Deep Residual Network for Remaining Useful Life Prediction  
Author Chun Fai Lui


Co-Author(s) Min Xie


Abstract Deep learning methods have been extensively used for prognostics health management (PHM) in recent years, especially for anomaly detection, fault diagnosis and remaining useful life (RUL) prediction. Although deep convolutional neural networks (DCNN) have shown promising capability in feature extraction, it requires heavy computation, which hinders the utility of CNN in PHM. In this article, we propose the use of deep residual network (DRN) for effective RUL prediction. With the residual structure, the proposed DRN is able to predict RUL effectively for system health prognostics. The effectiveness of the proposed DRN is demonstrated on the NASA C-MAPSS turbofan engine data subset FD001. The results of this study suggest that the DRN is an effective, stable approach for accurate RUL prediction.


Keywords Remaining useful life (RUL), degradation, prognostics health management (PHM), deep learning, convolutional neural network (CNN), deep residual network (DRN)
    Article #:  RQD28-247

Proceedings of 28th ISSAT International Conference on Reliability & Quality in Design
August 3-5, 2023