A hybrid model for turbofan engines RUL prediction based on multi-feature Fusion

In order to solve the problems such as insufficient temporal series feature extraction, suboptimal performance of single type feature types in prediction, and inadequate fault capture and generalization by single models in remaining useful life (RUL) prediction, this paper proposes a multi-feature fusion based on hybrid model for aircraft engine fault prediction. This method initially employs a bidirectional long short-term memory network (Bi-LSTM) to extract deep temporal features, and introduces an attention mechanism to precisely capture highly influential features. Secondly, a multi-feature fusion framework is proposed, which integrates representative handcrafted features to enhance the representation of complex fault degradation and improve the accuracy of aircraft engine RUL prediction. Finally, a GRU model is used for regression prediction, with the prediction model and feature extraction model operating independently to further boost overall performance. The effectiveness of the proposed method is validated using the C-MAPSS datasets, demonstrating superior accuracy and stability compared to existing methods.