Remaining useful life prediction based on intentional noise injection and feature reconstruction. (November 2021)

• Record Type: Journal Article
• Authors: Xiao, Lei; Tang, Junxuan; Zhang, Xinghui; Bechhoefer, Eric; Ding, Siyi
• Is Part Of: Reliability engineering & system safety. Volume 215(2021)
• Publication Date: 2021-11
• Subjects: Remaining useful life -- Feature reconstruction -- Noise injection -- Aero-engines; Reliability (Engineering) -- Periodicals; System safety -- Periodicals; Industrial safety -- Periodicals; Fiabilité -- Périodiques; Sécurité des systèmes -- Périodiques; Sécurité du travail -- Périodiques; 620.00452
• DOI: 10.1016/j.ress.2021.107871
• Languages: English
• ISSNs: 0951-8320

Abstract

Highlights: Propose a new RUL prediction method based on LSTM network for an engineered machine. Gaussian white noise is intentionally injected into the output of LSTM network. New degradation features are constructed based on correlation analysis and hypothesis testing among sensory measurements. The selection of degradation features is according to the trendability representation ability. The robustness and accuracy of the proposed method are improved. Abstract: The accurate remaining useful life (RUL) prediction is the foundation of prognostics and health management (PHM). The accuracy of RUL prediction model depends on not only the quality and quantity of degradation feature but also the prediction model. In most of the existing deep-learning based RUL prediction models, noise is considered harmful and has to be removed. Further, the correlation among sensory measurements is ignored. However, noise can boost the prediction performance if judiciously used. This paper proposes a new RUL prediction method where noise is intentionally added into a long short-term memory (LSTM) network. Additionally, correlation analysis is conducted among the sensory measurements to construct new degradation features as the inputs of the LSTM network. Validation of the proposed method was carried out on the C-MAPSS aero-engine lifetime dataset. Finally, the proposed RUL prediction model is compared to other the-state-of-the-art techniques.