Simulating Breaking Waves with the Reynolds Stress Turbulence Model

In the past two decades, previous computational fluid dynamics (CFD) studies of breaking waves have shown a marked tendency to severely over-estimate turbulence levels, both pre- and post-breaking (e.g. Brown et al. 2016). Larsen and Fuhrman (2018) have proved that this problem is due to the unconditional instability of two-equation turbulence closure models (both k-ω and k-ε types) in the potential flow core region beneath surface waves. A method for formally stabilizing two-equation models was derived in their work, and it was shown that stabilized two-equation models lead to pronounced improvement in the predicted turbulence and undertow velocity profiles prior to breaking and in the outer surf zone. However, even the stabilized two-equation models in Larsen and Fuhrman (2018) were still rather inaccurate in the inner surf zone (i.e. closer to the shoreline), thus seemingly requiring yet more advanced methods of achieving turbulence closure.