Experimental and Numerical Evaluation of Different Fluidization Parameters for Successful Target Lay

1
Experimental and Numerical Evaluation of
Different Fluidization Parameters for Successful
Target Layering
Kurt J. Boehm1,2 N. B. Alexander2, A. Bozek2, D.
Frey2, D. Goodin2, A. R. Raffray1 (1UCSD,
2General Atomics)
Goal Produce a uniform DT Layer on the Inside of
the Capsules while Maintaining a Smooth Outer
Surface
Sputter Coating -- Preparation of Studies to
Analyze Surface Roughening During Fluidization
Fluidized Bed Analysis Optimizing Fluidization
Parameters
Numerical Simulation using MFIX

• 
  The Optimal Frit Design
• Provides
• Spin rate and circulation rate higher than a
  certain threshold value for uniform layering
  (about 5Hz).
• (The spin rate is the more important performance
  criterion)
• A certain randomness of the orientation of the
  spin (hard to quantify)

MFIX (Multiphase Flow with Interphase eXchanges)
is a general-purpose computer code developed at
the National Energy Technology Laboratory (NETL)
for describing the hydrodynamics, heat transfer
and chemical reactions in fluids-solids systems.
MFIX has two significantly different models of
particle movement Model A - Particle movement
averaged ? computes the average
velocity for all particles within one fluid
cell Model B - Discrete Element Simulation (DES)
? each particle motion tracked
individually
Avoides Hard particle-
particle collisions to preserve smooth surface

• Development of a new method to produce a large
  number of smooth and uniformly coated shells
• Current techniques require further work since
• Number of targets produced needs to be increased
  to 100-1000
• Surface roughening during target coating needs
  to be eliminated

Previous set up bouncing pan
Proposed design spinning dish
Fluidized bed test stand at General Atomics
laboratory
Tracking random individual paths using DES
High Z target
Replicable Fluidized Bed Section
Targets roll up on one edge and down though the
middle
Closed loop plumbing circuit
Capsules
This run simulates a fluidized bed at 2 BE. The
simulated time was 1s, particle diameter 4mm, the
diameter of the fluidized bed was 15 cm.
Polonium Strip
Surface damage during bouncing pan
experiments Presented in High Z coatings for IFE
applications by Abbas Nikroo et alt., HAPL
Review Meeting, November 13-14, 2001, Pleasanton,
CA
Frit
- Reduction of surface damage expected, since
only soft collision will occur - Dish can be
loaded with larger number of shells

• We want to modify MFIX to calculate
• - time averaged circulation and spin rates of the
  pellets
• relation between spin and circulation rates at
  different locations in the bed
• ? in experiments only pellets next to the
  wall can be seen
• - Explore different bed sizes and fluidization
  parameters to predict optimized layering
• Model the layering process using the heat
  transfer capabilities
• Explore the effects of non-uniform mass density
  in target (unlayered condition)
• The following modifications will be necessary
• A suitable fluid - particle model for
  needs to be
  found.
• Tracking of the rotational motion and angular
  position of each target
• A collision model for unlayered targets
  (asymmetric mass distribution in capsule)

The performance of different frits during
fluidization was analyzed to determine the best
design for the given criteria.
Frit Performances
Analysis has been done by post experimental
processing of high speed videos (250-500 frames
per second)
Ablation of the gold layer was concluded to be a
result of collisions

• Observations
• - Flow in the lab shows a preference to circulate
  across the whole width of the bed (at BE gt 2.0)
• Pellets may behave differently near tube wall
  than in center of the bed, although
• Analysis of top and side view indicate
• similar behavior of the particle motion across
  the bed
• some degree of randomness in the spin
  orientation
• All bed designs show a spin rate above the
  threshold of 5Hz.
• The circulation rate shows a maximum for all
  frit designs between 2 and 4 BE.
• In the swirling bed, the particles roll along
  the inside of the tube causing a steep and steady
  increase in spin rate
• - The cone and the funnel are expected to provide
  best randomness of rotation

The sputter coater set up with a spinning dish
different dish sizes, angles of incline and the
rotational speeds can be chosen.
Air flow
Comparison between Numerical modeling and the
room temperature loop experiments
Schematic example of the observed flow using the
cone frit

• Problems and Challenges
• Soft pellet- pellet collisions lead to static
  charges
• Pellets stick to the pan and ultimately lead to a
  non-uniform layer
• Possible surface damage due to rolling over the
  dish surface and/or soft particle collisions
• Thickness of the sputtered layer can only be
  estimated.

• Using cell averaged model (A) in symmetric
  cylindrical coordinates
• Compared to room temperature experiment under
  same conditions
• At low flow speed
• Strong agreement for the bubbling flow pattern
• Bubbling frequency can be predicted quite well
• At higher flow speeds
• 3-D effect (circulation pattern) dominant in
  experiment
• ? 3D package of MFIX needed

Preferred motion of the pellet
A white foam ball makes it easier to track a
target
Screenshot of a side and a top view video with a
regular straight frit
Nominal Value of 2 BE
Nominal Value of 2 BE
T 0.00 s
T 0.12 s
T 0.04 s
T 0.08 s
All frits above critical value of 5 Hz
T 0.04 s
T 0.00 s
T 0.12 s
T 0.08 s
The circulation rate vs. bed expansion for the
different fluidized bed configurations
The spin rate vs. bed expansion for the different
fluidized bed configurations
Wyko surface data taken from a shell coated in
the rotating dish set up. The surface roughness
before the coating process was unknown.
Video of the room temperature fluidized bed
Bubbles can wander off center
Different particle void fractions in the fluid
cells over time