Development of effective thermal conductivity model for particle-type nuclear fuels randomly distributed in a matrix

Several advance nuclear reactors use particle-type Tristructural-Isotropic (TRISO) fuels randomly distributed in a matrix to allow aggressive operating conditions. Since those fuels are composites with randomly distributed TRISO particles in a matrix, suitable smearing methods are needed to obtain effective pellet-level thermomechanical properties for reactor design, and safety analysis. Currently available smearing methods for effective thermal conductivity assume uniform or no heat generation, thereby neglecting the random heat source distribution. By developing three-dimensional finite-element heat conduction models for randomly distributed heat generating kernels in a matrix, this study demonstrates that the consideration of a randomly distributed heat source is important in predicting the peak fuel temperature. In this study, (1) random packing of heat generating fuel particles introduces the statistical distribution of peak and average temperatures, and (2) those statistical temperature distributions are quantified. In light of this, thermal conductivity models with randomly distributed heat generating kernels are developed to predict peak pellet temperature for Fully Ceramic Microencapsulated (FCM) fuel and Cermet fuel for space propulsion. The developed methodology and models provide a practical methodology to predict statistically-informed peak and average fuel temperatures of nuclear fuel pellets of heat generating particles. The presented methodology is useful to quantify uncertainties in predicting nuclear fuel temperatures with TRISO particles. (C) 2018 Elsevier B.V. All rights reserved.