On numerical errors and turbulence modeling in tip vortex flow prediction

Effects of Turbulence Modeling on RANS Simulations of Tip Vortices

“…The FRSM fine grid simulation predicts both peak velocity components to within measurement uncertainty! The data also shows a big success for the very simple "ad hoc" [27] correction term, (referred to as SA w/ correction) particularly for predicting the axial velocity deficit. An interesting observation is that the FRSM coarse grid simulation yields better results (with regard to vortex resolution) than the SA models (both versions) on the fine grid, requires less computational resources, less time to convergence, and solves for Reynolds stresses.…”

“…Devenport et al [15] and Heyes et al [18] address this issue. Computational investigations are numerous as well [14,[22][23][24][25][26][27][28]. The majority of computational investigations are focused on grid generation and algorithm development.…”

“…But, other than using a vortical correction, little attention was given to the effect of turbulence modeling. By convecting a laminar isolated vortex Dacles-Mariani et al [27] agree that a fifth-order upwind biased scheme is required but that 15 -20 grid points are necessary for grid independence. Egolf et al [29] convect a Lamb-Oseen vortex and show that 14 grid points across the vortex core and fifth order scheme was required for "minimal diffusion for a long distance".…”

“…Dacles-Mariani et al [27] perform a computational companion study with Chow et al [20]. They first solve a subset problem involving convection of a laminar isolated vortex with the purpose of finding a suitable grid and differencing scheme.…”

“…where, Model constants: [27] and Duraisamy [22]. This modification is available in FLUENT CFD code as an option and is tested in this thesis.…”