sustainability

Article
A Deep Learning-Based Solar Power Generation Forecasting
Method Applicable to Multiple Sites
Seon Young Jang 1 , Byung Tae Oh 2, * and Eunsung Oh 3, *

1 Department of Electronics and Information Engineering, Korea Aerospace University,

Goyang-si 10504, Gyeonggi-do, Republic of Korea; tjsdud1551@[Link]
2 Department of Computer Engineering, Korea Aerospace University,
Goyang-si 10504, Gyeonggi-do, Republic of Korea
3 Department of Electrical Engineering, College of IT Convergence, Global Campus, Gachon University,
Seongnam-si 13120, Gyeonggi-do, Republic of Korea
* Correspondence: byungoh@[Link] (B.T.O.); esoh@[Link] (E.O.); Tel.: +82-2-300-0409 (B.T.O.);
+82-31-750-5346 (E.O.)

Abstract: This paper addresses the challenge of accurately forecasting solar power generation (SPG)
across multiple sites using a single common model. The proposed deep learning-based model is
designed to predict SPG for various locations by leveraging a comprehensive dataset from multiple
sites in the Republic of Korea. By incorporating common meteorological elements such as temperature,
humidity, and cloud cover into its framework, the model uniquely identifies site-specific features to
enhance the forecasting accuracy. The key innovation of this model is the integration of a classifier
module within the common model framework, enabling it to adapt and predict SPG for both known
and unknown sites based on site similarities. This approach allows for the extraction and utilization
of site-specific characteristics from shared meteorological data, significantly improving the model’s
adaptability and generalization across diverse environmental conditions. The evaluation results
demonstrate that the model maintains high performance levels across different SPG sites with
minimal performance degradation compared to site-specific models. Notably, the model shows
robust forecasting capabilities, even in the absence of target SPG data, highlighting its potential to
enhance operational efficiency and support the integration of renewable energy into the power grid,
thereby contributing to the global transition towards sustainable energy sources.
Citation: Jang, S.Y.; Oh, B.T.; Oh, E. A
Keywords: convolutional neural network; deep learning; domain estimation; long short-term memory;
Deep Learning-Based Solar Power
Generation Forecasting Method
machine learning; renewable; solar power generation
Applicable to Multiple Sites.
Sustainability 2024, 16, 5240. https://
[Link]/10.3390/su16125240
1. Introduction
Academic Editor: Hansang Lee
1.1. Motivation
Received: 9 April 2024 Renewable power generation has witnessed unprecedented growth in recent years,
Revised: 6 June 2024
with the largest growth observed for SPG. According to the International Renewable
Accepted: 18 June 2024
Energy Agency [1], solar power accounted for over 65% of global renewable capacity
Published: 20 June 2024
additions in 2023, underscoring its pivotal role in the transition towards sustainable energy
systems. This surge is attributed not only to the potential of solar power to address
pressing environmental concerns but also to its economic viability. Recent analyses indicate
Copyright: © 2024 by the authors.
that solar power’s levelized cost of energy (LCOE) has become increasingly competitive,
Licensee MDPI, Basel, Switzerland. outperforming coal and gas combined-cycle costs in many developed countries [2].
This article is an open access article However, integrating solar power into the energy grid introduces unique challenges
distributed under the terms and owing to its inherent variability. SPG is highly dependent on natural phenomena, such
conditions of the Creative Commons as sunlight intensity and duration, which fluctuate daily and seasonally. Accurate SPG
Attribution (CC BY) license (https:// forecasting ensures grid stability and maximizes solar energy utilization efficiency [3].
[Link]/licenses/by/
4.0/).

Sustainability 2024, 16, 5240. [Link] [Link]

Sustainability 2024, 16, 5240 2 of 15

1.2. Prior Works

Recent advancements in deep learning (DL) for SPG forecasting have led to the devel-
opment of more accurate and robust predictive models. Among these, the development
and application of various innovative DL frameworks and methodologies to enhance the
precision and reliability of solar power forecasts stand out.
According to the reviews in [4,5], for SPG forecasting, approximately 70% of studies
employ DL models, including Recurrent Neural Network (RNN) and Long Short-Term
Memory (LSTM) models, while 20% of the research specifically utilizes Convolutional Neu-
ral Network (CNN) models, which are also a type of DL model. Additionally, recent studies
have explored the combination of DL models with fuzzy logic or meta-heuristic approaches
to enhance prediction accuracy [6]. This distribution highlights the predominant reliance
on sequence-based models, such as RNNs and LSTMs, for forecasting tasks that involve
time series data of SPG, which require capturing temporal dynamics and dependencies
over time. These models are particularly suitable for handling sequential data because
of their ability to remember information over long periods, making them effective in pre-
dicting future values based on past trends. Dhaked et al. proposed an LSTM-based SPG
forecasting model considering weather conditions and demonstrated the effect of changing
the number of hidden layers and constraints on the forecast [7]. Wang et al. suggested
an SPG forecasting method based on an LSTM model with frequency-domain input data
decomposition [8]. In [8], Bayesian optimization was applied to optimize the hyperpa-
rameters of an LSTM model. Li et al. proposed an LSTM-based SPG forecasting method
using variational-mode data decomposition [9]. Similar to [8], in [9], the performance was
improved by creating multiple models for different states through data decomposition
and applying an artificial Gorilla troop optimizer. These studies indicate that for SPG
forecasting, the trend is not merely to apply classical LSTM models but rather to combine
LSTM models with optimizers and other techniques to fine-tune the model parameters
more effectively and enhance their performance.
The use of CNN models, traditionally known for their application in image processing
and recognition tasks, in 20% of the studies indicated the broader applicability of these
models in analyzing time series data. CNNs can capture spatiotemporal patterns in data,
which can be particularly useful in scenarios where the spatial distribution of inputs
(e.g., cloud cover and solar irradiance maps) plays a significant role in the forecasting
process. CNN models are often researched not as standalone solutions but rather in
combination with LSTM models to form hybrid models. Marinho et al. showed that a
CNN-LSTM model performs more accurately than standalone CNN and LSTM models [10].
Anu-Shalini et al. demonstrated that a CNN-bidirectional LSTM (biLSTM) model provides
more accurate predictions than standalone CNN, LSTM, and biLSTM models [11]. Similarly,
an autoencoder (AE)-biLSTM model was proposed in [12]. These studies indicate that
the biLSTM model is more efficient than the LSTM model for SPG forecasting due to its
ability to capture dependencies in both forward and backward directions, providing a more
comprehensive contextual understanding and enhancing prediction accuracy. Houran et al.
proposed a CNN–LSTM model based on the Coati optimization algorithm [13]. A Coati
optimizer determines the hyperparameters to improve the convergence and accuracy of the
CNN-LSTM model. Alharkan et al. proposed a dual-stem CNN-LSTM model integrated
attention mechanism [14]. In [14], the dual-stem design had two parallel pathways within
the model: one stem processed the input data through CNN layers, which were adept at
extracting spatial features and patterns, and the other stem utilized LSTM layers to capture
temporal dependencies and sequences in the data.
Research is also progressing on the application of transformers based on self-attention
mechanisms. Initially designed for natural language processing tasks, transformers have
shown remarkable success because of their ability to process sequences in parallel and
capture long-range dependencies in data through self-attention [15]. This mechanism
allows the model to weigh the importance of different parts of the input data relative
to each other, providing a dynamic method to focus on relevant information for making
Sustainability 2024, 16, 5240 3 of 15

