Digital twin-based prognostics and health management for fatigue-prone components and systems using metamodel-based approaches

Abstract

Prognostics and health management (PHM) is an interdisciplinary field that addresses challenges in reliability, availability, maintainability, and safety arising from design, manufacturing, environmental, and operational complexities. The primary objective of PHM is to ensure asset integrity while mitigating risks to mission reliability and safety. Enabled by the industrial internet-of-things and model-based systems engineering, digital twins (DTs) are transforming industries by creating accurate virtual replicas of physical assets. The DT technology significantly enhances PHM implementation by enabling reliable life reliability and health management, reducing costs, and improving system availability. To address challenges in life reliability prediction, metamodeling is used to support DTs. Traditional approaches to life reliability prediction rely on quantifying uncertainties and applying Monte Carlo simulations to finite element models. While effective, these methods are computationally intensive and time-consuming. In contrast, metamodeling improves prediction efficiency without sacrificing accuracy, as demonstrated in case studies involving turbine blisks and fan shafts. With the metamodel-based DT, the DT-PHM architecture enables life reliability prediction without requiring run-to-failure data, supporting cost-effective and availability-driven health management strategies, such as warranty design and predictive maintenance planning. The developed DT-PHM architecture is applicable to the health management of both single-unit components and multi-unit systems. For multi-unit systems, an opportunistic predictive maintenance has been designed, incorporating economic dependencies to optimize maintenance strategies effectively.