Bearings Remaining Useful Life Prediction with Combinatorial Feature Extraction Method and Gated Recurrent Unit Network

Remaining useful life (RUL) prediction is one of the most important technologies to implement the health management and predictive maintenance of rotating machinery. To predict precisely the RUL, a three-stage strategy is proposed. Firstly, twenty-four basic characteristics are extracted from vibration signal, which are reconstructed by combining those basic characteristics with complete ensemble empirical mode decomposition with adaptive noise (BC-CEEMDAN), and then the trend curves are extracted to reduce the fluctuation. Next, the most sensitive features are selected by employing a linear combination of monotonicity and correlation criteria. Finally, by input the selected features into the gated recurrent unit (GRU) neural network, we achieve the efficient health indicator with BC-CEEMDAN-GRU. To verify the effectiveness of the proposed approach, experiment on PRONOSTIA bearing datasets is carried out, and the advantage is emphasized by comparison with the six existing methods.