Parameter adaptation of k − ω SST turbulence model for improving resolution of moderately separated flows around 2D wing and 3D ship hulls via EnKF data assimilation

Parameters of turbulence model have rarely been changed to solve flow around ship hulls. Since these values are determined using relatively simple flows such as two-dimensional flat plate, there may be other parameter sets which are more suitable for resolving complex turbulent flow around ship hulls. Present study investigates suitable parameters of the k − ω SST model, one of the most popular turbulence models among ship flow simulations, throughout the data assimilation with ensemble Kalman filter aiming to improve computational accuracy of stern wake. Five parameters of k − ω SST model are first selected and are subjected to data assimilation using two-dimensional NACA4412 airfoil with moderate separation under 13.87 degrees of attack angle. Then, assimilated value “ \(\beta_{1} = 0.{16}\) ${\beta }_1=0.16$ (originally 0.075)” which controls the production and destruction of ω is applied to solve four different blunt hull forms under towed condition. Bulk flow and available turbulence quantities on the propeller plane are subjected to validation. Consequently, the model parameter “ \(\beta_{1} = 0.{16}\) ${\beta }_1=0.16$ (originally 0.075)” has been found to improve axial velocity distribution in model scale on the propeller plane in four different blunt hull forms maintaining the change of total resistance coefficient less than 0.8% difference compared to the original parameter. This research is of its first kind to bring the idea of parameter adaptation of turbulence model via data assimilation in the field of ship hydrodynamics.