predictions. Al-Ali et al. proposed a CNN-LSTM transformer model [16]. In [16], a CNN-
LSTM model was used to extract spatial and temporal features, and a transformer was
applied to generate forecast results from the features. Zhu et al. proposed a transformer-
based SPG forecasting method using data filtering [17]. In [17], the Savitzky–Golay filter
(SG) and Local Outlier Factor (LOF) filter were applied for data preprocessing to reduce
noise, and a transformer was employed for the SPG forecast model.
Table 1 summarizes DL-based SPG forecasting methods.

Table 1. DL-based SPG forecast methods.

Ref. Model Description

[7] LSTM LSTM with grid search-based hyperparameter fine-tuning
Frequency-domain data decomposition and LSTM with BO-based
[8] BO-LSTM
hyperparameter tuning
Fuzzy-based data decomposition and LSTM with AGT optimizer-based
[9] AGT-LSTM
hyperparameter tuning
[10] CNN-LSTM Hybrid CNN-LSTM model with statistical data model
[11] CNN-biLSTM Hybrid CNN-bidirectional LSTM model with statistical data model
[12] AE-biLSTM Hybrid Auto Encoder-bidirectional LSTM model with statistical data model
[13] CO-CNN-LSTM CNN-LSTM with CO-based hyperparameter tuning
[14] DOCNN-LSTM Dual-stem CNN-LSTM with an attention mechanism
[16] CNN-LSTM-Transformer Transformer combining CNN-LSTM
[17] SG-LOF-Transformer Transformer with SG-LOF data filtering

These studies indicate that the trend for SPG forecasting is not merely to apply classical
DL models but rather to combine models with optimizers and other techniques to fine-tune
the model parameters and enhance their performance more effectively.
However, the current research predominantly focuses on modeling individually im-
plemented SPG sites. This approach involves collecting and utilizing site-specific input
data, such as weather conditions and training data, to generate forecasts for a single site.
Consequently, this methodology implies that producing forecasts for multiple sites would
necessitate either the design of distinct models tailored to each site or separate training
sessions for models corresponding to each location. This presents a limitation regarding
scalability and efficiency when extending the forecasting capabilities across multiple sites.

1.3. Contribution
This study addresses the significant challenge of developing an accurate and scalable
SPG forecasting model that can be effectively applied across multiple sites using a single
common model. Traditional SPG forecasting methods often require site-specific models,
which can be inefficient and difficult to scale. Our research proposes a novel deep learning-
based common model designed to overcome these limitations by leveraging common
meteorological elements such as temperature, humidity, and cloud cover.
The key contributions of this study are as follows:
• The development of a common model: We propose a deep learning-based forecasting
model that can accurately predict SPG for various locations by utilizing common
meteorological data. This model is designed to extract site-specific characteristic
features from these shared data elements, thus enhancing the forecasting accuracy.
• The integration of a classifier module: To address the variability and unique charac-
teristics of different sites, we integrate a classifier module within the common model
framework. This module enables the model to adapt and predict SPG for both known
and unknown sites by identifying and leveraging site similarities.
• Improved adaptability and generalization: By extracting and utilizing site-specific
features from common meteorological data, the proposed model significantly im-
proves its adaptability and generalization capabilities across diverse environmental
conditions. This dual-layered approach allows for a better understanding of how dif-
Sustainability 2024, 16, x FOR PEER REVIEW 4 of 15

improves its adaptability and generalization capabilities across diverse environmen-

Sustainability 2024, 16, 5240 tal conditions. This dual-layered approach allows for a better understanding of4how of 15
different factors contribute to SPG outputs at each location, enabling a highly adapt-
able forecasting framework.
• Robustfactors
ferent performance across
contribute multiple
to SPG outputs sites: The location,
at each evaluation resultsa demonstrate
enabling that
highly adaptable
our model maintains
forecasting framework. high performance levels across different SPG sites, with mini-
• mal performance
Robust performance degradation compared
across multiple sites: to
Thesite-specific
evaluation models. Notably, thethat
results demonstrate model
our
exhibits
model robust forecasting
maintains capabilities
high performance levelseven in the
across absence
different SPGofsites,
targetwith
SPGminimal
data, high-
per-
lighting itsdegradation
formance potential tocompared
enhance operational efficiency
to site-specific [Link] supportthe
Notably, themodel
integration of
exhibits
renewable
robust energy capabilities
forecasting into the power
evengrid.
in the absence of target SPG data, highlighting its
potential to the
To validate enhance operational
effectiveness efficiency
of this and support
multifaceted the integration
approach, we employed of renewable
a compre-
energy
hensive into encompassing
dataset the power grid.a nationwide area in the Republic of Korea. Through rig-
orousTotesting
validate and
thevalidation,
effectivenesstheof model demonstrated
this multifaceted the capability
approach, we employed of being applied
a comprehen-
sive
acrossdataset encompassing
multiple a nationwide
sites as a common model area in the
with onlyRepublic
marginalofperformance
Korea. Through rigorous
degradation
testing
compared andtovalidation, the model
models trained demonstrated
specifically the capability
for individual of being applied
sites. In particular, using theacross
pro-
multiple
posed modelsites allows
as a common modelforecasting
us to obtain with only marginal performance
performance within adegradation compared
certain range, even in
to models
cases wheretrained specifically
the target SPG datafor individual
do not [Link]. Inverifies
This particular,
thatusing the proposed
the proposed model
model can
allows us to obtain forecasting performance within a certain range,
significantly improve operational efficiency and renewable energy integration into the even in cases where
the
powertarget SPG data do not exist. This verifies that the proposed model can significantly
grid.
improve operational efficiency and renewable energy integration into the power grid.
2. Methodology
2. Methodology
2.1. SPG Forecasting System Architecture
2.1. SPG Forecasting System Architecture
The objective of our proposed forecasting system is to accurately predict SPG by lev-
The objective of our proposed forecasting system is to accurately predict SPG by
eraging six key types of meteorological data: temperature (TMP), humidity (REH), pre-
leveraging six key types of meteorological data: temperature (TMP), humidity (REH),
cipitation probability (POP), cloudiness index (SKY), wind direction (VEC), and wind
precipitation probability (POP), cloudiness index (SKY), wind direction (VEC), and wind
speed (WSD).
speed (WSD). We We devised
devised aa dual-component
dual-component SPG SPG prediction
predictionframework
frameworkto toenhance
enhanceopera-
oper-
ational efficiency and feasibility, as illustrated in Figure 1. This framework
tional efficiency and feasibility, as illustrated in Figure 1. This framework is partitioned is partitioned
into two
into two distinct
distinct subsystems:
subsystems: one one dedicated
dedicated to to extracting
extracting features from raw
features from raw weather
weather data
data
and the other to utilize these features to forecast
and the other to utilize these features to forecast SPG. SPG.

