The first measurements were performed in the GALCIT Free Surface Water Tunnel (FSWT). A 20 W argon ion (Ar+) laser was used as the light source, producing a low-dispersion Gaussian light beam. The laser beam is fed into a 2-axis scanner (see figure below) used to scan a 3-D volume in the flow. The laser beam is scanned in a serpentine manner, first sweeping across the entire volume in the +x direction, then stepping 1/32 of the volume in the +z direction (depth), then sweeping across the entire volume in the x direction, stepping 1/32 along the +z direction and repeating these 4 steps 16 more times. At the end of 32 sweeps, the laser beam sweeps the full distance in the z direction, to repeat the process. Each sweep in the x direction across the field of view lasts 5 ms. The steps in the z direction also last 5 ms. When stepping in the z direction, the laser beam is blocked by the two beam stops. During this time, the camera reads out the image acquired during the previous sweep across the x-axis. The KFS camera system is synchronized to the sync outputs of the scanner controller.

A second Windows XP computer is used to measure the output of a Laser Doppler Velocimeter (LDV) that monitors the free-stream velocity of the FSWT.

A stainless steel grid with a wire size of 0.25 and a spacing of 1.00 between wires (1.25 between wire centerlines) is used to produce the 3-D turbulent flow. The camera center field of view is 25.7 downstream of the grid (20.6 grid spacings) to allow the turbulent flow and scalar dispersion to be developed. A dye injector injects Rhodamine 6G dye at the center of the grid in order to visualize the downstream turbulence. The KFS camera views the swept volume from the side of the FSWT (same perspective as the person viewing the figure above). An optical band-pass filter in front of the camera blocks the green laser light scattered from particles in the water while transmitting the yellow light (fluorescence) emitted by the Rhodamine 6G.

The first data sets derive from the relatively low Reynolds number turbulent flow that could be probed in the FSWT environment. A volume rendering of the first data set is reproduced below. Color has been assigned to denote local concentration.

The data reveal a surprising geometry and topology, in as much as previous, low-dimensionality transects had suggested a sheet-like topology, rather than the worm-like topology reminiscent of the vortical structures that have been revealed in large-scale computational fluid dynamics investigations. It is not clear if this is a consequence of the low Reynolds number characterizing these preliminary experiments.

A second-generation experiment is presently under design that will permit an increase in the flow Reynolds number by over an order of magnitude. Such an increase will permit straddling the mixing transition in turbulence (Dimotakis 2000, 2005).
ADDITIONAL RESOURCES

[ Offline Rendering of Large Movies ]
Parallel Visualization -> Volume Rendering Cluster

[ Data Wulf ]
Parallel Data Storage -> Datawulf

[ Enabling the T221 Display for Science ]
Parallel Visualization -> Parallel Display





[ Calculating the Needed Throughput ]
     Data Generated
        1000x1000 pixels per frame
        12 bits per pixel
        1000 frames per second
        1.5MB/frame
        1500 MB/s
        /8 ports = 200 MB/s/port
        slinks :: 120-160 MB/s
        raid :: 70 MB/s
    -------------------------------------
        ability: run for 1 hr
       (given lossless compressed data)
   


[ see PUBLICATIONS ]