Tuesday, December 23, 2008

Experimental Fluids

Last week I did not post because of today's post. Now I know this seems kind of silly, however this post will explain. Last week I was working on my final project for my Graduate Experimental Fluids class. It consisted of placing a model of the AIAA clubs airplane in a water tunnel. This year the Rutgers AIAA club is building an autonomous UAV that has a camera in the nose, because of the camera a ball turret is going to make up the front of the nose. We wanted to see how much this effected the flow, and to look at the vortex shedding from the ball.

This is the 3/4 scale model of the front section of the aircraft, the design has changed a bit since the model was initial built. However, the model took around 100 hours to build in a FDM (fused deposition modeling) rapid prototyping machine, so we did not have time to make another (also the model was screwed up 3 times due to power failure, one of which I spoke of here and here).
This is the entire set up as we ran the experiment. The experimental method we used is called PIV (Particle Image Velocimetry). It consists of a laser (green square, class 4 laser running at around 125-145 mJ and 532 nanometers wavelength) that is run through a series of optics (light blue square, spherical and cylindrical lenses) that makes the into a thin sheet. Then a mirror (yellow square) reflects the laser sheet around the model and a camera (red square, CCD camera) takes two pictures very close to each other timed to coincide with the lasers pulses with a time difference of in our case 2500-6000 microseconds so around 2.5-6 Milli seconds. The two pictures show illuminated seed particles (in this case I believe we where using glass beads, with a diameter of around 4 microns). Then computer software compares the two images and based upon where the particles moved to it creates a plot of the vector field, and vorticity.

This is the model as it looks before we had to invert it to get it in the right place with regards to the laser sheet.
This is what the whole set up looks like from the top, the only thing I haven't mentioned is the sychronizer, which synchronized the laser and the camera.This is the kind of vector field you get, in this case it is the mean velocity field from 50 individual vector fields. Note the stagnation point at 1.5 inches on the bottom axis (a stagnation point is a point where the flow speed is zero).
In the mean velocity plot for the 15 Hz pump speed the flow is moving a lot slower and as a result the stagnation point is now much farther to the right at pretty much the edge of the image. Also, there are two distinct vortexes (see if you can find them, one is a lot easier than the other. Hint: the shear layer/boundary layer propagates from where the flow separates from the sphere of shot to the left to the fuselage on the right) this indicates that for the slow flow there is a steady state solution.
This is what one of the instantaneous velocity fields looks like, there where 50 of these to make the average field in the image above. You can see a vortex in the approximately the same place as in the mean velocity plot and its pair.
Sorry there is no scale on this graph, also sorry about it not being bounded properly. At (0.75, 1.4) and at (1.25, 1) there is a clear vortex pair. A vortex pair consist of two counter rotating vortexes (see this animation).

Now you can see what I was talking about not posting last week, so this is just a post worthy of two weeks. To give you an idea, the experiment took two people (me and my friend Andrew) around 30 hours a piece to set up (after the model was made) and around an hour to take the data, such is the way experiments work. Hopefully we will be able to do more tests, looking at different sections of the flow, and trying to stop the vortex shedding, as it produces a ton of drag.


  1. There's a cool optical illusion in the first picture: Until I saw the others, I thought the spherical part was a bowl, not a ball.

  2. Do you have an actual (set of) picture(s) from the camera we can see?