Abstract:Variations in bearing operating conditions may cause a shift in the feature distribution of fault signals, weakening the generalization ability of the fault diagnosis model. Traditional methods with domain adaptation extract invariant cross-domain fault features, but they are mostly applied to a single-source domain. In the case of multi-source cross-domain transfer, however, determining which source domain has a better transfer effect to the target domain is difficult. In this paper, we aim to combine the transferable knowledge contained in all available source domains to improve the transfer performance. Low-correlation features are often more prone to bias under variable working conditions. The contribution of high-correlation features to the model must be increased. Accordingly, a multi-source alignment domain adaptation network with similarity measurement (MADASM) is proposed for this issue. First, domain adaptation is performed for each combination of source and target domains, and similarity measurement is introduced to constrain the similarity between the input and the central features. Second, the learned features are inputted into the corresponding domain-specific classifiers, of which the classifier discrepancy loss is then introduced to minimize the difference between the outputs of each domain-specific classifier, resulting in consistent prediction results for classifiers on the target domain. Finally, the average of output probabilities from all the classifier is calculated as the diagnosis result. Experimental results show that MADASM can fully utilize fault category information in multiple source domains to extract fault features with high correlation even in challeging senarios with unknown data labels and lacking of prior knowledge in target domain.

Multi-source subdomain negative transfer suppression and multiple pseudo-labels guidance alignment: A method for fault diagnosis under cross-working conditions

Multi-source Alignment Domain Adaptation with Similarity Measurement for Cross-domain Bearing Fault Diagnosis

Adversarial multi-domain adaptation for machine fault diagnosis with variable working conditions

Dual-weight Attention-based Multi-source Multi-stage Alignment Domain Adaptation for Industrial Fault Diagnosis

A new multi-source information domain adaption network based on domain attributes and features transfer for cross-domain fault diagnosis

Multi-source Partial Domain Adaptation Method Based on Pseudo-Balanced Target Domain for Fault Diagnosis

Intelligent fault diagnosis of double-aligned domain adaptation based on multi-structure fusion and multi-attention mechanism

Adversarial Domain Adaptation with Classifier Alignment for Cross-Domain Intelligent Fault Diagnosis of Multiple Source Domains

A multi-source information transfer learning method with subdomain adaptation for cross-domain fault diagnosis

A New Multisource Domain Bearing Fault Diagnosis Method With Adaptive Dual-Domain Obfuscation Weighting Strategy

Multisource cross-domain fault diagnosis of rolling bearing based on subdomain adaptation network

Multisource Domain Factorization Network for Cross-Domain Fault Diagnosis of Rotating Machinery: an Unsupervised Multisource Domain Adaptation Method

Multi-levels Domain Alignment Adaptation-Based Transfer Fault Diagnosis Method for Different Machines

A Balanced and Weighted Alignment Network for Partial Transfer Fault Diagnosis

Domain Adaptation for Intelligent Fault Diagnosis under Different Working Conditions

Multi-source Unsupervised Domain Adaptation for Machinery Fault Diagnosis under Different Working Conditions

Multisource-Refined Transfer Network for Industrial Fault Diagnosis under Domain and Category Inconsistencies

Multi-source Domain Self-supervised Enhanced Transfer Fault Diagnosis Approach with Source Sample Refinement Strategy

A Progressive Multi-Source Domain Adaptation Method for Bearing Fault Diagnosis

Duplex Adversarial Domain Discriminative Network for Cross-Domain Partial Transfer Fault Diagnosis

A Cross-Domain Intelligent Fault Diagnosis Method Based on Multi-source Domain Feature Adaptation and Selection