The Impact of Stress Redistribution on Structural Reliability Predictions of Deepwater Jackets

Structural reliability predictions for fixed jacket structures often involve calculating the probabilities of structural members failing and the overall collapse probabilities of the jackets existing in the corresponding damaged condition. It is current best practice to consider only single member failure conditions and to ignore multi member failure conditions since the probability of more than one member failing has been considered to be negligibly small. This approach assumes that the failure of a member is in no way related to the failure of another member, which implies that the two events are statistically independent. However, in reality the two events are not statistically independent since the failure of one member changes the stress state and therefore the failure probability of the remaining members. Two studies conducted to investigate the effects of stress redistribution on platform reliability predictions concluded that, by ignoring the effects of stress redistribution, platform reliability predictions could be significantly overestimated. However, they both concentrated on shallow water jackets and consequently a further study was performed to extend the existing work on stress redistribution to consider jacket structures in greater water depths. The study considered single and dual member failure conditions and compared the effects of including and excluding stress redistribution in reliability calculations. The findings of this work suggest that including stress redistribution reduces the predicted platform reliability. From the results, generic correction factors were derived for different inspection intervals, which account for stress redistribution and dual member failure. These correction factors can be applied to the probabilities derived using current best practice to determine a more realistic estimate of a structure’s reliability.