Statistical approach and benchmarking for modeling of multi-dimensional behavior in TRISO-coated fuel particles

The fundamental design for a gas-cooled reactor relies on the behavior of the coated particle fuel. The coating layers, termed the TRISO coating, act as a mini-pressure vessel that retains fission products. Results of US irradiation experiments show that many more fuel particles have failed than can be attributed to one-dimensional pressure vessel failures alone. Post-irradiation examinations indicate that multi-dimensional effects, such as the presence of irradiation-induced shrinkage cracks in the inner pyrolytic carbon layer, contribute to these failures. To address these effects, the methods of prior one-dimensional models are expanded to capture the stress intensification associated with multi-dimensional behavior. An approximation of the stress levels enables the treatment of statistical variations in numerous design parameters and Monte Carlo sampling over a large number of particles. The approach is shown to make reasonable predictions when used to calculate failure probabilities for irradiation experiments of the New Production – Modular High Temperature Gas Cooled Reactor Program.