Vibration feature extraction using local temporal self-similarity for rolling bearing fault diagnosis

This paper presents a new method for rolling bearing fault diagnosis. The novel vibration feature extraction is learned with local temporal self-similarities (TSS) continuously from collected vibration signals. The bag-of-words (BoW) scheme is then employed for fault classification taking advantages of these features. We investigated the effectiveness of the framework on the publicly-available Case Western Reserve University (CWRU) data set. We also compare the method with state-of-the-art approaches. The result demonstrates excellent performance of the proposed method, outperforming those compared state-of-the-art approaches.