Prediction of the Remaining Useful Lifetime of Turbofan Aircraft Engines Using Machine Learning: A Comparative Study

Predictive maintenance is a proactive strategy crucial for ensuring the optimal functioning of critical machinery, particularly in industries such as aerospace where equipment failure can have significant ramifications. Leveraging data analytics, machine learning, and statistical modeling, predictive maintenance aims to forecast machinery failures, enabling timely intervention to prevent costly downtimes and enhance operational efficiency. This paper presents a comprehensive comparative investigation of various established machine learning algorithms, including Convolutional Neural Network (CNN), Long Short-Term Memory Network (LSTM), Support Vector Regression (SVR), and Random Forest, for predicting the Remaining Useful Lifetime (RUL) of turbofan aircraft engines. The five machine learning algorithms were selected for comparison, encompassing a spectrum of techniques ranging from traditional regression methods to advanced deep learning approaches. The primary objective was to compare the performance of deep learning techniques, represented by CNN and LSTM, against classical machine learning algorithms such as SVR and Random Forest. The experimental studies demonstrated that LSTM and CNN-LSTM surpassed traditional machine learning models in predicting RUL. LSTM had RMSE values between 15.10 and 31.39, while CNN-LSTM had RMSE values between 13.92 and 31.08, showing competitive performance.