Axial-flow-induced vibration experiments on cantilevered rods for nuclear reactor applications

Highlights

• New experimental data for axial-flow-induced vibration of cantilevered rods in tubes.

• Simplified configuration relevant for water-cooled nuclear reactor cores.

• Rods vibration resolved with fast video-imaging, flow field measured with PIV.

• Turbulence buffeting and rods movement identified as sources of excitation.

• Data relevant for numerical fluid-structure models calibration and benchmarking.

Abstract

Axial-flow-induced vibration has been experimentally investigated with clamped-free cantilevered cylindrical rods confined in a tube and subjected to axial water flow directed from the rod free-end towards the clamped end: a simplified configuration relevant for water-cooled nuclear reactor cores. Non-contact optical techniques have been used to simultaneously detect the rods vibration and the flow field around the vibrating rods free-end. The source of excitation is turbulent buffeting at low flow velocity, while a movement induced excitation component is present at large flow velocities. The rods flow-induced vibration consists of a fuzzy period-1 motion: a periodic (period-1) motion with a chaotic component that increases in relative importance as the flow velocity is increased. The experimental data provided here are particularly suited for numerical fluid-structure model development and benchmarking, as they combine a rich fluid-structure multi-physics interaction with a relatively simple configuration and include both the flow field and the mechanical response of the vibrating rods.

Introduction

Structural vibrations induced by axial flow in slender cylindrical structures are generally of small amplitude, and therefore of little concern in most practical applications. Notable exceptions are nuclear reactor cores and other systems composed of closely-spaced cylindrical elements exposed to axial flows. Even small amplitude vibrations can consequently result in impacting between adjacent cylinders or between cylindrical elements and supports, thus causing fretting wear, fatigue and eventually structural cracking and system failure. Fretting wear is in fact responsible for the majority of fuel leaks observed in water-cooled nuclear reactors. This highlights the importance of analyzing and predicting axial-flow-induced vibrations for the safe and profitable operation of nuclear power stations. In most cases of practical interest involving cylindrical structures in axial flow, the observed flow-induced vibrations are so-called subcritical vibrations: low-amplitude vibrations whose dominant frequency corresponds to the first mode of the cylinder (Païdoussis, 2014). Divergence and flutter have been observed but are generally of little practical concern, as the flow velocities required for the inception of divergence and flutter are much higher than those typical of most engineering applications. With subcritical vibrations, the source of excitation is typically turbulent buffeting, so that these flow-induced vibration problems are characterized as Extraneously-Induced Excitations (EIE), as turbulence is largely independent of the movement of the structure, or of the way the structure affects the flow field. Turbulence in the flow induces pressure fluctuations along the cylinder surface that are spatially and temporally not uniform, thus giving rise to a random lateral load on the cylinder surface that causes lateral motion and triggers the vibration. These pressure fluctuations are typically wide-band, so that the cylinder can extract energy from the flow at frequencies corresponding to its natural frequencies, with the first natural frequency typically being dominant.

In nuclear fuel bundles, and similarly with other systems composed of closely spaced cylindrical elements, even small-amplitude vibrations can appreciably change the geometry of the flow passage and therefore affect the axial flow, thus adding a Movement-Induced source of Excitation (MIE) to the flow-induced vibration problem. The proximity of the fuel rods, and the relatively high density of the coolants used in nuclear applications, result in strong fluid coupling among different fuel rods, meaning that the vibration of one rod can propagate to the adjacent rods as well. Therefore, in nuclear reactor fuel bundles, as the fuel rods extract energy from the turbulent flow and vibrate, their movement affects the axial flow and adjacent rods influence each other through fluid coupling. These interactions result in a tightly coupled multi-body fluid-structure interaction problem, where the axial flow and the rod vibrations dynamically interact and affect each other. Secondary sources of rod excitation that add further complexity to these problems include: 1) far-field disturbances transmitted through the flow, such as pump noise and cavitation; 2) vibrational noise transmitted through the system support structures; 3) inlet conditions in the fuel assemblies and flow restructuring based on the design of the lower core support plate, debris bottom nozzle filters and lower plenum internals; 4) variation in distance among different fuel assemblies due to positioning; and 5) fuel rod bowing and side effects near the core edges.

Axial-flow-induced vibrations in nuclear reactor applications are currently investigated using coupled computational fluid dynamics (CFD) and structural dynamics (SD) simulations, as analytical modeling and purely experimental investigation alone are not feasible. Analytical modeling of axial-flow-induced vibrations is, in fact, hampered by the procurement of the detailed information needed as input to the models: fluctuating pressure fields are very difficult to characterize a priori, and far-field upstream effects are problem-specific and very challenging to incorporate. On the other hand, testing at prototypical nuclear reactor operating conditions is too expensive to rely on a purely experimental approach to address the problem. To be representative of reactor conditions, in fact, a test piece should include several fuel rods together with the spacer grids and support structures, and should replicate the inlet conditions that depend on the design of the lower core support plate, so that the generated data would be representative of one specific reactor core design. The high pressure and high temperature test rig required to run the experiments would be quite expensive to build and operate, and the high operating pressure and temperature would make the direct visualization of the flow very challenging, if at all possible. In coupled CFD/SD simulations, a high-resolution CFD calculation provides the instantaneous pressure field and associated turbulent fluid force on the fuel rod surface, which is then used as input to SD simulations that yield the fuel rod motion. Large computational grids (several tens of million cells) and advanced turbulent models (large eddy simulation) are typically required to properly capture the intricate geometry of nuclear fuel bundles and faithfully reproduce instantaneous pressure fields and mechanical loads (Conner et al., 2010, Yan et al., 2011, Elmahdi et al., 2011, Zhang and Yu, 2011, Delafontain and Ricciardi, 2012, Bakosi et al., 2013, Christon et al., 2016).

It is evident that experimental data are essential to properly back up the development and validation of CFD/SD simulations. However, most existing studies only focus on the mechanical vibration or on the flow field, while experimental investigations where the mechanical vibration and the flow field are simultaneously resolved are still missing. As already highlighted, the feedback of the rod vibration on the flow field can be significant in tightly packed nuclear reactor cores, and this effect is not yet properly captured in existing experimental studies. The present investigation was therefore conducted to specifically provide experimental data for the development and validation of CFD/SD models where both the flow field and the structural response are simultaneously resolved. In particular, we have investigated flow-induced vibrations with a clamped-free cantilevered cylindrical rod confined in a tube and subjected to axial flow directed from the rod free-end towards the clamped-end. The geometry of this system is considerably simplified, as spacer grids are not included and the single-rod configuration eliminates multi-body effects. Notably, this configuration yields relatively large rod displacements even at moderate flow velocities, thus making the two-way coupling between the fluid and the structure quite strong. This makes the present data particularly suited for CFD/SD model development and benchmarking, as they combine a rich fluid-structure interaction with a relatively simple configuration. The present configuration is actually the simplest one we could conceive that still retains relevance for CFD/SD model development for nuclear applications. Data generated with more prototypical configurations are clearly invaluable: however, the experimental complexity and the computational burden associated with actually using such data makes simplified configurations such as the one tested here valuable for CFD/SD models benchmarking. In addition, cantilevered flexible slender cylinders in axial annular flow directed from the free-end towards the clamped-end have not been extensively investigated so far (Rinaldi and Païdoussis, 2012), so the present study is actually also relevant in this more fundamental respect.