SavedAdd your journal

Journals / Balıkesir Üniversitesi Fen Bilimleri Enstitüsü Dergisi / 2026 / Cilt: 28 Sayı: 1

DENTELNet: A Novel Hybrid Deep Learning Method for Detecting Non-Technical Electricity Losses

DENTELNet: Teknik olmayan elektrik kayıplarının tespiti için yeni bir hibrit derin öğrenme yöntemi

Hamdullah Karamollaoğlu(Elektrik Üretim A.Ş.)0000-0001-6419-2249
Pages
81–98
Publication Date
2026-01-09
🏛️Bibliographic metadata and full text access for this article are provided via DergiPark.
DergiPark Record ↗
✨

AI Summary & Key Takeaways

AI Insight
⚡ TL;DR(Too Long; Didn't Read)

Experimental results demonstrate an accuracy of 92.153%, precision of 90.402%, recall of 92.153%, and an F1-score of 90.397%.

📌 3 Key Takeaways

🎯
Objective & Hypothesis:Electricity is one of the essential needs of modern societies, playing a critical role in economic growth and social welfare.
📊
Key Findings & Data:Non-technical losses, particularly due to unauthorized electricity consumption, meter tampering, and illegal connections, are among the most significant challenges faced by the energy sector.
💡
Conclusion & Impact:Experimental results demonstrate an accuracy of 92.153%, precision of 90.402%, recall of 92.153%, and an F1-score of 90.397%.
🔬Design:Modeling & Simulation
📍Scope:Comprehensive Sample
⚙️Variables:Teknik olmayan kayıplar · elektrik sahtekârlığı tespiti · derin öğrenme

Abstract

Electricity is one of the essential needs of modern societies, playing a critical role in economic growth and social welfare. However, losses occurring throughout the energy production and consumption process lead to inefficient use of resources and economic damages. Non-technical losses, particularly due to unauthorized electricity consumption, meter tampering, and illegal connections, are among the most significant challenges faced by the energy sector. In this study, an innovative hybrid deep learning model named DENTELNet is proposed for detecting non-technical losses. The model consists of Convolutional Neural Networks (CNN), Bidirectional Long Short-Term Memory (BiLSTM), and Liquid Neural Networks (LNN). To achieve feature reduction, a Time-Series Decomposition-based Principal Component Analysis method is used, and data imbalance is addressed using the Borderline Synthetic Minority Over-sampling Technique (BSMOTE) and Tomek Link (TLink) approach. Experimental results demonstrate an accuracy of 92.153%, precision of 90.402%, recall of 92.153%, and an F1-score of 90.397%. The ROC curve analysis, with high AUC values, reveals the model's discriminative power and balanced performance.

Özet

Elektrik enerjisi, modern toplumların temel ihtiyaçlarından biri olup, ekonomik büyüme ve toplumsal refah açısından kritik bir rol oynamaktadır. Ancak enerji üretiminden tüketime kadar geçen süreçte meydana gelen kayıplar, kaynakların verimsiz kullanımına ve ekonomik zararlara yol açmaktadır. Teknik olmayan kayıplar, özellikle usulsüz elektrik kullanımı, sayaç manipülasyonu ve kaçak bağlantılar gibi nedenlerle enerji sektörünün karşılaştığı en büyük zorluklardan biridir. Bu çalışmada, teknik olmayan kayıpların tespiti amacıyla yenilikçi bir hibrit derin öğrenme modeli olan DENTELNet önerilmiştir. Model, Evrişimli Sinir Ağları (Convolutional Neural Networks – CNN), İki Yönlü Uzun Kısa Vadeli Bellek (Bidirectional Long Short-Term Memory – BiLSTM) ve Sıvı Sinir Ağı (Liquid Neural Network – LNN) bileşenlerinden oluşmaktadır. Öznitelik indirgeme amacıyla Zaman Serisi Ayrışımı tabanlı Temel Bileşen Analizi yöntemi kullanılmış ve veri dengesizliği, Sınırda Sentetik Azınlık Aşırı Örnekleme Tekniği (Borderline SMOTE – BSMOTE) ve Tomek Bağlantısı Yöntemi (Tomek Link – TLink) yaklaşımı ile giderilmiştir. Yapılan deneysel çalışmalar sonucunda, model %92.153 doğruluk, %90.402 kesinlik, %92.153 duyarlılık ve %90.397 F1-skoru elde etmiştir. ROC eğrisi analizi, yüksek AUC değerleri ile modelin ayırt edici gücünü ve dengeli performansını göstermektedir.

🔖

Cite This Article

Karamollaoğlu, H. (2026). DENTELNet: A Novel Hybrid Deep Learning Method for Detecting Non-Technical Electricity Losses. Balıkesir Üniversitesi Fen Bilimleri Enstitüsü Dergisi, Cilt: 28 Sayı: 1, 81–98. https://doi.org/10.25092/baunfbed.1627699

📄 Full Text PDF Viewer

📄

PDF görüntüleyici hazırlanıyor...

