Remaining useful life estimation of aeroengine based on CNN-BiLSTM

Abstract

As the core propulsion system of an aircraft, the safety and reliability of aeroengines are directly related to flight safety and operational efficiency. Modern health monitoring systems increasingly rely on electromagnetic sensing technologies and multi-source signal acquisition to provide reliable condition information under complex operating environments, making accurate remaining useful life prediction essential for predictive maintenance and intelligent decision-making. To improve the prediction accuracy of aeroengine remaining useful life, this paper proposes a hybrid model integrating convolutional neural networks and bidirectional long short-term memory networks. The convolutional neural network is first employed to automatically extract spatial features from multi-dimensional sensor data, after which the bidirectional long short-term memory network captures temporal dependencies and degradation trends throughout the engine operating process. The proposed framework enables end-to-end mapping from multi-state monitoring parameters to remaining useful life without requiring explicit degradation modeling or manual feature engineering. Validation on the CMAPSS benchmark dataset demonstrates that the proposed CNN-BiLSTM model consistently outperforms CNN, DCNN, RNN, and BiLSTM approaches in terms of RMSE and MAE, providing more accurate and robust prediction results for aeroengine systems operating under complex degradation conditions. The proposed method not only enhances predictive maintenance capability for intelligent propulsion systems but also offers valuable methodological support for multi-sensor information fusion and electromagnetic-enabled condition monitoring in advanced aerospace applications.