Makine öğrenmesi kullanarak iletim hatlarında arıza türünün ve yerinin belirlenmesi

Abstract

Arka Plan—Günümüzde hızla artan enerji talebi ile enerjinin kesintisiz ve sürdürülebilir bir şekilde sağlanabilmesi zorunlu hale gelmiştir. Bu durum, iletim hatlarında meydana gelen arızaların hızlı ve doğru teşhis edilerek oluşan anormalliklerin giderilmesinin önemini artırmaktadır. Meydana gelen arızaların teşhisinde geleneksel yöntemler bazı durumlarda yetersiz kalabilmektedir. Bu nedenle yapay zekâ tabanlı yöntemlerin kullanılması büyük önem kazanmıştır.

Amaç—Bu çalışmanın amacı enerji iletim hatlarında oluşabilecek arızaların türünün ve yerinin tespitinde farklı makine öğrenmesi modelleri oluşturularak, performanslarının karşılaştırılması ve bunun sonucunda en güvenilir modeli belirlemektir. Bu analizlerle, modellerin performanslarının farklı işletme koşullarında da tutarlı olması hedeflenmiştir.

Yöntem— MATLAB / Simulink ortamında 400kV’luk bir iletim hattı modeli çeşitli arıza türü ve yeri için simüle edilmiş ve problemin gerekliliklerine göre üç fazlı gerilim, üç fazlı akım, sıfır bileşen akımı (I₀), zaman ve konum bilgilerinden oluşan kapsamlı veri kümeleri elde edilmiştir. Arıza türünün tespitinde destek vektör makineleri (SVM) ve karar ağacı (DT) algoritmaları kullanılırken arıza yeri tespitinde k-en yakın komşu (KNN) regresyon ve rastgele orman regresyon algoritmaları kullanılmıştır. Bu algoritmalarla oluşturulan modellerin performansı, sınıflandırma ve regresyon problemi için ayrı ayrı metriklerle değerlendirilmiştir.

Bulgular—Arıza türü tespiti sonuçlarına göre, her iki model de %99’un üzerinde doğruluk değerlerine ulaşmış; SVM modeli %99.70 doğruluk oranına ulaşırken, Karar Ağacı %99.81 doğruluk oranı ile en yüksek başarıyı yakalayan model olmuştur. Arıza yeri tespitinde ise ortalama mutlak hata (MAE) değerlerinde belirgin farklar gözlemlenmiştir. KNN regresyon modeli 0.416 km ortalama mutlak hata sergilerken, Rastgele Orman Regresyonu 0.114 km gibi oldukça düşük bir hata payı ile arıza konumunu çok daha hassas bir şekilde belirlemiştir. Değerlendirme sonucunda, oluşturulan model performansları başarılı ve güvenilir sonuçlar verirken sınıflandırma probleminde karar ağacının doğruluk oranının daha yüksek olduğu, yer tespitinde ise rastgele orman algoritmasının daha yavaş fakat daha az hatalı sonuçlar verdiği sonucuna varılmıştır. Modellerin farklı senaryolar altındaki genelleme yeteneğini değerlendirmek üzere; arıza direnci, hat parametreleri ve faz farkı değişimleri analiz edilmiştir. Sonuçlar, önerilen yapının değişken işletme koşulları altında bile yüksek hassasiyet, güvenilirlik ve kararlılık sunduğunu ortaya koymuştur.

Sonuç—Elde edilen bulgulara dayanarak, doğru öznitelik seçiminin ve problemin yapısına uygun model tercihinin genel sistem performansı üzerindeki önemi vurgulanmış ve böylece enerji iletim hatlarında arıza yönetim ve bakım süreçlerine entegre edilebilecek güvenilir bir çerçeve sunulmuştur.

Keywords

• Enerji iletim hatları
• makine öğrenmesi
• arıza türü tespiti
• arıza yeri tespiti
• kısa devre arızaları

Fault type and location identification in transmission lines using machine learning

Abstract

Context—With the rapidly increasing energy demand today, ensuring the uninterrupted and sustainable supply of energy has become essential. This situation increases the importance of quickly and accurately diagnosing faults occurring in transmission lines and eliminating the resulting abnormalities. Traditional methods used for fault diagnosis may be insufficient in some cases; therefore, the use of artificial intelligence-based methods has gained significant importance.

Objective—The aim of this study is to develop different machine learning models to identify the type and location of faults that may occur in power transmission lines, compare their performances, and determine the most reliable model. Through these analyses, it is also aimed to ensure that the performance of the models remains consistent under different operating conditions.

