Implementation of Domain Decomposition Parallel and gCMFD in MOCP: Method of Characteristics for Pebble-Bed

Abstract MOCP code is developed with Three-dimensional Method Of Characteristic (3D-MOC) for pebble-bed High Temperature Gas-cooled Reactor (HTGR) high-fidelity neutronics calculation. To further enhance its efficiency, domain decomposition parallel and gCMFD method are implemented in MOCP as alternatives to its previous calculation scheme, namely track grouping parallel and Macro-Track Transport Acceleration (MTTA) method. This paper provides parallel implementation details in track tracing, MOC sweeping and gCMFD solving. Parallel load balance is ensured by balancing subdomain volumes in each processor. Notably, Constructive Solid Geometry (CSG) based fine-coarse meshes mapping procedures are highlighted in this work. A simplified cylindrical model and HTR-10 whole-core model are used to demonstrate the performance of newly developed MOCP. It shows that MOCP parallel efficiency is around 80% using 160 MPI processors in the simplified cylindrical model. In HTR-10 whole-core model, the implemented domain decomposition parallel and gCMFD in newly developed MOCP could reduce around 70% core-hours compared with that using previous calculation scheme.