Enhanced Prediction of Aeroengine Remaining Useful Life via Fuzzy‐Based Deep Feature Representation

ABSTRACT This study develops an enhanced remaining useful life (RUL) prediction framework that combines deep feature extraction with fuzzy membership function and quantum‐behaved particle swarm optimisation (QPSO) optimisation. In the developed scheme, original features are transformed into membership‐based representations with fuzzy clustering. By tuning fuzzy parameters, additional informative features are generated to better capture potential patterns within the data. To address potential feature redundancy, a feature weight matrix is introduced and an objective function is established to minimise prediction error and QPSO is employed to optimise the feature weights effectively. Experimental results on the widely used commercial modular aero‐propulsion system simulation dataset validate the effectiveness of the proposed method in improving prediction accuracy, offering a novel and practical solution for data‐driven RUL estimation.