize
results.
proposed
was
tested
historical
hourly
data
from
four
transformer
substations
supplying
different
campus
areas
University
Beira
Interior,
Portugal.
training
were
January
2018
December
2021,
while
2022
used
testing.
results
demonstrate
effectiveness
one
day
ahead
its
potential
decision-making
processes
smart
grid
operations.
Energy,
Journal Year:
2023,
Volume and Issue:
283, P. 129213 - 129213
Published: Sept. 28, 2023
Accurate
load
forecasting
is
important
to
mitigate
the
negative
impact
of
Electric
vehicle
integration
into
existing
grid.
Previous
studies
mostly
focus
on
individual
or
aggregated
levels
without
specifying
accuracy
due
selection
different
spatial
and
lack
uncertainty
estimation
in
models.
To
address
these
issues,
this
study
compares
predictive
performance
a
Random
Forest
Artificial
Neural
Networks
at
with
15-min
resolution
data
across
case
(i)
2
Vehicles
charging
poles
3
users,
(ii)
75
poles,
8
rails
70
users.
The
outcome
shows
that
Vehicle
smaller
will
require
presence
calendar
information
Whereas
more
than
10
piles,
features
"previous
week's
power",
"hour
day"
"number
connections"
can
achieve
similar
results.
results
also
showed
was
accurate
piles.
Moreover,
plot
generated
for
90%
prediction
interval
estimates
were
reliable
large
numbers
Vehicles.
