Deep learning framework for gas turbine performance digital twin and degradation prognostics from airline operator perspective

Abstract

Digital twin technology has emerged as a research hotspot in the field of intelligent operation and maintenance of gas turbines. This paper proposes a data-driven Digital Twin approach for gas turbine performance monitoring and degradation prognostics from an airline operator perspective. The framework adopts a semi-supervised deep learning method to construct a data-driven Performance Digital Twin (PDT) rather than a physics-based performance model. The PDT-derived multi-dimensional health features is used to characterize the performance degradation and enhance the input features for the prognostics network. Specifically, domain knowledge from the asset operator perspective is incorporated into the prognostics model to improve the performance. The proposed approach is evaluated on real-world turbofan engines and the NCMAPSS dataset, achieving promising results compared to the state-of-art approaches. The developed data-driven prognostics framework provides a low-cost alternative to an expensive physics-based prognostics approach for gas turbine operators. It enables asset users to implement their own data-driven prognostics and maintenance strategies.