High-fidelity turbulence modeling for numerical wave tank applications: Evaluation of the scale-resolving PANS approach

Abstract

For analyzing vessel seakeeping, platform wave loads and similar tasks, potential flow-based methods are still the workhorse Numerical Wave Tank (NWT) used by industry. Recently, NWTs have employed more advanced viscous flow simulation methods. These are commonly based on the Reynolds-Averaged Navier-Stokes (RANS) approach. RANS is often preferred by industry due to it’s lower computational cost in relation to better scale resolving methods. The RANS approach however, has limits in NWT applications. Firstly, it is not suited for studying unsteady turbulence phenomena at different scales of wave and fluid motion. Secondly, RANS turbulence models assume that the energy transport in turbulent flows can be described as a dissipative process. In a NWT where large parts of the flow is unsteady, but not dominated by turbulent variations, this could lead to non-physical dissipation of waves. Focusing on the second issue, the applicability of the Partially-Averaged Navier-Stokes (PANS) approach in NWTs is investigated in this paper. PANS is a suite of turbulence models of varying modeled to resolved scale ratios ranging from RANS to Direct Numerical Simulation (DNS). PANS modeling allows for more direct control of how energy dissipation is treated. It is hypothesized that this will be beneficial in NWTs by reducing non-physical dissipation of waves. Initial results for a regular wave tank are promising. About 40% of the wave amplitude was lost downstream with the RANS approach versus less than 3% with the PANS approach; using the same computational grid and with only a slight increase in the computational effort. Similar results were obtained for a self-contained ”dam break” simulation.