International Journal of Power Electronics and Drive System (IJPEDS)

Vol. 16, No. 2, June 2025, pp. 1348~1355

ISSN: 2088-8694, DOI: 10.11591/ijpeds.v16.i2.pp1348-1355  1348

Wind turbine defect detection using deep learning

Deepa Somasundaram1, M. Vanitha2, T. Sathish Kumar3, I. Arul Doss Adaikalam4,

P. Kavitha5, R. Kalaivani6
1
Department of Electrical and Electronics Engineering, Panimalar Engineering College, Chennai, India
2
Department of Electronics and Communication Engineering, Saveetha Engineering College, Chennai, India
3
Department of Electrical and Electronics Engineering, S. A. Engineering College, Chennai, India
4
Department of Electrical and Electronics Engineering, Easwari Engineering College, Chennai, India
5
Department of Electrical and Electronics Engineering, R. M. K. Engineering College, Chennai, India
6
Department of Electrical and Electronics Engineering, Rajalakshmi Engineering College, Chennai, India

Article Info ABSTRACT

Article history: Wind turbines play a critical role in the generation of renewable energy, but
their maintenance and inspection, especially in large-scale wind farms,
Received Oct 5, 2024 present significant challenges. Traditionally, wind turbines have been
Revised Apr 11, 2025 inspected manually, a process that is not only time-consuming but also
Accepted May 6, 2025 costly and risky. Unmanned aerial vehicles (UAVs) have emerged as an
efficient alternative, offering a safer and more economical means of
gathering inspection data. However, the challenge lies in the manual analysis
Keywords: of the collected data, which demands expertise and considerable time. This
paper proposes using object detection algorithms, specifically YOLOv8, to
Deep learning automate the detection of wind turbines and their defects, streamlining the
Defect detection inspection process. The model is trained on wind turbine images to identify
Mean average precision potential faults such as cracks and corrosion. This approach aims to increase
Wind energy the accuracy and efficiency of wind turbine maintenance, ensuring prompt
YOLOv8 defect detection and reducing both operational costs and downtime.
This is an open access article under the CC BY-SA license.

Corresponding Author:
Deepa Somasundaram
Department of Electrical and Electronics Engineering, Panimalar Engineering College
Chennai, Tamil Nadu, India
Email: [email protected]

1. INTRODUCTION
Wind energy is a rapidly growing sector in the renewable energy landscape, playing a critical role in
reducing carbon emissions and promoting sustainable energy generation [1]-[3]. Wind turbines are key assets
in this domain, but their complex structures and constant exposure to harsh environmental conditions make
them prone to damage. Regular inspection and maintenance of wind turbines are essential to ensure their
operational efficiency and prevent unexpected failures. Traditional inspection methods, such as manual
climbing or using ground-based equipment, pose several challenges, including safety risks, high costs, and
limited detection accuracy [4]-[7]. As such, innovative solutions are needed to improve the efficiency and
safety of wind turbine maintenance operations.
Unmanned aerial vehicles (UAVs) have emerged as an innovative tool for wind turbine inspection.
UAVs can fly close to turbine structures, capturing high-resolution images of blades, towers, and nacelles,
even in challenging locations that are difficult for humans to reach [8]-[10]. This technology allows for safer,
faster, and more comprehensive inspections compared to traditional methods. However, as noted
by Lei et al. [1], while UAVs can collect a significant amount of visual data, the manual review of this data is
still time-consuming and requires expert analysis to identify potential issues, such as cracks or corrosion on

