A Semisupervised Deep Hybrid Multitask Model for RUL Prediction

Deep learning (DL) methods can be used to construct health indicators (HIs) for remaining useful life (RUL) prediction. The existing DL methods consider previous and current sensor signals and utilize labeled data, which are limited in practice. To leverage unlabeled data for extracting HIs, semisupervised methods, especially hybrid methods, can be employed. In this article, a semisupervised deep hybrid multitask model (DHMTM) for RUL prediction is developed. The DHMTM contains two temporal models for unlabeled and labeled multivariate time-series data, respectively. In the model training process, adding an extra task of prediction of future sensor signal values, the DHMTM can obtain His, which improve the RUL prediction accuracy. Besides, temporal dependency of sensor signals is captured in the proposed DHMTM. The effectiveness of the proposed model is validated using the commercial modular aero-propulsion system simulation (C-MAPSS) and the lithium-ion batteries datasets. The results show that using the proposed method, the prediction errors for the two datasets have been reduced by 2.5% and 23.5% on average, respectively, compared with the fully supervised regression model, and 17% and 44%, respectively, on average compared with three other widely used semisupervised methods.