Lagrangian interpolation algorithm for PIV data

Abstract

A new Lagrangian based interpolation scheme is proposed for recovering velocity field data from within masked regions in particle image velocimetry (PIV) experiments. As a first step, the mean field within the masked region is filled through an iterative convolution operation using an optimized kernel. Next, the Lagrangian interpolation scheme estimates velocity field data within the masked region using a weighted average of past and future snapshot data along a mean field streamline from outside of the masked region. The Lagrangian interpolation scheme is compared directly with a traditional bilinear interpolation approach for two datasets: the flow development over a circular cylinder within the laminar vortex shedding regime (ReD=150) and turbulent vortex shedding regime (ReD=3000). The results show that for both Reynolds numbers investigated, the Lagrangian approach provides up to a 5× improvement in average velocity reconstruction error. Application of the proposed interpolation scheme to the estimation of body forces was assessed. For both the momentum and impulse control volume approaches to estimate instantaneous forces, the Lagrangian interpolation scheme provides up to 10× improvement in the estimation of volume integral contributions to instantaneous drag and lift.