A cellular automaton method to simulate the microstructure and evolution of low-enriched uranium (LEU) U–Mo/Al dispersion type fuel plates

• Published in: Journal of nuclear materials 2014-10, Vol.453 (1-3), p.313-319
• Main Authors: Drera, Saleem S., Hofman, Gerard L., Kee, Robert J., King, Jeffrey C.
• Format: Article
• Language: English
• Subjects: Aluminum; Automation; Cellular; Cellular automata; Computer simulation; Fuels; NUCLEAR FUEL CYCLE AND FUEL MATERIALS; Nuclear power generation; Plates
• ISSN: 0022-3115; 1873-4820
• DOI: 10.1016/j.jnucmat.2014.07.016

•This article presents a cellular automata (CA) algorithm to synthesize the growth of intermetallic interaction layers in U–Mo/Al dispersion fuel.•The method utilizes a 3D representation of the fuel, which is discretized into separate voxels that can change identy based on derived CA rules.•The CA model is compared to ILT measurements for RERTR experimental data.•The primary objective of the model is to synthesize three-dimensional microstructures that can be used in subsequent thermal and mechanical modeling.•The CA model can be used for predictive analysis. For example, it can be used to study the dependence of temperature on interaction layer growth. Low-enriched uranium (LEU) fuel plates for high power materials test reactors (MTR) are composed of nominally spherical uranium–molybdenum (U–Mo) particles within an aluminum matrix. Fresh U–Mo particles typically range between 10 and 100μm in diameter, with particle volume fractions up to 50%. As the fuel ages, reaction–diffusion processes cause the formation and growth of interaction layers that surround the fuel particles. The growth rate depends upon the temperature and radiation environment. The cellular automaton algorithm described in this paper can synthesize realistic random fuel-particle structures and simulate the growth of the intermetallic interaction layers. Examples in the present paper pack approximately 1000 particles into three-dimensional rectangular fuel structures that are approximately 1mm on each side. The computational approach is designed to yield synthetic microstructures consistent with images from actual fuel plates and is validated by comparison with empirical data on actual fuel plates.