Attouri et al.

Journal of Engineering
Journal of Engineering and Applied Science (2023) 70:56
https://doi.org/10.1186/s44147-023-00227-3 and Applied Science

RESEARCH Open Access

Enhanced fault diagnosis of wind energy

conversion systems using ensemble learning
based on sine cosine algorithm
Khadija Attouri1, Khaled Dhibi2, Majdi Mansouri3* , Mansour Hajji1, Kais Bouzrara4 and Hazem Nounou3

*Correspondence:
[email protected] Abstract
1
Research Unit Advanced This paper investigates the problem of incipient fault detection and diagnosis (FDD) in
Materials and Nanotechnologies wind energy conversion systems (WECS) using an innovative and effective approach
(UR16ES03), Higher Institute called the ensemble learning-sine cosine optimization algorithm (EL-SCOA). The
of Applied Sciences
and Technology of Kasserine, evolved strategy involves two primary steps: first, a sine-cosine algorithm is used to
Kairouan University, extract and optimize features in order to only select the most descriptive ones. Second,
1200 Kasserine, Tunisia to further improve the capability, thereby providing the highest accuracy performance,
2
Chemical Engineering Program,
Texas A&M University at Qatar, the newly gathered dataset is introduced as input to an ensemble learning paradigm,
23874, Doha, Qatar which merges the benefits of boosting and bagging techniques with an artificial neu-
3
Electrical and Computer ral network classifier. The essential goal of the developed proposal is to discriminate
Engineering Program, Texas A&M
University at Qatar, 23874, Doha, between the diverse operating conditions (one healthy and six faulty conditions). Three
Qatar potential and frequent types of faults that can affect the system behaviors including
4
Laboratory of Automatic Signal short-circuit, open-circuit, and wear-out are considered and thereby injected at diverse
and Image Processing, National
Engineering School of Monastir, locations and sides (grid and generator sides) in order to evaluate the availability and
University of Monastir, performance of the proposed technique when compared to the conventional FDD
5019 Monastir, Tunisia methods. The diagnosis performance is analyzed in terms of accuracy, recall, precision,
and computation time. The acquired outcomes demonstrate the efficiency of the sug-
gested diagnostic paradigm compared to conventional FDD techniques (accuracy rate
has been successfully achieved 98.35%).
Keywords: Wind energy conversion (WEC) systems, Fault diagnosis, Machine learning
(ML), Sine-cosine optimization algorithm (SCOA), Feature selection, Feature extraction

Introduction
Year-on-year for the last few decades, wind energy has become one of the most prom-
ising, inexhaustible, green, clean, non-polluting, and sustainable energy sources.
From 2011 until 2020, its production capacity expanded from 220 GW to 733 GW
[1]. Unfortunately, this kind of energy is affected by several and various failures due
to its complexity, which leads to the loss of its efficiency and reliability. Generators [2,
3], gearboxes [4, 5], power converters [6, 7], and blades [8, 9] are typically the most
common faults. In this regard, several methods and techniques are investigated in

This is a U.S. Government work and not under copyright protection in the US; foreign copyright protection may apply 2023. Open Access
This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution
and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in
the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s
Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need
to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/.
The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made
available in this article, unless otherwise stated in a credit line to the data.
Attouri et al. Journal of Engineering and Applied Science (2023) 70:56 Page 2 of 16

the literature in order to ensure the safety, integrity, and performance of the opera-
tion of such systems [6, 10]. The study in [11] proposed two hybrid numerical weather
prediction models and an artificial neural network model for wind power forecasting
over extremely complicated terrain, the first model created predicts the energy output
of each wind turbine directly, while the second model forecasts first the wind speed
before converting it to power using a fitted power curve. By using an artificial neu-
ral network (ANN)-based distribution static compensator, the authors of [12] empha-
sized a new control strategy to enhance the power quality (DSTATCOM) in WECS.
Mansouri et al. [13] employed a detection and diagnosis strategy for diverse incipient
faults of the WECS under various states. In [14], the authors disposed of an advanced
fault detection and diagnosis (FDD) approach for wind energy conversion (WEC) sys-
tems based on reduced-gaussian process regression-based random forest (RGPR-RF).
Regarding the scientific community has been closely monitoring ensemble learning
(EL) approaches, which mix several and numerous machine learning models to cre-
ate the most optimal and best possible predictive model. The success of the ensemble
model can be attributed to a variety of factors, including statistical, computational,
and representation learning [15], bias-variance decomposition [16], and strength-cor-
relation [17]. There are numerous surveys in the literature that mostly concentrate on
the review of ensemble learning, such as the learning of ensemble models in classifi-
cation problems [18–21]; regression problems [22, 23]; and clustering problems [24].
Indeed, an effective neural network-based ensemble technique was employed in [25].
The authors of this paper used bagging, boosting, and random subspace combination
approaches together with an ensemble classifier constructed using neural network
techniques. The work in [26] employed the benefits of the support vector machine,
K-nearest neighbor, and the decision tree in an improved ensemble learning (EL)-
based intelligent fault diagnosis paradigm that aims to guarantee the high efficiency
of grid-connected photovoltaic (GCPV) systems. The initial step in data mining is
known as preprocessing [27], and it entails cleaning and arranging the dataset to suit
the requirements of the input for the subsequent stages.
Accordingly, one potential pre-processing step is feature selection (FS), which is a
method for keeping a subset of features from a dataset that can accurately represent the
data without outliers or redundancies [27]. In fact, several and numerous applications,
such as data classification [28–30], data clustering [31–33], image processing [34–36],
and text categorization [37, 38], deployed and utilized the FS technique. To examine the
FS issue, various and several distinct versions of SCOA have been emphasized [39–46].
Besides, various optimization algorithms, for instance, the genetic algorithm (GA) [47,
48], the backtracking search algorithm (BSA) [49], the coral reef optimization (CRO)
[50], the particle swarm optimization (PSO) [51], and the fruit fly optimization algo-
rithm (FOA) [52], are introduced to keep and depict the appropriate parameters for arti-
ficial intelligent (AI) methodologies.
This work proposes an improved and effective ensemble learning approach for
fault detection and diagnosis in wind conversion systems. The contribution of
this paper is threefold: firstly, pre-processing data is obtained. Secondly, a sine-
cosine optimization algorithm is performed in order to avoid redundant features
and select and extract only the more relevant observations from the entire set of
Attouri et al. Journal of Engineering and Applied Science (2023) 70:56 Page 3 of 16

