Experimental investigation of a turbulent boundary layer flow at high Reynolds number without and with separation using PIV. Design of the experiment, wall-laws of turbulence, and improvement of RANS models for maximum-lift prediction

Abstract

We present the design of a dedicated experiment and first results from corresponding PIV measurements of a turbulent boundary-layer flow at adverse pressure gradient without and with separation in a diffusor. The flow conditions are designed for the improvement of RANS turbulence models for flows around airfoils at high-lift conditions close to maximum lift. The design of the experiment and the wind-tunnel campaign are work of the DLR project RETTINA. RANS models are the state-of-the-art turbulence modelling approach in modern CFD codes for complete airplanes. However, they give reliable results only in the linear range of the lift-over-alpha curve. The reason is that they are derived for turbulent boundary-layer flows at small pressure gradients, as they reproduce the famous law-of-the-wall for the zero-pressure gradient case. For flows at a strong pressure gradient, there is no established law-of-the-wall in literature. The RANS model behaviour for such flows and in particular their predictions around maximum lift have significant uncertainties. We describe the design of the new experiment and the considerations taken into account, including the choice of PIV as measurement technique, the design of the diffusor geometry and the choice of the wind tunnel. The first aim of the present experiment is to obtain a dataset for studying the law-of-the wall for flows at a strong pressure gradient and as a reliable reference data set. The Reynolds number is expected to be high enough to have a significant overlap region between inner and outer part of the boundary layer (which is the log-layer in the case of a zero-pressure gradient flow). The wall-shear stress is measured directly using micro-PIV in the viscous sublayer. Integral boundary-layer quantities for scaling in outer coordinates are also measured. The question of breakdown of the law-of-the-wall can thus be investigated. The second aim of the work is to improve RANS models for reliable predictions around maximum lift without changing their well-established behaviour for flows at smaller incidence angles.