2021/12/03 by Zhaoyi Xu, Yanjie Guo, Joseph H. Saleh +1
Engineering · #Artificial intelligence #Benchmark (surveying) #Computer science #Data mining #Dependency (UML) #Engineering #Fault Detection and Control Systems #Gaussian #Gaussian process #Kriging #Machine Fault Diagnosis Techniques #Machine learning #Predictive modelling #Prognostics #Reliability and Maintenance Optimization #Robustness (evolution) #Turbofan
paper · doi:10.1109/tr.2021.3124944
openalex publication_date 2021/12/03 · crossref created 2021/12/03 · crossref issued 2022/03/01 · crossref published 2022/03/01 · crossref published-print 2022/03/01 · crossref deposited 2022/12/07 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/15 · crossref indexed 2026/08/04
Remaining useful life (RUL) refers to the expected remaining lifespan of a component or system. Accurate RUL prediction is critical for prognostic and health management and for maintenance planning. In this article, we address three prevalent challenges in data-driven RUL prediction, namely the handling of high-dimensional input features, the robustness to noise in sensor data and prognostic datasets, and the capturing of the time-dependency between system degradation and RUL prediction. We devise a highly accurate RUL prediction model with uncertainty quantification, which integrates and leverages the advantages of deep learning and nonstationary Gaussian process regression (DL–NSGPR). We examine and benchmark our model against other advanced data-driven RUL prediction models using the turbofan engine dataset from the NASA prognostic repository. Our computational experiments show that the DL-NSGPR predictions are highly accurate with root mean square error 1.7 to 6.2 times smaller than those of competing RUL models. Furthermore, the results demonstrate that RUL uncertainty bounds with the proposed DL-NSGPR are both valid and significantly tighter than other stochastic RUL prediction models. We unpack and discuss the reasons for this excellent performance of the DL-NSGPR.