Experimental and numerical investigation of turbulent boundary layers with strong pressure gradients

A detailled investigation of a turbulent boundary-layer flow subjected to a strong adverse pressure gradient (APG) is presented. The main goal is to define a test case for the validation and improvement of RANS-turbulence models from wind-tunnel measurement data collected over the course of multiple measurement campaigns, including volumetric Lagrangian Particle Tracking (LPT) and stereoscopic PIV (SPIV), and oil-film interferometry. The boundary layer at a zero-pressure gradient (ZPG) reference position upstream of the pressure gradient region is found to exhibit a mild deviation from a canonical flow in the sense that the boundary layer thickness and hence the Reynolds number based on the momentum loss thickness Re_θ are larger than for a canonical flow. Moreover a mild deviation in skin-friction coefficient and shape factor is found. The experimental data using LPT and SPIV in a spanwise domain around the centerplane show an increase of the boundary layer thickness compared to a canonical flow and a spanwise variability. This can possibly be attributed to the wake flow of the turning vanes upstream of the nozzle and the test-section. For the mean velocity profiles, this leads to a deviation in the law-of-the-wake region compared to canonical flows. The inner region, which is essential for the turbulence modelling and validation, is largely unaffected and agrees well with canonical flows. The Reynolds stresses are also in good agreement with canonical flows. Regarding the ultimate aim to define the computational set-up for RANS simulations, a pragmatic approach is pursued. The inlet length of the test-section is increased to account for the larger boundary layer thickness, corresponding to an adjustment of the virtual origin of the boundary layer. This leads to a good matching with the experimental mean velocity profile and the boundary layer parameters at the ZPG reference position. Downstream, in the pressure gradient region, which is the focus region for the improvement and validation of RANS turbulence models, the deviation between the RANS results and the experimental data is found to be almost insensitive with respect to minor changes in the computational set-up. In the strong APG region, the clearly most important deviation between the numerical predictions and the experimental data is due to the RANS turbulence models used.