TITLE: Bridging the Gap Between High-Fidelity and Deep-Water Wave Models with a Phase-Revolving Boussinesq-Type Numerical Model FUNWAVE (FULLY NONLINEAR BOUSSINESQ-TYPE WAVE)
PRESENTER: Matt Malej, Ph.D.,
Research Mathematician, Coastal and Hydraulics Laboratory (ERDC)
ABSTRACT: This presentation will showcase the FUNWAVE model and its various applications. Multiple real/physical case studies and USACE Civil Works projects will be discussed. In addition, new development and deployment efforts through the HPC Portal will be presented.
Modeling nonlinear nearshore coastal wave processes, such as overtopping/inundation, wave runup, bore propagation, tsunamia, harbor resonance, ship waves/wakes, and infragravity waves, requires efficient and accurate computing of the evolution of highly nonlinear time-dependent free surface wave fields in complex coastal environments. This is a challenging hydrodynamic problem. Most models commonly used for describing nonlinear surface waves are far from being complete. They rely on ad hoc models for the physical processes involved, such as nonlinear wave-wave interactions, energy dissipation due to wave breaking, or interplay between waves and currents.
An additional complication in modeling coastal waves is that there is a wide range of scales to be resolved (e.g, wind and infragravity waves). The coupling between various modes and thus the energy transfer between different spatial and temporal scales lacks thorough understanding. Operational phase-averaged wave-action balance models suffer from inaccurate predictions of the wave spectrum in shallow waters.
With improvements in the high performance computing (HPC), phase-resolving models like FUNWAVE are becoming more practical to apply. Their primary area of application thus far has been a shallow-water environment. Boussinesq-type models are especially attractive in these regimes, where strong nonlinearity and weak dispersion are prevalent. The USACE has a pressing need for a robust and computationally efficient phase-resolving numerical wave model. It is important that such a model be efficient and developed with HPC application in mind, without the immediate and cyclical need to resort to propriety software packages such as for example MATLAB®.