Selective maintenance of continuously degrading systems with non-identical and stochastically dependent components

Abstract

Selective maintenance (SM) is often performed on systems having limited maintenance resources to enhance mission reliability. In practical engineering, components within a system are typically distinct and correlated, leading to differences in the degradation processes and sensitivity to working conditions, as well as stochastic dependencies (S-dependencies) among the components. However, in most studies on the SM of continuously degrading systems (CDSs), these features are not considered, thereby limiting the accuracy of reliability evaluations and the efficiency of maintenance strategies. Therefore, considering the differences and S-dependencies of components, we investigate SM optimization for CDSs. Focusing on the diverse and stochastically dependent degradation processes of non-identical components, we first propose an extended degradation rate interaction model integrated with a general stochastic process to describe the multiple degradation processes of components and manifest the sensitivity differences to working conditions in the degradation states and rates. Next, a derivation method is developed to obtain explicit mission reliability functions for systems with arbitrary configurations. Subsequently, an SM optimization model that incorporates the effects of multiple differences and S-dependencies of the components is formulated and used to obtain an optimal maintenance strategy that considers resource and mission reliability constraints. Finally, the effectiveness of the proposed method is demonstrated using two numerical examples.