Transformatör sağlık endeksi tahmininde dalgacık ailesi seçiminin etkisi

Öz

Güç transformatörleri, iletim şebekelerinin kritik ve yüksek maliyetli bileşenleri olup, güvenilir çalışmaları şebeke kararlılığı ve ekonomik verimlilik açısından büyük önem taşımaktadır. Transformatör Sağlık Endeksi (TSE) çözünmüş gaz analizi, yağ kalitesi göstergeleri, su içeriği ve güç faktörü gibi temel teşhis parametrelerini bir araya getirerek, varlık durumu değerlendirmesi ve yatırım planlaması için önemli bilgiler sunmaktadır. Bu çalışmada, literatürde transformatör uygulamalarında yüksek tahmin doğruluğu ile bilinen Rastgele Orman algoritması; transformatör veri kümesi üzerinde, veri ön işleme ve filtreleme teknikleri ile kullanılarak TSE tahmini gerçekleştirilmiştir. Yazarların bilgisine göre, literatürde ilk kez, TSE tahmininde ön işleme aşamasında farklı dalgacık aileleri sistematik olarak karşılaştırılmış ve tahmin doğruluğu üzerindeki etkileri incelenmiştir. Sonuçlar, Symlet-2 konfigürasyonunun hem filtreli hem de filtresiz veri setlerinde diğer ailelere kıyasla tutarlı biçimde en yüksek performansı sağladığını, Coiflet-3 ve Coiflet-5’in ise boyut indirgeme yoluyla daha yüksek hesaplama verimliliği elde ederken Symlet-2’ye kıyasla R2 değerinde yaklaşık 0,09–0,10’luk bir düşüş yaşadığı görülmüştür. Bulgular, önişleme aşamasında seçilen dalgacık ailesinin hem özellik seçimi sonuçlarını hem de model performansını doğrudan etkilediğini ortaya koymakta olup yüksek doğruluklu transformatör durumu değerlendirme yaklaşımlarının geliştirilmesine yönelik değerli çıkarımlar sunmaktadır.

Anahtar Kelimeler

• Transformatör sağlık endeksi
• Dalgacık aileleri
• Pearson korelasyonu
• Rastgele Orman
• Durum izleme

Effect of wavelet family selection on transformer health index prediction

Abstract

In transmission systems, power transformers are key high-value components, and their consistent operation is fundamental for maintaining grid stability and achieving cost-effective performance. The Transformer Health Index (THI) integrates key diagnostic parameters including dissolved gas analysis, water content, oil quality indicators, and power factor providing essential insights for asset condition assessment and investment planning. In this study, THI prediction is conducted using the Random Forest algorithm, recognized in literature for its high predictive accuracy for transformer applications, in combination with data preprocessing and filtering techniques applied to transformer dataset. For the first time, to the best of our knowledge in the THI prediction literature, various wavelet families are systematically compared at the preprocessing stage to examine their influence on predictive accuracy. The results show that the Symlet-2 configuration consistently outperformed other families in both filtered and non-filtered datasets, while Coiflet-3 and Coiflet-5 achieved higher efficiency through dimensionality reduction but with an accuracy decrease of approximately 0.09–0.10 in R2 compared to Symlet-2. The findings demonstrate that the choice of wavelet family in the preprocessing phase directly impacts feature selection outcomes and model performance, offering valuable guidance for the development of high-accuracy transformer condition assessment frameworks.

Keywords

• Transformer Health Index
• wavelet families
• Pearson correlation
• Random Forest
• condition assessment

