Bias in particle tracking acceleration measurement

TL;DR

This paper identifies key bias errors in particle tracking acceleration measurements and proposes filtering and independent measurement techniques to reduce these biases, improving accuracy in turbulence studies.

Contribution

It introduces methods to eliminate bias errors in Lagrangian Particle Tracking acceleration data, validated through simulations and experimental data, applicable with minimal camera setups.

Findings

Bias errors mainly come from noise and selection biases.

Filtering and independent measurements effectively reduce biases.

Techniques enable accurate acceleration statistics with fewer cameras.

Abstract

We investigate sources of error in acceleration statistics from Lagrangian Particle Tracking (LPT) data and demonstrate techniques to eliminate or minimise bias errors introduced during processing. Numerical simulations of particle tracking experiments in isotropic turbulence show that the main sources of bias error arise from noise due to position uncertainty and selection biases introduced during numerical differentiation. We outline the use of independent measurements and filtering schemes to eliminate these biases. Moreover, we test the validity of our approach in estimating the statistical moments and probability densities of the Lagrangian acceleration. Finally, we apply these techniques to experimental particle tracking data and demonstrate their validity in practice with comparisons to available data from literature. The general approach, which is not limited to acceleration statistics, can be applied with as few as two cameras and permits a substantial reduction in the spatial resolution and sampling rate required to adequately measure statistics of Lagrangian acceleration.