features. Finally, the significant obtained features are fed to an ensemble learn-
ing algorithm to improve the classification performance and enhance the WECS
model’s reliability and ability to distinguish between the diverse operating modes.
In this work, therefore, we inserted frequent, potential, and diverse types of fail-
ures: wear-out faults, open-circuit faults, and short-circuit faults, at different sides
and locations (grid and generator sides) in order to examine the reliability of the
developed strategy compared to the state-of-the-art methods, including artificial
neural network (ANN), K-nearest neighbor (KNN), cascade forward neural net-
work (CFNN), feed forward neural network (FFNN), generalized regression neural
network (GRNN), and support vector machine (SVM). The rest of this paper is
arranged as follows:
The suggested ensemble learning-based sine-cosine optimization algorithm strategy
is highlighted in “Methods” section, and the concepts of each employed technique are
described. The proposed technique will be tested on wind energy conversion systems
in “Results and discussion” section, and the maintained results are analyzed and sum-
marized. “Conclusion” section of this paper offers a conclusion.

Methods
EL‑SCOA approach
The evolved strategy involves three major steps, including data processing and
treatment, feature optimization and selection, and fault detection and diagnosis
(FDD). The main goal of the suggested technique, called the ensemble learning-
based sine-cosine optimization algorithm (EL-SCOA), is to improve the fault diag-
nosis capabilities and efficiency of WECS. Unlike conventional diagnosis methods,
which apply the raw data directly, the established proposal extracts and selects the
best descriptive and intensive features from the original dataset and feeds them
as inputs to the classifier for diagnosis purposes. The classifier uses bagging and
boosting algorithms as ensemble techniques and ANN as a baseline classifier in
order to identify, classify, and discriminate between the various states that may
occur in the WECS.
The block diagram that illustrates the important steps of the evolved approach for
FDD purposes is shown in Fig. 1.

Fig. 1 The steps of the proposed approach for fault diagnosis

Attouri et al. Journal of Engineering and Applied Science (2023) 70:56 Page 4 of 16

Algorithm 1. EL-SCO AlgorithmThe EL-SCOA is divided into two major categories: the training set and the testing
set. The detailed descriptions are explained in Algorithm 1

Concept theoretical framework of artificial neural network (ANN)

Artificial NNs are computational models inspired by the networks of the human bio-
logical brain. These networks have gotten big attention until now [13, 48, 53]. The ANN
utilizes a network pattern to generate decisions. Indeed, the input, hidden, and output
layers are the three levels that make up the ANN structure. Each layer is made up of sets
of nodes. The output layer provides the network response after the information has been
processed by the hidden layer and received by the input layer. The number of inputs cor-
responds to the number of neurons in the input layer, similar to this, the number of out-
put layer neurons corresponds to the number of ANN outputs, for instance in our work
the ANN classifier is trained using WECS measurement variables ( x1 ...xm , m = 12) as
inputs, and ( N = 7) labels as their corresponding desired outputs as depicted in Fig. 2.
Nevertheless, the number of neurons in the hidden layer is determined experimentally.
It consists of various experiments by varying the number of neurons in the hidden layer
(10 hidden layers are employed in this study). In contrast to a complex ANN structure, a
simple ANN architecture provides accurate predictions. Moreover, a signal of weight wij
interconnects every two neurons of successive layers. Each neuron transfers the infor-
mation to the neurons in the next layer after processing it through an activation function
Attouri et al. Journal of Engineering and Applied Science (2023) 70:56 Page 5 of 16

Fig. 2 Structure ANN with m WECS inputs and their N labels’ corresponding outputs

( f ). The most frequently employed function is the sigmoid activation function since it
is a nonlinear function that can be differentiated [54]. This function, a logistic function
with a range of 0 to 1, has the following formula:

1
f = (1)
1 + e−x

The weight, signal weight adjustment, prediction error, and the output of neural net-
work equations are expressed in [55].

