Nonlinear stochastic models for ocean wave loads and responses of offshore structures and vessels

Second-order nonlinear models have been increasingly used to model the behavior of offshore structures. We apply such models to study random ocean waves and also apply it to study the nonlinear wave loads on offshore structures. To reduce the computational expense in a time-domain nonlinear analysis, we develop efficient methods to estimate structural fatigue damage.; Traditional linear models of wave loads tend to be inaccurate in structural response predictions. Second-order nonlinear models that provide an opportunity to better predict these loads, are based on a perturbation expansion of the linear form of the associated problem.; In modeling the random ocean waves as a second-order phenomenon, we develop convenient analytic formulae to predict the nonlinearities as function of the wave climate and of the water depth. We compare the model predictions to measured waves in wave tanks and in oceans.; We model a spar floating platform as a linear rigid-body with 6 degrees-of-freedom and with incident second-order wave forces. Of interest is the global response of the spar, which here is the total horizontal displacement near the spar deck. With such a model we find the predictions to reasonably agree with wave tank measurements.; We show an efficient method to estimate fatigue damage which is beneficial when the non-linear analysis is computationally expensive. We demonstrate this method for a ship where a two-dimensional strip theory is used to find the nonlinear wave loads. Here, damage estimated from a carefully selected set of sinusoidal waves from stochastic theory, seems to well approximate the damage estimated from a full time domain ship analysis in random waves.; Finally, we also show a fatigue reliability example for a ship structure where the probability of failure is found by integrating uncertainties at all three levels: wave climate, response given climate and material properties. Here we show an efficient method to select the associated design parameters in order to achieve a preselected reliability level.