A Systematic Literature Review on AI-Driven Predictive Maintenance and Fault Detection in Aircraft Systems

The increasing availability of onboard sensors and digital monitoring platforms has enabled the continuous acquisition of operational and health-related data in aircraft systems. In parallel, advances in Big Data analytics and Artificial Intelligence (AI) have driven significant progress in Predictive Maintenance (PdM), enabling earlier fault detection and more reliable estimations of Remaining Useful Life (RUL). This systematic literature review examines recent developments in AI-driven PdM and fault detection applied to aircraft over the last years. A total of 20 studies were selected based on predefined inclusion criteria and analyzed with respect to research trends, application domains, algorithmic approaches, and expected outputs. The findings indicate a strong research emphasis on civil aviation supported by accessible operational datasets, whereas military aviation research prioritizes fleet readiness and mission continuity, often with limited data transparency. Deep learning approaches, particularly hybrid models combining convolutional and recurrent architectures, dominate recent prognostic methodologies, while optimization and Model-Based Systems Engineering (MBSE) frameworks support decision-making integration. Despite these advancements, the transition from experimental models to operational deployment remains constrained by data heterogeneity, model explainability requirements, and regulatory certification processes. This review highlights current progress and identifies gaps and research opportunities to accelerate the adoption of robust and scalable PdM solutions in aviation.