Hostname: page-component-7b7567c8c-tnkh4 Total loading time: 0 Render date: 2026-11-16T04:53:44.568Z Has data issue: false hasContentIssue false

There is no critical mass ratio for galloping of a square cylinder under flow

Published online by Cambridge University Press:  26 November 2021

Peng Han*
Affiliation:
School of Marine Science and Technology, Northwestern Polytechnical University, Xi'an 710072, PR ChinaLadHyX, CNRS, Ecole polytechnique, Institut Polytechnique de Paris, 91120 Palaiseau, France
Emmanuel de Langre
Affiliation:
LadHyX, CNRS, Ecole polytechnique, Institut Polytechnique de Paris, 91120 Palaiseau, France
*
†Email address for correspondence: peng@ladhyx.polytechnique.fr

Abstract

The flow-induced vibration of square cylinders under flow is known to be caused by two distinct mechanisms of interaction: vortex-induced vibrations and galloping. In the present paper we address the issue of the apparent suppression of galloping when the mass ratio between the solid and the fluid is low enough. By using a reduced-order model that we validate on pre-existing results, we show that galloping is actually not suppressed, but delayed to higher values of the flow velocity. This is explained using a linear stability analysis where the competition between unstable modes is related to the transition between vortex-induced vibration and galloping. Direct numerical simulations coupled with a moving square cylinder confirm that galloping can be found even at very low mass ratios.

Information

Type
JFM Papers
Copyright
© The Author(s), 2021. Published by Cambridge University Press
Figure 0

Figure 1.Elastically supported square cylinder under flow, with a representation of the wake oscillator variable $q(t)$ and of the relative flow velocity that causes galloping.

Figure 1

Figure 2.(a) Amplitude and (b) frequency of motion of an elastically supported square cylinder under flow as a function of the dimensionless flow velocity. Comparison between the models used in the paper (nonlinear ROM and DNS) and other DNS results (Zhao et al.2013; Bhatt & Alam 2018; Li et al.2019). Dimensionless parameters: $Re=100$; $m_{s}^{*}=3$; $\zeta =0$.

Figure 2

Figure 3.(a) Mass ratio effect on the amplitude of motion of a square cylinder under flow. The results of both our ROM-NL and DNS are compared with those of Joly et al. (2012). (b) Magnitude of the vortex-induced lift as defined by Joly et al. (2012). Comparison between the value fitted by Joly et al. (2012) on their DNS results and that resulting from our ROM-NL. Dimensionless parameters: $Re=200$; $\zeta =0$.

Figure 3

Figure 4.Effect of the mass–damping parameter $S_{c}^{*}$ on the amplitude of response $\hat {y}$ obtained by ROM-NL and DNS, and on the growth rates $G$ of the modes obtained by ROM-LSA, as a function of the reduced velocity. Dimensionless parameter: total mass ratio $\mu =100$; structural mass ratio $m_{s}^{*}=98.8$.

Figure 4

Figure 5.Comparisons of frequency of response obtained by ROM-NL, ROM-LSA and DNS, as a function of the reduced velocity, at (a) $S_{c}^{*}=5$ for coupled and (b) $S_{c}^{*}=15$ for separate VIV–galloping motion.

Figure 5

Figure 6.Effect of mass ratio $\mu$ and of mass–damping parameter $S_{c}^{*}$ on the amplitude of response $\hat {y}$ and on the growth rates $G$ of modes obtained by ROM-LSA, as a function of the reduced velocity. (a,c,e) High damping, $S_{c}^{*}=15$ corresponding to separate VIV and galloping. (b,d,f) Low damping, $S_{c}^{*}=5$, corresponding to coupled VIV and galloping. Top line, amplitude of response using the ROM-NL and DNS. Middle and bottom lines growth rates of modes for large and small mass ratios, respectively. Dimensionless parameter: total mass ratio $\mu =5$, 100 and 500; structural mass ratio $m_{s}^{*}=3.8$, 98.8 and 498.8.

Figure 6

Figure 7.(a) Effect of mass ratio on the amplitude of motion of a square cylinder. The results of Joly et al. (2012) and our ROM-NL and DNS at $U_{r}=40$ are those already given in figure 3(a). Our DNS results at $U_{r}=66$ and $U_{r}=90$ show that there exists a transition to galloping at low mass ratio. (b) Time histories of the displacement at $U_{r}=66$ and $U_{r}=98$, at $m_{s}^{*}=2$, showing a VIV response and a galloping response, respectively. Panels (c,d) are respectively the amplitude $\hat {y}$ and the frequency $f^{*}$ of a square cylinder for a high $m_{s}^{*}=20$ and low $m_{s}^{*}=2$ mass ratios, as a function of the reduced velocity $U_{r}$.

Figure 7

Figure 8.Amplitude of motion of a $3:2$ rectangular cylinder obtained by the present ROM-NL, as a function of the reduced velocity for two mass ratios. Comparison with the experimental data of Mannini et al. (2018) for high mass ratio. The low mass ratio case shows extended VIV and delayed galloping, as for the square cylinder.

Figure 8

Figure 9.(a) Amplitude of lift coefficient, $R$, and (b) phase angle, $\varTheta$, of a square cylinder under forced vibrations ($y_0=0.1$). Experimental data are taken from Carassale et al. (2015).

Figure 9

Figure 10.A sensitivity analysis of the effects of the (a) lift coefficient, $C_{L0}$, and (b) Strouhal number, $S_{t}$, on the amplitude response $\hat {y}$ obtained by ROM-NL.