Deep transfer multi-wavelet auto-encoder for intelligent fault diagnosis of gearbox with few target training samples

• Published in: Knowledge-based systems 2020-03, Vol.191, p.105313, Article 105313
• Main Authors: He, Zhiyi, Shao, Haidong, Wang, Ping, Lin, Janet (Jing), Cheng, Junsheng, Yang, Yu
• Format: Article
• Language: English
• Subjects: Coders; Deep transfer multi-wavelet auto-encode; Drift och underhållsteknik; Fault diagnosis; Few target training samples; Gearbox fault; Gearboxes; Knowledge acquisition; Operation and Maintenance; Training; Transfer diagnosis; Variable working conditions; Wavelet analysis
• ISSN: 0950-7051; 1872-7409; 1872-7409
• DOI: 10.1016/j.knosys.2019.105313

Lack of typical fault samples remains a huge challenge for intelligent fault diagnosis of gearbox. In this paper, a novel approach named deep transfer multi-wavelet auto-encoder is presented for gearbox intelligent fault diagnosis with few training samples. Firstly, new-type deep multi-wavelet auto-encoder is designed for learning important features of the collected vibration signals of gearbox. Secondly, high-quality auxiliary samples are selected based on similarity measure to well pre-train a source model sharing similar characteristics with the target domain. Thirdly, parameter knowledge acquired from the source model is transferred to target model using very few target training samples. Transfer diagnosis cases for different fault severities and compound faults of gearbox confirm the feasibility of the proposed approach even if the working conditions have significant changes. •Multi-wavelet is used to design new-type deep auto-encoder.•Similarity measure is used to select high-quality auxiliary samples from source domain.•Parameter knowledge is transferred using very few target training samples.•Transfer diagnosis cases of fault severities and compound faults are used for verification.