Hierarchical Approach to Reliability-Based Inspection Planning of Hull Structures

Authors

Marc A. Maes and Michael H. Faber

Abstract

This paper addresses two important computational aspects of inspection and maintenance optimization of deteriorating ship structures or floating production/storage facilities: (1) the hierarchical modeling of a ship hull to allow for the efficient use of Bayesian probability nets (BPN) in the analysis required to develop optimal risk acceptance criteria; and (2) the use of hierarchical Bayes methods to update the condition states of spatially distributed hull components such as plates, girders, and connections. Using the hierarchical modeling of a hull described in this paper, a quantitative risk assessment model can be developed to quantify the risks associated with the degradation of the ship hull, using event and consequence scenario. Hierarchical Bayes models are discussed in this paper both for the case of spatial conditional independence and for the more general case where spatial correlation is considered. The latter is especially useful for decision making involving one- and two-dimensional components such as girders, pipes, vessels and plates subject to corrosion. The influence of correlation length between condition states is examined.

Published in/by

Proceedings COMPIT'07, Cortona, Italy, April 23-25, 2007.
http://www.compit07.insean.it/