Method—A 400 kV transmission line model was simulated in the MATLAB / Simulink environment for various fault types and locations, and comprehensive datasets consisting of three-phase voltage, three-phase current, zero-sequence current (I₀), time, and location information were obtained according to the requirements of the problem. For fault type detection, Support Vector Machine (SVM) and Decision Tree (DT) algorithms were used, while K-Nearest Neighbors (KNN) regression and Random Forest regression algorithms were used for fault location detection. The performance of the models developed using these algorithms was evaluated using separate metrics for classification and regression problems.

Results—According to the results of fault type detection, both models achieved accuracy values above 99%; while the SVM model reached an accuracy of 99.70%, the Decision Tree achieved the highest performance with an accuracy of 99.81%. In fault location detection, significant differences were observed in the Mean Absolute Error (MAE) values: the KNN regression model exhibited an MAE of 0.416 km, whereas the Random Forest regression determined the fault location much more precisely with a very low error margin of 0.114 km. As a result of the evaluation, although all developed models produced successful and reliable results, it was concluded that the Decision Tree had a higher accuracy rate in the classification problem, whereas the Random Forest algorithm provided slower but less erroneous results in fault location detection. To evaluate the generalization capability of the models under different scenarios, variations in fault resistance, line parameters, and phase angle differences were analyzed, and the results revealed that the proposed structure provides high accuracy, reliability, and stability even under varying operating conditions.

Conclusion—Based on the findings, the importance of correct feature selection and selecting a model appropriate to the structure of the problem for overall system performance was emphasized, and thus a reliable framework that can be integrated into fault management and maintenance processes in power transmission lines has been presented.

Keywords

• Energy transmission lines
• machine learning
• fault type identification
• fault location identification
• short circuit faults

References

1. [1] T. Goswami, U. B. Roy, “Predictive Model for Classification of Power System Faults using Machine Learning”, 2019 IEEE Region 10 Conference, Kochi, Hindistan, 17-20 October 2019, 1881-1885. https://doi.org/10.1109/TENCON.2019.8929264.
2. [2] A. Prasad, J. B. Edward, K. Ravi, “A review on fault classification methodologies in power transmission systems: Part—I”, Journal of Electrical Systems and Information Technology, 5(1), 48–60, 2018. https://doi.org/10.1016/j.jesit.2017.01.004.
3. [3] V. Venkatesh, “Fault Classification and Location Identification on Electrical Transmission Network Based on Machine Learning Methods”, M.S. thesis, Virginia Commonwealth University, Richmond, Virginia, USA, 2018.
4. [4] F. Heymann, H. Quest, T. Lopez Garcia, C. Ballif, M. Galus, “Reviewing 40 years of artificial intelligence applied to power systems – A taxonomic perspective”, Energy and AI, 15, 100322, 2024. https://doi.org/10.1016/j.egyai.2023.100322.
5. [5] M. Ünlü, H. Ergen, “Hava durumu verilerinin kısa dönem elektrik yük tahmini üzerine etkileri”, Pamukkale Üniversitesi Mühendislik Bilimleri Dergisi, 32(1), 115–130, 2026. https://doi.org/10.5505/pajes.2025.38921.
6. [6] R. Bayındır, Ş. Sağıroğlu, İ. Çolak, “Yapay Sinir Ağları Tabanlı Reaktif Güç Kompanzasyonu”, Politeknik Dergisi, 10(2), 129–135, 2007. https://izlik.org/JA35YA28MD.
7. [7] Z. Li, J. Yan, Y. Liu, W. Liu, L. Li, H. Qu, “Power system transient voltage vulnerability assessment based on knowledge visualization of CNN”, International Journal of Electrical Power & Energy Systems, 155, 109576, 2024. https://doi.org/10.1016/j.ijepes.2023.109576.
8. [8] Z. Mustaffa, M. H. Sulaiman, “Random forest based wind power prediction method for sustainable energy system”, Cleaner Energy Systems, 12, 100210, 2025. https://doi.org/10.1016/j.cles.2025.100210.

