The Parallel and Highly Innovative Simulation for INL Code System (PHISICS) has been under development at Idaho National Laboratory since 2010. It consists of several modules providing improved coupled core simulation capability: INSTANT (Intelligent Nodal and Semi-structured Treatment for Advanced Neutron Transport) (three-dimensional nodal transport core calculations); MRTAU (Multi- Reactor Transmutation Analysis Utility) (depletion and decay heat generation); and modules performing criticality searches, fuel shuffling, and generalized perturbation. Coupling of the PHISICS code suite to the thermal-hydraulic system code RELAP5-3D was finalized in 2013, and as part of the verification and validation effort, the first phase of the Organisation for Economic Co-operation and Development/Nuclear Energy Agency (OECD/NEA) MHTGR-350 benchmark has now been completed.

The theoretical basis and latest development status of the coupled PHISICS/RELAP5-3D tool are described in more detail in a concurrent paper. This paper provides an overview of the OECD/NEA MHTGR-350 benchmark and presents the results of exercises 2 and 3 defined for phase I. Exercise 2 required the modeling of a stand-alone thermal fluids solution at the end of equilibrium cycle for the Modular High Temperature Gas-Cooled Reactor (MHTGR). The RELAP5-3D results of four subcases are discussed, consisting of various combinations of coolant bypass flows and material thermophysical properties. Exercise 3 required a coupled neutronics and thermal fluids solution, and the PHISICS/RELAP5-3D code suite was used to calculate the results of two subcases.

The main focus of this paper is a comparison of results obtained with the traditional RELAP5-3D “ring” model approach against a much more detailed model that includes kinetics feedback on individual “block” level and thermal feedbacks on a triangular submesh. The higher fidelity that can be obtained by this block model is illustrated with comparison results on the temperature, power density, and flux distributions. It is shown that the ring model leads to significantly lower fuel temperatures (up to 10%) when compared with the higher-fidelity block model and that the additional model development and run-time efforts are worth the gains obtained in the improved spatial temperature and flux distributions.