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.

Multisource Domain Feature Adaptation Network for Bearing Fault Diagnosis Under Time-Varying Working Conditions

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

A Hybrid Adversarial Domain Adaptation Network for Bearing Fault Diagnosis Under Varying Working Conditions

Non-Uniformly Weighted Multisource Domain Adaptation Network For Fault Diagnosis Under Varying Working Conditions

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

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

One-Dimensional Multi-Scale Domain Adaptive Network for Bearing-Fault Diagnosis under Varying Working Conditions

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

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

A knowledge dynamic matching unit-guided multi-source domain adaptation network with attention mechanism for rolling bearing fault diagnosis

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

A Multiple Feature Space Subdomain Adaptation Network for Bearing Fault Diagnosis

A Dual-View Alignment-Based Domain Adaptation Network for Fault Diagnosis

Multiweight Adversarial Open-Set Domain Adaptation Network for Machinery Fault Diagnosis With Unknown Faults

A domain adaptation method for bearing fault diagnosis using multiple incomplete source data

Dual-Weight Consistency-Induced Partial Domain Adaptation Network for Intelligent Fault Diagnosis of Machinery

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

Deep Transfer Network with Multi-Space Dynamic Distribution Adaptation for Bearing Fault Diagnosis

A Fine-Grained Feature Decoupling Based Multi-Source Domain Adaptation Network for Rotating Machinery Fault Diagnosis

Bearing fault diagnosis based on partial domain adaptation adversarial network

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