Power Transformer Fault Diagnosis and Prediction Based on Deep Learning
DOI:
https://doi.org/10.67541/jdf2606Keywords:
Power transformer; Fault diagnosis; Remaining life prediction; Generative adversarial network; Temporal convolutional networkAbstract
Power transformers assume the core functions of power conversion and power flow regulation in the new power system, and its fault outage will lead to the chain risk of power grid. However, the diagnostic accuracy of traditional threshold methods and shallow machine learning is limited in the face of non‑stationary and strong‑noise1 signals, and the data distribution deviation caused by variable load conditions seriously restricts the field generalization ability of the model. To solve the above bottleneck, this paper proposes a complete technical system that integrates multi‑domain feature enhancement, dual‑branch domain‑adaptive diagnosis, and temporal adversarial prediction. In the preprocessing stage, an adaptive denoising strategy based on improved variational mode decomposition and permutation entropy is designed, and the time‑frequency‑energy 3D multi‑channel input tensor is constructed by using the synchrosqueezed S‑transform (SSST) combined with kurtosis, peak factor, and energy spectrum entropy, which effectively solves the problem of poor separability of fault features under strong noise. In the diagnosis stage, a dual‑branch adversarial domain‑adaptive convolutional network is proposed. With dilated EfficientNet as the backbone, global semantic and local sensitivity dual branches are set to extract wideband and transient features in parallel, and the dual‑domain adaptation mechanism of maximum mean discrepancy and an adversarial discriminator is introduced to achieve high‑precision diagnosis in the unlabeled target domain under variable working conditions. The experimental results show that the accuracy of the model reaches 94.7% in cross‑operating‑condition scenarios, which is 8.2 percentage points higher than that of the traditional domain‑adversarial network. In the prediction stage, a temporal generative adversarial network is constructed, which takes dilated causal convolution and self‑attention spatio‑temporal gating as its core, combines Monte Carlo dropout interval estimation guided by dynamic time warping morphological loss and discriminator confidence, and realizes probabilistic prediction of the remaining useful life (RUL), and the prediction interval coverage rate reaches 88.6%. The method in this paper has achieved significant performance improvement in the three dimensions of diagnosis accuracy, prediction reliability and computational efficiency, and provides complete technical support for the intelligent operation and maintenance of power transformers from state recognition to trend prediction.
References
[1] Shi, Z., Wu, G., Xin, D., Liu, K., Wei, S., Li, Y., ... & Gao, G. (2025). Insulation Characteristics and Failure Mechanisms of Ultra-High Voltage DC Bushings: A Review. IEEE Transactions on Dielectrics and Electrical Insulation. https://doi.org/10.1109/TDEI.2025.3623569
[2] Carlo Montanari, G., Morshuis, P., Seri, P., & Ghosh, R. (2020). Ageing and reliability of electrical insulation: the risk of hybrid AC/DC grids. High Voltage, 5(5), 620-627. https://doi.org/10.1049/hve.2019.0371
[3] Zhao, T., Chen, G., Suraphee, S., et al. (2025). A hybrid TCN-XGBoost model for agricultural product market price forecasting. PLoS One, 20(5), e0322496. https://doi.org/10.1371/journal.pone.0322496
[4] Zhao, T., Chen, G., Pang, C., & Busababodhin, P. (2025). Application and performance optimization of SLHS-TCN-XGBoost model in power demand forecasting. Comput. Model. Eng. Sci, 143(3), 2883-2917. http://dx.doi.org/10.32604/cmes.2025.066442
