Abstract:

Degradation modeling is an effective approach for reliability assessment, remaining useful life prediction, maintenance planning, and prognostics health management. Degradation models are usually developed based on degradation data and/or prior understandings of physics behind degradation processes of products or systems. Further, the effects of environmental or operational conditions on degradation processes and the knowledge about the dependency between degradation processes help improve the explanatory capabilities of degradation models. This paper presents a comprehensive review of existing degradation modeling approaches commonly used in engineering applications. To assist practitioners in understanding the concept of degradation modelling, the existing methods are classified into two broad categories: the data-driven approach and physics-based modelling approach. By systematically reviewing these approaches, we highlight their merits, useful applications and limitations. Finally, we provide a summary and indicate several future research challenges in this important area of reliability engineering.