ABSTRACT
Mean and turbulent characteristics are commonly evaluated by spatially averaging over a finite fluid volume. For Particle Image Velocimetry (PIV), the spatial averaging is accomplished over the interrogation area during image processing; for Particle Tracking Velocimetry (PTV), the spatial averaging is made in the sampling bins during post-processing. For both techniques, the obtained values are attributed to the center of the sampling volume under the assumption that the volume size can be set to be indefinitely small. In practice, however, a finite sampling volume has to be used in PIV to facilitate correlation analysis, and in PTV to collect enough samples for statistical analysis. While the spatial-averaging measurement approach is accurate for flow areas that are uniform, the presence of velocity gradient across the sampling volume will inherently introduce bias errors. This study presents both analytical and experimental evidence to show that the spatial averaging effect differs for PTV and PIV in quantifying turbulence intensity. Due to the spatial averaging effect, turbulence intensity is likely to be underestimated by PIV but overestimated by PTV. Therefore, care needs to be taken when results from various measuring techniques are compared to describe turbulence modifications.
