Computations of Turbulent Transition in Boundary Layer with LES

Abstract

Turbulent transition is a critical phenomenon in boundary layer flows. It is significant to be able to predict the transition in designing a moving object and improving the vehicle performance. In this study, laminar-to-turbulent transition in boundary layer flow is investigated using large-eddy simulation (LES) for high-fidelity computations. It is important to determine a proper subgrid-scale (SGS) model for LES. The wall-adapting local eddy-viscosity (WALE) model [1] is selected in this study because its capability in capturing turbulent transition. Although this model has been used for various flow conditions [2–6], how the model behaves in transitional boundary layer has not been fully investigated. Here, the performance of the WALE model is thoroughly investigated for the canonical case of zero-pressure-gradient boundary layer on a flat plate. The case of sub-harmonic resonance studied in the experiment [7] is reproduced in the current wall-resolved LES. Current LES computations are compared to to the current parabolized-stability equation (PSE) and direct-numerical simulation (DNS) data [8, 9]. Analysis of the WALE formulation reveals that the traceless, symmetric tensor of the square of the velocity gradient is active in the turbulent region, whereas the strain-rate tensor is dominant in viscous regions including laminar boundary layer and the viscous sublayer.