Multiscale and multiphysics simulation leveraging coupling techniques and state-of-the-art codes
Please login to view abstract download link
The simulation of complex systems, such as the primary circuit in a nuclear reactor, requires multiscale and multiphysics capabilities due to the inherent difficulty of dealing with many physical models, geometric complexity, and time scales. A full-scale simulation of the flow geometry would require extensive computational capabilities that are not always needed, especially at the design stage of a system when the configuration can iterate quickly or simulate long pipes that are best modeled with 1-D elements. In this framework, multiscale and multiphysics techniques can significantly reduce the computational effort and offer valuable insights for research and design purposes. Single-domain computational software has been developed for a long time and offers state-of-the-art capabilities and validation for an extensive range of applications. This paper details implementing a coupling technique suitable for multiscale and multiphysics frameworks that rely on existing codes by managing the data exchange between them directly in memory\cite{barbi2024numerical,coupling2}. The different computational domains can be coupled via various techniques, relying on overlapping domains, boundary data exchange, or the defective boundary approach. The latest technique is the most suited to exchange data between a complex 3-D domain representing the reactor pressure vessel and the external circuit and its additional components, such as heat exchangers and pumps, that can be modeled with 1-D meshes and 0-D components. The methodology is demonstrated on some simplified models that mimic the full-scale problem to address the accuracy, robustness, and performance of the proposed approach.