[en] This paper presents a method to asses hull girder ultimate bending moment distribution,
taking into account parameters uncertainty. This method is a part of a research activity
carried out at University of Liege. The method is able to compute four statistical
moments of ultimate strength distribution: average, standard deviation, skewness and
kurtosis. Geometry parameters, material parameters and initial imperfections (initial
deflections and residual stresses) are considered as random variables and a distribution
is associated to each of them. Ultimate bending moment algorithm is a progressive
collapse analysis based on Smith method and associated with the load-end shortening
model proposed by Bureau Veritas and Joint Bulker Project Rules (JBP). In order to
perform a reliability analysis, this algorithm is coupled with independent perturbations
method. The advantage is a lower number of simulations compared to the standard
Monte-Carlo procedure and a good use of the initial information that is available.
The paper contains also a comparative analysis between the presented method and a
linear sensitivity analysis performed by the ISSC 2000 “Ultimate Strength” committee
for a VLCC structure.