Ensemble learning theory

The ensemble learning methodology incorporates and combines various and several
individual models in order to generate one optimal predictive model, thereby upgrading
the performance and the classification results of the FDD techniques. Generally, boost-
ing and bagging strategies are the most well-known and used in the literature.

Boosting strategy
In ensemble models, the boosting methodology, often known as a sequential ensemble
[56], is used in ensemble models to improve the generalization of learning models that
have weak generalizations [57]. Boosting is an ensemble technique where the predictors
are created sequentially instead of independently. Indeed, boosting is based on the idea
that subsequent predictors should learn from their previous errors and, accordingly, the
obtained predictions become more accurate.

Bagging strategy
One of the most common techniques for generating ensemble-based algorithms is
bagging [58], also named bootstrap aggregating. A bagging technique is deployed to
enhance the performance of an ensemble classifier. Additionally, the intensive objective
of this technique is to generate a series of independent observations with the same size
and distribution as the raw dataset. Create a series of samples and generate an ensemble
predictor that is more precise than the single predictor generated on the raw dataset. In
fact, bagging concerns two tasks: the first is the creation of bagged observations and the
transfer of each bag of observations to the base models, and the second is a technique
for merging the predictions of the various predictors. Incorporating the output of the
Attouri et al. Journal of Engineering and Applied Science (2023) 70:56 Page 6 of 16

base predictors may differ because majority voting is utilized for classification issues and
averaging is used for regression issues in order to create the ensemble output.

Concept of the sine‑cosine optimization algorithm (SCOA)

The SCOA is a swarm-based optimization methodology that was first suggested by
Mirjalili in 2016 [57, 58]. It is based on periodic behaviors that use the sine and cosine
functions and is motivated by the transcendental function theory. Similarly to other
optimization techniques, SCOA performs optimization through the use of mathemati-
cal rules. It is probable that a number of initial randomnesses possible solutions varied
either away from or toward the final position (optimal solution). Some dynamic and ran-
domized parameters highlight search exploration and exploitation via diverse optimi-
zation milestones [59]. The two phases position update equation used by the SCOA is
demonstrated as follows:

pijk + r1 × sin(r2) × r3 × Pjk − pijk , r4 < 0.5

pijk+1 = (2)
pijk + r1 × cos(r2 ) × r3 × Pjk − pijk , r4 ≥ 0.5

Where p denotes the position of the ith individual in the jth dimension at the (k+1)th
iteration. P depicts the global best position in jth dimension at kth iteration. The param-
eter r1 decreases linearly with the iterative process, which is utilized to ensure the bal-
ance between exploration and exploitation. The parameter r1 is depicted as
σ
r1 = σ − k (3)
k

Where k indicates the maximum number of iterations, and σ is a constant number.

r2 is the random number uniformly distributed in [0, 2π ], r3 is the random number uni-
formly distributed in the range [0, 2], and r4 is a random number in [0, 1], which is used
to switch with an equal probability between sine and cosine trigonometric functions.
When r3 > 1, the exchange of information between Pjk and pijk increases; although when
r3 < 1 the influence between Pjk and pijk is reduced. Figure 3 depicts the SCOA search
mode diagram.

System description
In this research, a variable-speed wind turbine based on a squirrel cage induction gen-
erator (SCIG) is considered, as displayed in Fig. 4.
The squirrel cage induction machine (SCIG), which will be monitored and controlled
by the stator-side AC/DC converter, and the grid-side DC/AC converter sub-system
are the two major categories of the employed system’s model. This structure permits
an infinitely variable speed operation. Additionally, regardless of the machine’s rotation
speed, the required voltage is converted into direct current and voltage. Furthermore,
for this structure, the generator grid side is based on an Insulated Gate Bipolar Transis-
tor (IGBT), where its configuration is the same as that of the converter grid side. Table 1
illustrates the diverse properties of wind turbines.
Attouri et al. Journal of Engineering and Applied Science (2023) 70:56 Page 7 of 16

Fig. 3 Diagram of the SCOA search mode

Fig. 4 Variable speed wind turbine based on SCIG and converter topology

Table 1 Properties of wind turbine

Parameters Nomenclature Values

Nominal power of turbine Ptn 15 kW

Moment of inertia of turbine Jt 1000 kgm2
Stator resistance Rs 0.087 Ohm
Stator leakage inductance Is 0.8 mH
Rotor resistance Rr 0.228 Ohm
Rotor leakage inductance Ir 0.8 mH
Magnetizing inductance Lm 34.7 mH
Number of poles P 4
Moment of inertia of generator Jg 0.2 kgm2

