Remaining useful life prediction across different operational conditions based on domain adaptation feature transfer

To address the challenge of data distribution discrepancy in cross-condition remaining useful life (RUL) prediction, this paper proposes a domain adaptation feature transfer framework. The framework integrates the ensemble empirical mode decomposition (EEMD) with deep denoising autoencoders (DAE) to extract degradation-sensitive features. Subsequently, transfer component analysis (TCA) is adopted to align the feature distributions between the source and target domains. Experimental results on the IEEE PHM 2012 bearing dataset demonstrate that the proposed method outperforms traditional statistical features and existing domain adaptation approaches. Specifically, the root mean square error (RMSE) and mean absolute percentage error (MAPE) are reduced by 21.3% and 39.1%, respectively. Further analysis verifies the robustness of the proposed method under noisy and small-sample scenarios, which highlights its potential application prospects in industrial fields such as wind turbine gearbox monitoring and aircraft engine maintenance.