Harris, Jeffrey C. (Université Paris-Est) | Kuznetsov, Konstantin (Université Paris-Est) | Peyrard, Christophe (Université Paris-Est, EDF R&D) | Saviot, Sylvain (EDF R&D) | Mivehchi, Amin (University of Rhode Island) | Grilli, Stephan T. (University of Rhode Island) | Benoit, Michel (Aix-Marseille Univ.)
We report on recent developments of a three-dimensional (3D) model for wave propagation and wave-structure interaction. The velocity field is solved with a boundary element method (BEM), based on fully non-linear potential flow. This approach is efficiently parallelized on CPU clusters. Recent progress is presented for extending the model for the use of higher-order elements (i.e., cubic B-splines), and outline the future steps necessary to a high-order approach on completely arbitrary meshes necessary for complex industrial applications. Particular care is taken with regards to the corner compatibility condition along the intersection between the body and free-surface, which is necessary for high-accuracy modeling with the BEM approach. Applications are shown for academic tests as well as for the computation of wave-induced forces and moments on gravity-based foundations, where we compare numerical results against laboratory experiments. Such applications are of interest to the continued development of foundations for offshore wind farms, and extensions to this model are being implemented for simulating floating structures and coupling to other models including viscous effects, which can be important in some cases.
A large variety of ocean wave models have been applied to investigate wave-structure interaction; ever since the work of Longuet-Higgins and Cokelet (1976), the boundary integral approach based on potential flow theory has shown some interesting advantages, particularly as the calculations are only performed on the surfaces and not the interior of the domain. In the models, different ways to handle the free-surface have been proposed, both in frequency and time-domain, but in some cases where fully nonlinear effects are important, the standard approach has been to solve Laplace's equation for the velocity potential (mass conservation) at each time step (optionally multiple times, or for the time-derivative of the velocity potential), then updating the BEM mesh nodes and free surface boundary conditions with a mixed Eulerian-Lagrangian (MEL) approach. Tanizawa (2000) made a review to date of this technique.