[5] Feigl, M., Herrnegger, M., & Schulz, K. (2026). Distilling hydrological and land-surface model parameters from physio-geographical properties using text-generating AI. Nature Water, 4(2), 158-168. https://doi.org/10.1038/s44221-026-00583-3
[6] Wang, Z., Zhang, Z., Su, T., Ding, Z., & Zhao, T. (2024). Research on supply chain network optimisation based on the CNNs-BiLSTM model. In 2024 ITCEM (pp. 197-202). IEEE. https://doi.org/10.1109/ITCEM65710.2024.00044
[7] Zhao, T., Chen, G., Pang, C., et al. (2025). Hybrid convolutional neural network-graph attention network-gradient boosting decision tree model for seismic impedance inversion prediction. J Seismic Explor, 34(5), 81-98. https://doi.org/10.36922/JSE025310051
[8] Tran-Ngoc, H., Le Van, V., Nguyen Duc, L., Tran The, H., & Bui-Tien, T. (2026). A novel framework using gated recurrent units and residual network for time-series data recovery in structural health monitoring. European Journal of Environmental and Civil Engineering, 30(1), 1-28. https://doi.org/10.1080/19648189.2025.2551978
[9] Liu, Y., Gong, M., Chen, G., & Zhao, T. (2025). CNN-LSTM-based spatio-temporal prediction model for photovoltaic power generation. In Proceedings of the 2025 2nd International Conference on Big Data and Digital Management (pp. 570-575). https://doi.org/10.1145/3768801.3768892
[10] Tran-Ngoc, H., Le Van, V., Nguyen Duc, L., Tran The, H., & Bui-Tien, T. (2026). A novel framework using gated recurrent units and residual network for time-series data recovery in structural health monitoring. European Journal of Environmental and Civil Engineering, 30(1), 1-28. https://doi.org/10.1080/19648189.2025.2551978
[11] Zhao, T., Chen, G., & Busababodhin, P. (2026). An intelligent expert system for logistics disruption prediction and mitigation in global supply chains. Scientific Reports. https://doi.org/10.1038/s41598-026-61617-0
[12] Wang, Z., Wan, Y., Chen, G., et al. (2025). Multiscale Time Series Prediction of Crude Oil Prices Based on Transformer and CNN. In International Conference on Artificial Intelligence and Communication Technology (pp. 430-443). Springer. https://doi.org/10.1007/978-3-032-06372-4_36
[13] Zhang, Z., Qiu, T., Liao, Z., et al. (2025). Research on global power equipment trade prediction and logistics optimization based on 5G and edge computing. In ICOIT 2025 (Vol. 13665, pp. 302-310). SPIE. https://doi.org/10.1117/12.3071144
[14] Yan, J., Zhao, T., Pang, C., & Chen, G. (2025). Global crude oil trade flow prediction based on graph neural networks. In Proceedings of the 2025 2nd International Conference on Big Data and Digital Management (pp. 564-569). https://doi.org/10.1145/3768801.3768891
[15] He, C., Cheng, X., Xiong, L., Xu, X., Yu, L., & Yang, Z. (2025). A novel maximum impulse amplitude deconvolution for rolling bearing weak fault identification enhanced by spectrum information guided VMD. Mechanical Systems and Signal Processing, 237, 112973. https://doi.org/10.1016/j.ymssp.2025.112973
[16] Xia, S., Yang, J., Cai, W., Zhang, C., Hua, L., & Zhou, Z. (2021). Adaptive complex variational mode decomposition for micro-motion signal processing applications. Sensors, 21(5), 1637. https://doi.org/10.3390/s21051637
[17] Zhou, X., Li, Y., Jiang, L., & Zhou, L. (2021). Fault feature extraction for rolling bearings based on parameter-adaptive variational mode decomposition and multi-point optimal minimum entropy deconvolution. Measurement, 173, 108469. https://doi.org/10.1016/j.measurement.2020.108469
[18] Ur Rehman, N., & Aftab, H. (2019). Multivariate variational mode decomposition. IEEE Transactions on signal processing, 67(23), 6039-6052. https://doi.org/10.1109/TSP.2019.2951223
[19] Guo, Y., & Zhang, Z. (2021). Generalized variational mode decomposition: A multiscale and fixed-frequency decomposition algorithm. IEEE Transactions on Instrumentation and Measurement, 70, 1-13. https://doi.org/10.1109/TIM.2021.3076569
