13th Int. Symp on Appl. Laser Techniques to Fluid Mechanics, Lisbon, Portugal, June 26 – 29, 2006
Application of 3D-PTV to track particle in fluid mechanics experiments
Antonio Cenedese1, John H. Cushman2 and Moroni Monica1
1: DITS – University of Rome “La Sapienza”, Italy
2: Mathematics Department - Purdue University, Italy
Keywords: 3D-PTV processing, particle tracking
1. Introduction
3. Results
There exist a number of imaging-based measurement
techniques for determining 3D velocity fields in an
observation volume. Among these are:
• scanning techniques;
• holographic techniques;
• defocusing techniques;
• photogrammetric techniques;
• Feature-tracking techniques FT.
We modify photogrammetric 3D-PTV by using feature
tracking. The reconstruction of position and object shapes
from photographs requires the camera/test section geometry
be known. The system must be calibrated, then
correspondences must detected to determine the 3D position
of the tracer particles and finally, 3D trajectories are
reconstructed via a tracking algorithm similar to the one
used in the 2D reconstruction of Moroni and Cushman
(2001). The time integral scales for two sets of data are
presented in Tab. 1 while the velocity correlation functions
are presented in figure 2.
We will focus our attention on 3D-PTV which is an
experimental technique based on reconstructing 3D
trajectories of reflecting generic image features through a
photogrammetric recording of image sequences. Feature
coordinates are determined first and then trajectories are
defined. 3D-PTV requires a lighted volume of the test
section as opposed to 2D techniques that require a light
sheet.
1/e
3 giri
6 giri
Txx (s)
0.8392
1.7791
Tyy (s)
0.6657
1.5873
Tzz (s)
4.3893
2.1154
2. Experimental set-up
Outlined in Figure 1 is the hexagonal cell (side length 24
cm, height 60 cm). As we are using an optical system based
on visible light, the porous matrix within the test section and
the fluid that bathes the matrix must be transparent to visible
light. Moreover, both the matrix and the fluid must have the
same refractive index so that matrix- fluid interface does not
scatter light. For a tracer we will continue the use of small
(<0.1 m) air bubbles for our Pyrex/glycerol mixtures and
fluorescent micro-beads for the cryolite/water systems.
Fig. 2 Comparison of velocity correlations along the
principal axes for two sets of data
Fig. 1 Hexagonal test section
16.5