Kaynakça

1. F. E. Bezerra, F. A. Z. Garcia, S. I. Nabeta, G. F. M. de Souza, I. E. Chabu, J. C. Santos, S. Nagao Junior, and F. H. Pereira, Wavelet-like transform to optimize the order of an autoregressive neural network model to predict the dissolved gas concentration in power transformer oil from sensor data. Sensors, 20(9), 2730, 2020. https://doi.org/10.3390/s20092730.
2.    S. Li, X. Li, Y. Cui and H. Li, Review of transformer health index from the perspective of survivability and condition assessment. Electronics, 12(11), 2407, 2023. https://doi.org/10.3390/electronics12112407.
3.    H. Guo and L. Guo, Health index for power transformer condition assessment based on operation history and test data. Energy Reports, 8, 9038–9045, 2022. https://doi.org/10.1016/j.egyr.2022.07.041.
4.    N. El-Rashidy, Y. A. Sultan and Z. H. Ali, Predicting power transformer health index and life expectation based on digital twins and multitask LSTM-GRU model. Scientific Reports, 15, 1359, 2025. https://doi.org/10.1038/s41598-024-83220-x.
5.    A. Alqudsi and A. El-Hag, Application of machine learning in transformer health index prediction. Energies, 12(14), 2694, 2019. https://doi.org/10.3390/en12142694.
6.    M. A. A. Putra, Suwarno and R. A. Prasojo, Improving transformer health index prediction performance using machine learning algorithms with a synthetic minority oversampling technique. Energies, 18(9), 2364, 2025. https://doi.org/10.3390/en18092364.
7.    M. M. Islam, G. Lee and S. N. Hettiwatte, Application of a general regression neural network for health index calculation of power transformers. International Journal of Electrical Power and Energy Systems, 93, 308–315, 2017. https://doi.org/10.1016/j.ijepes.2017 .06.008.
8.    I. B. Taha, Power transformers health index enhancement based on convolutional neural network after applying imbalanced-data oversampling. Electronics, 12(11), 2405, 2023. https://doi.org/ 10.3390/electronics12112405.

