Quality and Reliability Engineering International, Год журнала: 2025, Номер unknown

Опубликована: Фев. 18, 2025

ABSTRACT As the critical technology of prognostics and health management (PHM), remaining useful life (RUL) prediction has received much attention. Deep learning algorithms based on data‐driven stand out among various methods. However, convolutional neural networks/recurrent networks (CNNs/RNNs) suffer from design gaps, making it challenging to achieve parallel processing long‐time series data. In addition, most methods focus only point but lack uncertainty assessment results. Therefore, a recurrence multi‐information time‐frequency (RMTF) Transformer network is proposed in this paper, which kind slice memory mechanism (SMRM), with multiple feature extraction encoders, ability extract features. RMTF can realize effective fusion long time information, cross‐time period multi‐source information. Bayesian (BNN) variational inference used predict interval RUL. The advancedness our method verified compared by aero‐engines dataset tool wear dataset. particular, demonstrates significant advantages over mainstream advanced models under complex operating conditions two sub‐datasets CMAPSS, FD002, FD004. Specifically, for FD002 dataset, RMSE reduced 8.66%, SF 23.37%. For FD004 lowered 6.36%, decreased 13.83%. experimental results show that effectively RUL assess

Язык: Английский