Towards online data-driven prognostics system

Abstract

Complex engineering systems are always working in harsh operating conditions. Internal wears and tears of such systems can lead to catastrophic failures which affect system operation and endanger human lives in many cases. To avoid sudden failures, continuous monitoring, fault diagnosis, and failure prognosis are required. Prognostics as a discipline plays the major rule in impending failure prevention. Offline prognostics system focuses on maintenance and logistics operations whereas online prognostics focuses on maintaining safe operation. In a previous work we successfully developed an offline prognostics system for aircraft turbofan engines' remaining useful life (RUL) estimation. In this paper we will show how we used the offline prognostics system as the baseline definition towards creating an online data-driven prognostics system to enable informed real-time decisions.

Physics-Embedded Deep Learning to Predict Real-Time Flow Parameters in Complex Thermodynamic Machinery

“…A more recent strategy involves combining physical principles and traditional algorithms with deep learning in a more nuanced manner to create more powerful models in the burgeoning field of scientific deep learning [34]. An effective way to create a hybrid model is to synergistically combine data-driven and model-based approaches [35]. Another approach that has received significant attention is the use of physics-informed neural networks that can be used to solve forward and inverse problems related to differential equations [36,37].…”

Failure Prognostic of Turbofan Engines with Uncertainty Quantification and Explainable AI (XIA)

“…In PHM, three essential activities consisting of prognostic, anomaly detection and diagnostic are carried out. Prognostic is the act of defining the Remaining Useful Life (RUL) or the remaining operational time of an industrial assets before failure (Elattar et al, 2018;Akpudo&Hur, 2020). Anomaly detection refers to the identification of unusual patterns going against the normal behaviour of operational parameter measurement (Gurkan&Burak, 2020;Liu &Gryllias, 2020).…”