9.    R. Zemouri, Power transformer prognostics and health management using machine learning: A review and future directions. Machines, 13(2), 125, 2025. https://doi.org/10.3390/machines13020125.
10. S. S. M. Ghoneim and I. B. M. Taha, Comparative Study of Full and Reduced Feature Scenarios for Health Index Computation of Power Transformers. IEEE Access, 8, 181326-181339, 2020. https://doi.org/10.1109/ACCESS.2020.3028689
11. A. S. Mogos, X. Liang and C. Y. Chung, Enhancing transformer health index prediction using dissolved gas analysis data through integration of LightGBM and robust EM algorithms. IEEE Access, 12, 108472–108483, 2024. https://doi.org/10.1109/ACCESS.2 024.3439248.
12. N. Islam, R. Khan, S. K. Das, S. K. Sarker, M. M. Islam, M. Akter and S. M. Muyeen, Power transformer health condition evaluation: A deep generative model aided intelligent framework. Electric Power Systems Research, 218, 109201, 2023. https://doi.org/10.1016/ j.epsr.2023.109201.
13. S. T. Zahra, S. K. Imdad, S. Khan, S. Khalid and N. A. Baig, Power transformer health index and life span assessment: A comprehensive review of conventional and machine learning based approaches. Engineering Applications of Artificial Intelligence, 139, Part A, 109474, 2025. https://doi.org/10.1016/j.engappai.2024 .109474.
14. D. Rediansyah, R. A. Prasojo, Suwarno and A. Abu-Siada, Artificial intelligence-based power transformer health index for handling data uncertainty. IEEE Access, 9, 150637–150648, 2021. https://doi.org/10.1 109/ACCESS.2021.3125379.
15. S. H. Syed and V. Muralidharan, Feature extraction using discrete wavelet transform for fault classification of planetary gearbox: A comparative study. Applied Acoustics, 188, 108572, 2022. https://doi.org/10.1016/ j.apacoust.2021.108572.
16. N. Dhobale, S. S. Mulik and S. P. Deshmukh, Naïve Bayes and Bayes net classifier for fault diagnosis of end mill tool using wavelet analysis: A comparative study. Journal of Vibration Engineering & Technologies, 10, 1721–1735, 2022. https://doi.org/10.1007/s42417-022-00478-z.
17. Z. Abda, M. Chettih and B. Zerouali, Assessment of neuro-fuzzy approach based different wavelet families for daily flow rates forecasting. Modeling Earth Systems and Environment, 7, 1523–1538, 2021. https://doi.org/10.1007/s40808-020-00855-1 .
18. J. A. Domínguez-Navarro, T. B. Lopez-Garcia and S. M. Valdivia-Bautista, Applying wavelet filters in wind forecasting methods. Energies, 14(11), 3181, 2021. https://doi.org/10.3390/en14113181.
19. T. Guo, T. Zhang, E. Lim, M. López-Benítez, F. Ma and L. Yu, A review of wavelet analysis and its applications: Challenges and opportunities. IEEE Access, 10, 58869-58903, 2022. https://doi.org/10.1 109/ACCESS.2022.3179517.
20. A. M. Dahman, A. A. Abou El-Ela, M. I. Zaki and R. A. El Sehiemy, A proposed wavelet analysis based fault diagnosis scheme of power transformers using fault signatures and CT saturation. Results in Engineering, 105820, 2025.
21. Failure Analysis in Power Transformers, https://www.kaggle.com/code/sradha92/failure-analy sis-in-power-transformers, Accessed 15 August 2025.
22. İ. Öz, Comparative analysis of wavelet families in image compression, featuring the proposed new wavelet. Turkish Journal of Science and Technology, 19(1), 279–294, 2024. https://doi.org/10.55525/tjst.14 28424.
23. A. K. Karaev, O. S. Gorlova, V. V. Ponkratov, M. L. Sedova, N. S. Shmigol and M. L. Vasyunina, A comparative analysis of the choice of mother wavelet functions affecting the accuracy of forecasts of daily balances in the treasury single account. Economies, 10(9), 213, 2022.
24. N. Hussain, M. Hasanzade, D. W. Breiby and M. N. Akram, Performance comparison of wavelet families for noise reduction and intensity thresholding in fourier ptychographic microscopy. Optics Communications, 519, 128400, 2022.
25. F. Vatansever., F. Uysal and A. Uzun, Ayrık Dalgacık Dönüşümü ile Gürültü Süzme. 338-342, 2002. https://www.emo.org.tr/ekler/7841cc9e552bd5c_ek.pdf
26. Ç. K. Çavaş, Protein verilerinin ayrık dalgacık dönüşümü ile analizi. Bayburt Üniversitesi Fen Bilimleri Dergisi, 6(1), 20–29, 2023.  https://doi.org /10.55117/bufbd.1192229
27. C. Vonesch, T. Blu and M. Unser, Generalized Daubechies Wavelet Families. IEEE Transactions on Signal Processing, 55(9), 4415-4429, 2007, https:// doi.org/10.1109/TSP.2007.896255
28. X. Wang, G. Gong and N. Li, Automated recognition of epileptic EEG states using a combination of symlet wavelet processing, gradient boosting machine, and grid search optimizer. Sensors, 19(2), 219, 2019.
29. A. Antoniadis, Smoothing noisy data with coiflets, Statistica Sinica, 651–678, 1994.
30. R. Szewczyk, K. Grabowski, M. Napieralska, W. Sankowski, M. Zubert and A. Napieralski, A reliable iris recognition algorithm based on reverse biorthogonal wavelet transform. Pattern Recognition Letters, 33(8), 1019–1026, 2012.
31. S. Mallat, A Wavelet Tour of Signal Processing, Elsevier, 1999.
32. D. Domingo, A. B. Kareem, C. N. Okwuosa, P. M. Custodio and J. W. Hur, Transformer core fault diagnosis via current signal analysis with Pearson correlation feature selection. Electronics, 13(5), 926, 2024. https://doi.org/10.3390/electronics13050926.
33. V. Satopaa, J. Albrecht, D. Irwin and B. Raghavan, Finding a "Kneedle" in a Haystack: Detecting Knee Points in System Behavior. 31st International Conference on Distributed Computing Systems Workshops, pp. 166-171, Minneapolis, MN, USA, 2011. https://doi.org/10.1109/ICDCSW.2011.20 .
34. J. Wang, G. Li and W. Zhang, Combine-net: an improved filter pruning algorithm. Information, 12(7), 264, 2021. https://doi.org/10.3390/info12070264.
35. X. Chen, H. Cui and L. Luo, Fault Diagnosis of Transformer Based on Random Forest, 2011 Fourth International Conference on Intelligent Computation Technology and Automation, pp. 132-134, Shenzhen, China, 2011, https://doi.org/10.1109/ICICTA.2011.40
36. M. E. A. Senoussaoui, M. Brahami and I. Fofana, Transformer oil quality assessment using random forest with feature engineering. Energies, 14(7), 1809, 2021, https://doi.org/10.3390/en14071809.
37. D. Rediansyah, R. A. Prasojo and Suwarno, Study on Artificial Intelligence Approaches for Power Transformer Health Index Assessment. 2021 International Conference on Electrical Engineering and Informatics (ICEEI), pp. 1-4, Kuala Terengganu, Malaysia, 2021. https://doi.org/10.1109/ICEEI52609. 2021.9611109
38. G. Chintia, R. A. Prasojo and Suwarno, Power Transformer Insulation System Health Index with Missing Data Prediction using Random Forest. 2023 IEEE 3rd International Conference in Power Engineering Applications (ICPEA), pp. 5-8, Putrajaya, Malaysia, 2023, https://doi.org/10.1109/ICPEA56918. 2023.10093216 .