SACGNet: A Remaining Useful Life Prediction of Bearing with Self-Attention Augmented Convolution GRU Network
In recent years, the development of deep learning-based remaining useful life (RUL) prediction methods of bearings has flourished because of their high accuracy, easy implementation, and lack of reliance on a priori knowledge. However, there are two challenging issues concerning the prediction accur...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-02-01
|
| Series: | Lubricants |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2075-4442/10/2/21 |
| _version_ | 1827654261448638464 |
|---|---|
| author | Juan Xu Shiyu Duan Weiwei Chen Dongfeng Wang Yuqi Fan |
| author_facet | Juan Xu Shiyu Duan Weiwei Chen Dongfeng Wang Yuqi Fan |
| author_sort | Juan Xu |
| collection | DOAJ |
| description | In recent years, the development of deep learning-based remaining useful life (RUL) prediction methods of bearings has flourished because of their high accuracy, easy implementation, and lack of reliance on a priori knowledge. However, there are two challenging issues concerning the prediction accuracy of existing methods. The run-to-failure sequential data and its RUL labels are almost inaccessible in real-world scenarios. Meanwhile, the existing models usually capture the general degradation trend of bearings while ignoring the local information, which restricts the model performance. To tackle the aforementioned problems, we propose a novel health indicator derived from the original vibration signals by combining principal components analysis with Euclidean distance metric, which was motivated by the desire to resolve the dependency on RUL labels. Then, we design a novel self-attention augmented convolution GRU network (SACGNet) to predict the RUL. Combining a self-attention mechanism with a convolution framework can both adaptively assign greater weights to more important information and focus on local information. Furthermore, Gated Recurrent Units are used to parse the long-term dependencies in weighted features such that SACGNet can utilize the important weighted features and focus on local features to improve the prognostic accuracy. The experimental results on the PHM 2012 Challenge dataset and the XJTU-SY bearing dataset have demonstrated that our proposed method is superior to the state of the art. |
| first_indexed | 2024-03-09T21:34:51Z |
| format | Article |
| id | doaj.art-f44a993d885c411d9030d1cd9824cdab |
| institution | Directory Open Access Journal |
| issn | 2075-4442 |
| language | English |
| last_indexed | 2024-03-09T21:34:51Z |
| publishDate | 2022-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Lubricants |
| spelling | doaj.art-f44a993d885c411d9030d1cd9824cdab2023-11-23T20:47:39ZengMDPI AGLubricants2075-44422022-02-011022110.3390/lubricants10020021SACGNet: A Remaining Useful Life Prediction of Bearing with Self-Attention Augmented Convolution GRU NetworkJuan Xu0Shiyu Duan1Weiwei Chen2Dongfeng Wang3Yuqi Fan4Key Laboratory of Knowledge Engineering with Big Data, School of Computer Science and Information Engineering, Hefei University of Technology, Hefei 230009, ChinaKey Laboratory of Knowledge Engineering with Big Data, School of Computer Science and Information Engineering, Hefei University of Technology, Hefei 230009, ChinaShanghai Aerospace Control Technology Institute, Shanghai 201109, ChinaLuoyang Bearing Research Institute Co., Ltd., Luoyang 471033, ChinaSchool of Computer and Information, Hefei University of Technology, Hefei 230009, ChinaIn recent years, the development of deep learning-based remaining useful life (RUL) prediction methods of bearings has flourished because of their high accuracy, easy implementation, and lack of reliance on a priori knowledge. However, there are two challenging issues concerning the prediction accuracy of existing methods. The run-to-failure sequential data and its RUL labels are almost inaccessible in real-world scenarios. Meanwhile, the existing models usually capture the general degradation trend of bearings while ignoring the local information, which restricts the model performance. To tackle the aforementioned problems, we propose a novel health indicator derived from the original vibration signals by combining principal components analysis with Euclidean distance metric, which was motivated by the desire to resolve the dependency on RUL labels. Then, we design a novel self-attention augmented convolution GRU network (SACGNet) to predict the RUL. Combining a self-attention mechanism with a convolution framework can both adaptively assign greater weights to more important information and focus on local information. Furthermore, Gated Recurrent Units are used to parse the long-term dependencies in weighted features such that SACGNet can utilize the important weighted features and focus on local features to improve the prognostic accuracy. The experimental results on the PHM 2012 Challenge dataset and the XJTU-SY bearing dataset have demonstrated that our proposed method is superior to the state of the art.https://www.mdpi.com/2075-4442/10/2/21self-attentiongated neural networkremaining useful life predictionhealth indicator |
| spellingShingle | Juan Xu Shiyu Duan Weiwei Chen Dongfeng Wang Yuqi Fan SACGNet: A Remaining Useful Life Prediction of Bearing with Self-Attention Augmented Convolution GRU Network Lubricants self-attention gated neural network remaining useful life prediction health indicator |
| title | SACGNet: A Remaining Useful Life Prediction of Bearing with Self-Attention Augmented Convolution GRU Network |
| title_full | SACGNet: A Remaining Useful Life Prediction of Bearing with Self-Attention Augmented Convolution GRU Network |
| title_fullStr | SACGNet: A Remaining Useful Life Prediction of Bearing with Self-Attention Augmented Convolution GRU Network |
| title_full_unstemmed | SACGNet: A Remaining Useful Life Prediction of Bearing with Self-Attention Augmented Convolution GRU Network |
| title_short | SACGNet: A Remaining Useful Life Prediction of Bearing with Self-Attention Augmented Convolution GRU Network |
| title_sort | sacgnet a remaining useful life prediction of bearing with self attention augmented convolution gru network |
| topic | self-attention gated neural network remaining useful life prediction health indicator |
| url | https://www.mdpi.com/2075-4442/10/2/21 |
| work_keys_str_mv | AT juanxu sacgnetaremainingusefullifepredictionofbearingwithselfattentionaugmentedconvolutiongrunetwork AT shiyuduan sacgnetaremainingusefullifepredictionofbearingwithselfattentionaugmentedconvolutiongrunetwork AT weiweichen sacgnetaremainingusefullifepredictionofbearingwithselfattentionaugmentedconvolutiongrunetwork AT dongfengwang sacgnetaremainingusefullifepredictionofbearingwithselfattentionaugmentedconvolutiongrunetwork AT yuqifan sacgnetaremainingusefullifepredictionofbearingwithselfattentionaugmentedconvolutiongrunetwork |