Generalization of Deep Neural Networks for Imbalanced Fault Classification of Machinery Using Generative Adversarial Networks
Mechanical fault datasets are always highly imbalanced with abundant common mechanical fault samples but a paucity of samples from rare fault conditions. To overcome this weakness, the simulation of rare fault signals is proposed in this paper. Specifically, frequency spectra are employed as model s...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2019-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/8742570/ |
| _version_ | 1818566469625053184 |
|---|---|
| author | Jinrui Wang Shunming Li Baokun Han Zenghui An Huaiqian Bao Shanshan Ji |
| author_facet | Jinrui Wang Shunming Li Baokun Han Zenghui An Huaiqian Bao Shanshan Ji |
| author_sort | Jinrui Wang |
| collection | DOAJ |
| description | Mechanical fault datasets are always highly imbalanced with abundant common mechanical fault samples but a paucity of samples from rare fault conditions. To overcome this weakness, the simulation of rare fault signals is proposed in this paper. Specifically, frequency spectra are employed as model signals, then Wasserstein generative adversarial network (WGAN) is implemented to generate simulated signals based on a labeled dataset. Finally, the real and artificial signals are combined to train stacked autoencoders (SAE) to detect mechanical health conditions. To validate the effectiveness of the proposed WGAN-SAE method, two specially designed experiments are carried out and some traditional methods are adopted for comparison. The diagnosis results show that the proposed method can deal with imbalanced fault classification problem much more effectively. The improved performance is mainly due to the artificial fault signals generated from the WGAN to balance the dataset, where the signals that are lacking in training dataset are effectively augmented. Furthermore, the learned features in each layer of the generator network are also analyzed via visualization, which may help us understand the working process of the WGAN. |
| first_indexed | 2024-12-14T01:54:05Z |
| format | Article |
| id | doaj.art-55955c9542624bee9a700b986ce88ab3 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-14T01:54:05Z |
| publishDate | 2019-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-55955c9542624bee9a700b986ce88ab32022-12-21T23:21:16ZengIEEEIEEE Access2169-35362019-01-01711116811118010.1109/ACCESS.2019.29240038742570Generalization of Deep Neural Networks for Imbalanced Fault Classification of Machinery Using Generative Adversarial NetworksJinrui Wang0https://orcid.org/0000-0001-8690-0672Shunming Li1Baokun Han2Zenghui An3Huaiqian Bao4Shanshan Ji5College of Mechanical and Electronic Engineering, Shandong University of Science and Technology, Qingdao, ChinaCollege of Energy and Power Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, ChinaCollege of Mechanical and Electronic Engineering, Shandong University of Science and Technology, Qingdao, ChinaCollege of Energy and Power Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, ChinaCollege of Mechanical and Electronic Engineering, Shandong University of Science and Technology, Qingdao, ChinaCollege of Mechanical and Electronic Engineering, Shandong University of Science and Technology, Qingdao, ChinaMechanical fault datasets are always highly imbalanced with abundant common mechanical fault samples but a paucity of samples from rare fault conditions. To overcome this weakness, the simulation of rare fault signals is proposed in this paper. Specifically, frequency spectra are employed as model signals, then Wasserstein generative adversarial network (WGAN) is implemented to generate simulated signals based on a labeled dataset. Finally, the real and artificial signals are combined to train stacked autoencoders (SAE) to detect mechanical health conditions. To validate the effectiveness of the proposed WGAN-SAE method, two specially designed experiments are carried out and some traditional methods are adopted for comparison. The diagnosis results show that the proposed method can deal with imbalanced fault classification problem much more effectively. The improved performance is mainly due to the artificial fault signals generated from the WGAN to balance the dataset, where the signals that are lacking in training dataset are effectively augmented. Furthermore, the learned features in each layer of the generator network are also analyzed via visualization, which may help us understand the working process of the WGAN.https://ieeexplore.ieee.org/document/8742570/Fault diagnosisimbalanced classificationWasserstein generative adversarial network (WGAN)stacked autoencoders (SAE)artificial signals |
| spellingShingle | Jinrui Wang Shunming Li Baokun Han Zenghui An Huaiqian Bao Shanshan Ji Generalization of Deep Neural Networks for Imbalanced Fault Classification of Machinery Using Generative Adversarial Networks IEEE Access Fault diagnosis imbalanced classification Wasserstein generative adversarial network (WGAN) stacked autoencoders (SAE) artificial signals |
| title | Generalization of Deep Neural Networks for Imbalanced Fault Classification of Machinery Using Generative Adversarial Networks |
| title_full | Generalization of Deep Neural Networks for Imbalanced Fault Classification of Machinery Using Generative Adversarial Networks |
| title_fullStr | Generalization of Deep Neural Networks for Imbalanced Fault Classification of Machinery Using Generative Adversarial Networks |
| title_full_unstemmed | Generalization of Deep Neural Networks for Imbalanced Fault Classification of Machinery Using Generative Adversarial Networks |
| title_short | Generalization of Deep Neural Networks for Imbalanced Fault Classification of Machinery Using Generative Adversarial Networks |
| title_sort | generalization of deep neural networks for imbalanced fault classification of machinery using generative adversarial networks |
| topic | Fault diagnosis imbalanced classification Wasserstein generative adversarial network (WGAN) stacked autoencoders (SAE) artificial signals |
| url | https://ieeexplore.ieee.org/document/8742570/ |
| work_keys_str_mv | AT jinruiwang generalizationofdeepneuralnetworksforimbalancedfaultclassificationofmachineryusinggenerativeadversarialnetworks AT shunmingli generalizationofdeepneuralnetworksforimbalancedfaultclassificationofmachineryusinggenerativeadversarialnetworks AT baokunhan generalizationofdeepneuralnetworksforimbalancedfaultclassificationofmachineryusinggenerativeadversarialnetworks AT zenghuian generalizationofdeepneuralnetworksforimbalancedfaultclassificationofmachineryusinggenerativeadversarialnetworks AT huaiqianbao generalizationofdeepneuralnetworksforimbalancedfaultclassificationofmachineryusinggenerativeadversarialnetworks AT shanshanji generalizationofdeepneuralnetworksforimbalancedfaultclassificationofmachineryusinggenerativeadversarialnetworks |