Abstract
Turbulent convection velocities in a turbulent boundary layer at a Reynolds number of R e 𝜃 = 2250 are examined via the use of a high repetition rate particle image velocimetry measurement undertaken in a water tunnel. Multiple cameras are used to improve the spatial dynamic range of the measurement and reduce the bias towards large-scale structures while simultaneously capturing a wall-normal domain of 0.06δ to 1.7δ. The impact of measurement noise is minimized via careful temporal and spatial filtering of the velocity fields as guided by the comparison of temporal and spatial velocity power spectra with spatially filtered direct numerical simulation data, enabling an estimation of the effective noise-limited spatial and temporal dynamic range of the present experimental measurement. Space-time correlations and phase-spectra are used to estimate the mean and streamwise wave-number dependent convection velocities at various heights above the wall. Results reveal convection velocities greater than the local mean velocity in the lower log layer, decreasing to a level 3.5 % lower than the mean velocity in the upper log and wake regions. The convection velocity is shown to depend on the streamwise length scale and is found to decrease at higher wave-numbers for all wall-normal locations. Comparison between the measured and reconstructed spatial fields show that Taylor’s hypothesis can only be applied over short streamwise distances of less than 1δ in the buffer and inner log-layer, while larger projection distances (≥3δ) are possible in the outer-log and wake region of the turbulent boundary layer.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call