Dynamic Distribution Adaptation Based Transfer Network for Cross Domain Bearing Fault Diagnosis

• Published in: Chinese journal of mechanical engineering Vol. 34; no. 1; pp. 1 - 10
• Main Authors: Liao, Yixiao, Huang, Ruyi, Li, Jipu, Chen, Zhuyun, Li, Weihua
• Format: Journal Article
• Language: English
• Published: Singapore Springer Singapore 01-12-2021; Springer Nature B.V; SpringerOpen
• Edition: English ed.
• Subjects: Adaptation; Cross domain fault diagnosis; Diagnosis & Prognosis; Domains; Dynamic distribution adaptation; Electrical Machines and Networks; Electronics and Microelectronics; Engineering; Engineering Thermodynamics; Fault diagnosis; Heat and Mass Transfer; Instance-weighted dynamic MMD; Instrumentation; Intelligent Manufacturing Technology; Machines; Manufacturing; Mechanical Engineering; Original Article; Power Electronics; Processes; Special Issue on AI-Enabled Monitoring; Theoretical and Applied Mechanics; Transfer learning; Dynamic distribution adaptation; Instance-weighted dynamic MMD; Transfer learning; Cross domain fault diagnosis
• ISSN: 1000-9345; 2192-8258
• DOI: 10.1186/s10033-021-00566-3

In machinery fault diagnosis, labeled data are always difficult or even impossible to obtain. Transfer learning can leverage related fault diagnosis knowledge from fully labeled source domain to enhance the fault diagnosis performance in sparsely labeled or unlabeled target domain, which has been widely used for cross domain fault diagnosis. However, existing methods focus on either marginal distribution adaptation (MDA) or conditional distribution adaptation (CDA). In practice, marginal and conditional distributions discrepancies both have significant but different influences on the domain divergence. In this paper, a dynamic distribution adaptation based transfer network (DDATN) is proposed for cross domain bearing fault diagnosis. DDATN utilizes the proposed instance-weighted dynamic maximum mean discrepancy (IDMMD) for dynamic distribution adaptation (DDA), which can dynamically estimate the influences of marginal and conditional distribution and adapt target domain with source domain. The experimental evaluation on cross domain bearing fault diagnosis demonstrates that DDATN can outperformance the state-of-the-art cross domain fault diagnosis methods.