Grid converter and generator converter are the two levels of the power conversion topol-
ogy used in the wind chain. Each converter has a total of three arms. Each arm is made up
of high and low IGBTs, as shown in Fig. 5.
Attouri et al. Journal of Engineering and Applied Science (2023) 70:56 Page 8 of 16

Fig. 5 Variable speed wind turbine based on SCIG and converter topology

Table 2 Description of the diverse labeled failures injected in the WEC system
Fault side Symbol Fault description

Generator side HIGBT1GeS SC1: A short circuit fault that impacts

only the first high IGBT (HIGBT) in
the first arm of the converter genera-
tor side.
OC1: An open circuit fault that
impacts the first high IGBT in the first
arm of the converter generator side.
WO1: A wear-out fault that affects
the first high IGBT in the first arm of
the converter generator side.
Grid side HIGBT1GrS SC2: A short circuit fault that affects
only the first high IGBT in the first
arm of the converter grid side.
OC2: An open circuit fault that
affects only the first high IGBT in the
first arm of the converter grid side.
WO2: A wear-out fault that affects
the first high IGBT in the first arm of
the converter grid side.

Results and discussion

Data collection
In this study, we utilize data obtained from a healthy WTCS, which are then injected
with several faulty scenarios: open-circuit (OC), short-circuit (SC), and wear-out. In
other words, we initially considered how the system behaves in a healthy condition,
and then we independently injected each faulty scenario, considering how each failure
impacts and affects the system’s behavior in that situation. In fact, we do not take into
account the transitional regime that appears when we switch from a healthy state to a
faulty condition. These faults are thoroughly described in Table 2. The internal resist-
ance of two ohms is used to indicate the final fault (WO fault). Accordingly, each mode
behavior is adequately described over 2000 10-time-lagged samples within a second-
time duration with 20 KHz as the sampling frequency.
Seven operating modes of WECS are used in this study, including one healthy case
(designated as C1) and six different faulty states (C2… C7). Twelve measurement vari-
ables, which are used to represent these seven scenarios, are shown in further detail in
Table 3.
Attouri et al. Journal of Engineering and Applied Science (2023) 70:56 Page 9 of 16

Table 3 Labeling, description and ranges of the measured and monitored system variables
Variables Descriptions Ranges

x1 Cm: Mechanical torque (Nm) [−346 , 731.12]

x2 Ng: Generator speed (tr/m) [4.01 , 966.91]
[−423.23 , 110.58]
x3 isag: Generator current phase a (A) [−285.44 , 445.71]
x4 isbg: Generator current phase b (A) [−401.49 , 362.89]
x5 isd: Generator current along d-axis (A) [−466.33 , 318.59]
[−0.0034 , 0.00126]
x6 isq: Generator current along q-axis (A) [−0.00127 , 0.000197]
x7 VDC: Bus voltage (V ) [−564.29 , 349.93]
x8 Pout: Output power (W ) [−404.01 , 407.88]
[−33.3 , 568.85]
x9 isar: Grid current phase a (A) [−239.54 , 347.24]
x
10 isbr: Grid current phase b (A)
x
11 isd: Grid current along d-axis (A)
x
12 isq: Grid current along q-axis(A)

Table 4 Construction of database for fault detection and diagnosis system

Label State Training data Testing data

C1 Healthy 2000 2000

C2 HIGBT1GeS 2000 2000
C3 2000 2000
C4 2000 2000
C5 HIGBT1GrS 2000 2000
C6 2000 2000
C7 2000 2000

The generator variables isd , isq , and the grid variables isd , isq , isar , isbr can be calculated
and obtained using the Park transformation, with θ (rad),
 � �  
� �
2 cosθ cos θ − 2π
cos(θ + 2π
) isa
isd 3 � 3 
isq
=  � isb  (4)
3 sinθ sin θ − 2π sin(θ + 2π ) isc
3 3

Table 4 depicts the distinct operating scenarios. In both the training and testing
phases, we used 50% of the observations.
Certain electrical and mechanical variables under various faulty situations are dis-
played in the following figures.

Evaluation metrics
Different metrics, often known as performance metrics or evaluation metrics are used to
fully assess the effectiveness or quality of the model. These performance metrics enable
us to evaluate how well our model performed the supplied data. In this manner, we can
improve the model’s performance by tuning the hyperparameters. The approved criteria
are accuracy (%), which denotes the rate of samples that are correctly predicted over
the total number observations. Recall (%) which denotes, in the pertinent class, the rate
of positive samples correctly predicted to the observations. Precision (%) denotes the
Attouri et al. Journal of Engineering and Applied Science (2023) 70:56 Page 10 of 16

Fig. 6 Mechanical torque for different operating modes

Fig. 7 Generator speed for different operating modes

number of positive samples correctly predicted divided by the number of total predicted
positive observations. Computation time (CT(s)) represents the time required to carry
out the algorithm Figs. 6, 7, 8, 9 and 10.

