Enhanced bearing RUL prediction based on dynamic temporal attention and mixed MLP
Abstract Bearings are critical components in machinery, and accurately predicting their remaining useful life (RUL) is essential for effective predictive maintenance. Traditional RUL prediction methods often rely on manual feature extraction and expert knowledge, which face specific challenges such...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Springer
2025-01-01
|
| Series: | Autonomous Intelligent Systems |
| Subjects: | |
| Online Access: | https://doi.org/10.1007/s43684-024-00088-4 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841544418525446144 |
|---|---|
| author | Zhongtian Jin Chong Chen Aris Syntetos Ying Liu |
| author_facet | Zhongtian Jin Chong Chen Aris Syntetos Ying Liu |
| author_sort | Zhongtian Jin |
| collection | DOAJ |
| description | Abstract Bearings are critical components in machinery, and accurately predicting their remaining useful life (RUL) is essential for effective predictive maintenance. Traditional RUL prediction methods often rely on manual feature extraction and expert knowledge, which face specific challenges such as handling non-stationary data and avoiding overfitting due to the inclusion of numerous irrelevant features. This paper presents an approach that leverages Continuous Wavelet Transform (CWT) for feature extraction, a Channel-Temporal Mixed MLP (CT-MLP) layer for capturing intricate dependencies, and a dynamic attention mechanism to adjust its focus based on the temporal importance of features within the time series. The dynamic attention mechanism integrates multi-head attention with innovative enhancements, making it particularly effective for datasets exhibiting non-stationary behaviour. An experimental study using the XJTU-SY rolling bearings dataset and the PRONOSTIA bearing dataset revealed that the proposed deep learning algorithm significantly outperforms other state-of-the-art algorithms in terms of RMSE and MAE, demonstrating its robustness and accuracy. |
| format | Article |
| id | doaj-art-2e41034ed6e745bba3d314313375138a |
| institution | Kabale University |
| issn | 2730-616X |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Springer |
| record_format | Article |
| series | Autonomous Intelligent Systems |
| spelling | doaj-art-2e41034ed6e745bba3d314313375138a2025-01-12T12:33:05ZengSpringerAutonomous Intelligent Systems2730-616X2025-01-015111610.1007/s43684-024-00088-4Enhanced bearing RUL prediction based on dynamic temporal attention and mixed MLPZhongtian Jin0Chong Chen1Aris Syntetos2Ying Liu3Department of Mechanical Engineering, School of Engineering, Cardiff UniversityGuangdong Provincial Key Laboratory of Cyber-Physical System, Guangdong University of TechnologyPARC Institute of Manufacturing Logistics and Inventory, Cardiff Business School, Cardiff UniversityDepartment of Mechanical Engineering, School of Engineering, Cardiff UniversityAbstract Bearings are critical components in machinery, and accurately predicting their remaining useful life (RUL) is essential for effective predictive maintenance. Traditional RUL prediction methods often rely on manual feature extraction and expert knowledge, which face specific challenges such as handling non-stationary data and avoiding overfitting due to the inclusion of numerous irrelevant features. This paper presents an approach that leverages Continuous Wavelet Transform (CWT) for feature extraction, a Channel-Temporal Mixed MLP (CT-MLP) layer for capturing intricate dependencies, and a dynamic attention mechanism to adjust its focus based on the temporal importance of features within the time series. The dynamic attention mechanism integrates multi-head attention with innovative enhancements, making it particularly effective for datasets exhibiting non-stationary behaviour. An experimental study using the XJTU-SY rolling bearings dataset and the PRONOSTIA bearing dataset revealed that the proposed deep learning algorithm significantly outperforms other state-of-the-art algorithms in terms of RMSE and MAE, demonstrating its robustness and accuracy.https://doi.org/10.1007/s43684-024-00088-4Deep learningRemaining useful lifePrognostic and health managementTransformer network |
| spellingShingle | Zhongtian Jin Chong Chen Aris Syntetos Ying Liu Enhanced bearing RUL prediction based on dynamic temporal attention and mixed MLP Autonomous Intelligent Systems Deep learning Remaining useful life Prognostic and health management Transformer network |
| title | Enhanced bearing RUL prediction based on dynamic temporal attention and mixed MLP |
| title_full | Enhanced bearing RUL prediction based on dynamic temporal attention and mixed MLP |
| title_fullStr | Enhanced bearing RUL prediction based on dynamic temporal attention and mixed MLP |
| title_full_unstemmed | Enhanced bearing RUL prediction based on dynamic temporal attention and mixed MLP |
| title_short | Enhanced bearing RUL prediction based on dynamic temporal attention and mixed MLP |
| title_sort | enhanced bearing rul prediction based on dynamic temporal attention and mixed mlp |
| topic | Deep learning Remaining useful life Prognostic and health management Transformer network |
| url | https://doi.org/10.1007/s43684-024-00088-4 |
| work_keys_str_mv | AT zhongtianjin enhancedbearingrulpredictionbasedondynamictemporalattentionandmixedmlp AT chongchen enhancedbearingrulpredictionbasedondynamictemporalattentionandmixedmlp AT arissyntetos enhancedbearingrulpredictionbasedondynamictemporalattentionandmixedmlp AT yingliu enhancedbearingrulpredictionbasedondynamictemporalattentionandmixedmlp |