Journal homepage: http://ijpeds.iaescore.com

Int J Pow Elec & Dri Syst ISSN: 2088-8694  1349

turbine blades. The need for automating this process has led to the exploration of advanced algorithms in
machine learning and computer vision [11]-[13].
Fault diagnosis using machine learning algorithms is gaining traction in wind energy research. For
instance, Qu et al. [2] developed a fault detection system based on fuzzy logic for turbines, demonstrating the
potential of AI-driven techniques in this field. Similarly, the study by Rezaei et al. [3] employed modal-based
damage identification to address nonlinearities in wind turbine blades. While these studies primarily focused
on the mechanical aspects of wind turbines, they highlight the growing role of AI in predictive maintenance
and fault detection [14]-[19]. Building on these concepts, integrating object detection algorithms into UAV-
based wind turbine inspection offers the potential to automate the identification of structural defects, thus
reducing human intervention.
Object detection models, such as the You Only Look Once (YOLO) series, have revolutionized real-
time object detection by providing fast and accurate results. YOLOv8, the latest iteration, combines speed
with improved detection accuracy, making it suitable for applications like wind turbine defect detection.
According to Sun et al. [4], identifying damage in turbine blades using advanced machine learning
techniques can significantly enhance maintenance processes by pinpointing issues early. Leveraging
YOLOv8 in this context could automate the detection of critical defects, such as cracks, corrosion, or blade
misalignments, based on the visual data captured by UAVs, thus improving the efficiency of turbine
inspections [20]-[23].
Recent advancements in AI, particularly in the fields of object detection and anomaly detection,
have shown promising results in wind turbine maintenance. Wang et al. [5] proposed a two-stage anomaly
detection model to enhance fault detection in wind turbines, illustrating the benefits of combining machine
learning with real-time monitoring systems. By training object detection algorithms like YOLOv8 on wind
turbine images, it becomes possible to automate the identification of common issues such as blade damage or
structural wear. This approach not only reduces the time and expertise required for manual inspections but
also increases the overall reliability of the maintenance process.
Several previous studies have explored the integration of drone technology and computer vision for
turbine inspection. For example, Foster et al. [10] demonstrated the potential of drone footage for detecting
surface damage on turbines, showing the effectiveness of UAVs in capturing relevant data. However, the
application of cutting-edge object detection models like YOLOv8 for this specific task is still in its early
stages, making it a promising area of research. By refining these algorithms for wind turbine inspection,
operators can achieve higher precision and better maintenance outcomes, leading to reduced downtime and
enhanced energy output [24]-[25].
In summary, UAV-based inspection, combined with advanced object detection algorithms like
YOLOv8, offers a powerful solution to the challenges associated with traditional wind turbine maintenance.
This paper seeks to explore the potential of YOLOv8 in automating the defect detection process for wind
turbines. By leveraging recent developments in AI and drone technology, this approach could significantly
reduce the time, cost, and safety risks involved in turbine inspection, while improving the reliability of
renewable energy generation.

2. METHODOLOGY
The object detection system for wind turbines was implemented using the YOLOv8 algorithm. The
dataset consisted of wind turbine images with various defects, including cracks and surface corrosion. The
images were annotated to define the bounding boxes around turbine components and labeled for defect types.
The dataset was divided into training, validation, and testing sets, with a focus on training YOLOv8 to detect
defects efficiently.
The YOLOv8 model was configured using pre-trained weights and fine-tuned on the wind turbine
dataset. The data preprocessing included resizing images to a standard size to ensure compatibility with the
YOLO architecture, while maintaining the aspect ratios of the turbine components. Data augmentation
techniques, such as random rotation, flipping, and scaling, were applied to increase the variability of the
training set and improve model robustness.
The training process utilized a batch size of 1 and a learning rate optimized through grid search. The
model was trained for five epochs with a GPU accelerator to speed up computations. YOLOv8's architecture,
which consists of convolutional layers and anchor-based detection, enables it to quickly identify objects in
real time. After training, the best-performing model was saved and evaluated using the validation dataset to
measure its performance metrics, including precision, recall, and mean average precision (mAP).
The evaluation metrics provided insights into the model's ability to accurately detect wind turbine
defects. A confusion matrix was generated to assess the false positives and false negatives in defect detection.
The final model was deployed for inference on the test dataset to evaluate its real-world applicability and
effectiveness in detecting turbine defects. The object detection system for wind turbine defect identification
Wind turbine defect detection using deep learning (Deepa Somasundaram)
1350  ISSN: 2088-8694

using YOLOv8 involves several key steps, each enhanced by mathematical formulations and data-driven
optimization techniques. Figure 1 shows the flowchart of proposed method.

