Advanced AI Technologies for Aero-Engine Fault Detection and Diagnosis

Advanced AI Technologies for Aero-Engine Fault Detection and Diagnosis

Farid Choobtashani1 Sajad Heidarian2 Morteza Karkhaneh3

1) Ph.D. Student in Aerospace Engineering, Department of Aerospace Engineering, Sharif University of Technology, Tehran, Iran.
2) Ph.D. Student in Computer Engineering, Faculty of Electrical and Computer Engineering, Razi University, Kermanshah, Iran.
3) M.Sc. Student in Aerospace Engineering, Faculty of Aerospace Engineering, Malek Ashtar University of Technology, Tehran, Iran.

IOI Code : ABEB-ACAGZB
Publication : 4th International Conference on Recent Advances in Engineering, Innovation and Technology - Belgium(eitconf.com)
Abstract :
    As critical propulsion components, aero-engines are vulnerable to mechanical and performance-related faults arising from high temperatures and pressures, vibration, and varying operating conditions. Early fault detection is essential for flight safety, reduced unplanned downtime, and effective maintenance planning. This review examines artificial intelligence (AI) techniques for aero-engine fault diagnosis, including conventional machine learning, convolutional neural networks, sequential models, XGBoost, and multimodal learning. It also discusses signal processing, feature extraction and selection, multisensory data fusion, explainable AI, and knowledge distillation. The reviewed studies indicate that adaptive feature selection and sensor fusion can help identify weak fault signatures, while compression of deep models may reduce computational requirements. Nevertheless, the scarcity of real failure data, class imbalance, discrepancies between laboratory and flight conditions, distribution shifts, and nonunique fault signatures constrain generalization. A synthesis of the literature suggests that integrating engine physics with data-driven algorithms, together with independent validation under representative operating conditions, offers a promising route toward reliable aero-engine fault diagnosis systems.

Keywords : artificial intelligence aero engine fault diagnosis deep learning health monitoring sensor fusion