CBM Policy Optimization Based on L\'{e}vy Copula Modeling of Latent Degradation

Within practical systems, the continuous buildup of wear and tear is an unavoidable reality. Industries subject to such systems must account for degradation to maximize runtime while minimizing costs. Among common methods, Condition-Based Maintenance (CBM) uses observable system conditions to make decisions. This is considered both cost effective and reliable with regard to planning system maintenance. Latent degradation refers to any degradation that is hidden within a system and can cause breakdowns without warning. CBM can use correlated performance characteristics, called markers, to make predictions about latent degradation for decision making. Recent evidence suggests that a copula-based approach may be better at capturing the dependence structure between latent degradation and markers compared to previous methods. This study investigates the use of the Lévy copula function to model the dependence between observed markers and latent degradation, enabling prediction of latent degradation from system markers.

Keywords: Condition-Based Maintenance, Latent Degradation, Marker Process, Lévy Copula.