Figure 1.
Figure 1. SPG
SPG forecasting system architecture.
forecasting system architecture.

The proposed two-tiered model assumes that the initial feature extraction from the
input data operates independently of the subsequent SPG forecasting forecasting phase.
phase. Given that
the input
input weather
weatherdata dataare
areintrinsically
intrinsically linked
linked to specific
to specific locallocal geological
geological factors
factors and
and con-
considering
sidering thatthat
SPGSPG output
output is predominantly
is predominantly determined
determined by these
by these input
input parameters—
parameters—as-
assuming a consistentSPG
suming a consistent SPGsystem
systemconfiguration—it
configuration—itstands standsto toreason
reason that
that segmenting the
overall system into two discrete units and independently optimizing each for its intended
purpose would result
result in
in increased
increased efficiency.
efficiency.
The feature extraction subsystem
feature extraction subsystem employs
employs a combination
a combination of CNN
of CNNand LSTM
and LSTMnetworks
net-
for the analysis. Specifically, it processed 24 h of weather data to ensure
works for the analysis. Specifically, it processed 24 h of weather data to ensure efficient efficient and
practical information
and practical handling.
information At each
handling. At timestamp, the data
each timestamp, pertaining
the to all sixtoweather
data pertaining all six
variables were considered.
weather variables In addition,
were considered. we factored
In addition, we in the solar
factored elevation
in the (SE) and (SE)
solar elevation azimuth
and
angle (AA) data to directly account for variations in solar irradiance.
azimuth angle (AA) data to directly account for variations in solar irradiance.
Let xi,t be the input element, which is the numerical weather prediction (NWP) for the
SPG site i at the timestamp t. It is defined as a vector:

xi,t = [ AAi,t , SEi,t , TMPi,t , REHi,t , POPi,t , SKYi,t , VECi,t , WSDi,t ] T , (1)
Sustainability 2024, 16, 5240 5 of 15

where [·] T is the transpose of a matrix. The input data for the forecasting of the SPG site i
are expressed as the 24-h time series of the input element as:

NWPi = [ xi,0 , xi,1 , · · · , xi,t , · · · , xi,22 , xi,23 ] T . (2)

To summarize, we constructed a 24 × 8 dimensional data matrix, where each row

represents the NWP data for each hour.
To manage this data structure, we propose a CNN architecture with a kernel size of 3,
allowing the CNN to incorporate temporal adjacency (i.e., data from the preceding and
succeeding time points) to ensure stability and robustness during the feature extraction
phase. This subsystem utilizes two convolutional layers complemented by batch normaliza-
tion and a Rectified Linear Unit (ReLU) activation function. The resultant features are then
further refined through a biLSTM module that considers the sequential nature of the 24-h
weather data chosen for its inherent temporal correlation. The biLSTM module comple-
ments the preceding CNN module by fully considering the sequential and non-sequential
flows for 24 h. Thus, the first subsystem serves as a Feature Encoder, distilling essential
information from the weather data into a usable format for forecasting.
The characteristics of the feature encoder are summarized as follows:
• CNN and LSTM Interaction: The CNN layers are responsible for extracting spatial
features from the input weather data. These features are then fed into the LSTM layers,
which analyze temporal patterns. The interaction is triggered sequentially: once the
CNN extracts the spatial features, these features are used as inputs in the LSTM layers
for further temporal analysis.
• Interaction Trigger: The trigger for the interaction between CNN and LSTM layers is
the sequential data flow within the model. After the CNN layers process the input
data to capture spatial correlations, the resulting feature maps are fed into the LSTM
layers to capture temporal dependencies.
• Justification for biLSTMs: The use of biLSTMs is justified by its ability to capture
dependencies in both forward and backward directions. This capability is particularly
useful for SPG forecasting as it allows the model to consider the influence of past and
future weather conditions on the current prediction. Bidirectional LSTMs provide a
more comprehensive temporal analysis, leading to improved prediction accuracy.
As stated in Section 1.2, the use of biLSTMs is known to be effective for power
generation forecasting.
The second subsystem, the Regressor, interprets the encoded features to predict SPG
output. This prediction mechanism was configured straightforwardly using a conventional
multilayer perceptron (MLP) structure comprising two layers.
This architecture underscores our commitment to developing a sophisticated yet
pragmatic solution for SPG forecasting. By dissecting the process into distinct feature
extraction and regression phases, our system not only optimizes the utilization of weather
data for solar power prediction but also sets a new standard for forecasting models in the
renewable energy sector.

2.2. Implicit Site Estimation using the Classification Module

A pivotal aspect of our research was addressing the significant challenge of predict-
ing SPG amid high variability across local sites. The variance in weather information
attributed to geographical, environmental, and seasonal factors complicates the collection
of a sufficient and reliable dataset for training. This variability often results in unstable and
unreliable training samples, ultimately impairing the forecasting accuracy for sites that are
not represented in the training data.
To mitigate these challenges, we introduce a novel system design that reduces the
sample bias inherent to each local site through a feature similarity-based classification
mechanism. This approach aggregates similar features to form coherent groups that
facilitate more accurate predictions. Specifically, we integrated a site classification module
Sustainability 2024, 16, x FOR PEER REVIEW 6 of 15

To mitigate these challenges, we introduce a novel system design that reduces the
Sustainability 2024, 16, 5240 sample bias inherent to each local site through a feature similarity-based classification 6 of 15
mechanism. This approach aggregates similar features to form coherent groups that facil-
itate more accurate predictions. Specifically, we integrated a site classification module
within theEncoder
within the Encodersystem,
system,asasshown
shownininFigure
Figure2. 2. This
This classifier
classifier waswas ingeniously
ingeniously designed
designed to
to compel
compel thethe encoder
encoder to implicitly
to implicitly assimilate
assimilate the the types
types of weather
of weather data,data, thereby
thereby embed-
embedding
ding
the the site
local localcharacteristics
site characteristics intomodel’s
into the the model’s understanding.
understanding. Consequently,
Consequently, our our sys-
system
tem exhibits
exhibits enhanced
enhanced robustness
robustness andand adaptability
adaptability andand can
can generatereliable
generate reliableforecasts
forecasts for
for
previously unknown
previously unknown sitessites by
by leveraging
leveraging feature
feature similarity
similarity toto infer
infer local
local environmental
environmental
conditions.
conditions.

Figure 2. Training
Figure 2. Trainingsystem
systemusing
usingthe
theclassifier
classifierfor
forsite
siteestimation.
estimation.