Figure 1. The flowchart of proposed method

- Dataset collection: The dataset consists of high-resolution wind turbine images. Let I represent the image
set, where each image. Let I represent the image set, where each image I i. I is an array of pixel values.
The images contain defects such as cracks and corrosion, represented by bounding boxes
𝐵𝑖 = {(𝑥1, 𝑦1), (𝑥2, 𝑦2)} which are manually annotated.
- Image annotation: The images are annotated with bounding boxes Bi, and defect types are labeled as Li.
The labels are assigned to each bounding box such that 𝐷 = {(𝐵𝑖, 𝐿𝑖)} where D is the dataset of defect-
labeled bounding boxes.
- Data preprocessing: Each image Ii undergoes preprocessing. The resizing of images is crucial to ensure
compatibility with YOLOv8’s input size, represented by the transformation function:

𝐼𝑖 = 𝑓𝑟𝑒𝑠𝑖𝑧𝑒(𝐼𝑖, ℎ, 𝑤) (1)

where h and w are the height and width of the resized image. The aspect ratio is preserved during this
process. Preprocessing also includes normalization to scale pixel values between 0 and 1.
- Data augmentation: Data augmentation is applied to increase the variability of the training set.
Augmentation techniques such as random rotations 𝑅(θ)) and scaling (S(Sx, Sy), S are used, where:

𝐼𝑖 ′′ = 𝑆(𝑆𝑥, 𝑆𝑦) ⋅ 𝑅(𝜃) ⋅ 𝐼𝑖′ (2)

this enhances model robustness by creating diverse training examples from the original images.
- Model training: The YOLOv8 model is trained using a loss function L i, which combines classification
loss Lclass, bounding box regression loss Lbbox, and object confidence loss Lconf. The total loss is given by:

𝐿 = 𝜆1𝐿𝑐𝑙𝑎𝑠𝑠 + 𝜆𝑖𝐿𝐶𝑙𝑎𝑠𝑠 + 𝜆2𝐿𝑏𝑏𝑜𝑥 + 𝜆3𝐿𝑐𝑜𝑛𝑓 (3)

where λ1, λ2, and λ3 are hyperparameters that balance the contributions of each component. The training is
performed using a batch size of 1 and an optimized learning rate η, which is fine-tuned using grid search.
- Model evaluation: The model is evaluated using performance metrics such as precision (P), recall (R),
and mean Average Precision (mAP). Precision and recall are computed as:

𝑃 = (𝑇𝑟𝑢𝑒 𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒𝑠)/(𝑇𝑟𝑢𝑒 𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒𝑠 + 𝐹𝑎𝑙𝑠𝑒 𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒𝑠)

𝑅 = (𝑇𝑟𝑢𝑒 𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒𝑠)/(𝑇𝑟𝑢𝑒 𝑃𝑜𝑠𝑖𝑡𝑖𝑣𝑒𝑠 + 𝐹𝑎𝑙𝑠𝑒 𝑁𝑒𝑔𝑎𝑡𝑖𝑣𝑒𝑠)

the mAP is calculated by averaging the precision over different recall levels.
- Inference on test data: The trained model is deployed for inference, where the bounding box predictions
Bi and their corresponding confidence scores Ci are generated for each test image Ii. The detection is

Int J Pow Elec & Dri Syst, Vol. 16, No. 2, June 2025: 1348-1355
Int J Pow Elec & Dri Syst ISSN: 2088-8694  1351

considered successful if the intersection over union (IoU) between the predicted and ground-truth
bounding boxes exceeds a threshold τ calculated as (4).

