Table of Contents
|
III. THE EXPERIMENT
Two coaxial PlexiglasTM cylinders having radii
R0=15 cm and R1=7.5 cm are separately rotated about a common
axis. The annular test space is the region bounded below by the jet
port plate and above by a transparent plastic window. The window is
called a witness plate because it allows one to view the rising jet
through an optically flat surface. Figure 1 shows the experiment with
its rotational axis vertical, and Fig. 2 is a schematic of the
apparatus. For the small port discussed below, the annular height was
12.5 cm and for the large port it was 10 cm. The outer cylinder was
rotated at 1/2 Hz, resulting in a peripheral velocity of ~ 50 cm/s. The pulsed jets were injected with a velocity of ~ 10 cm/s. Consequently, the Reynolds number for the Ω-flow is Re
≅ 105 and Re ≅ 104 for the jets. We designed the
experiment to have the largest Re feasible in order to simulate
turbulent effects, but also designed for the ease of video recording
at 30 Hz. These modest velocities resulted in negligible centrifugal
and pulsed jet drive pressures. This is in contrast to the liquid
sodium experiment where these pressures approach the structural limit
of materials.
Figure 1:
Figure 1 shows the Pulsed Jet Rotation Experiment with its
camera mounts, electrical communications, air supply, and the cylinder
drive systems.
Figure 2:
Figure 2 shows a schematic of the Pulsed Jet Rotation
Experiment. An air pressure pulse forces a pulse of water out of the
inner cylinder, through the plenum ports and then through the jet
port. This forms a rising vortex ring that simulates an expanding
plume. The inner and outer cylinder can be rotated together and
separately. The pulsed jet can be observed from the side and
co-rotating with the outer cylinder from the top.
Two different diameters of jet ports were used. The small port
diameter, dSP=3.3 cm, was chosen such that the entraining jet
would not interact with the cylinder walls prior to impinging upon the
witness plate. This allowed the pulsed jet to fully develop by
entrainment and radial divergence independent of wall interference.
The large port diameter, dLP=4.8 cm, was chosen such that when
the jet impinged on the witness plate, its diameter equaled the width
of the annulus. This was the maximum size possible that would allow
an entraining jet to rise symmetrically through the annular space,
impinge on the witness plate, and rotate. Since the experiment was
designed to simulate the flow fields that we believe are requisite for
the sodium dynamo, the pulsed jet produced by dLP would then
maximize the magnetic flux displaced and rotated by the jet. The jet
is driven by a pulse of air exerting pressure on the water within the
inner cylinder. The water is then forced out through radial ports at
the base of the inner cylinder into the plenum and then through an
orifice as a pulsed jet. At the termination of the pressure pulse,
gravity forces the water, now at a different height in the annular and
inner cylinder volumes, to return slowly through the same ports,
reaching equilibrium in a few seconds. For calibration purposes a
pressure transducer measured the pressure of the air applied on the
fluid within the inner cylinder.
The outer cylinder assembly consists of the port plate and the
hydrogen electrolysis anode and cathode for flow visualization by
pulsed hydrogen electrolysis from a tungsten wire. The tungsten wire
is stretched across the jet port on the top of the port plate for the
large port and at 2.5 cm above the port for the small port. The
distance that the jet travels from the wire to the witness plate was
therefore 10 cm in either case. The wire is centered across the jet
port and is tangent to a circle about the cylinder's axis. The outer
cylinder also provides mechanical support for the camera mounts.
A DC voltage supply, 200-400 volts, was used to supply pulsed voltage,
of 0.1 and 0.5 second duration, for the hydrogen electrolysis. The
short pulse produced a thin sheet, ~ 1 cm in height, of bubbles.
The long pulse allows a wider range, ~ 5 cm, of pulsed jet length
to be monitored in order to demonstrate the integral behavior of it.
The relative delay between the pulsed jet initiation and the
electrolysis pulse allowed for the observation of different jet
regions.
A TV and VHS/VCR system for real-time viewing and secondary image
recording was used in conjunction with a DV camcorder. The camcorder
records a digital image internally, and in addition, the analog output
of the DV camcorder was transmitted via slip rings and a coax cable to
the real-time viewing monitor. This system permitted convenient,
single-frame replay capability of the studio-quality VHS/VCR and
allowed for quantitative analysis of individual pulsed jets.
Two flow visualization methods were employed in the
experiment: pulsed hydrogen electrolysis from a 50 μm tungsten
wire in an aqueous Na2SO4 solution (Merzkirch13) and
dispersed guanidine14 in water. Hydrogen electrolysis creates a
thin sheet of ~ 50 μm diameter bubbles. Hydrogen
electrolysis was used primarily to reveal the average behavior of the
near laminar flow at the head of the jet. The pattern of these small
bubbles that outlines the jet is rapidly lost in the turbulence within
the jet. Guanidine dispersed in water was used to emphasize the
turbulently entraining fluid. Guanidine, a pearlescent material of
micron-size platelets, produces an image of the entire pulsed jet.
The modulation of this image emphasizes the the largest shear in the
flow. Since this largest shear occurs primarily at the boundary of the
eddies, guanidine outlines the turbulence within the jet and thus it
outlines the entrainment boundary of the jet. In addition, it shows
the turbulence as the jet strikes the witness plate. The electrolysis
bubbles are even more random in pattern in this region. The guanidine
images show the average differential rotation of the turbulent fluid
in the pulsed jet.
Both methods reveal a coherent counter-rotation of the flow, either
within the laminar flow ahead of the rising jet, or in the turbulent
divergent flow when the jet strikes the witness plate. The linear
array of bubbles produced by electrolysis remains clearly a line
during rotation of the jet when it is embedded in the near-laminar
flow at the head of the jet. We will show in forthcoming images the
clear sequential rotation of the line of bubbles. In contrast, when
the jet strikes the witness plate only the guanidine shows the average
behavior of the flow. The average rotation of the turbulence was
readily discerned from analysis of the original 30 Hz image series.
These two different measurements are important to the sodium dynamo
experiment. First, the toroidal magnetic flux will be lifted ahead of
the pulsed jet and will be rotated by a large, but finite
angle. Second, the flux embedded in the turbulent jet will also be
lifted and rotated coherently upon striking the end wall.
|