Nuclear reactor cores are inherently multi-physics systems. The interplay of the fluid flow, the conjugate heat transfer and the neutron distribution govern the behavior of the reactor. Accurate simulations of such systems are both computationally challenging and expensive. Due to system size and the multiple concurrent fields of physics, the problem is often solved using a so called divide-and-conquer approach, where single purpose codes are typically applied for one field of physics or one spatial resolution of the problem. In a later stage, the results of multiple such codes are re-combined to determine the coupled behavior.
In the DREAM4SAFER project we aim to solve the described reactor core problem in a single software, multi-scale and multi-physics manner. By including both coarse and fine scales in the same solver, the reactor core behavior can potentially be solved with a higher accuracy and in a more efficient manner as compared to existing code frameworks. For the multi-physics problem, an open-source finite-element software is extended with new algorithms and modules to capture both the distribution of neutrons at fine and coarse scales, as well as the fluid flow and heat transfer at the coarse scales. Special attention is given to the computational aspects of the coupling of the different scales and the different fields of physics. Furthermore, a separate methodology is used for the high-resolution fluid behavior, which is typically solved with a commercial or open-source CFD software.
In the FIRE project we aim to directly solve the multi-physics problem with a high-resolution in space and time. Due to the computational burden of such calculations, only a sub-geometry of the real reactor core is considered. In contrast to the DREAM4SAFER project, the primary aim is to investigate the behavior of the multi-physics couplings directly at the high-resolution level without recovering the global core behavior. In this project, open-source CFD software is used and extended to cover the multi-physics application. The project focuses on numerical methods as well computational performance of the developed algorithms.
All calculation schemes proposed and utilized in the presented projects are based on deterministic methods. Typically the finite-element or the finite-volume methods are used to discretize the systems of equations and the formulated problems are solved with matrix solvers.