Simultaneous-fault diagnosis considering time series with a deep learning transformer architecture for air handling units
An advanced deep learning-based method that employs transformer architecture is proposed to diagnose the simultaneous faults with time-series data. This method can be directly applied to transient data while maintaining the accuracy without a steady-state detector so that the fault can be diagnosed...
| Main Authors: | , , , |
|---|---|
| Other Authors: | |
| Format: | Journal Article |
| Language: | English |
| Published: |
2022
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/161886 |
| _version_ | 1826124979080527872 |
|---|---|
| author | Wu, Bingjie Cai, Wenjian Cheng, Fanyong Chen, Haoran |
| author2 | School of Electrical and Electronic Engineering |
| author_facet | School of Electrical and Electronic Engineering Wu, Bingjie Cai, Wenjian Cheng, Fanyong Chen, Haoran |
| author_sort | Wu, Bingjie |
| collection | NTU |
| description | An advanced deep learning-based method that employs transformer architecture is proposed to diagnose the simultaneous faults with time-series data. This method can be directly applied to transient data while maintaining the accuracy without a steady-state detector so that the fault can be diagnosed in its early stage. The transformer architecture adopts a novel multi-head attention mechanism without involving any convolutional and recurrent layers as in conventional deep learning methods. The model has been verified by an on-site air handling unit with 6 single-fault cases, 7 simultaneous-fault cases, and normal operating conditions with satisfactory performances of test accuracy of 99.87%, Jaccard score of 99.94%, and F1 score of 99.95%. Besides, the attention distribution reveals the correlations between features to the corresponding fault. It is found that the length of the sliding window is key to the model performance, and a trade-off is made for the window length between the model performance and the diagnosis time. Based on the similar idea, another sequence-to-vector model based on the gated recurrent unit (GRU) is proposed and benchmarked with the transformer model. The results show that the transformer model outperforms the GRU model with a better Jaccard score and F1 score in less training time. |
| first_indexed | 2024-10-01T06:29:14Z |
| format | Journal Article |
| id | ntu-10356/161886 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T06:29:14Z |
| publishDate | 2022 |
| record_format | dspace |
| spelling | ntu-10356/1618862022-09-23T02:37:18Z Simultaneous-fault diagnosis considering time series with a deep learning transformer architecture for air handling units Wu, Bingjie Cai, Wenjian Cheng, Fanyong Chen, Haoran School of Electrical and Electronic Engineering SJ-NTU Corporate Lab Engineering::Electrical and electronic engineering Fault Diagnosis Transformer Architecture An advanced deep learning-based method that employs transformer architecture is proposed to diagnose the simultaneous faults with time-series data. This method can be directly applied to transient data while maintaining the accuracy without a steady-state detector so that the fault can be diagnosed in its early stage. The transformer architecture adopts a novel multi-head attention mechanism without involving any convolutional and recurrent layers as in conventional deep learning methods. The model has been verified by an on-site air handling unit with 6 single-fault cases, 7 simultaneous-fault cases, and normal operating conditions with satisfactory performances of test accuracy of 99.87%, Jaccard score of 99.94%, and F1 score of 99.95%. Besides, the attention distribution reveals the correlations between features to the corresponding fault. It is found that the length of the sliding window is key to the model performance, and a trade-off is made for the window length between the model performance and the diagnosis time. Based on the similar idea, another sequence-to-vector model based on the gated recurrent unit (GRU) is proposed and benchmarked with the transformer model. The results show that the transformer model outperforms the GRU model with a better Jaccard score and F1 score in less training time. 2022-09-23T02:37:18Z 2022-09-23T02:37:18Z 2022 Journal Article Wu, B., Cai, W., Cheng, F. & Chen, H. (2022). Simultaneous-fault diagnosis considering time series with a deep learning transformer architecture for air handling units. Energy and Buildings, 257, 111608-. https://dx.doi.org/10.1016/j.enbuild.2021.111608 0378-7788 https://hdl.handle.net/10356/161886 10.1016/j.enbuild.2021.111608 2-s2.0-85121618060 257 111608 en Energy and Buildings © 2021 Elsevier B.V. All rights reserved. |
| spellingShingle | Engineering::Electrical and electronic engineering Fault Diagnosis Transformer Architecture Wu, Bingjie Cai, Wenjian Cheng, Fanyong Chen, Haoran Simultaneous-fault diagnosis considering time series with a deep learning transformer architecture for air handling units |
| title | Simultaneous-fault diagnosis considering time series with a deep learning transformer architecture for air handling units |
| title_full | Simultaneous-fault diagnosis considering time series with a deep learning transformer architecture for air handling units |
| title_fullStr | Simultaneous-fault diagnosis considering time series with a deep learning transformer architecture for air handling units |
| title_full_unstemmed | Simultaneous-fault diagnosis considering time series with a deep learning transformer architecture for air handling units |
| title_short | Simultaneous-fault diagnosis considering time series with a deep learning transformer architecture for air handling units |
| title_sort | simultaneous fault diagnosis considering time series with a deep learning transformer architecture for air handling units |
| topic | Engineering::Electrical and electronic engineering Fault Diagnosis Transformer Architecture |
| url | https://hdl.handle.net/10356/161886 |
| work_keys_str_mv | AT wubingjie simultaneousfaultdiagnosisconsideringtimeserieswithadeeplearningtransformerarchitectureforairhandlingunits AT caiwenjian simultaneousfaultdiagnosisconsideringtimeserieswithadeeplearningtransformerarchitectureforairhandlingunits AT chengfanyong simultaneousfaultdiagnosisconsideringtimeserieswithadeeplearningtransformerarchitectureforairhandlingunits AT chenhaoran simultaneousfaultdiagnosisconsideringtimeserieswithadeeplearningtransformerarchitectureforairhandlingunits |