TP + TN
Accuracy = (5)
TP + TN + FP + FN

TP
Recall = (6)
TP + FN
Attouri et al. Journal of Engineering and Applied Science (2023) 70:56 Page 11 of 16

Fig. 8 Generator current for different operating modes

Fig. 9 Bus voltage for different operating modes

TP
Precision = (7)
TP + FP

Where TP (true positive) is the number of observations that are correctly iden-
tified, TN (true negative) represents the number of observations that are correctly
dismissed, FP (false positive) is the number of observations that are incorrectly dis-
missed and FN (false negative) is the number of observations that are incorrectly
identified.
Attouri et al. Journal of Engineering and Applied Science (2023) 70:56 Page 12 of 16

Fig. 10 Grid current for different operating modes

Table 5 Performance evaluations of various classification methods

Methods Global performance
Accuracy Recall Precision CT(s)

EL-SCOA 98.35 98.35 98.37 12.00

EL 98.88 98.88 98.88 23.74
ANN 93.70 93.70 93.73 3.53
KNN 88.30 88.31 88.31 0.91
CFNN 97.17 97.17 97.16 8.36
FFNN 97.17 97.17 97.8 8.14
GRNN 97.06 97.06 97.06 7.06
SVM 92.14 92.14 92.16 15.69

Discussion
In order to demonstrate, and show the effectiveness of the presented approach in terms
of diagnostic recall, precision, accuracy, and computation time, a number of methods,
including CFNN, FFNN, ANN, GRNN, KNN, and SVM, have been employed. The
diverse existing methods are modeled and tested in a MATLAB toolbox. To evaluate
the overall effectiveness of the provided strategies, the accuracy was calculated using a
10-fold cross-validation metric. For the FFNN, CFNN, GRNN, and ANN, 10 hidden lay-
ers with a total of 50 hidden neurons in each layer were chosen. In order to introduce
non-linearity, a sigmoid function is used in the hidden layers. The K and C parameters
for SVM are set with the lowest RMSE value, and the K value for KNN is equal to 3. The
maximum number of iterations for the SCOA is 100, and the number of solutions that
are chosen is 10.
The comparison analysis in Table 5 showed that the proposed strategy EL, which com-
bines the bagging and boosting strategies with the ANN classifier, performed much
better than the ANN and the other methodologies in terms of accuracy (98.88%) and
Attouri et al. Journal of Engineering and Applied Science (2023) 70:56 Page 13 of 16

Table 6 The selected features and the performance evaluations of the evolved classification
strategies
Global performance
Methods Number of Selected features Accuracy Recall Precision CT(s)
features

EL 12 1, ...12 98.88 98.88 98.88 23.74

EL-SCOA 6 2, 3, 4, 7, 8,10 98.35 98.35 98.37 12.00

Table 7 Confusion matrix for the EL-based SCOA in the testing phase
Predicted classes
True classes C1 C2 C3 C4 C5 C6 C7 Recall

C1 1858 0 0 0 142 0 0 92.90

C2 0 2000 0 0 0 0 0 100
C3 0 0 20 0 0 0 0 100
C4 0 0 0 1991 900 0 0 99.55
C5 75 0 0 100 1924 0 0 96.20
C6 4 0 0 0 0 1996 0 99.80
C7 0 0 0 0 0 0 2000 100
Precision 95.92 100 100 99.95 92.72 100 100 98.35

outperformed the other models. In spite of the fact that the suggested EL approach per-
forms better and produces good results in terms of classification accuracy compared to
conventional techniques, it still suffers from a difficult training phase and a high time
complexity. To deal with this drawback, we actually employed a sine-cosine optimization
algorithm (SCOA) in order to pick and select the best descriptive features and reduce
the computation time, which represents a significant challenge in the fault diagnosis
domain, as well as accelerate the learning and the classification tasks. As a result, the
computation time is significantly decreased from 23.74 s to 12.00 s, with only a minor
difference in its accuracy (by 0.53%). The inefficient KNN and SVM classification out-
comes are attributable to the direct usage of raw data, demonstrating the success of the
suggested approach that selects the more significant features before performing the clas-
sification task. Six features (out of 12) of the developed EL-SCOA strategy were best
selected, as shown in the following table (Table 6).
Table 7 illustrates the obtained testing classification outcomes of the diverse classes
by the use of a confusion matrix (CM) to further demonstrate the effectiveness of the
evolved methodology. Indeed, the samples that were successfully labeled to the healthy
condition (C1) and the various faulty operating states (C2 to C7) as well as the samples
that were incorrectly labeled, are both displayed in this matrix. Specifically, the X and Y
axes highlight the true classes and the projected conditions, respectively.
Table 7 demonstrates that the EL-based SCOA strategy correctly identifies the 2000
observations from the 2000 true positives for the conditions operating modes (C2, C3,
and C7), indicating that these modes are correctly classified and there was no misclas-
sification. However, there is a misclassification for the healthy state (C1), faulty modes
3 (C4), 4 (C5), and 5 (C6), as evidenced by the classification of 142 observations from the
Attouri et al. Journal of Engineering and Applied Science (2023) 70:56 Page 14 of 16

healthy class as the class (C5), 9 observations from the class (C4) as the class (C5), and 4
samples from the class (C6) as the class (C1).

