Unscheduled maintenance, in-flight anomalies, and increasing emissions are the critical challenges that the aviation industry is experiencing in safety management, operation efficiency, and environmental sustainability. Old methods of reacting and isolated optimization systems do not tackle these intertwined problems as a whole. This study introduces an Integrated Aviation Predictive Analytics Framework, which uses artificial intelligence (AI) and machine learning (ML) to make simultaneous contributions to better safety, operational efficiency, and sustainability using eight core contributions. The framework applies: (1) an LSTM-based early-warning system detecting in-flight anomalies with 94.3% accuracy and 8.5 minutes lead time; (2) explainable AI (XAI) via SHAP/LIME to demonstrate clear decision-making; (3) Random Forest ensemble models to predictive maintenance and Remaining Useful Life (RUL); (4) federated learning to predict RUL across airlines, no data transfers; (5) intelligent maintenance scheduling to reduce Aircraft on Ground (AOG) time by 30%; The implemented framework is an interactive Streamlit dashboard with the ability to monitor results in real-time, which allows showing quantifiable results, with 35% fewer safety risks, 39.1% goals of AOG, 12% of costs in CO2 emission saved, and a cost reduction of 3.3M annually. The modular design of its architecture allows the system to seamlessly add predictive insights to the safety, maintenance, and sustainability spheres, offering aviation operators with the decision support they can take, yet ensuring the regulatory transparency of AI and privacy of data.
Lama Talal Khrais (Fri,) studied this question.