Argonne’s Computational Science Division performs research and development on particle transport methods as applied to problems in nuclear energy and high energy physics. To that end, researchers in CPS contribute to a number of open source software packages for modeling and simulation. One of our most visible products in the this area is OpenMC, a community developed framework for Monte Carlo particle transport simulations. OpenMC is capable of simulating neutrons and photons using either a constructive solid geometry or CAD representation. The code is implemented with hybrid MPI/OpenMP parallelism and has demonstrated near-linear scaling up to hundreds of thousands of cores on leadership-class supercomputers. OpenMC features a rich Python API that enables programmatic pre- and post-processing of jobs; parsing, analysis, and generation of nuclear data files; and automated generation of multigroup cross sections for use in downstream deterministic transport codes.
A number of ongoing projects are developing capabilities for the coupled solution of particle transport and other systems of equations, such as heat transfer and fluid dynamics. Through the Exascale Computing Project (ECP), researchers in CPS are working on developing a software package called ENRICO that couples OpenMC to the Nek5000 computational fluid dynamics code. This code is being using to study the behavior of small modular nuclear reactors and will be deployed on Aurora, Argonne’s future supercomputer. Several projects are also underway to couple OpenMC to the MOOSE parallel finite element framework, which will enable the coupled solution of particle transport, fuel performance calculations, and other physics fields.