Higher-order statistical moments of wave-induced response of offshore structures via efficient sampling techniques

Abstract

Nonlinear wave loading leads to non-Gaussian offshore structural response so that higher-order statistical moments are often necessary for its probabilistic description. The existing models for determination of these moments are computationally very demanding. Consequently, the distributed wave loading on the structure is idealised by a relatively small number of nodal loads, requiring care and experience in the representation of the continuous loading on (complex) structures with many structural elements. However, importance sampling enhanced by a so-called 'Gaussian Booster', as described herein, has led to a dramatic reduction in computational effort so that the distributed loading can be idealised more realistically by a large number of nodal loads. The effectiveness of the proposed procedures, which have arisen from a UK EPSRC-sponsored project, are demonstrated by applying them to a test structure under different environmental conditions.