Remaining Life Prediction Method for Rolling Bearings Based on RLMD-SCINet

In order to address the problem that existing deep learning methods are not sufficient for feature mining, which leads to low prediction accuracy, a remaining life prediction method combining robust locally mean decomposition with sample convolutional interaction network is proposed. Firstly, the Root Mean Square (RMS) is extracted from the bearing vibration signals as a health indicator reflecting the overall degradation performance of the bearings, and then the health indicators are decomposed by the Robust Locally Mean Decomposition (RLMD) so as to explore the potential information inside them, and then the decomposed components of the health indicators are imported into the SCINet network model for the prediction of the lifespan. Finally, the validation is carried out on the public dataset PHM2012 dataset and compared with the Long Short-Term Memory Network (LSTM) and Temporal Convolutional Network (TCN) models. The experimental results show that the proposed method can fully exploit the potential characteristics of bearings and effectively improve the accuracy of bearing remaining life prediction.