ENVIRONMENTAL CONTOURS BASED ON A DIRECT SAMPLING APPROACH AND THE IFORM APPROACH: CONTRIBUTION TO A BENCHMARK STUDY

摘要

Environmental contours are often applied in probabilistic structural reliability analysis to identify extreme environmental conditions that may give rise to extreme loads and responses. It represents an approximate method for performing long-term extreme response analyses in cases where full long-term analyses are not feasible due to computationally heavy and time-demanding response calculations. There are various methods for deriving environmental contours given a set of metocean data. These relate to different approaches for modelling the joint behaviour of the metocean variables, i.e., a joint distribution function fitted to the data, but also different ways of establishing the environmental contour given a joint distribution for the environmental variables. In light of this, a benchmark exercise was announced at OMAE 2019 [1], asking for contributions from different practitioners involved with environmental contours. Various bivariate datasets are provided and two exercises are specified for which different solutions are elicited. The first part of the exercise concerns the estimation of the actual contours, whereas the second part relates to the uncertainty characterization of the contours in light of sampling variability. This paper is a response to this announcement and provides one contribution to these benchmark exercises; environmental contours based on a direct sampling approach as well as contours based on the IFORM approach will be presented. Both sets of contours are based on the same models for the joint distribution of the environmental variables, i.e., a conditional model where the joint distribution is modelled as a product of a marginal model for one variable and a conditional model for the other. Both the joint modelling of the environmental variables and the different approaches to estimate environmental contours are described in this paper and the results for the provided datasets are shown.