Intelligent cross-machine fault diagnosis approach with deep auto-encoder and domain adaptation

• Published in: Neurocomputing (Amsterdam) Vol. 383; pp. 235 - 247
• Main Authors: Li, Xiang, Jia, Xiao-Dong, Zhang, Wei, Ma, Hui, Luo, Zhong, Li, Xu
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 28.03.2020
• Subjects: Auto-encoder; Deep learning; Fault diagnosis; Model generalization; Rolling bearing; Deep learning; Fault diagnosis; Auto-encoder; Model generalization; Rolling bearing
• DOI: 10.1016/j.neucom.2019.12.033
• ISSN: 0925-2312

•The challenging cross-machine transfer learning problem in fault diagnosis is investigated.•The machine-invariant features are extracted using deep auto-encoder, and domain adaptation is used for feature alignment.•The practical scenarios in fault diagnosis are considered where only the target-machine data in healthy state are available .•Different fault locations and severities are both considered in the cross-machine fault diagnosis.•Experiments on three rotating machinery datasets validate the effectiveness of the proposed method. Recently, due to the rising industrial demands for intelligent machinery fault diagnosis with strong generalization, transfer learning techniques have been used to enhance adaptability of data-driven approaches. Particularly, the domain shift problem where training and testing data are sampled from different operating conditions of the same machine is well addressed. However, it is still difficult to prepare sufficient labeled data on the tested machine. Therefore, the idea of transferring fault diagnosis knowledge learned from one machine to different but related machines is motivated, and that is realized through a deep learning-based method in this paper. Features of different equipments are first projected into the same subspace using an auto-encoder structure, and cross-machine adaptation algorithm is adopted for knowledge generalization, where the distribution discrepancy between data from different machines is minimized. Experiments on three rolling bearing datasets are implemented to validate the proposed method. The results suggest it is feasible to transfer fault diagnosis knowledge across different machines, and the proposed method offers a novel and promising approach for knowledge generalization.