𝐼𝑜𝑈 = (∣ 𝐵𝑝𝑟𝑒𝑑 ∩ 𝐵𝑡𝑟𝑢𝑒 ∣)/(∣ 𝐵𝑝𝑟𝑒𝑑 ∪ 𝐵𝑡𝑟𝑢𝑒 ∣) (4)

The final model’s real-world applicability is assessed based on these metrics. The flowchart displayed
outlines the complete process, from dataset collection to inference on test data, emphasizing each stage's role
in the wind turbine defect detection system. This approach integrates mathematical equations at each stage to
optimize the model's performance and ensure accurate defect identification. Figure 2 shows the visualized
sample images with corresponding annotations.

Figure 2. Visualizing sample images with corresponding annotations

3. RESULTS AND DISCUSSION

The trained YOLOv8 model demonstrated strong performance in detecting defects on wind turbines
with high precision and recall. The model achieved a train box loss of 0.04, which indicates that the predicted
bounding boxes closely matched the actual defect locations. Similarly, the train class loss was reduced to
0.02, suggesting that the model was able to classify turbine defects accurately. These results indicate the
effectiveness of YOLOv8 in identifying wind turbine defects from UAV-captured images. Figure 3 shows
the training metrics and loss.
The model's performance was further evaluated using the mAP50 metric, which measures the
average precision across different object categories at a 50% intersection over union (IoU) threshold. The
model achieved an mAP50 score of 0.87, demonstrating its ability to accurately detect turbine defects across
varying images. Additionally, the mAP50-95 score was recorded at 0.76, highlighting the model’s robustness
across a wide range of IoU thresholds, thus confirming its capability for accurate localization of defects.
One of the major challenges in wind turbine defect detection is distinguishing between false
positives and true positives, especially in cases of small or subtle defects. By analyzing the confusion matrix,
it was observed that the model had a false positive rate of 3%, which is acceptable given the complexity of
turbine images. This low rate of false positives indicates that the model can reliably detect actual defects
without over-predicting them.
The ability to deploy this object detection model in real-world scenarios was also explored. UAVs
equipped with high-resolution cameras can capture large volumes of data across wind farms, and the
YOLOv8 model can process these images in real time to identify potential defects. This system can help
operators focus on turbines that require immediate attention, streamlining the maintenance process. The
results suggest that this approach can significantly reduce inspection time and costs, allowing operators to
optimize turbine performance and reduce downtime. Figure 4 shows the simulated output results.
Moreover, the model's ability to generalize across different wind turbine environments was tested by
introducing new images from various wind farms. The YOLOv8 model demonstrated consistent
performance, detecting defects even in images taken under different lighting conditions and perspectives.
This flexibility highlights the practical utility of this system for diverse wind farm environments.

Wind turbine defect detection using deep learning (Deepa Somasundaram)

1352  ISSN: 2088-8694

Figure 3. The training metrics and loss

Figure 4. The simulated output

Int J Pow Elec & Dri Syst, Vol. 16, No. 2, June 2025: 1348-1355
Int J Pow Elec & Dri Syst ISSN: 2088-8694  1353

4. CONCLUSION
In conclusion, the application of YOLOv8 for wind turbine object detection offers a promising
solution to the challenges of maintaining and inspecting wind farms. The model’s high accuracy and real-
time detection capabilities can reduce the need for manual inspections, improving safety and reducing
operational costs. Furthermore, this approach allows wind farm operators to quickly identify and address
potential defects, ensuring continued optimal performance of turbines. With further improvements and
integration into UAV systems, YOLOv8 can revolutionize wind turbine maintenance, promoting the wider
adoption of renewable energy.

FUNDING INFORMATION
The authors confirm that the research was carried out independently without financial influence.

AUTHOR CONTRIBUTIONS STATEMENT

This journal uses the Contributor Roles Taxonomy (CRediT) to recognize individual author
contributions, reduce authorship disputes, and facilitate collaboration.

