Transfer fault diagnosis of bearing installed in different machines using enhanced deep auto-encoder

• Published in: Measurement : journal of the International Measurement Confederation Vol. 152; p. 107393
• Main Authors: Zhiyi, He, Haidong, Shao, Lin, Jing, Junsheng, Cheng, Yu, Yang
• Format: Journal Article
• Language: English
• Published: London Elsevier Ltd 01.02.2020; Elsevier Science Ltd
• Subjects: Bearing; Bearing fault; Bearings; Coders; Coding theory; Different machines; Drift och underhållsteknik; Enhanced deep auto-encoder model; Fault diagnosis; Heat transfer; Limited labeled information; Operation and Maintenance Engineering; Parameter modification; Structural engineering; Training; Transfer diagnosis; Vibration; Vibration analysis; Vibration measurement; Enhanced deep auto-encoder model; Bearing fault; Different machines; Limited labeled information; Transfer diagnosis
• ISSN: 0263-2241; 1873-412X; 1873-412X
• DOI: 10.1016/j.measurement.2019.107393

[Display omitted] •Scaled exponential linear unit is used to normalize the raw vibration data.•Nonnegative constraint and correntropy are adopted to modify the loss function.•Parameter transfer is introduced into the enhanced deep auto-encoder.•Two transfer diagnosis cases confirm the feasibility of the proposed approach. The collected vibration data with labeled information from bearing is far insufficient in engineering practice, which is challenging for training an intelligent diagnosis model. For this purpose, enhanced deep transfer auto-encoder is proposed for fault diagnosis of bearing installed in different machines. First, scaled exponential linear unit is used to improve the quality of the mapped vibration data collected from bearing. Second, nonnegative constraint is adopted for modifying the loss function to improve reconstruction effect. Then, the parameter knowledge of the well-trained source model is transferred to the target model. Finally, target training samples with limited labeled information are adopted for fine-tuning the target model to match the characteristics of the target testing samples. The proposed approach is applied for analyzing the measured vibration signals of bearings installed in different machines. The analysis results show that the proposed approach holds better transfer diagnosis performance compared with the existing approaches.