Conclusions
This paper developed an enhanced fault detection and diagnosis approach called an
ensemble learning-based sine-cosine optimization algorithm (EL-SCOA) for wind
energy conversion (WEC) systems. The presented methodology was addressed so that
the sine-cosine algorithm is proposed in order to optimize, select, and extract the most
informative features from the raw data, where the maintained selected features are fed
to the classification technique for diagnosis purposes. The classification method incor-
porates bagging and boosting as ensemble methods and an ANN as a baseline classi-
fier. The proposed paradigm attempted to discriminate between various operating states
(short circuit, open circuit, and wear-out faults) introduced at various locations and
sides (generator and grid sides). As compared to other existing methods including ANN,
KNN, CFNN, FFNN, GRNN, and SVM, the experimental outcomes show that the sug-
gested strategy performs very well. As a result, the effectiveness of the suggested tech-
nique inspires us to further examine its computation time and memory storage in future
research. In order to simultaneously improve diagnosis accuracy and decrease WEC sys-
tem execution time, a strategy that combines data size reductions and the aforemen-
tioned technique will be proposed.

Abbreviations
SCA Sine cosine algorithm
SCOA Sine cosine optimization algorithm
EL Ensemble learning
WECS Wind energy conversion system
WT Wind turbine
FDD Fault detection and diagnosis
FE Feature extraction
FS Feature selection
ANN Artificial neural network
NN Neural network
RNN Recurrent NN
FFNN Feed-forward NN
CFNN Cascade forward NN
GRNN Generalized regression NN
RF Random forest
KNN K-nearest neighbors
SC Short circuit
OC Open circuit
WO Wear out
SVM Support vector machine
CT Computation time
CM Confusion matrix

Acknowledgements
The publication is the result of the Qatar National Research Fund (QNRF) research grant.

Authors’ contributions
KA wrote the original draft and worked on the software. KD worked on the software. MM and MH defined the methodol-
ogy, reviewed, and edited the manuscript. Kais KB and HN supervised the work. All authors read and approved the final
manuscript.

Funding
Funding provided by the Qatar National Library.

Availability of data and materials

Data will be made available on request.
Attouri et al. Journal of Engineering and Applied Science (2023) 70:56 Page 15 of 16

Declarations
Competing interests
The authors declare that they have no competing interests.

