Book excerpt: This report summarizes the results of preliminary work at Oak Ridge National Laboratory (ORNL) to demonstrate application of computational fluid dynamics modeling to the scale-up of a Fluidized Bed Chemical Vapor Deposition (FBCVD) process for nuclear fuels coating. Specifically, this work, referred to as Modeling Scale-Up Phase I, was conducted between January 1, 2006 and March 31, 2006 in support of the Advanced Gas Reactor (AGR) Program. The objective was to develop, demonstrate and "freeze" a version of ORNL's computational model of the TRI ISOtropic (TRISO) fuel-particle coating process that can be specifically used to assist coater scale-up activities as part of the production of AGR-2 fuel. The results in this report are intended to serve as input for making decisions about initiating additional FBCVD modeling work (referred to as Modeling Scale-Up Phase II) in support of AGR-2. The main computational tool used to implement the model is the general-purpose multiphase fluid-dynamics computer code known as MFIX (Multiphase Flow with Interphase eXchanges), which is documented in detail on the DOE-sponsored website http://www.mfix.org. Additional computational tools are also being developed by ORNL for post-processing MFIX output to efficiently summarize the important information generated by the coater simulations. The summarized information includes quantitative spatial and temporal measures (referred to as discriminating characteristics, or DCs) by which different coater designs and operating conditions can be compared and correlated with trends in product quality. The ORNL FBCVD modeling work is being conducted in conjunction with experimental coater studies at ORNL with natural uranium CO (NUCO) and surrogate fuel kernels. Data are also being obtained from ambient-temperature, spouted-bed characterization experiments at the University of Tennessee and theoretical studies of carbon and silicon carbide chemical vapor deposition kinetics at Iowa State University. Prior to the current scale-up activity, considerable effort has gone in to adapting the MFIX code to incorporate the unique features of fuel coating reactors and also in validating the resulting simulation features with experimental observations. Much of this work is documented in previous AGR reports and publications (Pannala et al., 2004, Pannala et al., 2005, Boyalakuntla et al., 2005a, Boyalakuntla et al., 2005b and Finney et al., 2005). As a result of the previous work described above, the ORNL coater model now has the capability for simulating full spatio-temporal details of the gas-particle hydrodynamics and gas-particle heat and mass transfer in the TRISO coater. This capability provides a great deal of information about many of the processes believed to control quality, but the model is not yet sufficiently developed to fully predict coating quality for any given coater design and/or set of operating conditions because the detailed chemical reaction kinetics needed to make the model fully predictive are not yet available. Nevertheless, the model at its current stage of development already provides the most comprehensive and detailed quantitative information available about gas flows, solid flows, temperatures, and species inside the coater during operation. This level of information ought to be highly useful in expediting the scale-up process (e.g., in correlating observations and minimizing the number of pilot-scale tests required). However, previous work had not yet demonstrated that the typical design and/or operating changes known to affect product quality at the lab scale could be clearly discriminated by the existing model. The Modeling Scale-Up Phase I work was initiated to produce such a demonstration, and two detailed examples are discussed in this report.