The
The goal
goal ofof our
our research
researchwaswas to
to develop
develop aa common
common model
model for
for SPG
SPG forecasting
forecasting across
across
multiple
multiple sites. Therefore, to train the proposed network, NWP data from multiple
sites. Therefore, to train the proposed network, NWP data from multiple sites
sites
were
wereused
usedsimultaneously.
simultaneously. The subscript
The i represents
subscript the site
𝑖 represents thelabel, consistent
site label, with its
consistent usage
with its
in Equations
usage (1) and(1)
in Equations (2).and
Moreover, we used
(2). Moreover, wethe following
used loss functions:
the following loss functions:
L T𝐿= =
α L𝛼R𝐿+ β 𝛽
LC𝐿, , (3)

where L𝐿R and

where andLC𝐿 represent
represent the
the regressionand
regression andclassification
classificationerrors, respectively.α 𝛼
errors,respectively. andand
β
𝛽 are
are thethe weight
weight factors
factors between
between the the regression
regression andand the the classification.
classification. In our
In our experiment,
experiment, we
weαset
set = β.𝛼 = 𝛽.
The regression
The regression error
error was
was determined
determined byby measuring
measuring the the total
total absolute
absolute error
error between
between
the
the predicted
predicted SPG
SPG and
and the
the target
target SPG
SPG values,
values, i.e.,
i.e., the
the actual
actual hourly
hourly SPG
SPG data,
data, across
across all
all
sites I 𝐼and
sites timestampsT:𝑇:
andtimestamps

L𝐿R ==∑ ∑ | pred

𝑝𝑟𝑒𝑑 , −
i,t − 𝑡𝑎𝑟𝑔𝑒𝑡
target i,t |. , . (4)
(4)
I t∈ T∈
i ∈∈

The common
The common cross-entropy
cross-entropyloss
losswas
wasadopted
adoptedfor
forthe
theclassification
classificationerror:
error:

LC𝐿==∑ qi log
𝑞 log
(Ci𝐶
) (5)
i∈ I ∈

whereq𝑞indicates
where indicates the
the true
true label
label and
and Ci 𝐶 indicates
indicates thethe predicted
predicted probability.
probability.
i
This innovation represents a significant advancement in SPG
This innovation represents a significant advancement in SPG forecasting forecasting and
and allows
allows
our model
our model to to overcome
overcome the thelimitations
limitationsofoftraditional
traditionalforecasting
forecastingmethods.
methods. By
By equipping
equipping
the encoder with the ability to discern and adapt to each site’s unique characteristics,
the encoder with the ability to discern and adapt to each site’s unique characteristics, we we
enabled more generalized and accurate predictions across a diverse array of
enabled more generalized and accurate predictions across a diverse array of SPG sites. This SPG sites.
This approach
approach not broadens
not only only broadens the applicability
the applicability of ourofforecasting
our forecasting
modelmodel butsets
but also alsoasets
newa
new standard for addressing the complexities of renewable energy prediction
standard for addressing the complexities of renewable energy prediction in a geographically in a geo-
graphically
varied varied landscape.
landscape.
Sustainability 2024, 16, x FOR PEER REVIEW 7 of 15
Sustainability 2024, 16, 5240 7 of 15

3. Results and and

3. Results Discussion
Discussion
3.1. Dataset
3.1. Dataset
ThisThis
study waswas
study conducted
conducted using SPG
using SPGdata measured
data measuredatatseven
sevensites
sitesdistributed
distributedna- nation-
tionwide
wide across the Republic
Republic ofof Korea,
Korea, asas shown
shownin inFigure
[Link]
TheSPG
SPGdata
datawere
wererecorded
recorded by
by the
theKorea
KoreaSouthern
SouthernPower
PowerCorporation
Corporationfrom from2013
2013toto
2022. The
2022. Thedata
datawere
werepublished
published onon a
a public data
public dataportal
portalmanaged
managedby bythe
theMinistry
Ministryof ofInterior
Interior and
and Safety,
Safety, Republic of of Korea
Korea [18].
[18].Data
Datafrom
from2013
2013to to 2020
2020 were
were usedused to train
to train the proposed
the proposed model,model, and
and the the remaining
remaining data were
dataused
weretoused to calculate
calculate the results.
the results. Each Each SPGhad
SPG site siteahad a different
different installation
installation capacity,
capacity, ranging
ranging
fromfrom
93 to93187
to 187
kWp. kWp.
ForFor a clear
a clear comparison,
comparison, thethe installation
installation capacities
capacities ofofallall siteswere
sites
werenormalized
normalizedtoto100100kWp
kWpininthe
theexperimental
experimentalresults.
results.

Figure
Figure 3. SPG
3. SPG site map.
site map.

The weather information utilized in this study was derived from day-ahead NWP data
The weather information utilized in this study was derived from day-ahead NWP
provided by the National Climate Data Center of the Republic of Korea Meteorological
data provided by the National Climate Data Center of the Republic of Korea Meteorolog-
Administration [19].
ical Administration [19].
The datasets used in this study are publicly available and were collected from various
The datasets used in this study are publicly available and were collected from various
data sources within public data portals [18,19]. The data include solar generation and mete-
data sources within public data portals [18,19]. The data include solar generation and me-
orological elements such as temperature, humidity, precipitation probability, cloudiness
teorological elements
index, wind such and
direction, as temperature,
wind speed,humidity, precipitation
collected from various probability, cloudiness
sites across the Republic of
index, wind direction, and wind speed, collected
Korea, as described in Tables 2 and 3. from various sites across the Republic of
Korea, as described in Tables 2 and 3.
Table 2. Description of the used SPG data.
Table 2. Description of the used SPG data.
Generation
Site ID Date Generation
Site ID Date 00 h ... 08 h 09 h ... 22 h 23 h
00 h … 08 h 09 h … 22 h 23 h
Site 1 2013-01-01 0.00 ... 9.88 32.60 ... 0.00 0.00
Site 1Site 12013-01-01
2013-01-02
0.00 0.00 … ...
9.88 10.52 32.60 39.02 … ...
0.00 0.00 0.00 0.00
Site 1 . 2013-01-02 . 0.00 .. … .. 10.52 . 39.02 . … .. 0.00 . 0.00 .
.. .. . . .. .. . .. .
⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ .

