A review of artificial intelligence applications for fault detection in aerospace hydraulics systems

The integration of advanced technologies into aerospace systems, par- particularly within hydraulic systems such as landing gear mechanisms, has introduced new dimensions of complexity and vulnerability. This paper discusses the methods of fault detection from a historical perspective to the current state of the art adopted in industry for real-time detection of faults in hydraulic systems deployed in aerospace. A review of fault detection methods ranging from the manual method, the threshold method, the statistical method, the model-based method, the signal-based method, the knowledge-based method and time frequency analysis method, and the artificial intelligence method, Explainable AI (XAI) was explored and discussed in this paper. The review reiterated that while detection methods, such as manual inspection and threshold-based monitoring, are straightforward to implement, they fail to deliver precise results when detecting complex faults. Model-based and AI-driven advanced techniques enhance precision at the expense of demanding greater computational power and sufficient data availability.