<ul><li>A novel framework, MCSFormer, is introduced for precise estimation of the Remaining Useful Life (RUL) of rolling bearings.</li><li>The framework combines wavelet-based denoising method, Wavelet Packet Decomposition (WPD), and a multi-channel Swin Transformer model with attention mechanisms for feature fusion.</li><li>MCSFormer outperformed state-of-the-art models in intra-condition experiments and demonstrated superior generalization in cross-condition testing on the PRONOSTIA dataset.</li><li>The model's focus on accurate early detection and customized loss function for differentiating early and late predictions makes it a reliable predictive maintenance tool for industrial applications.</li></ul>

Multi-Channel Swin Transformer Framework for Bearing Remaining Useful Life Prediction

Discover more