Predictive group maintenance for multi-system multi-component networks

Predictive maintenance has become highly popular in recent years due to the emergence of novel condition monitoring and data analysis techniques. However, the application of predictive maintenance at the network-level has not seen much attention in the literature. This paper presents a model for predictive group maintenance for multi-system multi- components networks (MSMCN). These networks are composed of multiple systems that are, in turn, composed of multiple components. In particular, the hierarchical structure of the MSMCN enables different representations of dependences at the network and system levels. The key novelty in the paper is that the designed approach combines analytical and numerical techniques to optimize the predictive group maintenance policy for MSMCNs. Moreover, we introduce a genetic algorithm with agglomerative mutation (GA-A) that enables a more effective evolution of the predictive group maintenance policy. Application of this model on a case study of a two-bridge network made of 23 different components shows a potential 11.27% reduction in maintenance cost, highlighting the model’s practical significance.This research was funded by the Engineering and Physical Sciences Research Council (UK) and Innovate UK through the Innovation and Knowledge Centre for Smart Infrastructure and Construction (Grant EP/N021614/1). This work was partially supported by Talent recruitment Funds of Tsinghua University grant NO.113052