Remaining Useful Life Prediction of Turbofan Engine in Varied Operational Conditions Considering Change Point: A Novel Deep Learning Approach with Optimum Features

Abstract

In the era of Internet of Things (IoT), remaining useful life (RUL) prediction of turbofan engines is crucial. Various deep learning (DL) techniques proposed recently to predict RUL for such systems have remained silent on the effect of environmental changes on machine reliability. This paper aims (i) to identify the change point in RUL trends and patterns (ii) to select the most relevant features, and (iii) to predict RUL with the selected features and identified change points. A two-stage feature-selection algorithm was developed, followed by a change point identification mechanism, and finally, a Bidirectional Long Short-Term Memory (BiLSTM) model was designed to predict RUL. The study utilizes NASA’s C-MAPSS data set to check the performance of the proposed methodology. The findings affirm that the proposed method enhances the stability of DL models, resulting in a 27.8% improvement in RUL prediction compared to popular and cutting-edge DL models.

Keywords

• remaining useful life (RUL)
• Internet of Things (IoT)
• sensors
• Bidirectional Long Short-Term Memory (BiLSTM)
• feature engineering
• change point