Investigation of turbulent/non-turbulent interfaces in high Reynolds number adverse pressure gradient boundary layers

Abstract

This study utilises high-magnification two-dimensional Particle Image Velocimetry (2D-PIV) to investigate adverse pressure gradient (APG) turbulent boundary layers at high friction Reynolds numbers, which evolve from a canonical zero pressure gradient (ZPG) upstream condition. The primary focus is on comparing the turbulent/non-turbulent interface (TNTI) for ZPG and APG at high Reynolds number. Local kinetic energy (LKE) and spanwise vorticity methods are considered for TNTI detection, with Joint Probability Density Functions (JPDFs) for LKE/vorticity and wall-normal distance used to determine the respective thresholds and estimate corresponding TNTI heights. A sensitivity analysis of the mean TNTI height with respect to the selected threshold for distinguishing turbulent and non-turbulent regions is conducted for both methods, with the LKE method demonstrating a lower sensitivity to the threshold compared to vorticity. The results confirm the inadequacy of using in-plane vorticity for the current experimental data while also highlighting the limitations of the LKE method. Overall, the present findings based on relatively small thresholds support a decrease in normalised TNTI height with increasing adverse pressure gradients at high Reynolds number, demonstrating consistency with previous low Reynolds number results. Conditional averaging analysis is conducted for the instantaneous streamwise and wall-normal velocities based on TNTI height, confirming that the APG outer wake region is distinctly different to the ZPG case. For the APG flow, the spread in the conditional averaged velocity curves is larger than in the ZPG case, with the start of the deviation from the mean velocity profile occurring closer to the normalised position from the wall.