A hybrid PIV/PSV approach for a large dynamic velocity range

Abstract

Particle image velocimetry (PIV), particle tracking velocimetry (PTV) and particle streak velocimetry (PSV) are well-established image-based measurement techniques for flow velocity fields; the main difference among the techniques lies in the timing sequence, i.e., the choice of illumination/exposure times and the time between image frames. These are chosen with respect to the expected range of seeding particle movement in time and the desired temporal resolution of flow fluctuations. Difficulties arise when very large differences in velocity occur within a single planar field of view, such that any single timing sequence is inappropriate for certain regions of the observed field. If for instance, the illumination time is adjusted to suit a PIV measurement at low velocities, high velocities in other flow regions will result in streaks. One question to be addressed is therefore, to what extent PIV processing algorithms can properly cope with streaks, which are possibly truncated in either of the consecutive images? In this context, the notion of ‘truncation bias’ will be introduced and evaluated. The present study investigates the possibility of combining image processing algorithms of PIV and PSV in a hybrid manner, while using a single timing sequence for image recording. After image recording, the images are segmented into regions according to which processing algorithm is most appropriate—PIV or PSV—and the respective algorithm is applied to each of the segments. This is followed by a recombination of the computed velocity field from each segment to obtain a result for the total field of view. The performance of this hybrid approach will be evaluated by generating and processing synthetic input data from a direct numerical simulation of a transitional boundary layer. An example laboratory flow is then used to demonstrate the hybrid technique in practice.