Name of Author C M So Va Fo I R D O E Vi Su P Fu
Deepa Somasundaram ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
M. Vanitha ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
T. Sathish Kumar ✓ ✓ ✓ ✓ ✓ ✓ ✓
I. Arul Doss ✓ ✓ ✓ ✓ ✓ ✓ ✓
Adaikalam
P. Kavitha ✓ ✓ ✓ ✓ ✓ ✓ ✓
R. Kalaivani ✓ ✓ ✓ ✓ ✓ ✓ ✓

C : Conceptualization I : Investigation Vi : Visualization

M : Methodology R : Resources Su : Supervision
So : Software D : Data Curation P : Project administration
Va : Validation O : Writing - Original Draft Fu : Funding acquisition
Fo : Formal analysis E : Writing - Review & Editing

CONFLICT OF INTEREST STATEMENT

Authors state no conflict of interest.

DATA AVAILABILITY
Raw data is not publicly available due to privacy or institutional restrictions.

REFERENCES
[1] J. Lei, C. Liu, and D. Jiang, “Fault diagnosis of wind turbine based on long short-term memory networks,” Renewable Energy,
vol. 133, pp. 422–432, Apr. 2019, doi: 10.1016/j.renene.2018.10.031.
[2] F. Qu, J. Liu, H. Zhu, and B. Zhou, “Wind turbine fault detection based on expanded linguistic terms and rules using non-
singleton fuzzy logic,” Applied Energy, vol. 262, p. 114469, Mar. 2020, doi: 10.1016/j.apenergy.2019.114469.
[3] M. M. Rezaei, M. Behzad, H. Moradi, and H. Haddadpour, “Modal-based damage identification for the nonlinear model of
modern wind turbine blade,” Renewable Energy, vol. 94, pp. 391–409, Aug. 2016, doi: 10.1016/j.renene.2016.03.074.
[4] S. Sun, T. Wang, H. Yang, and F. Chu, “Damage identification of wind turbine blades using an adaptive method for compressive
beamforming based on the generalized minimax-concave penalty function,” Renewable Energy, vol. 181, pp. 59–70, Jan. 2022,
doi: 10.1016/j.renene.2021.09.024.
[5] A. Wang, Y. Pei, Z. Qian, H. Zareipour, B. Jing, and J. An, “A two-stage anomaly decomposition scheme based on multi-variable
correlation extraction for wind turbine fault detection and identification,” Applied Energy, vol. 321, p. 119373, Sep. 2022, doi:
10.1016/j.apenergy.2022.119373.
[6] S. A. Boyer, SCADA: Supervisory control and data acquisition. Research Triangle Park, NC, United States: International Society
of Automation, 2009.
[7] K.-S. Choi, Y.-H. Huh, I.-B. Kwon, and D.-J. Yoon, “A tip deflection calculation method for a wind turbine blade using
temperature compensated FBG sensors,” Smart Materials and Structures, vol. 21, no. 2, p. 025008, Feb. 2012, doi: 10.1088/0964-
1726/21/2/025008.
[8] X. Dai et al., “Dynamic head: Unifying object detection heads with attentions,” in Proceedings of the IEEE/CVF International
Conference on Computer Vision (ICCV), 2021, pp. 7373–7382.
[9] J. Deng, W. Li, Y. Chen, and L. Duan, “Unbiased mean teacher for cross-domain object detection,” in Proceedings of the
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021, pp. 4091–4101.
Wind turbine defect detection using deep learning (Deepa Somasundaram)
1354  ISSN: 2088-8694

