Three-dimensional Particle Tracking Velocimetry for Turbulence Applications: Case of a Jet Flow

The goals of this experiment are to set up and optimize a particle tracking velocimetry system for turbulence applications and use it to study jet flow. This method can help to answer key questions in the turbulence field using Lagrangian description of the flow. The main advantage of this techniques is its ability to track a large set of particles in time and space.

Though this method can provide insight into a turbulence and Lagrangian frame of reference it can be applied to other application such as motion capture or motion tracking. The set up for this experiment is for the study of the intermediate flow field of a pipe jet. This water filled tank has within it a flume to create a jet with the diameter of one centimeter.

This schematic provides an overview of the setup before the addition of equipment for three dimensional particle tracking velocimetry. To prepare for particle tracking, identify a region of study, the interrogation volume. In this experiment it is a cube with its faces parallel to those of the larger volume and its closest point 15 centimeters from the end of the flume.

Once the interrogation volume is chosen, begin with positioning the primary mirror. The mirror is pyramid shaped and mounted on a vertical mounting post along the side of the experiment tank. Slide the mirror along the post to level it with the center of the interrogation volume then fix it in place.

Next, mount the camera behind the primary mirror. Set the center of the camera image to be coincident with the center of the primary mirror. Now, move to the computer to check its connection with the camera.

Continue by setting up the interface between the camera and recording software. Begin with the camera view of the primary mirror and its surroundings. Use the software's region of interest setting to adjust the width and height of the camera view to just cover the primary mirror.

The next step is to place a calibration target in the tank to set up the secondary mirrors. This target is custom designed to encompass the entire interrogation volume. The target has one millimeter diameter target points for calibration.

First, place an adjustable height platform into the tank to support the target. Position the calibration target on the platform and orient it to face the camera. Adjust the height of the platform so as to match the center of the calibration target with the center of the calibration volume.

Proceed by completing the four view splitter. Adjust the position of the primary mirror so that it captures the full interrogation volume. The secondary mirrors are mounted on vertical mounting posts on either side of the primary mirror.

Mount each secondary mirror so that it is roughly aligned with the camera view from the closest side of the primary mirror. Secure the secondary mirror in place. The final arrangement should be geometrically symmetric with respect to the primary mirror.

Use a laser pointer and direct it onto a mirror in turn to visualize the image path. Make final adjustments to the secondary mirror using the beam path as a guide so the view encompasses the entire calibration target. Repeat the visualization of the image path and make adjustments for each of the remaining secondary mirrors.

After all mirrors have been adjusted, monitor the sub-images of the camera view. At the computer, observe the camera view of the images. At the view splitter, move one of the secondary mirrors to investigate sub-image overlap.

If only one view changes as the mirror moves there is negligible overlap. With the four way splitter in place, turn off the camera and move on to work with the lighting. Place light sources directly facing the interrogation volume from above.

In this experiment a magnifying lens is put in position directly under the light sources to enhance light intensity. Turn the light sources on and visually check that light is uniformly distributed over the entire interrogation volume. Begin optimizing the set up by returning to work with the camera.

Adjust the lens focus until the reflection through the primary mirror is equally focused in all four views of the secondary mirrors. Ensure the images from the view splitter are symmetric and show the interrogation volume. Move on to adjust the f-stop to be able to capture both the closest and furthest point on the calibration target.

The camera will only capture tracer particles in the interrogation volume. At this point, turn to the recording software. Set the desired frame rate.

550 Hertz is used for this experiment. Then maximize the light sensitivity in accordance with the chosen frame rate. Now, prepare to take calibration images of the calibration target.

Attach an LED flashlight to the tank directly above the target for illumination. Capture several images of the target via the four view splitter for later calibration steps. When done remove the calibration target, its support stand and the light in order to begin adjusting the illumination.

Prepare a small quantity of seeding particles for optimizing the illumination. Get a small beaker containing water and add the seeding particles to it. Wait for the particles to diffuse.

The particles are 100 micrometer silver-coated hollow ceramic spheres. Pour the mixture into the tank to assist with optimizing the illumination and camera settings. Continue with the room lights off and the light sources on.

