Advanced Multi-Model Prediction of Aircraft Engine Remaining Useful Life with Random Sampling-Based Class Balancing and Voting-Based Features Selection

In Industry 4.0, predictive maintenance for critical systems like aircraft engines relies on accurate Remaining Useful Life (RUL) estimation to prevent unexpected failures and optimize maintenance schedules. However, existing models face several limitations that hinder their effectiveness in real-world applications. Common challenges include data imbalance, which can lead to biased predictions; suboptimal feature selection, which may overlook important predictive variables; and limitations in model accuracy, particularly when handling complex, high-dimensional datasets. These issues often reduce the reliability and generalizability of RUL predictions across varied operational contexts. To address these challenges, this study proposes a novel data-driven approach that integrates random-sampling class balancing, a voting-based feature selection method, and advanced machine learning techniques, including ensemble and deep learning models. Applied to the Commercial Modular Aero-Propulsion System Simulation dataset, our model demonstrates significant improvements in RUL prediction, achieving Root Mean Square Error values of 11.10% for FD001, 12.09% for FD002, 11.80% for FD003, and 12.88% for FD004. These results highlight the model’s robustness and its potential to enhance predictive maintenance in aeronautical engineering.