Pragmatic modelling of axial flow-induced vibration (FIV) for nuclear fuel rods

Abstract

Flow-induced vibration (FIV) at the spacer grid in the fuel assembly of a Light Water Reactor (LWR) is the leading cause of fuel failure. This project aims to produce a simulation benchmark on the experimental campaign at MACE on axial FIV of a cantilever beam in an annular tube, that mimics the configuration and environment of a typical LWR. The nuclear fuel rod, which consists of fuel pellets filled in Zirconium alloy cladding is modelled in the experiment as a steel rod filled with lead shots that closely approximates the filling density of the fuel pellet. To reduce the complexity and increase the efficiency of the simulation, further simplification is applied by on the geometry and assuming the solid domain as a single material instead of multiple materials. The first mode of frequency of the rod vibrating in quiescent water, which had been validated via the Euler-Bernoulli beam theory against experimental measurements, was used to design the single-material solid domain. Two models were proposed, firstly a solid rod with lower density and stiffness (SLE), and secondly an empty cladding with high density and low stiffness (EHD). Both solid and fluid domains were discretised using the cell-centred finite volume (FV) method and coupled with strong two-way fluid-structure interaction (FSI). Results on the frequency of vibration in quiescent water and in axial flow showed good agreement with experimental measurement, and the computational efficiency is analyzed for different rod models and changes in parameters of the solid domains.

URANS simulation and experimental validation of axial flow-induced vibrations on a blunt-end cantilever rod for nuclear applications

“…The URANS model closely matches the experimental results, registering just slightly above the first mode of vibration in still water. This near alignment reflects the limited impact of flow on the vibration frequency, further emphasising their dependence on the rod's properties [9][10][11][12][13][14][15].…”

“…Recent work by Salachna et al (2023) [14] has demonstrated the effectiveness of the URANS Reynolds stress model proposed by Launder, Reece, and Rodi (LRR) [18] in successfully reproducing both the amplitude and the frequency of axial-FIVs observed in experiments with a cantilever rod exposed to turbulent water flow at a Reynolds number of 16.1k [16]. This same methodology was then successfully extended to a higher Reynolds number of 35.1k [15]. This paper further develops the axial-FIV numerical methodology originally proposed by Salachna et al (2023) [14] by expanding the range of Reynolds numbers investigated (up to 61.7k), by incorporating new highresolution experimental data which have been generated specifically to support the further development of the numerical methodology, and by implementing the k-ω SST turbulence model in addition to the URANS LRR model previously employed.…”

“…Furthermore, when comparing the single-material model featuring solid cylinder geometry against the hollow cylinder geometry, similar to the thickness of the steel cladding, both demonstrate a similar first mode of vibration. However, the hollow cylinder geometry achieves significantly greater computational efficiency [15]. Therefore, considering these two factors, the geometry of the single-material rod is chosen to be similar to that of the steel cladding.…”

“…High-fidelity LES turbulence modelling has been successfully used in a number of studies to reproduce the flow unsteadiness [3][4][5], which has then be coupled with structural solvers to predict the subsequent structural vibration [6][7][8]. The significant computational cost of LES simulations has motivated further investigations with the more cost-effective URANS turbulence models, which have been employed in a number of recent studies [9][10][11][12][13][14][15] While most URANS studies have been quite successful in capturing the frequency of the structural vibration, reproducing the amplitude of vibration has proven to be considerably more challenging. In fact, the frequency of vibration mostly depends on the mechanical properties of the structure [16,17], and is only minimally affected by the turbulent flow.…”