Prognostics and health management for predictive maintenance: A review

In the pursuit of smart manufacturing, predictive maintenance (PdM) holds significant importance as it allows manufacturing firms to effectively mitigate avoidable downtime and maintenance costs. Prognostics and health management (PHM) is a viable solution for achieving PdM by evaluating the health condition and predicting the remaining useful life (RUL) of industrial assets. The advancement of digital technologies has facilitated the implementation of PHM through two distinct methods: life consumption monitoring (LCM) and system health monitoring (SHM) methods. Both PHM methods can be applied utilizing knowledge-based and data-driven approaches. This paper comprehensively reviews the application of knowledge-based and data-driven approaches in LCM and SHM methods. After the review, the strengths, weaknesses, and potential advancements of the knowledge-based and data-driven approaches are analyzed. It also assesses the benefits and drawbacks of LCM and SHM methods and clarifies how these two methods are deployed in real-time PdM. This paper is valuable for academics seeking guidance in establishing and improving PHM models. In addition, it provides practitioners with new insights into practical ways to address challenges encountered in PHM and PdM projects.