Table 3. Description
Table of the
3. Description ofused NWP
the used [Link].
NWP

Site ID
Site ID Date DateHour TMP (°C) TMPREH
Hour (◦ C) (%) REH POP
(%) (%)POP (%) SKY SKY VEC (°)
VEC (◦ )WSDWSD
(m/s)(m/s)
Site 1 2013-01-01 00 7.3 52 15 2 279 4.9
Site 1 2013-01-01 00 7.3 52 15 2 279 4.9
Site 1Site 12013-01-01 01
2013-01-01 01 5.8 5.8 50 50 12 12 2 2 281 281 5.8 5.8
⋮ . ⋮ . ⋮ .. ⋮ .. ⋮ .. ⋮ .. ⋮ .. ⋮ .. ⋮ .
.. .. . . . . . . ..
Sustainability 2024, 16, 5240 8 of 15

3.2. Experimental Setups

The experiments were performed on a system equipped with dual NVIDIA GeForce
RTX 4090 GPUs, leveraging the PyTorch framework for the optimal tensor processing of
GPU architectures. The initial learning rate was set to 0.001 and managed by the LambdaLR
scheduler as an epsilon value of 1 × 10−8 and a weight decay of 1 × 10−3 to dynamically
adjust the rate for optimal convergence and model performance.
A batch size of 30 was used to balance the memory constraints and computational
efficiency. The training plan was structured to run for 300 epochs within each iteration,
focusing on exposing the model to various data samples. The epoch yielding the smallest
test loss over 50 iterations was chosen to indicate model performance. This approach,
chosen after preliminary testing, aimed to maximize training effectiveness within comput-
ing resources, ensure a thorough learning process, and minimize overfitting against the
background of task complexity and the selected model architecture.
We evaluated the performance of our proposed system with three performance metrics.
The primary evaluation metric used was the Mean Absolute Error (MAE). The MAE was
calculated as follows:
1
MAEi = ∑ |ŷi,t − yi,t |, (6)
T t∈ T
where ŷi,t and yi,t are the predicted SPG value and actual SPG value of site i at timestamp i
and T is the length of observation periods, e.g., 24 h. This metric provides a measure of the
average magnitude of errors in the predictions, without considering their direction.
To provide a more comprehensive evaluation of our model, we also included the Mean
Squared Error (MSE) and Root Mean Square Error (RMSE) as additional metrics. The MSE
is defined as:
1
MSEi = ∑ (ŷi,t − yi,t )2 , (7)
T t∈ T
The MSE metric emphasizes larger errors more than the MAE due to the squaring of
the error terms. The RMSE, being the square root of the MSE, is calculated as:
s
1
RMSEi =
T ∑ (ŷi,t − yi,t )2 , (8)
t∈ T

The RMSE provides a measure that maintains the units of the original data, making it
interpretable in the context of the original values.

3.3. Site-Specific Performance

Initially, we evaluated the performance of our proposed system on a site-specific basis;
that is, we trained and tested the model independently for data from each site. In this
scenario, the site-specific model operates without the need for a classification module
because the model focuses solely on forecasting for a singular known site.
Table 4 lists the forecasting performances of the site-specific model.

Table 4. Forecast performance of the proposed site-specific model.

Model Loss Site 1 Site 2 Site 3 Site 4 Site 5 Site 6 Site 7 Avg.
MAE 3.16 3.33 3.35 3.34 6.08 1.72 3.65 3.52
Site-specific MSE 50.34 51.19 50.77 51.97 173.55 14.06 59.58 64.49
RMSE 7.10 7.15 7.13 7.21 13.17 3.75 7.72 7.60

The MAE across the seven sites was 3.43, with a standard deviation of 1.11. Given that
the installation capacity of each site was normalized to 100 kWp, the proposed method
achieved a forecast error of 3.5% or less, according to the installation capacity. This level of
accuracy is particularly noteworthy in the context of the Republic of Korea’s regulatory
Sustainability 2024, 16, 5240 9 of 15

framework for renewable energy grid integration, which stipulates a participation thresh-
old based on forecasting accuracy that requires an MAE of no more than 8% for SPG [20].
Therefore, the proposed forecasting method in a site-specific scenario demonstrated excel-
lent performance for SPG utilization, significantly surpassing the regulatory requirements
and showcasing its potential to contribute effectively to the integration of renewable energy
sources into the power grid.
In Table 4, site 5 exhibits outlier performance, with an MAE of 5.5, notably higher
than the other sites. This deviation can be attributed to the fact that site 5 is located
in an island region, as shown in Figure 3, where the accuracy of weather forecasting is
generally low because of the unique meteorological conditions of the area. Despite this
challenge, the performance at site 5 still met the regulatory requirement of remaining
within an 8% MAE threshold for participation in the Republic of Korea’s renewable energy
generation forecasting system. This underscores the robustness of the proposed forecasting
method, demonstrating its capability to deliver satisfactory forecast performance, even in
geographically challenging locations where weather prediction is inherently less precise.
The MSE across the seven sites averaged 64.49, with a standard deviation reflecting the
variance in prediction errors. The MSE metric emphasizes larger errors more than the MAE
due to the squaring of the error terms. This means that higher MSE values indicate that some
predictions had significantly larger errors. Site 5 again shows a much higher MSE (173.55)
compared to other sites, reinforcing the impact of its unique meteorological conditions. The
relatively higher MSE at site 5 suggests that the model’s predictions occasionally deviate
significantly from actual values, which is consistent with the challenges of forecasting in
this specific region.
The RMSE, which provides a measure that maintains the units of the original data,
averaged 7.60 across all sites. The RMSE is particularly useful as it directly relates to the
magnitude of errors in the same units as the predicted and actual SPG values. Site 5 had
the highest RMSE (13.17), indicating that the errors at this site were not only frequent
but also substantial in size. Despite this, the average RMSE value of 7.60 across all sites
demonstrates that the proposed model maintains a reasonable error margin, even when
considering the challenging conditions at site 5.
Overall, the proposed forecasting method shows strong performance across most sites,
significantly exceeding regulatory requirements. The higher errors observed at site 5 high-
light areas for potential improvement, particularly in regions with unique meteorological
conditions. The combination of MAE, MSE, and RMSE metrics provides a comprehensive
evaluation of the model’s accuracy and robustness, confirming its effectiveness for SPG
forecasting across diverse geographical locations. The results in Table 4 will be used as a
baseline for subsequent experiments.

3.4. Multisite Performance

Subsequently, we assessed the performance of a common model designed for multisite
applications. The weather data used to train the model were aggregated from all sites by
combining the datasets into a single comprehensive dataset. This process involved normal-
izing and standardizing the data to ensure uniformity across different sites. Specifically,
the weather data from 2013 to 2020 were used for training, while the data from 2021 to 2022
were reserved for testing.
To aggregate the data for training, we combined the meteorological elements from each
site into a unified dataset. Each site’s data were pre-processed to handle any missing values
and ensure consistency in the input features. This aggregated dataset was then used to train
the common model, allowing it to learn from a diverse range of environmental conditions.
For testing, the trained common model was evaluated using the individual site data
from 2021 to 2022. This approach ensured that the model’s performance was assessed on
data that were not seen during the training phase, providing an unbiased evaluation of its
generalization capabilities across different sites.
Sustainability 2024, 16, 5240 10 of 15

Investigating the performance of a common model is crucial as it represents a poten-

tially straightforward solution for forecasting at unknown sites. By leveraging a compre-
hensive dataset that encompasses the variability across different geographical locations, the
common model aims to generalize the forecasting capability, thereby facilitating accurate
predictions, even for sites not explicitly represented in the training data. This approach sim-
plifies the forecasting process for new or unknown sites and tests the model’s adaptability
and scalability across diverse environmental conditions.
Table 5 presents a comparison of the forecasting performance of the proposed models.
In the table, “Common w/o cls.” refers to the performance of the common model trained
on data from sites 1 to 7 without including a classifier, while “Common w cls.” shows the
performance outcomes of the proposed method with the applied classifier. The common
model without a classifier exhibited an average performance decrease of 22% compared
to the site-specific model results listed in Table 4. By contrast, the proposed method with
the classifier showed an average performance decrease of 15%, indicating a 7% improve-
ment over the basic common model without the classifier. This highlights the efficacy
of incorporating a classifier to enhance the forecasting accuracy of the common model,
demonstrating significant mitigation of performance loss when extending the model to
multisite applications.

