Conference Agenda

Flow-induced vibrations (FIV) of key nuclear reactor components, such as fuel rods and steam generator (SG) tubes, may lead to wear and damage. Hence understanding and being able to predict the vibrational behavior of these components is crucial to mitigating any potential risks and preventing undesired outages. Fuel rods primarily vibrate due to the turbulent axial flow, requiring generally scale-resolving models to properly study their vibrations numerically. SG tubes on the other hand are exposed to cross-flow, with vibrations being a result of a combination of turbulence-induced and vortex-induced vibrations, possibly resulting in fluid-elastic instability. This cross-flow nature of the problem may make it possible to study it using computationally less expensive numerical techniques, such as those based on the Unsteady Reynolds-Averaged Navier-Stokes (URANS) approach or hybrid turbulence models. The current paper attempts to give an overview of where we are in terms of our understanding of FIV of a multiple tube configuration in cross-flow and how well these problems can be modelled using medium-resolution numerical approaches. This is done by first considering two well studied problems, being the numerical benchmark of Turek & Hron of a flexible flap attached to a fixed cylinder, and a single cylinder in cross-flow. The former allows one to validate properly the FSI framework used to study cylinders subjected to cross-flow, while the latter serves as a canonical problem fundamental to understanding tube bundles in cross-flow. Following these two cases, two-cylinder systems, with cylinders positioned either inline or side-by-side, and tube bundles are discussed. In general, a lot of data coming from experiments is available for all these cases, allowing one to validate and study them numerically in detail. Also, medium-resolution simulations do provide reasonable predictions for single and two-cylinder configurations, but struggle to recover vibration amplitudes in the lock-in regime. For tube bundles though, the amplitudes are generally overpredicted. This may be caused by a lack of turbulence that is actually resolved, although more detailed benchmark data is needed to further investigate this.