9. [9] B. Guo, B. Yang, S. Wang, W. Shi, F. Yang, D. Wang, “Machine learning-based fault diagnosis and classification of three-phase transmission lines with RFE and domain knowledge”, Electric Power Systems Research, 247, 111777, 2025. https://doi.org/10.1016/j.epsr.2025.111777.
10. [10] O. Turanlı, Y. Benteşen Yakut, “Classification of Faults in Power System Transmission Lines Using Deep Learning Methods with Real, Synthetic, and Public Datasets”, Applied Sciences, 14(20), 9590, 2024. https://doi.org/10.3390/app14209590.
11. [11] S. Budak, B. Akbal, “Fault Location Estimation by Using Machine Learning Methods in Mixed Transmission Lines”, European Journal of Science and Technology, 245-250, 2020. https://doi.org/10.31590/ejosat.802888.
12. [12] S. Budak, B. Akbal, “Görüntü İşleme ve Yapay Sinir Ağları ile İletim Hatlarında Arıza Yeri Belirleme”, Konya Mühendislik Bilimleri Dergisi, 8(3), 678–692, 2020. https://izlik.org/JA97HF75EN.
13. [13] N. Q. Minh, N. T. Khiem, V. H. Giang, “Fault classification and localization in power transmission line based on machine learning and combined CNN-LSTM models”, Energy Reports, 12, 5610–5622, 2024. https://doi.org/10.1016/j.egyr.2024.11.061.
14. [14] D. Akmaz, M. S. Mamiş, M. Arkan, “Transmission line fault location using traveling wave frequencies and extreme learning machine”, Electric Power Systems Research, 155, 1–7, 2018. https://doi.org/10.1016/j.epsr.2017.09.019.
15. [15] D. Deepika, M. M. Charan, S. C. Nossam, P. V. Manitha, “Advanced Machine Learning Models for Electrical Fault Detection and Classification in Transmission Lines”, 2025 3rd International Conference on Intelligent Data Communication Technologies and Internet of Things, Bengaluru, Indian, 05-07 February 2025, 1338–1341. https://doi.org/10.1109/IDCIOT64235.2025.10914898.
16. [16] V. S. S. Reddy, C. Divyasri, B. Panimathi, M. Venugopalan, “Classification of Types of Faults and Their Location in Power Transmission Lines”, 2025 7th International Conference on Intelligent Sustainable Systems, India, 12-14 March 2025, 125-130. https://doi.org/10.1109/ICISS63372.2025.11076561.
17. [17] S. Ekici, F. Unal, U. Ozleyen, “Comparison of different regression models to estimate fault location on hybrid power systems”, IET Generation, Transmission & Distribution, 13(20), 4756–4765, 2019. https://doi.org/10.1049/iet-gtd.2018.6213.
18. [18] Z. El Mrabet, N. Sugunaraj, P. Ranganathan, S. Abhyankar, “Random Forest Regressor-Based Approach for Detecting Fault Location and Duration in Power Systems”, Sensors, 22(2), 458, 2022. https://doi.org/10.3390/s22020458.
19. [19] T. Anwar, C. Mu, M. Z. Yousaf, W. Khan, S. Khalid, A.O. Hourani, I. Zaitsev, “Robust fault detection and classification in power transmission lines via ensemble machine learning models”, Scientific Reports, 15(1), 2549, 2025. https://doi.org/10.1038/s41598-025-86554-2.
20. [20] R. Gupta, D. Suhavan, P. P. Vuppuluri, “Effective Machine Learning Models for Fault Detection and Classification in Power Transmission Lines”, 2025 7th International Conference on Energy, Power and Environment, Sohra, India, 9-11 May 2025. https://doi.org/10.1109/ICEPE65965.2025.11139445.
21. [21] M. O. F. Goni, M. Nahiduzzaman, M. S. Anower, M. M. Rahman, M. R. Islam, M. Ahsan, J. Haider, M. Shahjalal, “Fast and Accurate Fault Detection and Classification in Transmission Lines using Extreme Learning Machine”, e-Prime - Advances in Electrical Engineering, Electronics and Energy, 3, 100107, 2023. https://doi.org/10.1016/j.prime.2023.100107.
22. [22] J. Hu, W. Hu, J. Chen, D. Cao, Z. Zhang, Z. Liu, “Fault Location and Classification for Distribution Systems Based on Deep Graph Learning Methods”, Journal of Modern Power Systems and Clean Energy, 11(1), 35–51, 2023. https://doi.org/10.35833/MPCE.2022.000204.
23. [23] J. Cervantes, F. Garcia-Lamont, L. Rodríguez-Mazahua, A. Lopez, “A comprehensive survey on support vector machine classification: Applications, challenges and trends”, Neurocomputing, 408, 189–215, 2020. https://doi.org/10.1016/j.neucom.2019.10.118.