Table 5. Forecast performance of the common model. The common models are trained by the data in
sites 1–7.

Model Loss Site 1 Site 2 Site 3 Site 4 Site 5 Site 6 Site 7 Avg.
MAE 3.90 4.30 3.34 3.85 5.99 3.47 4.45 4.19
Common
MSE 62.88 75.96 48.89 63.59 170.25 52.70 79.62 79.13
w/o cls.
RMSE 7.93 8.72 6.99 7.97 13.05 7.26 8.92 8.69
3.93
MAE 4.37 3.56 3.47 3.67 5.78 2.45 4.24
(7%)
Common 55.54
MSE 50.96 41.19 46.83 56.17 113.21 14.89 65.55
w cls. (42%)
7.21
RMSE 7.14 6.42 6.84 7.49 10.64 3.86 8.1
(21%)

In particular, for site 6, which exhibited one of the best forecast performances in the
site-specific model analysis, the MAE of the basic common model without a classifier
approximately doubled. However, in the case of the proposed model with the classifier, the
increase in the MAE was within 50%. This demonstrates that including the classification
model enables the encoder to incorporate site-specific information to a certain extent. This
incorporation significantly aids the regressor in making more accurate predictions by
providing it with contextually relevant features tailored to the characteristics of each site.
The improvement in forecasting accuracy for site 6 underlines the value of integrating
site-specific nuances into the common model framework, showcasing the classifier’s role in
enhancing the adaptability and predictive capability of the model across diverse locations.
The MSE and RMSE values further substantiate these findings. The MSE across all sites
for the common model without a classifier averaged 79.13, while for the model with a clas-
sifier, it averaged 55.54, indicating a 42% improvement. This substantial reduction in MSE
highlights the classifier’s role in mitigating larger errors, which is crucial for improving the
overall robustness of the forecasting model. The RMSE, which averaged 8.69 for the model
without a classifier and 7.21 for the model with a classifier, shows a 21% improvement. The
consistent improvement in both MSE and RMSE metrics underscores the effectiveness of
the classifier in enhancing the predictive performance of the common model.

3.5. Unknown-Site Performance

To further demonstrate the advantages of the proposed system, additional experiments
were conducted by segregating the seven sites into known and unknown categories. We
Sustainability 2024, 16, 5240 11 of 15

then train the common model exclusively using data from known sites. For instance,
we trained the common model using data from known sites, namely, 1, 2, 3, and 7, and
subsequently tested the model on unknown sites, namely, 4, 5, and 6.
Tables 6 and 7 present the forecast performances of the common model for unknown
sites, i.e., 4, 5, and 6. The common model was trained using data from known sites, i.e., 1, 2,
3, and 7 (Table 6) and sites 1, 3, and 7 (Table 7). The purpose of using different training sets
for the common models in Tables 6 and 7 was to evaluate the robustness and adaptability
of our model under varying conditions. By training the model on different combinations of
sites, we aimed to test its ability to generalize and predict SPG accurately across unknown
sites with diverse environmental conditions.

Table 6. Forecast performance of the common model for unknown sites. The common models are
trained by the data in known sites (1, 2, 3, and 7).

Model Loss Site 4 Site 5 Site 6 Avg.

MAE 4.82 6.76 7.63 6.40
Common
MSE 94.27 198.87 187.86 160.33
w/o cls.
RMSE 9.71 14.1 13.71 12.51
6.03
MAE 4.54 6.63 6.93
(6%)
Common 147.79
MSE 85.35 194.4 163.62
w cls. (8%)
11.99
RMSE 9.23 13.94 12.79
(4%)

Table 7. Forecast performance of the common model for unknown sites. The common models are
trained by the data in known sites (1, 3, and 7).

Model Loss Site 4 Site 5 Site 6 Avg.

MAE 5.21 6.95 7.31 6.49
Common
MSE 106.86 212.97 178.12 165.98
w/o cls.
RMSE 10.34 14.59 13.35 12.76
6.32
MAE 5.12 6.74 7.11
(3%)
Common 161.87
MSE 105.18 204.63 175.88
w cls. (3%)
12.61
RMSE 10.26 14.3 13.26
(1%)

As expected, a significant drop in the prediction accuracy for unknown sites was
observed compared to the baseline performance of the site-specific models. However, in
all the cases of the common model presented in Tables 6 and 7, the performance still met
the 8% MAE threshold required for participation in the Republic of Korea’s renewable
energy generation forecast system. Notably, implementing the proposed method resulted
in approximately 3–6% performance improvement for the basic common model. This
enhancement can be attributed to the classifier’s utilization of stored information from
the trained sites to assist in forecasting unknown sites. Even without direct historical
data for these unknown locations, the classifier leverages similarities with known sites to
make more accurate predictions, demonstrating the efficacy of incorporating site-specific
characteristics through classification to improve the forecasting accuracy across new and
diverse locations.
In addition to MAE, MSE and RMSE metrics provided further insights into the model’s
performance. The MSE values in Table 6 for the common model without a classifier
averaged 160.33, whereas the model with a classifier achieved an average MSE of 147.79,
indicating an 8% improvement. Similarly, in Table 7, the MSE values were 165.98 without
a classifier and 161.87 with a classifier, showing a reduction in error by approximately
Sustainability 2024, 16, 5240 12 of 15

