AI-Powered Grid Resilience: Hybrid CNN-Transformer for Predictive Fault Detection and Real-Time Optimization

Author (s)

Gibrilla Deen Kamara, Saio Alusine Marrah, Abu Bakarr Koroma, Nyallay Sheku, Mabel Ernestine Cole, Samuel Idriss Kargbo, & Eric Komba Foyoh Mani

Abstract

Modern power grids are increasingly complex due to growing integration of renewable energy sources, power‑electronic devices, and bidirectional power flows challenging traditional fault detection and stability monitoring methods. In this study, we propose a hybrid convolutional neural networks -sCNN‑Transformer model for predictive fault detection and real‑time grid optimization. Using synthetic time‑series data representing normal operation and a variety of fault conditions (including amplitude spikes, harmonics, and voltage sag/swell), we perform four comprehensive analyses: time‑domain waveform and spectrogram inspection; frequency‑ and harmonic‑domain analysis; deep model feature and attention‑map visualization; and grid‑level stability simulation (load, voltage, frequency). Our results show the hybrid model successfully distinguishes faults from normal conditions, separates different fault types in latent space, and yields robust classification performance under noisy and distorted signals. The study demonstrates the feasibility of combining spatial feature extraction and temporal sequence modeling for smart‑grid fault detection, and highlights the potential for real‑time monitoring and proactive grid management. Future work will target real‑world grid datasets, renewable‑integration scenarios, and extension to multi‑class/multi‑fault localization tasks.

Key words: Smart grid, Fault detection, Hybrid CNN‑Transformer, Time-frequency analysis, Power system resilience.

Download PDF

Cite This Article:

Kamara, G. D., Marrah, S. A., Koroma, A. B., Sheku, N., Cole, M. E., Kargbo, S. I., & Mani, E. K. F. (2026). AI-Powered Grid Resilience: Hybrid CNN-Transformer for Predictive Fault Detection and Real-Time Optimization, Journal of Scientific Reports, 12(1), 163-176.  DOI: https://doi.org/10.58970/JSR.1162

About Author (s)

Gibrilla Deen Kamara, School of Computer Science and Engineering, University of Electronic Science and Technology of China (UESTC), Chengdu, Sichuan 611731, China. & Faculty of Information Technology, Limkokwing University of Creative Technology, Freetown, Sierra Leone. ORCID: https://orcid.org/0009-0005-1732-6407

Saio Alusine Marrah (Corresponding Author), Faculty of Information Technology, Limkokwing University of Creative Technology, Freetown, Sierra Leone. & School of Software Engineering, University of Electronic Science and Technology of China (UESTC), Chengdu, Sichuan 611731, China. ORCID: https://orcid.org/0009-0008-4727-2480

Abu Bakarr Koroma, School of Economics and Management (SEM), University of Electronic Science and Technology of China (UESTC), Chengdu, Sichuan 611731, China. & Faculty of Business and Entrepreneurship, Ernest Bai Koroma University of Science and Technology, Sierra Leone. ORCID: https://orcid.org/0009-0005-3613-104X

Nyallay Sheku, College of Software Engineering, Nankai University, Tianjin 300350, China. ORCID: https://orcid.org/0009-0000-9885-7261

Mabel Ernestine Cole, Faculty of Information Technology, Limkokwing University of Creative Technology, Freetown, Sierra Leone. & School of Economics and Management (SEM), University of Electronic Science and Technology of China (UESTC), Chengdu, Sichuan 611731, China. ORCID: https://orcid.org/0009-0002-1956-4474

Samuel Idriss Kargbo, College of Software Engineering, Nankai University, Tianjin 300350, China. ORCID: https://orcid.org/0009-0005-4264-0801

Eric Komba Foyoh Mani, Faculty of Business and Entrepreneurship Studies, Department of Social Work, Eastern Technical University of Sierra Leone, Sierra Leone. ORCID: https://orcid.org/0009-0001-1934-0668

Download PDF

DOI: https://doi.org/10.58970/JSR.1162

Summarize this article with: