Development of Large-Scale Particle Tracking Velocimetry: A Comparison Between 2D and 3D Techniques

Abstract

This study compares the three-dimensional version of a previously proposed large-scale particle tracking velocimetry (LS-PTV) measurement system with a simpler, less computationally expensive, two-dimensional version for use in in-situ measurements. The LS-PTV (3D) system was used as a baseline for measuring wind flow over a flat roof in a 3m x 3m x 1.5m volume. The LS-PTV (2D) system covered a 1.5m x 3m plane within the same volume. A 3D ultrasonic anemometer was used outside of the measurement volume as a means to validate the LS-PTV measurements. An analytical model of the bubbles used as tracer particles for the LS-PTV measurements determined that the bubbles were able to resolve fluid length-scales 0.25m or larger, which was within the size of the measurement volume. Despite being more susceptible to errors due to incomplete background image subtraction, which was responsible for noise in the tracking algorithm, the LS-PTV (2D) system was able to track approximately 80% more bubbles and at a lower spatial uncertainty than the LS-PTV (3D) system. The LS-PTV (2D) system was unable to measure an exact projection of the LS-PTV (3D) results, which was a result of the LS-PTV systems tracking different bubbles within the same flow field. Both LS-PTV systems were found to measure the same velocities and Reynolds stresses within the field, and the missing vertical component from the LS-PTV (2D) measurements was deemed negligible.