Fatigue reliability of offshore Mooring chains under influence of mean load and corrosion

Abstract

A number of full-scale fatigue tests have been conducted for offshore mooring chain in the last decade for both new and used chains, retrieved after operation in the North Sea. The surface condition of the used chains vary considerably, from minor corrosion and slightly increased surface roughness to severe pitting corrosion. These tests have indicated that the fatigue performance of the chains depends strongly on mean load and corrosion condition, in such a way that the fatigue life increases when the mean load is reduced while pitting corrosion has a detrimental effect. In fatigue calculations following the current standards and industry practice, mean load is disregarded whereas corrosion is accounted for in a simplified way that is not representative of its actual impact on fatigue. The present thesis addresses the effects of mean load and corrosion on the fatigue capacity of studless mooring chain and, subsequently, on the fatigue reliability of chain segments. Our work builds on the data from these fatigue tests, conducted at mean loads from 6% to 20% of the chains’ minimum breaking load (MBL). The used chain samples have been categorized from visual inspection and assigned a corrosion grade on a custom scale, as a quantitative measure of their surface condition and the degree of pitting corrosion. An extended S–N model is proposed to model the fatigue capacity of the chains, with the S–N intercept parameter expressed as a log-linear function of mean load and corrosion grade. The parameters of this model are estimated from linear regression on the test data. A hierarchical linear model is used for the regression analysis to account for, and quantify, correlation in fatigue performance between samples within subsets of the data. The resulting model describes substantial effects from mean load and corrosion on the fatigue capacity of chains. Based on the new capacity model, we develop a reliability formulation for fatigue failure of chain segments. We initially assume independent fatigue capacity between chain links in a segment, then develop the model further to account for within-segment correlation for consistency with the hierarchical linear model. This reliability model allows accounting for the effects of mean load and degradation due to corrosion. In a generic case study, we demonstrate their influence on the probability of fatigue failure. We first conduct a global sensitivity analysis and show that the inferential uncertainty of the capacity model is non-influential for the estimated fatigue damage in the presented case. The subsequent parameter study shows that both mean load and corrosion have significant impacts. Although neglecting either of them have opposite effects, neglecting both may lead to over- or underestimated failure probabilities.