Episodic Training and Feature Orthogonality-Driven Domain Generalization for Rotating Machinery Fault Diagnosis Under Unseen Working Conditions
In recent years, domain generalization-based fault diagnosis (DGFD) methods have shown significant potential in rotating machinery fault diagnosis in unseen target domains. However, these methods focus on learning domain-invariant representations via feature distribution adaptation. The generalizati...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-06-01
|
| Series: | Machines |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2075-1702/13/7/563 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849733442320728064 |
|---|---|
| author | Yixiao Liao Songbin Zhou Yisen Liu Kunkun Pang Jing Li Chang Li Lulu Zhao |
| author_facet | Yixiao Liao Songbin Zhou Yisen Liu Kunkun Pang Jing Li Chang Li Lulu Zhao |
| author_sort | Yixiao Liao |
| collection | DOAJ |
| description | In recent years, domain generalization-based fault diagnosis (DGFD) methods have shown significant potential in rotating machinery fault diagnosis in unseen target domains. However, these methods focus on learning domain-invariant representations via feature distribution adaptation. The generalization of classifiers and the orthogonality between fault-related and domain-related features have not been thoroughly explored, which hinders further improvements in DGFD performance. To address these limitations, an episodic training and feature orthogonality-driven domain generalization (EODG) method is proposed. In this method, episodic training is introduced to jointly improve the generalization capabilities of both the feature extractor and fault classifier, while a novel feature transfer loss is proposed for learning domain-invariant representations. Furthermore, the orthogonality between fault-related and domain-related features is enhanced by minimizing their cosine similarity, thereby improving the generalization capability of the DGFD model. The experimental results validated the effectiveness and superiority of the proposed method on domain generalization-based fault diagnosis tasks. |
| format | Article |
| id | doaj-art-aef40a8f7bd5488297c0da5cabdf85f8 |
| institution | DOAJ |
| issn | 2075-1702 |
| language | English |
| publishDate | 2025-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Machines |
| spelling | doaj-art-aef40a8f7bd5488297c0da5cabdf85f82025-08-20T03:08:01ZengMDPI AGMachines2075-17022025-06-0113756310.3390/machines13070563Episodic Training and Feature Orthogonality-Driven Domain Generalization for Rotating Machinery Fault Diagnosis Under Unseen Working ConditionsYixiao Liao0Songbin Zhou1Yisen Liu2Kunkun Pang3Jing Li4Chang Li5Lulu Zhao6Institute of Intelligent Manufacturing, Guangdong Academy of Sciences, Guangzhou 510070, ChinaInstitute of Intelligent Manufacturing, Guangdong Academy of Sciences, Guangzhou 510070, ChinaInstitute of Intelligent Manufacturing, Guangdong Academy of Sciences, Guangzhou 510070, ChinaInstitute of Intelligent Manufacturing, Guangdong Academy of Sciences, Guangzhou 510070, ChinaInstitute of Intelligent Manufacturing, Guangdong Academy of Sciences, Guangzhou 510070, ChinaInstitute of Intelligent Manufacturing, Guangdong Academy of Sciences, Guangzhou 510070, ChinaInstitute of Intelligent Manufacturing, Guangdong Academy of Sciences, Guangzhou 510070, ChinaIn recent years, domain generalization-based fault diagnosis (DGFD) methods have shown significant potential in rotating machinery fault diagnosis in unseen target domains. However, these methods focus on learning domain-invariant representations via feature distribution adaptation. The generalization of classifiers and the orthogonality between fault-related and domain-related features have not been thoroughly explored, which hinders further improvements in DGFD performance. To address these limitations, an episodic training and feature orthogonality-driven domain generalization (EODG) method is proposed. In this method, episodic training is introduced to jointly improve the generalization capabilities of both the feature extractor and fault classifier, while a novel feature transfer loss is proposed for learning domain-invariant representations. Furthermore, the orthogonality between fault-related and domain-related features is enhanced by minimizing their cosine similarity, thereby improving the generalization capability of the DGFD model. The experimental results validated the effectiveness and superiority of the proposed method on domain generalization-based fault diagnosis tasks.https://www.mdpi.com/2075-1702/13/7/563fault diagnosisrotating machinerydomain generalizationepisodic trainingfeature orthogonality |
| spellingShingle | Yixiao Liao Songbin Zhou Yisen Liu Kunkun Pang Jing Li Chang Li Lulu Zhao Episodic Training and Feature Orthogonality-Driven Domain Generalization for Rotating Machinery Fault Diagnosis Under Unseen Working Conditions Machines fault diagnosis rotating machinery domain generalization episodic training feature orthogonality |
| title | Episodic Training and Feature Orthogonality-Driven Domain Generalization for Rotating Machinery Fault Diagnosis Under Unseen Working Conditions |
| title_full | Episodic Training and Feature Orthogonality-Driven Domain Generalization for Rotating Machinery Fault Diagnosis Under Unseen Working Conditions |
| title_fullStr | Episodic Training and Feature Orthogonality-Driven Domain Generalization for Rotating Machinery Fault Diagnosis Under Unseen Working Conditions |
| title_full_unstemmed | Episodic Training and Feature Orthogonality-Driven Domain Generalization for Rotating Machinery Fault Diagnosis Under Unseen Working Conditions |
| title_short | Episodic Training and Feature Orthogonality-Driven Domain Generalization for Rotating Machinery Fault Diagnosis Under Unseen Working Conditions |
| title_sort | episodic training and feature orthogonality driven domain generalization for rotating machinery fault diagnosis under unseen working conditions |
| topic | fault diagnosis rotating machinery domain generalization episodic training feature orthogonality |
| url | https://www.mdpi.com/2075-1702/13/7/563 |
| work_keys_str_mv | AT yixiaoliao episodictrainingandfeatureorthogonalitydrivendomaingeneralizationforrotatingmachineryfaultdiagnosisunderunseenworkingconditions AT songbinzhou episodictrainingandfeatureorthogonalitydrivendomaingeneralizationforrotatingmachineryfaultdiagnosisunderunseenworkingconditions AT yisenliu episodictrainingandfeatureorthogonalitydrivendomaingeneralizationforrotatingmachineryfaultdiagnosisunderunseenworkingconditions AT kunkunpang episodictrainingandfeatureorthogonalitydrivendomaingeneralizationforrotatingmachineryfaultdiagnosisunderunseenworkingconditions AT jingli episodictrainingandfeatureorthogonalitydrivendomaingeneralizationforrotatingmachineryfaultdiagnosisunderunseenworkingconditions AT changli episodictrainingandfeatureorthogonalitydrivendomaingeneralizationforrotatingmachineryfaultdiagnosisunderunseenworkingconditions AT luluzhao episodictrainingandfeatureorthogonalitydrivendomaingeneralizationforrotatingmachineryfaultdiagnosisunderunseenworkingconditions |