Jiandong Zhou1, Duane D. Bruns1, C. Stuart Daw2, Sreekanth Pannala3, and Charles E. A. Finney4. (1) Chemical Engineering, University of Tennessee, Knoxville, 419 Dougherty Hall, Knoxville, TN 37996, (2) Fuels, Engines and Emissions Research Center, Oak Ridge National Laboratory, 2360 Cherahala Blvd, NTRC, Knoxville, TN 37932, (3) Computer Science and Mathematics Division, Oak Ridge National Laboratories, Oak Ridge, TN 37831, (4) Engineering Science and Technology Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831
U. S. Department of Energy (DOE) has a large initiative to develop options for generation IV nuclear reactors. By 2010, the goal is to certify the process for making several types of nuclear reactor fuels that are inherently safe with no melt down issues, have a life cycle for all of the initial fuel and fission products by utilizing additional technologies and are hardened against proliferation. The fuel and process for the AGR (Advanced Gas Reactor, sometimes called the AHTGR, Advanced High Temperature Gas Reactor) is the nearest term option. The AGR will use 350-micron uranium fuel particles (UCO) that are coated by 4 layers: an amorphous (soft) carbon layer, then a pyrolytic (hard) carbon layer, then Silicon Carbide (SiC), and finally another pyrolytic carbon layer. This results in a particle with a diameter from 500 to 700-micron in diameter and referred to as TRISO particle. The SiC layer provides a containment vessel for the radioactive fission products including gases. The coating process uses a spouted bed CVD (Chemical Vapor Deposition) process at temperatures typically from 1200 to 1800 degree F. For the particles to meet quality specifications less than 1 in 100,000 can be defective. In order to produce large quantities of fuel to support AGRs, the current 5 cm diameter spouted bed coaters will have to be scaled-up. A power reactor will require around 5 to 15 x 10
9 particles. The technology and art used with the 5 cm coaters allows around 1 x 10
5 to be coated over six hours. There is no universal way to scale-up fluidized beds even for hydrodynamics [1]. The state-of-the-art in fluidized bed design and scale-up is to establish the hydrodynamics from cold bed studies without reaction. In the high temperature coaters, it is nearly impossible to optically observe and monitor the coater due to opaque carbon wall of the bed and dense carbon soot inside the bed. In general, it is also very expensive to set up and run a high-temperature surrogate coater. Given these constraints, in this paper we correlate the results obtained from the ambient room-temperature coater to those of the high-temperature ones using dimensionless groups. The cold bed studies are used to validate simulations, study this process at other conditions, then incorporate reaction into the simulation and optimize the operation parameters. Experimental work is underway at the University of Tennessee in Knoxville in 5.08, 8.89, 15.24 and 22.86cm cold beds for five particles sizes, at many gas flow rates, with temperatures from room temperature to as high a temperature (~ 500 deg C), 4 cone angles, different throat diameters, different particle inventories (weight of particles in bed), and designed distributors (more sophisticated than just a conical bottom). Spherical Zirconium Oxide (ZrO2) particles are used as surrogates for the uranium fuel particles since they have similar physical properties in terms of their high density compared to particles normally presented in the literature for spouted beds. The work reported in this paper includes two parts, experimental and modeling. The experimental part focuses on fountain height correlations for ZrO2 particles. During the coating process, the particle size grows from 350 to 700 micron with the feed gas composition being changed for each coating layer. The particle size and density change dramatically during the coating process. For the spouted bed to run smoothly inputs to the process should be adjusted continuously, generally the gas velocity is adjusted. On the one hand, the gas velocity has to be above the minimum spouted velocity, Ums, which increases as particle size grows. Note that the increasing particle diameter dominates the generally decreasing particle density. On the other hand, the gas velocity cannot be too high as the spout dynamics and particle collisions can damage the fragile carbon coatings. The experimental fountain height correlations for ZrO2 are obtained with a 5 cm bed with a 0.4 cm throat diameter. The experimental matrix includes four cone angles (30, 45, 60 and 75 degrees inclusive), particle radius, gas flow rate, and particle inventory. The modeling part concentrates on lower order modeling of the spouted bed. The current models, either DEM (discrete element method) models or continuum phase models, costs are high in computation power and time. Presently, it is impossible to use online model based prediction and control with these models. It is practically meaningful to develop a lower order model to simulate the behavior of the solid particles whose hydrodynamics keep evolving during the coating process. This allows the bed behavior to be predicted and online control implemented in real time. Additionally the low order model demands that the critical characteristics and nonlinearities of the process be identified and understood. The lower order model combines five interacting features: • granular flow in the annular region • solid particles entrained from the bottom of the annular region by high velocity gas flow into the bottom of the spout • solid particles carried by gas flow in the spout region • particle rising region in the fountain • free falling region in the fountain. This model is tested by the experimental hydrodynamic correlations of minimum spouting velocity, of gas pressure drop across the spouted bed, and fountain height.