Extreme structural responses by nonlinear system identification for fixed offshore platforms

Abstract

Offshore structures are exposed to random wave loading in the ocean environment. The response is commonly non-Gaussian due to the nonlinearity of the drag component of Morison's wave loading and also the load intermittency on members in the splash zone. The previous study shows that the developed finite-memory nonlinear system (FMNS) model can efficiently simulate the response of an offshore structure to random Morison wave loading. However, the FMNS models minimise the computational effort, and the estimates are not particularly good for lower significant wave height value. To overcome this deficiency, a modified version of FMNS models (MFMNS) is used to compute the extreme response values which increase the accuracy but is computationally less efficient than the current FMNS models. In this study, the probability distributions of extreme responses from both models are compared with corresponding distributions from the conventional time simulation technique to establish their level of accuracy.