Computations of active flow control via steady blowing over a NACA-0018 airfoil: Implicit LES and RANS validated against experimental data

In this work, Computational Fluid Dynamics (CFD) simulation of active flow control via steady blowing is carried out on a thick symmetric airfoil (NACA-0018) at low Reynolds numbers (65000 − 250000). Within this Reynolds number range, the flow over the airfoil is characterized by leading edge separation, including the presence of a Laminar Separation Bubble (LSB) and associated transitional flow that is challenging to numerically predict. Indeed, the prediction of the exact location of the separation, transition and reattachment of the LSB has been the focus of many recent works.^{5, 6, 8, 9} Consequently, three approaches are examined in this work. Namely, the 2D and 3D Compressible Navier-Stokes (CNS) equations on unstructured grids using an in-house, high-order Flux-Reconstruction solver and 2D Reynolds-Averaged-Navier-Stokes (RANS) with the commercial package, STAR-CCM+. For the 3D computations, no explicit turbulence model is used and hence the calculations can be considered to be an implicit-LES (iLES) approach. The RANS equations are solved with the Realizable k − ɛ turbulence model. The results from each approach are compared with experimental data by Müller-Vahl et al.¹ where available. Care is taken to appropriately model the plenum slots through which a jet of fixed velocity (steady blowing) is used to control the separation on the airfoil surface. The control slots have a non-negligible effect on the aerodynamic characteristics of the basic NACA-0018 profile and thus this work differs from previous studies^{14, 15} using structured curvilinear grids to study the effect of blowing and/or suction on airfoils. Another novelty of this work is that in addition to validating the CFD results for reducing flow separation, we also consider the more challenging case of increased separation under specific jet blowing velocities as observed by Müller-Vahl et al.¹ Our results show that the 2D-CNS and RANS calculations using STAR-CCM+ are adequate for low angles of attack in the range of Reynolds numbers and jet blowing velocities investigated. Additionally, even for separated flows, the 2D calculations can be used when flow control is used to reattach the separated flow. The 3D-iLES approach on unstructured grids is the most accurate but has the drawback of requiring large computational resources. Nevertheless, it has the potential to supplement experiments as it enables an investigation into the unsteady mechanism of jet interaction near the plenum slots that can be used to either control or induce flow separation.