Nonlinear ship motion models to predict capsize in regular beam seas

The problem of vessel capsize in regular beam seas has been studied in this thesis through numerical simulations and laboratory experiments. A fully nonlinear time-domain ship motion model based on the mixed Euler-Lagrange time stepping scheme coupled with the desingularized boundary element method is applied to the computation of three-degree-of-freedom motions of a two-dimensional body floating in the incident beam seas. The projections of the state variables obtained from fully nonlinear computations onto the roll phase plane are used to validate the feasibility of the basic assumptions employed in a multi-degree-of-freedom nonlinear dynamics model. Numerical simulations with three-DOF blended model have been developed to account for the effects of water on deck on the global responses of a floating two-dimensional barge in the incident beam waves which are modelled equivalent to those generated in a laboratory gravity wave tank. This three-DOF blended model proved capable of predicting extreme roll dynamics of vessels including capsize in regular beam seas. Through extensive numerical simulations with the three-DOF blended model by systematically varying initial conditions imposed on the two-dimensional barge model, numerical fractal capsize basin boundaries have been constructed. Correlation with the laboratory experimental capsize basin boundaries was excellent with an unexpected level of engineering accuracy. Extensive capsize experiments on a two-dimensional barge model in a laboratory wave tank have been conducted to investigate the sensitivity of initial conditions to the responses of the subject model in the incident beam waves. The laboratory experiments also revealed the important role of the dynamic behavior of the water on deck on the global responses of the current dynamical system. Experimental capsize basin boundaries have been established through the collections of initial Poincaré samplings. These experimental capsize basin boundaries are expected to be used as a solid foundation to validate nonlinear ship motion models in predicting vessel capsize in beam seas.