Author(s): Jose M. Grassa
Linked Author(s): José María Grassa Garrido
Keywords: Maritime hydraulics; Wavemaker; Wave forces; Numerical model; SPH
Abstract: Knowledge of wave forces on a wavemaker during experiments is an important question for the design of structural and driving system design of such elements. Also, it may be useful for active wave absorption of reflected waves and it may be considered as a very simple case of study of wave forces on a structure. Within a systematic study of a numerical SPH wave flume, this paper uses a simple procedure for obtaining such forces representing the boundary of the flume and, hence, the wave maker, by means of repulsive force boundary particles as proposed by Monaghan (1994). While other formulations for force boundaries do exist that provide a smooth boundary force field, Monaghan (1999) and other approaches like mirror particles may enable a better approximation of essential mathematical boundary conditions, this simplest implementation has some interest because it mimics in a physical way molecular forces, is easy to implement and may be applied in a generalized system without the need to consider special cases. Some aspects of the noise that appears in the simulations is discussed, and numerical results are compared with analytic results concluding that SPH may provide a rather appropriate evaluation of the force for the studied conditions, if the distance between boundary particles is made shorter than the initial distance between fluid particles while keeping a radius of influence of the boundary force of the order of magnitude of fluid particles separation, in order to reduce the added mass effect associated with this boundary condition. Numerical experiments have been conducted at three resolutions obtaining satisfactorily converging results that may imply that not an extreme resolution is needed at least in this case, an important fact given the computational cost of the method. Even if the boundary method requires an increased number of boundary particles, the dimensionality of the boundary is lower than that of the experiment and therefore this may not be a severe computational constraint in most cases.
Year: 2007