3%. The reduction in MSE demonstrates the classifier’s effectiveness in minimizing larger
errors, which is crucial for maintaining robust forecasting performance across diverse sites.
The RMSE values also reflect the classifier’s impact on model performance. In Table 6,
the RMSE for the common model without a classifier averaged 12.51, while the model with
a classifier had an RMSE of 11.99, representing a 4% improvement. In Table 7, the RMSE
values were 12.76 without a classifier and 12.26 with a classifier, indicating a 1% reduction.
The impact of the classifier is also evidenced by the comparative forecast performance
depicted in Tables 6 and 7, where we observed a decrease in performance in Table 7 relative
to that in Table 6. Specifically, Tables 6 and 7 differ in the number of sites used to train
the common model, with four sites considered in Table 6 and only three sites considered
in Table 7. This reduction in the number of training cases directly affected the model’s
domain generalization capability. The fewer sites used in training, the less diverse the
data the model must learn from, which can limit its ability to generalize across new or
unknown domains accurately. This outcome highlights the critical role that the breadth of
the training data plays in enhancing the robustness of the model and its ability to adapt
to varied geographical and environmental conditions, underscoring the importance of
incorporating as much site-specific data as possible into the training phase.
Furthermore, we explored the potential of a transfer learning (TL) scenario to enhance
the adaptability of our forecasting model to specific sites. Transfer learning is particularly
valuable when limited weather information is available for unknown sites [21]. By applying
TL, we fine-tuned the network parameters, making the model more tailored and responsive
to the conditions of a specific site. In detail, the Feature Encoder was frozen, and the Regressor
was only retrained with the same hyperparameters as in the original experiment to preserve
the parameter features trained with large-scale data. This approach ensures that the general
features learned from the comprehensive dataset are retained while adapting the model to
the unique characteristics of each site. To evaluate the effectiveness of the TL scenario, we
posited a situation in which four months of weather data (equating to approximately 4%
of the total training data) were accessible for each site’s training. This setup allowed us to
investigate how even a small subset of site-specific data can significantly affect a model’s
performance by adjusting its parameters via TL. The incorporation of TL is designed
to bridge the knowledge gap between known and unknown sites, leveraging available
site-specific information to improve forecasting accuracy and model generalization across
various environments.
Tables 8 and 9 illustrate the outcomes of retraining the system using TL. Consistent
with expectations, applying TL generally enhanced the prediction accuracy across all
sites. Notably, site 6, which demonstrated strong performance in the site-specific model
but experienced a decline in accuracy within the common model framework, showed a
dramatic improvement because of site-specific fine-tuning. This significant enhancement
underscores the necessity of a synergistic approach that combines the broad applicability of
the common model with the tailored precision of site-specific models, depending on each
site’s unique characteristics.

Table 8. Forecast performance of the common model with TL for unknown sites. The common
models are trained by the data in known sites (1, 2, 3, and 7).

Model Loss Site 4 Site 5 Site 6 Avg.

MAE 3.71 4.97 1.91 3.53
Common + TL
MSE 59.4 111.51 16.32 62.41
w/o cls.
RMSE 7.71 10.56 4.04 7.45
3.39
MAE 3.39 4.94 1.83
(4%)
Common + TL 61.42
MSE 56.17 113.21 14.89
w cls. (2%)
7.33
RMSE 7.49 10.64 3.86
(2%)
Sustainability 2024, 16, 5240 13 of 15

Table 9. Performance comparison of the common model to TL for unknown sites. The common
models are trained by the data in known sites (1, 3, and 7).

Model Loss Site 4 Site 5 Site 6 Avg.

MAE 3.65 4.88 1.88 3.47
Common + TL
MSE 59.26 109 16.19 61.48
w/o cls.
RMSE 7.7 10.44 4.02 7.39
3.41
MAE 3.52 4.86 1.85
(2%)
Common + TL 61.39
MSE 57.73 110.32 16.13
w cls. (0.1%)
7.37
RMSE 7.6 10.5 4.02
(0.2%)

Moreover, across all outcomes, the proposed method yielded an additional 2–4%
MAE performance improvement over the basic common model. This finding highlights
the classifier module’s efficacy in refining the model’s capability to adapt and generalize
across various sites, further reinforcing the argument for integrating such a module into the
forecasting system. The role of the classifier in leveraging site-specific information, even
when only a small amount of data is available, is crucial for enhancing the overall accuracy
and adaptability of the model in a TL scenario.
However, when examining the MSE and RMSE metrics, we observed that the im-
provements are less pronounced, especially in Table 9. The MSE values in Table 8 show a
slight improvement of about 2% with the inclusion of the classifier, while Table 9 shows
a negligible improvement of 0.1%. Similarly, the RMSE values indicate minimal changes,
with only a 0.2% improvement observed in Table 9.
This discrepancy between MAE and MSE/RMSE improvements can be attributed
to the nature of these metrics. The MAE provides a linear measure of the average error
magnitude, while the MSE and RMSE emphasize larger errors due to the squaring of
differences. The less significant improvements in the MSE and RMSE suggest that while
the overall average prediction error (as indicated by the MAE) decreased, the variance in
error magnitude remained relatively unchanged. In other words, although the model with
the classifier reduced the average error, it did not significantly mitigate the impact of larger
prediction errors.
This outcome highlights the complexity of improving the forecasting accuracy across
all error metrics and underscores the need for further refinement of the model to specifically
address larger errors. Enhancing the model’s ability to consistently predict across all error
magnitudes will be essential for achieving more robust performance improvements in the
MSE and RMSE.

4. Conclusions
This study proposes an innovative DL-based approach for SPG forecasting that
demonstrates considerable potential across multiple sites. Our research successfully nav-
igated the complexities associated with SPG forecasting by leveraging a comprehensive
dataset from various locations throughout the Republic of Korea and employing DL tech-
niques to enhance forecasting accuracy by integrating common meteorological data and
site-specific features.
A key contribution of our study is developing a common model framework that incor-
porates a classification module, enabling the effective adaptation and prediction of both
known and unknown sites. This methodology significantly enhances the model’s adaptabil-
ity and generalization capabilities across diverse environmental conditions. The evaluation
results demonstrate that our model maintains high performance levels across different SPG
sites with minimal performance degradation compared to site-specific models.
By addressing these key areas and providing a robust evaluation, we have shown
that our developed solution is both effective and adaptable, making it a valuable tool for
Sustainability 2024, 16, 5240 14 of 15

improving SPG forecasting accuracy and supporting the integration of renewable energy
into the power grid.
Although the proposed system shows promising results for SPG forecasting, it still
has limitations. First, the system architecture was not fully optimized due to data collection
issues. We found that a larger system architecture is necessary to deal with the complex
behavior of the SPG system. Additionally, data augmentation schemes could be applied
to improve model robustness. Moreover, it is necessary to take into account more mean-
ingful information, such as the geographical characteristics of the site. Incorporating this
information could enrich the feature extraction stages and ultimately improve the overall
forecasting accuracy.
Future research should explore how the site set’s composition influences the common
model’s performance. This involves investigating the problem of identifying an optimal
set of information for configuring a common model to ensure that it can be effectively
generalized across multiple sites. In addition, the development of hybrid models that
combine the strengths of common and site-specific models is a promising avenue. Such
hybrid approaches can leverage the broad applicability of common models while incorpo-
rating the precision and adaptability of site-specific models, potentially leading to superior
forecasting accuracy and robustness under diverse environmental conditions.
Moreover, while our current model incorporates the solar elevation and azimuth angle
data to account for basic seasonal variations in solar irradiance and utilizes multi-year
weather data (2013 to 2020 for training and 2021 to 2022 for testing) to inherently consider
seasonal effects, future research could significantly benefit from a more explicit seasonal
approach. This would involve collecting and analyzing data across different seasons to
better understand and predict the impact of seasonal changes on SPG. Implementing such
seasonal approaches could enhance the accuracy and robustness of SPG forecasting models,
thereby improving their practical application in diverse climatic conditions.