Received: 20 March 2023 Accepted: 17 May 2023

References
1. Murgas B, Henao A, Guzman L (2021) Evaluation of investments in wind energy projects, under uncertainty. state of
the art review. Appl Sci 11(21):10213
2. Singh G, Sundaram K (2022) Methods to improve wind turbine generator bearing temperature imbalance for
onshore wind turbines. Wind Eng 46(1):150–159
3. Xu Y, Nascimento NMM, de Sousa PHF, Nogueira FG, Torrico BC, Han T, Jia C, Rebouças filho PP (2021) Multi-sensor
edge computing architecture for identification of failures short-circuits in wind turbine generators. Appl Soft Com-
put 101:107053
4. López-Uruñuela FJ, Fernandez-Diaz B, Pagano F, López-Ortega A, Pinedo B, Bayón R, Aguirrebeitia J (2021) Broad
review of “white etching crack” failure in wind turbine gearbox bearings: Main factors and experimental investiga-
tions. Int J Fatigue 145:106091
5. Wang L, Zhang Z, Long H, Xu J, Liu R (2016) Wind turbine gearbox failure identification with deep neural networks.
IEEE Transact Industrial Inform 13(3):1360–1368
6. Kouadri A, Hajji M, Harkat M-F, Abodayeh K, Mansouri M, Nounou H, Nounou M (2020) Hidden markov model
based principal component analysis for intelligent fault diagnosis of wind energy converter systems. Renew Energ
150:598–606
7. Xiao C, Liu Z, Zhang T, Zhang X (2021) Deep learning method for fault detection of wind turbine converter. Appl Sci
11(3):1280
8. Ravikumar K, Subbiah R, Ranganathan N, Bensingh J, Kader A, Nayak SK (2020) A review on fatigue damages in
the wind turbines: Challenges in determining and reducing fatigue failures in wind turbine blades. Wind Eng
44(4):434–451
9. Mishnaevsky L Jr (2022) Root causes and mechanisms of failure of wind turbine blades: Overview. Materials. 15:2959
10. Fezai R, Dhibi K, Mansouri M, Trabelsi M, Hajji M, Bouzrara K, Nounou H, Nounou M (2020) Effective random forest-
based fault detection and diagnosis for wind energy conversion systems. IEEE Sensors J 21(5):6914–6921
11. Donadio L, Fang J, Porté-Agel F (2021) Numerical weather prediction and artificial neural network coupling for wind
energy forecast. Energies 14(2):338
12. Irfan MM, Malaji S, Patsa C, Rangarajan SS, Hussain SS (2022) Control of dstatcom using ann-bp algorithm for the
grid connected wind energy system. Energies 15(19):6988
13. Mansouri M, Dhibi K, Nounou H, Nounou M (2022) An effective fault diagnosis technique for wind energy conver-
sion systems based on an improved particle swarm optimization. Sustainability 14(18):11195
14. Mansouri M, Fezai R, Trabelsi M, Nounou H, Nounou M, Bouzrara K (2021) Reduced gaussian process regression
based random forest approach for fault diagnosis of wind energy conversion systems. IET Renew Power Gener
15(15):3612–3621
15. Dietterich TG (2000) Ensemble methods in machine learning. Multiple Classifier Systems: First International Work-
shop, MCS 2000 Cagliari, Italy, June 21–23, 2000 Proceedings 1. Springer, Cagliari, Italy, pp 1–15
16. R. Kohavi, D. H. Wolpert, et al., (1996). Bias plus variance decomposition for zero-one loss functions, in: ICML, Vol. 96,
Citeseer, 275–83.
17. Breiman L (2001) Random forests. Machine Learn 45:5–32
18. Zhao J, Gao X, Yang (2005). A survey of neural network ensembles, in: 2005 international conference on neural
networks and brain, vol 1. IEEE, Beijing, pp 438–442
19. Rokach L (2010) Ensemble-based classifiers. Artif Intell Rev 33:1–39
20. Gopika D, Azhagusundari B (2014) An analysis on ensemble methods in classification tasks
21. Yang P, Hwayang Y, Zhou BB, Zomaya AY (2010) A review of ensemble methods in bioinformatics. Current Bioinfor-
matics. 5(4):296–308
22. Mendes-Moreira J, Soares C, Jorge AM, Sousa JFD (2012) Ensemble approaches for regression: a survey. ACM Com-
put Surveys (csur) 45(1):1–40
23. Ren Y, Suganthan P, Srikanth N (2015) Ensemble methods for wind and solar power forecasting—a state-of-the-art
review. Renew Sustain Energ Rev 50:82–91
24. Vega-Pons S, Ruiz-Shulcloper J (2011) A survey of clustering ensemble algorithms. Int J Pattern Recogn Artif Intell
25(03):337–372
25. Dhibi K, Mansouri M, Bouzrara K, Nounou H, Nounou M (2022) Reduced neural network based ensemble approach
for fault detection and diagnosis of wind energy converter systems. Renew Energ. 194:778–787
26. Dhibi K, Mansouri M, Bouzrara K, Nounou H, Nounou M (2021) An enhanced ensemble learning-based fault detec-
tion and diagnosis for grid-connected pv systems. IEEE Access 9:155622–155633
27. Guyon I, Elisseeff A (2003) An introduction to variable and feature selection. J Machine Learn Res. 3:1157–1182
28. Hua J, Tembe WD, Dougherty ER (2009) Performance of feature-selection methods in the classification of high-
dimension data. Pattern Recogn 42(3):409–424
29. Gómez-Verdejo V, Verleysen M, Fleury J (2009) Information-theoretic feature selection for functional data classifica-
tion. Neurocomputing 72(16–18):3580–3589
Attouri et al. Journal of Engineering and Applied Science (2023) 70:56 Page 16 of 16