On the computer, monitor the live camera view of the primary mirror and check the illumination in each view by observing the particle density difference between views. The geometry in this experiment causes the top two mirrors to receive less light. At the setup remedy this by adding a flat mirror below the light source at the bottom of the tank.

Under the light sources this addition helps reduce the variation across the view. At the camera adjust the gain and black levels each in the range of about zero to 500. The goal is to better capture light scatter from the particles.

In this sequence, three stills using different gain and black levels are compared. A gain setting of 300 and a black setting of 500 was used in this experiment. Turn off the light sources and turn on the room lighting.

At the computer open software for particle tracking velocimetry. The idea is to load one of the calibration images taken previously. Use the software to divide this image into four independent sub-images that correspond to the different secondary mirror views.

Save each of these sub-images to a different file. Each image will be used as if from a separate camera. Proceed with calibration steps using coordinate data and steps for multiple cameras.

Prepare for data collection by turning on the velocimetry light sources. Next, get the seeding particles and add them to the tank. Wait for steady-state flow to develop in the system.

With the room lights off monitor the camera view to verify the correct particle seeding density as in this slowed video of jet flow. When satisfied, record the desired number of flow images. After the data has been collected begin data processing by dividing the collected images into four sub-images and storing them.

With this done, the open PTV screen appears with menu choices across the top and a display for different camera views at the right. Move to the start tab in the menu items and click to reveal the options. Click image restart to load the four camera views of the first raw image in the data.

Move the mouse cursor to the run directory and click on it. From there click main parameters. On the new window go to the general tab.

Enter the number of cameras, in this case four. Continue entering parameters under the other tabs. When done, move the mouse to click on the preprocess menu item.

There click high pass filter. The result is to intensify the light scattering in the four images. Also under the preprocess menu item, find and click on particle detection to determine the centroid of each detected particle.

Then again under the preprocess menu item, click correspondences to establish stereoscopic correspondences. Next, move to the 3D positions menu item. Under it select 3D positions to determine the three dimensional position of detected particles based on the calibration.

Under the sequence menu item find and click sequence without display to repeat the preprocessing steps for all of the collected image sequences. Now, enter the parameters to be used to identify particles for tracking. Continue by selecting the tracking menu item.

From there choose tracking without display to correlate adjacent frames. And choose detected particles to visualize 3D reconstructed particles. When the tracking is complete go to the tracking menu item and select it.

Click show trajectories to visualize trajectories in each camera view. Each of the camera views now has the particle trajectories depicted. In addition, a data file has been created for post-processing.

This is a sample of three representative particle trajectories in the intermediate field region around the plane at 16 times the pipe jet diameter. When the distance from the jet core is small, longer trajectories are observed in the one second time interval of these data. As the distance from the jet core increases, the trajectory length decreases.

At three times the pipe jet diameter the tracer particles exhibit short and complex trajectories. Here are all of the successfully reconstructed particle trajectories that cross the same plane. The distribution of velocities ranges from zero to 60%of the jet's exit velocity.

Low velocities are in the blue range of the spectrum and high velocities are in the red range. Analysis reveals that particles near the jet core exhibit smoother trajectories. Particle tracking allows the calculation of the curvature along the particle trajectory.

This is a plot of the mean curvature times the pipe jet diameter versus the scaled radial distance for two planes. One at 16 and the other at 17 pipe jet diameters. Both increase monotonically but the plane further from the pipe outlet has a reduced mean curvature outside the jet core.

This plot provides information on the mean streamwise velocity distribution of the jets in the plane's 16 pipe jet diameters beyond the pipe outlet. Once mastered this technique can be done in six hours if it is performed properly excluding the processing time using the software. While attempting this procedure it is important to remember to have uniform illumination, correct camera settings and a sufficient amount of seeding particles.

After watching this video you should have a good understanding of how to set up the view splitter, assemble the light source, control the camera parameters and optimize the PTV system. After its development this technique paved the way for researchers in the field of turbulence to explore experimental studies in Lagrangian flow dynamics.