Axial Flow-Induced Vibrations over a Blunt-Ended Cantilevered Rod for Nuclear Applications

Abstract

Fretting wear at the spacer grid in fuel assemblies, owing to flow-induced vibration (FIV), is a leading cause of fuel failure in Light Water Reactors (LWRs). Accurate FIV prediction is therefore vital, necessitating an efficient computational approach. To this end, this thesis employs the Unsteady Reynolds-Averaged Navier-Stokes (URANS) method as an effective alternative to the computationally intensive Large Eddy Simulation (LES). Focusing on strong two-way Fluid-Structure Interaction (FSI) coupling, the study benchmarks axial FIV in nuclear settings by comparing simulations with experiments conducted at the University of Manchester Thermo-Fluid Labs. Two key aspects of self-excited FIV, namely the dominant vibration frequency and the root-mean-square (RMS) amplitude of vibration, are examined. The former depends on optimising the solid domain and FSI coupling, whereas the latter depends on the fluid solver's capacity to replicate accurate unsteady flow behaviour, especially where flow separation and turbulent boundary layers interact. Initial challenges involved overcoming computational limitations to obtain sufficient vibrations to reliably predict the RMS amplitude and first mode vibrations. Most computational bottlenecks were observed in the solid domain; optimisation strategies included simplifying fuel rod geometry and enhancing the FSI coupling algorithm. Various URANS models and convection schemes were evaluated for their ability to reproduce unsteady flow behaviour. Both the eddy viscosity model (EVM) k-ω SST and the Reynolds Stress Model (RSM) Launder, Reece, and Rodi (LRR) successfully predict the RMS amplitude and frequency of vibrations for annulus Reynolds numbers range between 16,400 and 61,700. Challenges arise at higher Reynolds numbers, particularly in maintaining fluid solver stability and mesh quality. The numerical methodology was further validated through numerical experiments involving different solid-to-fluid density ratios and flow directions. While first mode vibrations are accurately modelled in all cases, reproducing precise unsteady flow behaviour proves more elusive, particularly in fixed-free configuration. The study reaffirms the importance of resolving flow in areas where separation and turbulent boundary layers occur to achieve accurate unsteady flow behaviour. Future work could address these limitations and broaden the methodology's applications, benefiting not only the nuclear industry, but also the aerospace and construction sectors.

Date of Award	1 May 2025
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Hector Iacovides (Supervisor)