Author Contributions: Conceptualization, S.Y.J., B.T.O. and E.O.; methodology, S.Y.J., B.T.O. and
E.O.; software, S.Y.J.; validation, S.Y.J., B.T.O. and E.O.; formal analysis, B.T.O. and E.O.; investigation,
S.Y.J., B.T.O. and E.O.; resources, S.Y.J., B.T.O. and E.O.; data curation, S.Y.J., B.T.O. and E.O.; writing—
original draft preparation, S.Y.J.; writing—review and editing, B.T.O. and E.O.; visualization, S.Y.J.,
B.T.O. and E.O.; supervision, B.T.O. and E.O.; project administration, E.O.; funding acquisition, E.O.
All authors have read and agreed to the published version of the manuscript.
Funding: This research was supported by the Korea Electric Power Corporation (Grant number:
R22XO02-23) and the Ministry of Science and ICT (MSIT), Republic of Korea, through the Information
Technology Research Center (ITRC) Support Program supervised by the Institute for Information &
Communications Technology Planning & Evaluation (IITP) under Grant RS-2023-00259004.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Publicly available datasets were analyzed in this study. These data can
be found at [Link] (accessed on 8 April 2024).
Conflicts of Interest: The authors declare no conflicts of interest.

References
1. Camera, F.L. Renewable Capacity Statistics 2023; International Renewable Energy Agency (IRENA): Masdar City, Abu Dhabi, 2023;
ISBN 978-92-9260-525-4.
2. Mandys, F.; Chitnis, M.; Silva, S.R.P. Levelized cost estimates of solar photovoltaic electricity in the United Kingdom until 2035.
Patterns 2023, 4, 100735. [CrossRef] [PubMed]
3. Massidda, L.; Bettio, F.; Marrocu, M. Probabilistic day-ahead prediction of PV generation. A comparative analysis of forecasting
methodologies and of the factors influencing accuracy. Sol. Energy 2024, 271, 112422. [CrossRef]
4. Rajasundrapandiyanleebanon, T.; Kumaresan, K.; Murugan, S.; Subathra, M.; Sivakumar, M. Solar energy forecasting using
machine learning and deep learning techniques. Arch. Comput. Methods Eng. 2023, 30, 3059–3079. [CrossRef]
5. Benti, N.E.; Chaka, M.D.; Semie, A.G. Forecasting renewable energy generation with machine learning and deep learning: Current
advances and future prospects. Sustainability 2023, 15, 7087. [CrossRef]
Sustainability 2024, 16, 5240 15 of 15

6. Gao, J.; Heng, F.; Yuan, Y.; Liu, Y. A novel machine learning method for multiaxial fatigue life prediction: Improved adaptive
neuro-fuzzy inference system. Int. J. Fatigue 2024, 178, 108007. [CrossRef]
7. Dhaked, D.K.; Dadhich, S.; Birla, D. Power output forecasting of solar photovoltaic plant using LSTM. Green Energy Intell. Transp.
2023, 2, 100113. [CrossRef]
8. Wang, L.; Mao, M.; Xie, J.; Liao, Z.; Zhang, H.; Li, H. Accurate solar PV power prediction interval method based on frequency-
domain decomposition and LSTM model. Energy 2023, 262, 125592. [CrossRef]
9. Li, G.; Wei, X.; Yang, H. Decomposition integration and error correction method for photovoltaic power forecasting. Measurement
2023, 208, 112462. [CrossRef]
10. Marinho, F.P.; Rocha, P.A.; Neto, A.R.; Bezerra, F.D. Short-term solar irradiance forecasting using CNN-1D, LSTM, and CNN-LSTM
deep neural networks: A case study with the Folsom (USA) dataset. J. Sol. Energy Eng. 2023, 145, 041002. [CrossRef]
11. Anu Shalini, T.; Sri Revathi, B. Hybrid power generation forecasting using CNN based BILSTM method for renewable energy
systems. Autom. Časopis Autom. Mjer. Elektron. Računarstvo Komun. 2023, 64, 127–144. [CrossRef]
12. Khan, N.; Ullah, F.U.M.; Haq, I.U.; Khan, S.U.; Lee, M.Y.; Baik, S.W. AB-net: A novel deep learning assisted framework for
renewable energy generation forecasting. Mathematics 2021, 9, 2456. [CrossRef]
13. Abou Houran, M.; Bukhari, S.M.S.; Zafar, M.H.; Mansoor, M.; Chen, W. COA-CNN-LSTM: Coati optimization algorithm-based
hybrid deep learning model for PV/wind power forecasting in smart grid applications. Appl. Energy 2023, 349, 121638. [CrossRef]
14. Alharkan, H.; Habib, S.; Islam, M. Solar power prediction using dual stream CNN-LSTM architecture. Sensors 2023, 23, 945.
[CrossRef] [PubMed]
15. Vaswani, A.; Shazeer, N.; Parmar, N.; Uszkoreit, J.; Jones, L.; Gomez, A.N.; Kaiser, Ł.; Polosukhin, I. Attention is all you need.
Adv. Neural Inf. Process. Syst. 2017, 30, 1–11.
16. Al-Ali, E.M.; Hajji, Y.; Said, Y.; Hleili, M.; Alanzi, A.M.; Laatar, A.H.; Atri, M. Solar energy production forecasting based on a
hybrid CNN-LSTM-transformer model. Mathematics 2023, 11, 676. [CrossRef]
17. Zhu, J.; Zhao, Z.; Zheng, X.; An, Z.; Guo, Q.; Li, Z.; Sun, J.; Guo, Y. Time-Series Power Forecasting for Wind and Solar Energy
Based on the SL-Transformer. Energies 2023, 16, 7610. [CrossRef]
18. Public Data Portal. Ministry of the Interior and Safety, South Korea. 2024. Available online: [Link]
(accessed on 8 April 2024).
19. Open MET Data Portal. Korea Meteorological Administration, South Korea. 2024. Available online: [Link]
resources/html/en/[Link] (accessed on 8 April 2024).
20. Yu, H.-U.; Kim, S.; Wi, Y.-M.; Lee, J. An Offer Method for Photovoltaic Power Plants with ESSs Considering Incentives for
Forecasting Accuracy of Renewable Generation. Trans. Korean Inst. Electr. Eng. 2022, 71, 1076–1083. [CrossRef]
21. Miraftabzadeh, S.M.; Colombo, C.G.; Longo, M.; Foiadelli, F. A day-ahead photovoltaic power prediction via transfer learning
and deep neural networks. Forecasting 2023, 5, 213–228. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.