Accurate Nodal Diffusion Modelling on the High Temperature Test Reactor (HTTR)

Abstract

The High Temperature Test Reactor is a 30-MWth helium-cooled, graphite-moderated, prismatic-type gas reactor developed by the Japan Atomic Energy Agency (JAEA). The hexagonal shape of fuel blocks in the HTTR core combined with complex inner structures containing TRISO particles results in the double-heterogeneity effect that increases the simulation challenge of the reactor. This research has a goal to accurately model the HTTR fuel blocks employing the standard two-step procedure of a reactor analysis: employ a lattice physics calculation to generate homogeneous cross sections and use them in a nodal diffusion calculation. The implementation of the diffusion approximation results in a faster calculation with acceptable accuracy compared to the high-resolution lattice calculation. An advanced method called Triangular Polynomial Expansion Nodal (TriPEN) method was used in this work for the nodal diffusion calculation to accurately model the flux discontinuity effect between blocks by generating discontinuity factors in each surface and corner point. To do this, the heterogeneous solutions obtained from the lattice physics calculation, which is done by Serpent Monte Carlo in this case, are utilized by TriPEN to generate the discontinuity factors. Due to its capability to simulate any reactor geometry with a high-resolution, the results generated from Serpent calculation were also used as the reference cases for this work. In this work, the TriPEN method is implemented in the PARCS core neutronic module inside AGREE, a U.S. N.R.C. Multiphysics code system for the High Temperature Reactor (HTR). Test cases conducted for this method involved the original design of the HTTR. The 4-hex model built consisted of the central control-rod block of HTTR together with 6 of the half-fuel-blocks surround it. The differences in material composition in each assembly block in HTTR resulted in flux discontinuity effects on the assembly block interfaces. To correct this discrepancy, discontinuity factors were applied in order to make the homogenous solutions from the nodal calculation agree with the heterogeneous solutions from the lattice physics calculation. Applying this procedure to the HTTR nodal models, TriPEN is able to produce exact results in the eigenvalue compared to Serpent calculation, the rod worth calculated perfectly matched the reference, and the flux and power distribution only has negligible discrepancies.