Loading web-font TeX/Main/Regular
A Multichannel Long-Term External Attention Network for Aeroengine Remaining Useful Life Prediction | IEEE Journals & Magazine | IEEE Xplore

A Multichannel Long-Term External Attention Network for Aeroengine Remaining Useful Life Prediction


Impact Statement:In the deep learning-based (DL-based) RUL prediction method, it is not necessary to construct a complex physical engine degradation model. Therefore, the DL-based method ...Show More

Abstract:

Accurately estimating the remaining useful life (RUL) of aircraft engines can effectively prevent aircraft crashes and human casualties. In some RUL prediction methods, p...Show More
Impact Statement:
In the deep learning-based (DL-based) RUL prediction method, it is not necessary to construct a complex physical engine degradation model. Therefore, the DL-based method is becoming increasingly popular among scholars. However, because of the complex operating conditions, the engine degradation cannot be fully characterized, resulting in poor prediction accuracy. In this study, a new RUL prediction method is proposed to improve the prediction accuracy by considering the complex operating condition information, extracting effective temporal features and improving the generalizability. The related experiments show that the performance of the proposed method is improved by 10.4%\sim26.2% compared with the listed methods.

Abstract:

Accurately estimating the remaining useful life (RUL) of aircraft engines can effectively prevent aircraft crashes and human casualties. In some RUL prediction methods, particularly for aircraft engines running under complex conditions, they are difficult to comprehensively characterize the engine degradation process, resulting in poor predicted RUL. To address the above challenge, a multichannel long-term external attention network (MLEAN) is proposed for the RUL prediction of turbofan engines. First, the preprocessed samples are transformed to enable MLEAN to focus on learning inter-sensor correlations within the same degradation stage. To improve the feature representation capability of the network, multichannel time attention network (MTANet) is then designed to realize multiscale and multifrequency feature learning, which effectively achieves multiperspective analysis of long-term dependencies in different channels. Then, external attention block (EAB) is introduced to memorize im...
Published in: IEEE Transactions on Artificial Intelligence ( Volume: 5, Issue: 10, October 2024)
Page(s): 5130 - 5140
Date of Publication: 15 May 2024
Electronic ISSN: 2691-4581

Funding Agency:


References

References is not available for this document.