[10] A. Foster, O. Best, M. Gianni, A. Khan, K. Collins, and S. Sharma, “Drone footage wind turbine surface damage detection,” in
2022 IEEE 14th Image, Video, and Multidimensional Signal Processing Workshop (IVMSP), IEEE, Jun. 2022, pp. 1–5, doi:
10.1109/IVMSP54334.2022.9816220.
[11] C.-Y. Fu, W. Liu, A. Ranga, A. Tyagi, and A. C. Berg, “DSSD: deconvolutional single shot detector,” arXiv:1701.06659, 2017,
[Online]. Available: http://arxiv.org/abs/1701.06659
[12] T. Gao, X. Yao, and D. Chen, “SimCSE: simple contrastive learning of sentence embeddings,” Proceedings of the 2021
Conference on Empirical Methods in Natural Language Processing, 2021, pp. 6894–6910, doi: 10.18653/v1/2021.emnlp-
main.552.
[13] R. Girshick, “Fast R-CNN,” in Proceedings of the IEEE International Conference on Computer Vision (ICCV), 2015, pp. 1440–
1448.
[14] R. Girshick, J. Donahue, T. Darrell, and J. Malik, “Rich feature hierarchies for accurate object detection and semantic
segmentation,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2014, pp. 580–587,
doi: 10.1109/CVPR.2014.81.
[15] T. Guo et al., “Nacelle and tower effect on a stand-alone wind turbine energy output—A discussion on field measurements of a
small wind turbine,” Applied Energy, vol. 303, p. 117590, Dec. 2021, doi: 10.1016/j.apenergy.2021.117590.
[16] K. He, X. Zhang, S. Ren, and J. Sun, “Spatial pyramid pooling in deep convolutional networks for visual recognition,” IEEE
Transactions on Pattern Analysis and Machine Intelligence, vol. 37, no. 9, pp. 1904–1916, Sep. 2015, doi:
10.1109/TPAMI.2015.2389824.
[17] T.-Y. Hsu, S.-Y. Shiao, and W.-I. Liao, “Damage detection of rotating wind turbine blades using local flexibility method and
long-gauge fiber Bragg grating sensors,” Measurement Science and Technology, vol. 29, no. 1, p. 015108, Jan. 2018, doi:
10.1088/1361-6501/aa97f0.
[18] S. Hwang, Y.-K. An, and H. Sohn, “Continuous-wave line laser thermography for monitoring of rotating wind turbine blades,”
Structural Health Monitoring, vol. 18, no. 4, pp. 1010–1021, Jul. 2019, doi: 10.1177/1475921718771709.
[19] M. Kharrich et al., “Optimal design of an isolated hybrid microgrid for enhanced deployment of renewable energy sources in
Saudi Arabia,” Sustainability, vol. 13, no. 9, p. 4708, Apr. 2021, doi: 10.3390/su13094708.
[20] J. A. Carballo, J. Bonilla, L. Roca, and M. Berenguel, “New low-cost solar tracking system based on open source hardware for
educational purposes,” Solar Energy, vol. 174, pp. 826–836, Nov. 2018, doi: 10.1016/j.solener.2018.09.064.
[21] J. A. Carballo, J. Bonilla, M. Berenguel, J. Fernández-Reche, and G. García, “New approach for solar tracking systems based on
computer vision, low cost hardware and deep learning,” Renewable Energy, vol. 133, pp. 1158–1166, Apr. 2019, doi:
10.1016/j.renene.2018.08.101.
[22] X W. Li, W. Zhao, T. Wang, and Y. Du, “Surface defect detection and evaluation method of large wind turbine blades based on
an improved Deeplabv3+ deep learning model,” Structural Durability & Health Monitoring, vol. 18, no. 5, pp. 553–575, 2024,
doi: 10.32604/sdhm.2024.050751.
[23] X. Sun, G. Wang, L. Xu, H. Yuan, and N. Yousefi, “Optimal estimation of the PEM fuel cells applying deep belief network
optimized by improved archimedes optimization algorithm,” Energy, vol. 237, p. 121532, Dec. 2021, doi:
10.1016/j.energy.2021.121532.
[24] W. Abitha Memala, C. Bhuvaneswari, S. M. Shyni, G. Merlin Sheeba, M. S. Mahendra, and V. Jaishree, “DC-DC converter based
power management for go green applications,” International Journal of Power Electronics and Drive Systems, vol. 10, no. 4, pp.
2046–2054, Dec. 2019, doi: 10.11591/ijpeds.v10.i4.pp2046-2054.
[25] M. Aqib, R. Mehmood, A. Alzahrani, I. Katib, A. Albeshri, and S. M. Altowaijri, “Smarter traffic prediction using big data, in-
memory computing, deep learning and GPUs,” Sensors, vol. 19, no. 9, p. 2206, May 2019, doi: 10.3390/s19092206.