30. R. Z. Al-Abdallah, A. S. Jaradat, I. A. Doush, Y. A. Jaradat. (2017) .A binary classifier based on firefly algorithm. Jordan J
Comput Inform Technol (JJCIT); 3(3):172-185
31. Liu H, Yu L (2005) Toward integrating feature selection algorithms for classification and clustering. IEEE Transact
Knowledge Data Eng 17(4):491–502
32. Boutemedjet S, Bouguila N, Ziou D (2008) A hybrid feature extraction selection approach for high-dimensional non-
gaussian data clustering. IEEE Transact Pattern Analysis Machine Intell 31(8):1429–1443
33. ElMustafa S, Jaradat A, Doush IA, Mansour N (2017) Community detection using intelligent water drops optimisa-
tion algorithm. Int J Reason-based Intell Syst 9(1):52–65
34. Huang K, Aviyente S (2008) Wavelet feature selection for image classification. IEEE Transact Image Process
17(9):1709–1720
35. Chen B, Chen L, Chen Y (2013) Efficient ant colony optimization for image feature selection. Sign Process
93(6):1566–1576
36. Sawalha R, Doush IA (2012) Face recognition using harmony search-based selected features. Int J Hybrid Inform
Technol 5(2):1–16
37. Shang W, Huang H, Zhu H, Lin Y, Qu Y, Wang Z (2007) A novel feature selection algorithm for text categorization. Exp
Syst Appl 33(1):1–5
38. Zheng Z, Wu X, Srihari R (2004) Feature selection for text categorization on imbalanced data. ACM Sigkdd Explor
Newslett 6(1):80–89
39. Neggaz N, Ewees AA, Abd Elaziz M, Mafarja M (2020) Boosting salp swarm algorithm by sine cosine algorithm and
disrupt operator for feature selection. Exp Syst Appl 145:113103
40. Sindhu R, Ngadiran R, Yacob YM, Zahri NAH, Hariharan M (2017) Sine–cosine algorithm for feature selection with
elitism strategy and new updating mechanism. Neural Comput Appl 28:2947–2958
41. Eid MM, El-kenawy E-SM, Ibrahim A (2021) A binary sine cosine modified whale optimization algorithm for feature
selection. 2021 National Computing Colleges Conference (NCCC). IEEE, Taif, Saudi Arabia, pp 1–6
42. Hussain K, Neggaz N, Zhu W, Houssein EH (2021) An efficient hybrid sine-cosine harris hawks optimization for low
and high-dimensional feature selection. Exp Syst Appl 176:114778
43. M. E. Abd Elaziz, A. A. Ewees, D. Oliva, P. Duan, S. Xiong. (2017). A hybrid method of sine cosine algorithm and differ-
ential evolution for feature selection, in: Neural Information Processing: 24th International Conference, ICONIP 2017,
Guangzhou, China, November 14–18, 2017, Proceedings, Part V 24, Springer:145–155.
44. Abualigah L, Dulaimi AJ (2021) A novel feature selection method for data mining tasks using hybrid sine cosine
algorithm and genetic algorithm. Cluster Comput 24:2161–2176
45. Abd Elaziz M, Oliva D, Xiong S (2017) An improved opposition-based sine cosine algorithm for global optimization.
Expert Syst Appl 90:484–500
46. R. Sindhu, R. Ngadiran, Y. M. Yacob, N. A. Hanin Zahri, M. Hariharan, K. Polat. (2019). A hybrid sca inspired bbo for
feature selection problems. Math Problems Eng; 2019.
47. Dhunny A, Timmons D, Allam Z, Lollchund M, Cunden T (2020) An economic assessment of near-shore wind farm
development using a weather research forecast-based genetic algorithm model. Energy 201:117541
48. Hichri A, Hajji M, Mansouri M, Abodayeh K, Bouzrara K, Nounou H, Nounou M (2022) Genetic-algorithm-based
neural network for fault detection and diagnosis: Application to grid-connected photovoltaic systems. Sustainability
14(17):10518
49. Kartite J, Cherkaoui M (2017) Improved backtracking search algorithm for renewable energy system. Energy Proce-
dia 141:126–130
50. Salcedo-Sanz S, Gallo-Marazuela D, Pastor-Sánchez A, CarroCalvo L, Portilla-Figueras A, Prieto L (2014) Offshore wind
farm design with the coral reefs optimization algorithm. Renew Energy. 63:109–115
51. He Z, Chen Y, Shang Z, Li C, Li L, Xu M (2019) A novel wind speed forecasting model based on moving window and
multi-objective particle swarm optimization algorithm. Appl Math Model 76:717–740
52. Zhang Q, Qian H, Chen Y, Lei D (2020) A short-term traffic forecasting model based on echo state network opti-
mized by improved fruit fly optimization algorithm. Neurocomputing 416:117–124
53. Hajji M, Yahyaoui Z, Mansouri M, Nounou H, Nounou M (2023) Fault detection and diagnosis in grid-connected pv
systems under irradiance variations. Energ Rep 9:4005–4017
54. Hippert HS, Pedreira CE, Souza RC (2001) Neural networks for short term load forecasting: a review and evaluation.
IEEE Transact Power Syst 16(1):44–55
55. Jamii J, Mansouri M, Trabelsi M, Mimouni MF, Shatanawi W (2022) Effective artificial neural network-based wind
power generation and load demand forecasting for optimum energy management. Front Energy Res. 10:898413
56. Ganaie MA, Hu M, Malik A, Tanveer M, Suganthan P (2022) Ensemble deep learning: a review. Eng Appl Artificial
Intell 115:105151
57. Zhang W, Jiang J, Shao Y, Cui B (2020) Snapshot boosting: a fast ensemble framework for deep neural networks. Sci
China Inform Sci 63:1–12
58. Breiman L (1996) Bagging predictors. Machine Learn 24:123–140
59. Shang C, Zhou T-T, Liu S (2022) Optimization of complex engineering problems using modified sine cosine algo-
rithm. Sci Rep 12(1):20528

Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.