Fluid viscosity is known to influence hydrodynamic forces on an oscillating body when its motion amplitude is large and its shape is bluff. An experimental program is developed to measure the roll inertia and damping of a floating rectangular cylinder, a problem of intrinsic interest in evaluating the resonant motion of ships in waves. Results from the experiments are compared with those from a recently developed method for modeling vortical flow in the presence of a free surface. This free-surface random-vortex method (FSRVM, Yeung &; Vaidhyanathan, 1994) is observed to yield predictions consistent with the measured results. However, cases exist where both results produce added roll inertia that differ from those measured by Vugts (1968). A promising roll hydrodynamic loading model that includes both linear (wave-related) and nonlinear (flow-separation related) damping is also proposed and evaluated in the paper.
In order to predict the roll motion of marine vehicles with accuracy, viscous-flow models need to be developed. While a number of empirical methods exist for engineering purposes (see Himeno, 1981), predictions based on rational-mechanics methods are only emerging. When a vehicle undergoes large-amplitude motion, separation occurs around the bilges, and it is conceivable that this would change the hydrodynamic character of the flow from what is normally predicted by inviscid-fluid flow theory (Wehausen, 1971). The nature of this inviscid and viscous coupling has received some attention (Yeung & Wu, 1991, Yeung & Ananthakrishnan, 1992, Gentaz et al., 1997), even though a complete understanding is yet to be achieved. In a sequel of works (see Yeung & Cermelli, 1998 for a short review), an effective grid-free numerical method was developed to model viscous flow in the presence of a free surface. The method is based on the "Random Vortex" algorithm, but properly improved to accommodate bodies of arbitrary shapes and reformulated to include the effects of free surface.