[20] Huang, Z. L., Zhang, J., Zhao, T. H., & Sun, Y. (2015). Synchrosqueezing S-transform and its application in seismic spectral decomposition. IEEE Transactions on Geoscience and Remote Sensing, 54(2), 817-825. https://doi.org/10.1109/TGRS.2015.2466660
[21] Sun, X., Yang, Y., Chen, C., Tian, M., Du, S., & Wang, Z. (2025, January). A multi-branch convolution and dynamic weighting method for bearing fault diagnosis based on acoustic–vibration information fusion. In Actuators (Vol. 14, No. 1, p. 17). MDPI. https://doi.org/10.3390/act14010017
[22] Li, J., Ye, Z., Gao, J., Meng, Z., Tong, K., & Yu, S. (2024). Fault transfer diagnosis of rolling bearings across different devices via multi-domain information fusion and multi-kernel maximum mean discrepancy. Applied Soft Computing, 159, 111620. https://doi.org/10.1016/j.asoc.2024.111620
[23] Yu, X., Wang, Y., Liang, Z., Shao, H., Yu, K., & Yu, W. (2023). An adaptive domain adaptation method for rolling bearings’ fault diagnosis fusing deep convolution and self-attention networks. IEEE Transactions on Instrumentation and Measurement, 72, 1-14. https://doi.org/10.1109/TIM.2023.3246494
[24] Tang, Z., Hou, X., Huang, X., Wang, X., & Zou, J. (2024). Domain adaptation for bearing fault diagnosis based on SimAM and adaptive weighting strategy. Sensors, 24(13), 4251. https://doi.org/10.3390/s24134251
[25] Şevik, U., & Mutlu, O. (2025). Automated multi-class classification of retinal pathologies: A deep learning approach to unified ophthalmic screening. Diagnostics, 15(21), 2745. https://doi.org/10.3390/diagnostics15212745
[26] Golkarieh, A., Afjehsoleymani, B., Kiashemshaki, K., & Boroujeni, S. R. (2026). Advanced deep learning techniques for classifying dental conditions using panoramic X-ray images. BMC Oral Health, 26(1), 472. https://doi.org/10.1186/s12903-026-07727-7
[27] Zhao, T., Chen, G., Pang, C., Seenoi, P., Papukdee, N., & Busababodhin, P. (2025). Time-lapse earthquake difference prediction based on physics-informed long short-term memory coupled with interpretability boosting. Journal of Seismic Exploration, 34(3), 25.https://doi.org/10.36922/JSE025310049
[28] Chen, G., Zhao, T., Pang, C., & Busababodhin, P. (2026). Integrated CNN–LSTM–XGBoost hybrid model predicts shale oil seismic attributes and global oil price trends. Scientific Reports,16, 25588 (2026). https://doi.org/10.1038/s41598-026-55910-1
[29] Mundu, M. M., Sempewo, J. I., Goparaju, A., & Uti, D. E. (2026). Comparative analysis of model evaluation metrics in energy systems, environmental modeling, and sustainability science. International Journal of Energy Research, 2026(1), 6170467. https://doi.org/10.1155/er/6170467
[30] Zhao, T., & Xu, H. (2025). A Novel STCA-LightGBM-SVR Hybrid Model for Port Cargo Throughput Prediction. In Artificial Intelligence and Applications. https://doi.org/10.47852/bonviewAIA62027055
[31] Shi, Z., Liang, H., & Dinavahi, V. (2017). Direct interval forecast of uncertain wind power based on recurrent neural networks. IEEE Transactions on Sustainable Energy, 9(3), 1177-1187. https://doi.org/10.1109/TSTE.2017.2774195
[32] Quan, H., Srinivasan, D., & Khosravi, A. (2014). Uncertainty handling using neural network-based prediction intervals for electrical load forecasting. Energy, 73, 916-925. https://doi.org/10.1016/j.energy.2014.06.104
[33] Wang, X., Kang, Y., Petropoulos, F., & Li, F. (2022). The uncertainty estimation of feature-based forecast combinations. Journal of the Operational Research Society, 73(5), 979-993. https://doi.org/10.1080/01605682.2021.1880297
Published
Data Availability Statement
The data that support the findings of this study are available upon request from the corresponding authors, H.Y.
Issue
Section
License
Copyright (c) 2026 Haoyu You, Longtai Hua, Tianyi Su (Author)
This work is licensed under a Creative Commons Attribution 4.0 International License.
