Generalization of the PIV loss-of-correlation formula introduced by Keane and Adrian

Abstract

In 2D particle image velocimetry (PIV), the loss-of-correlation due to out-of-plane motion or light-sheet mismatch has two effects. First, it reduces the probability of detecting a valid vector. Second, it increases the uncertainty measured in velocity fields. The loss-of-correlation is commonly determined by the \(F_\mathrm {O}\) factor, which was initially proposed by Keane and Adrian (Appl Sci Res 49:191–215, 1992). However, the present study demonstrates that the validity of the original \(F_\mathrm {O}\) definition is confined to cases with identical laser intensity profiles. As light sheets usually differ in width and shape, the proposed definition is of limited use in reality. To overcome this restriction, a new definition for \(F_\mathrm {O}\) is proposed which covers the effect of light-sheet pairs with different shapes and widths. The proposed improvement was validated by means of synthetic PIV images based on various light-sheet profiles. The loss-of-correlation for the images was compared to the theoretical solution based on the light-sheet profiles. The results show that the new definition of \(F_\mathrm {O}\) accurately predicts the loss-of-correlation for all tested laser mismatches and agrees with the old definition for the ideal case involving identical light sheets. Based on the revised \(F_\mathrm {O}\) definition, prediction of loss-of-correlation due to light-sheet mismatches and misalignment is now possible using a laser profiling camera. This allows the optimization of a laser prior to any PIV measurements. For the case of out-of-plane motion, the loss-of-correlation can also be estimated from the correlation function of the PIV images. Thus, it is possible to optimize the laser alignment to match the flow conditions while setting up an experiment. These findings help experimentalists to understand and control the sources of errors associated with out-of-plane effects and help to minimize the measurement uncertainty.