Hybrid Turbulence Model Simulations of Blade Section in Dynamic Stall

Dynamic stall on helicopter blades increases power consumption and causes unsteady aerodynamic loads, vibrations, and noise. The prediction of dynamic stall onset is of interest to designers and operators of helicopters. For Reynolds numbers up to one million and higher, dynamic stall can be initiated by the bursting of a laminar separation bubble. This has been demonstrated in experiments and simulations. The accurate prediction of bubble bursting in simulations requires that the separation bubble physics are captured accurately. While wall-resolved large-eddy simulations can fulfill that requirement, they are computationally expensive. This paper explores the feasibility of computationally more affordable hybrid turbulence modeling strategies based on the Spalart-Allmaras and Menter shear-stress transport model in combination with the Coder transition model for predicting dynamic stall onset. The case under consideration is a blade section with NACA 0012 airfoil that is subjected to a harmonic pitching motion and a constant-rate pitch-up maneuver. The chord Reynolds number is one million and five hundred thousand, respectively. Hybrid simulations with Menter shear-stress transport and transition model capture the unsteady flow physics with reasonable accuracy. With Spalart-Allmaras model, bubble bursting is delayed. Two-dimensional simulations and three-dimensional simulations with narrow spanwise domain extent can capture bubble bursting but fail to accurately represent the post-stall aerodynamics.