There is growing interest worldwide in high temperature gas-cooled reactors (HTGRs) as candidates for next generation reactor systems. Either in a pebble type or in a prismatic type HTGR, coated particle fuel (TRISO fuel) appears to be the most promising fuel candidate to be used. For design and analysis of such a reactor, transport models, in particular, stochastic models that permit the simulation of neutron transport through the stochastic mixture of fuel and moderator materials, are becoming essential and gaining importance.
Naturally, the Monte Carlo methods have been used for this situation. However, the methods reported in the literature all have their own deficiencies. In this thesis, we propose a new Monte Carlo method named fine lattice stochastic (FLS) modeling that is distinct from others. This method is based on fine lattice system in which a lattice circumscribes a fuel particle. Once the problem is given, an interface Fortran code gives out the TRISO particle fuel configurations (a set of lattice center points only) for MCNP input. The number of available lattice center points is far larger than the number of fuel particles according to packing fraction of the fuel element. We apply discrete random sampling here to choose a certain number of lattices to fill with fuel particles. In this aspect, FLS modeling allows more realistic fuel particle distributions. In this thesis, only simple cube (SC) structure is used in cubic lattice. However, FLS model can be easily extended to BCC, FCC structures or hexagonal prism type lattice.
The criticality calculations for our FLS modeling were first tested on a small cube problem and compared with other models. The results indicate that the new stochastic model is an accurate and efficient approach to analyze TRISO particle fuel configurations. Then the FLS modeling was performed to analyze HTGR fuel elements for both pebble type and prismatic type and the results were also good as expected.