Continuous-energy COMET solution to the stylized AHTR benchmark problem

In this paper, the continuous energy coarse mesh transport code COMET was used to provide the whole core transport solution to a set of stylized Advanced High Temperature Reactor (AHTR) benchmark problems. This set includes two full length single assembly and three full core configurations. Solving this type of reactor core is a significant challenge to high and low order neutron transport methods because of its high heterogeneities due to TRISO fuel and burnable absorber particles imbedded in graphite plates with Y-shaped control blades and molten salt as coolant. The eigenvalue and fission density distributions calculated by COMET were compared with the direct Monte Carlo solutions. In the single assembly calculations, it was found that eight-bin expansion in energy and 2nd order expansion in the other phase space variables used by COMET are sufficient for calculation of the eigenvalue and the assembly averaged fission density distribution, while 4th order expansion in space is necessary to obtain the stripe-wise fission density distribution with accuracy comparable to that of the continuous energy Monte Carlo method. It was also found that it is impractical to use Monte Carlo to generate the detailed solutions (stripe-wise fission density distribution) to the full core benchmark problems because of the prohibitive computation resource requirement (estimated to be about 2200 days on an 84-CPU cluster). It took only about 4.5 h on a single CPU for COMET to calculate the detailed solution including the stripe-wise fission density distribution in the entire core. Given that COMET agreed very well with Monte Carlo in terms of assembly averaged and stripe-wise fission density distributions in the single assembly problems as well as in terms of the assembly averaged fission density distribution in the full core problems, it was concluded that the continuous energy COMET can provide the reference solution to the stylized AHTR benchmark problems.