Anais da Academia Brasileira de Ciências,
Journal Year:
2024,
Volume and Issue:
96(4)
Published: Jan. 1, 2024
Accurate
wind
power
prediction
can
effectively
alleviate
the
pressure
of
system
peak
frequency
regulation,
and
is
more
conducive
to
economic
dispatch
system.
To
enhance
forecasting
accuracy,
a
hybrid
approach
for
interval
proposes
in
this
study.
Firstly,
an
Improved
Complete
Ensemble
Empirical
Mode
Decomposition
with
Adaptive
Noise
(ICEEMDAN)
applied
decompose
initial
sequence
into
multiple
modes,
Variational
used
further
high-frequency
non-stationary
components.
Next,
Fuzzy
Entropy
(FE)
utilized
assess
complexity
post-decomposed
Intrinsic
Functions
(IMFs),
different
methods
are
employed
accordingly,
point
predictions
were
obtained
by
linearly
summing
component
predictions.Additionally,
improved
sparrow
search
algorithm
(ISSA)
seek
optimal
hyperparameters
algorithm.
Finally,
intervals
constructed
using
results
based
on
kernel
density
estimation
(KDE).
The
root
mean
square
errors
(RMSE)
deterministic
2.8458
MW
1.8605
MW,
uncertainty
coverage
rates
95.83%
97.92%
at
95%
confidence
level.