References (27)

  1. Saqib, S. M., Mazhar, T., Iqbal, M., Shahazad, T., Almogren, A., Ouahada, K. ve Hamam, H., Deep learning-based electricity theft prediction in non-smart grid environments, Heliyon, 10, 15, e35167, (2024).
  2. Khan, N., Shahid, Z., Alam, M. M., Sajak, A. A. B., Nazar, M. ve Mazliham, M. S., A novel deep learning technique to detect electricity theft in smart grids using AlexNet, IET Renewable Power Generation, 18, 6, 941–958, (2024).
  3. Souza, M. A., Gouveia, H. T., Ferreira, A. A., de Lima Neta, R. M., Nóbrega Neto, O., da Silva Lira, M. M., Torres, G. L. ve Aquino, R. R., Detection of Non-Technical Losses on a Smart Distribution Grid Based on Artificial Intelligence Models, Energies, 17, 7, 1729, (2024).
  4. Shehzad, F., Javaid, N., Aslam, S. ve Javed, M. U., Electricity theft detection using big data and genetic algorithm in electric power systems, Electric Power Systems Research, 209, 107975, (2022).
  5. Tripathi, A. K., Pandey, A. C. ve Sharma, N., A new electricity theft detection method using hybrid adaptive sampling and pipeline machine learning, Multimedia Tools and Applications, 83, 18, 54521–54544, (2024).
  6. Shehzad, F., Ullah, Z., Alhussein, M., Aurangzeb, K. ve Aslam, S., Deep learning-based meta-learner strategy for electricity theft detection, Frontiers in Energy Research, 11, 1232930, (2023).
  7. Nawaz, A., Ali, T., Mustafa, G., Rehman, S. U. ve Rashid, M. R., A novel technique for detecting electricity theft in secure smart grids using CNN and XG-boost, Intelligent Systems with Applications, 17, 200168, (2023).
  8. Fei, K., Li, Q., Zhu, C., Dong, M. ve Li, Y., Electricity frauds detection in low-voltage networks with contrastive predictive coding, International Journal of Electrical Power and Energy Systems, 137, 107715, (2022).
  9. Zheng, Z., Yang, Y., Niu, X., Dai, H.-N. ve Zhou, Y., Wide and deep convolutional neural networks for electricity-theft detection to secure smart grids, IEEE Transactions on Industrial Informatics, 14(4), 1606–1615, (2017).
  10. SGCC dataset: Analysis on Abnormal Behavior of Electricity Customers. https://www.datafountain.cn/competitions/241 (06.05.2025).
  11. Wen, Q., Gao, J., Song, X., Sun, L., Xu, H. ve Zhu, S., RobustSTL: A robust seasonal-trend decomposition algorithm for long time series, Proceedings of the AAAI Conference on Artificial Intelligence, 33(01), 5409–5416, (2019).
  12. Kherif, F. ve Latypova, A., Principal component analysis. İçinde A. Mechelli & V. Sandra (Ed.), Machine learning (ss. 209–225). Elsevier, (2019).
  13. Sulandari, W. ve Utami, H., Forecasting time series with trend and seasonal patterns based on SSA, 3rd International Conference on Science in Information Technology, 648–653, (2017).
  14. Al Majzoub, H., Elgedawy, I., Akaydın, Ö. ve Köse Ulukök, M., HCAB-SMOTE: A hybrid clustered affinitive borderline SMOTE approach for imbalanced data binary classification, Arabian Journal for Science and Engineering, 45(4), 3205–3222, (2020).
  15. Smiti, S. ve Soui, M., Bankruptcy prediction using deep learning approach based on borderline SMOTE, Information Systems Frontiers, 22(5), 1067–1083, (2020).
  16. Pereira, R. M., Costa, Y. M. ve Silla Jr, C. N., MLTL: A multi-label approach for the Tomek Link undersampling algorithm, Neurocomputing, 383, 95–105, (2020).
  17. Swana, E. F., Doorsamy, W. ve Bokoro, P., Tomek link and SMOTE approaches for machine fault classification with an imbalanced dataset, Sensors, 22(9), 3246, (2022).
  18. He, H. ve Garcia, E. A., Learning from imbalanced data, IEEE Transactions on Knowledge and Data Engineering, 21(9), 1263–1284, (2009).
  19. Kabakus, A. T., DroidMalwareDetector: A novel Android malware detection framework based on convolutional neural network, Expert Systems with Applications, 206, 117833, (2022).
  20. Riyaz, B. ve Ganapathy, S., A deep learning approach for effective intrusion detection in wireless networks using CNN, Soft Computing, 24(22), 17265–17278, (2020).
  21. Sri Vidhya, G. ve Nagarajan, R., A novel bidirectional LSTM model for network intrusion detection in SDN-IoT network, Computing, 106, 2613–2642, (2024).
  22. Chahine, M., Hasani, R., Kao, P., Ray, A., Shubert, R., Lechner, M., Amini, A. ve Rus, D., Robust flight navigation out of distribution with liquid neural networks, Science Robotics, 8(77), eadc8892, (2023).
  23. Kumar, K., Verma, A., Gupta, N. ve Yadav, A., Liquid neural networks: A novel approach to dynamic information processing, International Conference on Advances in Computation, Communication and Information Technology (ICAICCIT), 725–730, (2023).
  24. Depuru, S. S. S. R., Wang, L. ve Devabhaktuni, V., Support vector machine based data classification for detection of electricity theft, Power Systems Conference and Exposition, 1–8, (2011).
  25. Hasan, M. N., Toma, R. N., Nahid, A. A., Islam, M. M. ve Kim, J.-M., Electricity theft detection in smart grid systems: A CNN-LSTM based approach, Energies, 12(17), 3310, (2019).
  26. Munawar, S., Asif, M., Kabir, B., Pamir, Ullah, A. ve Javaid, N., Electricity theft detection in smart meters using a hybrid bi-directional GRU–bi-directional LSTM model, 15th International Conference on Complex, Intelligent and Software Intensive Systems, 297–308, (2021).
  27. Pereira, J. ve Saraiva, F., Convolutional neural network applied to detect electricity theft: A comparative study on unbalanced data handling techniques, International Journal of Electrical Power and Energy Systems, 131, 107085, (2021).