Assessing external sulfate attack on thin-shell artificial reef structures under uncertainty

Thin-shell artificial reef (AR) structures with spatial internal volumes have demonstrated superior stock recruitment ability and material efficiency than many gravity-based reef blocks, and cementitious materials, given the easy-to-tailor nature, remains the most popular in reef constructions to date. However, under constant seawater immersion, external sulfate attack (ESA) introduces a major and uncertain reliability concern to this type of AR system, due to the inherent material randomness. This study is concerned with developing a novel stochastic modelling framework for assessing the ESA under material uncertainty. In this paper, the main difficulty associated with the stochastic ESA modelling is identified for the first time, and a novel machine learning aided chemophysical modelling approach is proposed. The performance of the developed framework is carefully examined through the analyses on two types of cementitious materials under ESA.