Explainable CNN–GRU learning on Mel-spectrogram acoustic signals for bearing fault diagnosis under small-sample experimental conditions

Emrah Aslan, Yıldırım Özüpak, Department of Computer Engineering, Faculty of Engineering, Dicle University, Silvan Vocational School, Dicle University

Proceedings of the Institution of Mechanical Engineers, Part G: Journal of Aerospace Engineering

Published online on July 23, 2026

Abstract

Proceedings of the Institution of Mechanical Engineers, Part G: Journal of Aerospace Engineering, Ahead of Print.
Reliable bearing fault diagnosis is essential for predictive maintenance of rotating machinery, particularly in applications where contact-based vibration sensors are difficult to install or maintain. This study proposes an explainable acoustic fault ...