Yifei Wang, Rui He, Xingkai Yang, Zhigang Tian
Advancements in monitoring techniques have facilitated the use of prognostic information to evaluate the health status of rotating machinery, enabling proactive mitigation of potential failures and thereby reducing maintenance costs in manufacturing systems. However, the application of prognostic techniques incurs substantial costs, mainly attributable to sensor acquisition, scheduled replacement, and reinstallation requirements. Consequently, the economic trade-offs of implementing prognostic techniques for continuous condition monitoring remain insufficiently explored in existing maintenance strategies. To bridge this gap, a novel asset-criticality-guided maintenance strategy is developed to maximize the expected revenue of manufacturing systems. Compared to reported works, three key contributions are made. First, the proposed maintenance strategy incorporates machine criticality as a key decision variable within the optimization framework, utilizing actual maintenance records to inform maintenance decision-making. Second, the proposed decision-making model addresses the critical challenge of identifying specific assets for continuous monitoring to maximize net revenue. Third, the strategy details component-level repairs for diverse failure modes within a degraded working efficiency model. This framework enables probabilistic assessment of policy benefits and quantifies uncertainty in maintenance planning. A numerical example and a real-world case study from a pulp mill are provided to demonstrate and validate the proposed method.