Data augmentation based on diffusion probabilistic model for remaining useful life estimation of aero-engines

Wei Wang, Honghao Song, Shubin Si, Wenhao Lu and Zhiqiang Cai

Reliability Engineering and System Safety, 2024, vol. 252, issue C

Abstract: Predicting the remaining useful life (RUL) of aero-engines is essential for their prognostics and health management (PHM). Deep learning technologies are effective in this area, but their success depends critically on acquiring sufficient engine monitoring data from operation to failure, a process that is expensive and challenging in practice. Insufficient data limit the training of deep learning methods, thereby affecting their predictive performance. To address this issue, this study proposes a novel method named DiffRUL for augmenting multivariate engine monitoring data and generating high-quality samples mimicking degradation trends in real data. Initially, a specialized degradation trend encoder is designed to extract degradation trend representations from monitoring data, which serve as generative conditions. Subsequently, the diffusion model is adapted to the scenario of generating multivariate monitoring data, reconstructing and synthesizing data from Gaussian noise through a reverse process. Additionally, a denoising network is developed to incorporate generative conditions and capture spatio-temporal correlations in the data, accurately estimating noise levels during the reverse process. Experimental results on C-MAPSS and N-CMAPSS datasets show that DiffRUL successfully generates high-fidelity multivariate monitoring data. Furthermore, these generated data effectively support the RUL prediction task and significantly enhance the predictive ability of the underlying deep learning models.

Keywords: Prognostics and health management; Remaining useful life estimation; Aero-engines; Deep learning; Data augmentation; Diffusion model (search for similar items in EconPapers)
Date: 2024
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0951832024004666
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:reensy:v:252:y:2024:i:c:s0951832024004666

DOI: 10.1016/j.ress.2024.110394

Access Statistics for this article

Reliability Engineering and System Safety is currently edited by Carlos Guedes Soares

More articles in Reliability Engineering and System Safety from Elsevier
Bibliographic data for series maintained by Catherine Liu ().