Accurate determination of acceleration statistics in a tur-bulent jet using locally adapted TrackFit

Abstract

In order to derive accurate acceleration statistics from experiments involving Lagrangian Particle Tracking, it is essential to filter the position noise from the tracks. The ‘TrackFit’ filter is espe-cially well-suited for this task. However, this filter is predicated on the estimation of temporal filter parameters from a model of the global amplitude spectral density of the measured tracks. Consequently, its effectiveness is diminished in inhomogeneous flows with a large range of ac-celeration variances depending on local flow regions (e.g. free turbulent jets in quiescent fluid). This paper presents a methodology for the automatic determination of requisite filter parameters at a local level, employing the local mean acceleration variance. This approach enables the utili-zation of the merits of the Trackfit approach, particularly its autonomy from track length and its efficacious noise suppression even in the high frequency band, in statistically stationary, inho-mogeneous (turbulent) flows. The results of an experimental study of a round turbulent jet in air utilizing helium-filled soap bubbles illuminated by LED arrays as tracer particles demonstrate the potential of the presented TrackFit version to outperform conventional polynomial-based filters, such as Savitzky-Golay, in an unparalleled range of fluid flows.