Very-large-scale structures in pipe flows are characterized using an extended Proper Orthogonal Decomposition (POD)-based estimation. Synchronized non-time-resolved Particle Image Velocimetry (PIV) and time-resolved, multi-point hot-wire measurements are integrated for the estimation of turbulent structures in a pipe flow at friction Reynolds numbers of 9500 and 20000. This technique enhances the temporal resolution of PIV, thus providing a time-resolved description of the dynamics of the large-scale motions. The experiments are carried out in the CICLoPE facility. A novel criterion for the statistical characterization of the large-scale motions is introduced, based on the time-resolved dynamically-estimated POD time coefficients. It is shown that high-momentum events are less persistent than low-momentum events, and tend to occur closer to the wall. These differences are further enhanced with increasing Reynolds number.

The present study proposes a data-driven strategy to extract the dynamics of wave-packets from low-speed velocity field measurements. The flow field under study is a subsonic turbulent round jet at Re = 33000, measured over a 20-nozzle-diameters domain in the axial direction using low-repetition-rate planar Particle Image Velocimetry. The turbulent features in the jet have been extracted by Proper Orthogonal Decomposition of the jet velocity field over the whole extension of the domain. The modes produced describe the evolution of the turbulent features in the axial and radial directions of the jet. The temporal evolution of turbulent flow structures and the associated pressure fluctuations, which are not directly accessible from the low-speed measurements, have been estimated using an advection-based Galerkin projection model. The extracted velocity/pressure modes describe a set of modulated waves in the axial direction, which shares many similarities with wave-packets already observed in literature. This behaviour suggests that the employed strategy is effective in retrieving the dynamics of wave-packets extending over the entire measurement domain, paving the way to the estimation of their sound emission from low-speed measurements.