Creating a Platform for Autonomous Mesh Refinement and Solution Verification

Illinois Rocstar is excited to announce a contract to continue development of the NEMoSys© toolkit for autonomous mesh refinement and solution verification. Modeling and simulation (M&S) are critical for cost-effective design and advanced safety modeling of nuclear reactors. Appropriate verification and validation (V&V) of the codes, models, and solutions is critical to successful acceptance of modern high-fidelity methods by regulators and the industry. This contract will automate the solution verification step of V&V through the addition of graphical and programmatic interfaces, new and existing meshing tools, and streamlined automated workflows to the existing Illinois Rocstar NEMoSys© toolkit. At completion, an open-source, web-enabled, modular, and extensible platform for mesh generation and solution transfer for complex reactor geometries will modified and improved using custom and open-source tools.

Nuclear reactors, whether light water reactors or liquid metal cooled fast reactors, are complex machines that require detailed analyses in complex multiphysics scenarios for design, licensing, and operation. Prior to the advent of high-performance computing, reactor modeling and simulation were limited to one-dimensional models including lumped parameter methods, homogenization, and empirical correlations. The introduction of high-performance computing, computational fluid dynamics, and high fidelity neutronics methods led to the formation of advanced modeling and simulation tools. These tools leverage state-of-the-art computational assets to provide physical insight into processes that were previously accessible only through experiment. Use of these tools can be challenging in some instances due to their demanding requirements, such as fully hexahedral grids that are difficult to produce, or lack of attention paid to user experience. The NEMoSys© toolkit will focus on the development of automated hexahedral meshing support for geometries of interest to nuclear modeling and simulation, as well as the overall user experience, to improve the use of modern high-fidelity methods in industry and beyond.