A widely-applicable structural maintenance decision-analytic modelling approach assisted by information value computation

Abstract

Improving the quality of maintenance decision-making under uncertainty is important for safety, reliability and cost management of marine assets. It is worth studying how condition inspections can be objective-oriented and inspection information be fully utilized in the maintenance decision-making process. Herein a new maintenance decision-analytic modelling approach assisted by value of information (VoI) computation is proposed and applied to a safety-critical marine structure. Three maintenance decision-making cases and three inspection strategies are tested. Below advantages of the proposed approach represent contribution to the current literature. The approach can yield optimal maintenance decisions in various decision-making cases, while condition based maintenance (CBM) is efficient only in the cases where inspections add value. The approach yields maintenance decisions associated with the same or lower life cycle costs (LCC) than decision rule-based approaches (e.g. CBM), and gives more improved decisions (associated with lower LCC) with more in-depth inspections. The advantages are benefited from VoI computation, decision-analytic modelling, and probabilistic modelling of uncertainty, crack growth and inspection outcomes. Maintenance decisions are derived based on integrated crack information via Bayesian updating, rather than only on inspection results. The approach can be used to optimize inspection and maintenance strategies for structural assets.