A three dimensional (3D) Boundary Element Method (BEM) model has been developed to simulate Ii numerical wave tank. Using Taylor series expansions about the mean position of the boundary and perturbation techniques for the variables at the boundary provides a second order problem formulated in a time-invariant domain. Hence, the computational costs will be significantly reduced compared to those of a fully nonlinear model. A new technique has been used to absorb the outgoing waves of the fluid domain. The technique is based on active wave absorption by using an array of wavemakers. Numerical results for absorption of oblique incident waves are given.
A major challenge in wave tank experiments (physical as well as numerical) is to model open boundaries i.e. to simulate that there is no boundary at all. Several techniques for that purpose involving e.g. radiation conditions, sponge layers and active: wave absorbers can be found in the literature. Active absorption of waves by using a wavemaker has been performed in physical wave flume experiments during many years. In numerical models the active wave absorption has also been considered as a method for letting waves progress out of the fluid domain (see e.g. Clement, 1988, Brorsen and Frigaard, 1992, Skourup and Schäffer, 1995, Clement and Domgin, 1995). Common for these models is that they are all two dimensional (2D). For three dimensional (3D) wave tanks the active wave absorption technique has not yet been developed. However, one step towards active wave absorption in a wave tank is to use a quasi-3D wave absorber which is simulated by using an array of independently controlled 2D active wave absorbers. This idea is pursued in the present paper. Among the numerical methods for potential theoretical simulations of numerical wave flumes (2D) and wave tanks (3D) is the Boundary Integral Equation Method (BIEM) which is solved numerically and then referred to as the Boundary Element Method (BEM).