24. [24] V. Kecman, “Support Vector Machines – An Introduction”, Support Vector Machines: Theory and Applications - Studies in Fuzziness and Soft Computing, L. Wang, Eds., Berlin, Heidelberg: Springer, 2005, Chapter 1, 1–47. https://doi.org/10.1007/10984697_1.
25. [25] T. V. Saradhi, “A Study on Hyperparameter Tuning in Support Vector Machines and its Impact on Model Accuracy”, Global Journal of Engineering Innovations and Interdisciplinary Research, 5(1), 11, 2025. https://dx.doi.org/10.33425/3066-1226.1063.
26. [26] I. H. Sarker, “Machine Learning: Algorithms, Real-World Applications and Research Directions”, SN Computer Science, 2(3), 160, 2021. https://doi.org/10.1007/s42979-021-00592-x.
27. [27] S. F. Sabbeh, “Machine-Learning Techniques for Customer Retention: A Comparative Study”, International Journal of Advanced Computer Science and Applications, 9(2), 273-281, 2018, https://dx.doi.org/10.14569/IJACSA.2018.090238.
28. [28] scikit-learn, “DecisionTreeClassifier”, https://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeClassifier.html (25.09.2025).
29. [29] A. A. Karcioğlu, H. Bulut, “Performance Evaluation of Classification Algorithms Using Hyperparameter Optimization”, 2021 6th International Conference on Computer Science and Engineering, Ankara, Türkiye, 15-17 September 2021, 354-358. https://doi.org/10.1109/UBMK52708.2021.9559003.
30. [30] Ö. F. Ertuğrul, M. E. Tağluk, “A novel version of k nearest neighbor: Dependent nearest neighbor”, Applied Soft Computing, 55, 480–490, 2017. https://doi.org/10.1016/j.asoc.2017.02.020.
31. [31] scikit-learn, “KNeighborsRegressor”, https://scikit-learn.org/stable/modules/generated/sklearn.neighbors.KNeighborsRegressor.html (25.09.2025).
32. [32] scikit-learn, “distance_metrics”, https://scikit-learn.org/stable/modules/generated/sklearn.metrics.pairwise.distance_metrics.html (25.09.2025).
33. [33] J. Alzubi, A. Nayyar, and A. Kumar, “Machine Learning from Theory to Algorithms: An Overview”, Journal of Physics Conference Series, 1142, 012012, 2018. https://doi.org/10.1088/1742-6596/1142/1/012012.
34. [34] R. Choudhary, H. K. Gianey, “Comprehensive Review On Supervised Machine Learning Algorithms”, 2017 International Conference on Machine Learning and Data Science, Bangalore, India, 14-15 December 2017, 37-43. https://doi.org/10.1109/MLDS.2017.11.
35. [35] scikit-learn, “RandomForestRegressor”, https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestRegressor.html (25.09.2025).
36. [36] H. Yuliana, Iskandar, Hendrawan, S. Basuki, M. R. Hidayat, A. Charisma, H. Vidyaningtyas, "Hyperparameter optimization of random forest algorithm to enhance performance metric evaluation of 5G coverage prediction", Buletin Pos dan Telekomunikasi, 22(1), 75–90, 2024. https://doi.org/10.17933/bpostel.v22i1.390.
37. [37] G. Varoquaux, O. Colliot, “Evaluating Machine Learning Models and Their Diagnostic Value”, Machine Learning for Brain Disorders, O. Colliot, Eds., Humana, New York, USA, 2023, Chapter 20, 601-631. https://doi.org/10.1007/978-1-0716-3195-9_20.
38. [38] M. Steurer, R. J. Hill, and N. Pfeifer, “Metrics for evaluating the performance of machine learning based automated valuation models”, Journal of Property Research, 38(2), 99–129, 2021 https://doi.org/10.1080/09599916.2020.1858937.
39. [39] S. Mandal, A. K. Bohre, “Fault Classification in Electrical Systems Using Machine Learning Algorithms”, 2022 Third International Conference on Intelligent Computing Instrumentation and Control Technologies, Kannur, India, 11-12 August 2022, 1412-1416. https://doi.org/10.1109/ICICICT54557.2022.9917976.
40. [40] P. S. P. Eboule, J. H. C. Pretorius, N. Mbuli, “Machine learning techniques applied on a nine-phase transmission line for faults classification and location”, Energy Reports, 8, 801–810, 2022. https://doi.org/10.1016/j.egyr.2021.11.062.