BIOGRAPHIES OF AUTHORS

Dr. Deepa Somasundaram received her B.E. from K.S.R. College of

Technology, affiliated to Periyar University, in 2003, and M.E. from Annamalai University in
2005. She completed her Ph.D. degree from Sathyabama University in 2013. Presently, she is
working as a professor in the Department of EEE at Panimalar Engineering College, Chennai.
She has published more than 30 papers in international and national journals. Her area of
interest is power systems and optimization techniques. She has more than 15 years of
experience in teaching field. She can be contacted at email: [email protected].

Dr. M. Vanitha is a Professor in the Department of ECE, has over 22 years of

teaching experience and specializes in Wireless Networks and Embedded Systems. She has
published 28 journal papers, holds 6 published patents, and 1 granted patent. She is currently
guiding 7 Ph.D. scholars and has served as a keynote speaker and session chair in several
international conferences. She has completed multiple NPTEL and Coursera-certified courses
and organized global conferences, including ICATEM24 in Dubai. She can be contacted
at email: [email protected].

Int J Pow Elec & Dri Syst, Vol. 16, No. 2, June 2025: 1348-1355
Int J Pow Elec & Dri Syst ISSN: 2088-8694  1355

T. Sathish Kumar is an Assistant Professor in the Department of Electrical and

Electronics Engineering at S.A. Engineering College, Tamil Nadu, India. He specializes in
Power System Engineering and has contributed to various reputed journals. His research work
spans multiple years, showcasing consistent academic involvement. He has collaborated with
numerous researchers in the field. His scholarly work reflects both depth and diversity in
engineering research. He can be contacted at email: [email protected].

I. Arul Doss Adaikalam received B.E. degree in Electrical and Electronics

Engineering and an M.E degree in Power System Engineering, both from Anna University,
Chennai, in 2001 and 2009, respectively. He received the Ph.D. degree in Electrical
Engineering from Anna University, Chennai, in 2020. He has 14 years of work experience in
the field of teaching from various Reputed Academic Organizations across Tamil Nadu, India,
since the year of 2009. He is currently working as an Assistant Professor at Chennai Institute of
Technology, Chennai. He is a life member of the Indian Society for Technical Education and
other Professional Societies. He can be contacted at email: [email protected].

Dr. P. Kavitha working as an associate professor in the Department of Electrical

and Electronics Engineering R.M.K. Engineering College. She has about 27 years of teaching
experience. She received her B.E. degree in Electrical and Electronics Engineering with first
class and M.E. degree in Control and Instrumentation with distinction from Anna University,
Chennai. She has published 17 research papers in refereed international journals and various
international conferences. Her areas of research include power controllers, machine design, and
renewable energy systems. She can be contacted at email: [email protected].

R. Kalaivani obtained her B.E. degree in Electrical and Electronics Engineering in

the year 2000 from National Institute of Technology, Trichy, M.E. degree in Power Systems in
the year 2002 from National Institute of Technology, Trichy. She obtained her Ph.D. degree in
the year 2013 from Anna University, Chennai. She is currently working as a Professor at
Rajalakshmi Engineering College in the Department of Electrical and Electronics Engineering.
Her current research areas are smart grid, renewable energy, and electric vehicles. She can be
contacted at email: [email protected].

Wind turbine defect detection using deep learning (Deepa Somasundaram)