Benchmarking of the MIT High Temperature Gas-cooled Reactor TRISO-coated particle fuel performance model

Abstract

MIT has developed a Coated Particle Fuel Performance Model to study the behavior of TRISO nuclear fuels. The code, TIMCOAT, is designed to assess the mechanical and chemical condition of populations of coated particles and to determine the failure probability of each of the structural coating layers. With this, the code determines the overall particle failure rate. TIMCOAT represents a significant advancement over earlier codes as it includes a pyrocarbon crack induced particle failure mechanism, which applies probabilistic fracture mechanics. As part of the International Atomic Energy Agency (IAEA) Coordinated Research Program (CRP) on coated particle fuel technology, a code benchmark has been developed by the Idaho National Laboratory (INL). The benchmark includes simple analytic studies and more complex simulations of TRISO particles from past and future experiments. For each study a large variety of particle parameters are specified. This thesis reports TIMCOAT's results from the benchmark study. As this was a blind benchmark, no other results are available for comparison. However, TIMCOAT is able to independently calculate several of the parameters specified by the benchmark, and comparisons are made between results obtained using IAEA parameters and TIMCOAT calculated parameters. The material properties which describe the effects of irradiation on pyrolytic carbon are the focus of the comparison. The rates of irradiation induced creep and swelling calculated by TIMCOAT and the rates provided in the benchmark differed by a factor of two to four in some cases and led to differences in particle failure rate by several orders of magnitude. In addition, varying the uncertainties in these and other parameters was found to have a large impart on the failure probability.

(cont.) It is concluded that accurate modeling of TRISO particles depends on having very high accuracy data describing material properties and a very good understanding of the uncertainties in those measurements.