Hostname: page-component-7b7567c8c-tnkh4 Total loading time: 0 Render date: 2026-11-16T19:10:58.108Z Has data issue: false hasContentIssue false

Sound source and pseudo-sound in the near field of a circular cylinder in subsonic conditions

Published online by Cambridge University Press:  01 June 2021

Shuai Li*
Affiliation:
Department of Mechanical and Materials Engineering, Queen's University, Kingston, ON K7L 2V9, Canada
David E. Rival
Affiliation:
Department of Mechanical and Materials Engineering, Queen's University, Kingston, ON K7L 2V9, Canada
Xiaohua Wu
Affiliation:
Department of Mechanical and Materials Engineering, Queen's University, Kingston, ON K7L 2V9, CanadaDepartment of Mechanical and Aerospace Engineering, Royal Military College of Canada, Kingston, ON K7K 7B4, Canada
*
†Email address for correspondence: shuai.li@queensu.ca

Abstract

It is well known that the pressure fluctuations on both sides of a cylinder and those in its oscillating near-wake region are both sound sources at low Reynolds and Mach numbers. However, assessment of the propagating capacity and quantification of the radiating versus non-radiating components of these two sound sources are not currently available for this important benchmark aeroacoustic problem. Here, we isolate the radiating acoustic sound sources from the non-radiating hydrodynamic pseudo-sounds by applying the wavelet decomposition technique of Mancinelli et al. (J. Fluid Mech., vol. 813, 2017), previously used in subsonic jet-noise experiments, to decompose the cylinder near-field pressure fluctuations obtained from our direct numerical simulations. Rigorous independence and convergence analyses of the wavelet decomposition procedure are performed. It is found that the radiating acoustic component strongly dominates over the non-radiating hydrodynamic component at near-field locations above and upstream of the cylinder. In the oscillating near-wake region, the hydrodynamic component dominates over the acoustic component at most frequencies, except at the vortex shedding frequency where they exhibit comparable strengths. Furthermore, within the oscillating near-wake region, the overall sound pressure level associated with the hydrodynamic pressure fluctuations exceeds that associated with the acoustic pressure fluctuations. Away from the oscillating near-wake region, the hydrodynamic pressure fluctuations decrease dramatically while the acoustic counterparts decay slowly, demonstrating that the hydrodynamic pressure fluctuation does not propagate, and that the acoustic pressure fluctuation is the only component to propagate to the far field.

Information

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

Figure 1.Instantaneous dilatation field of a circular cylinder at $Re=3900$ and $Ma=0.4$ with contour levels between $-0.01 c_{\infty }/D$ and $0.01 c_{\infty }/D$: (a) far field; (b) near field.

Figure 1

Figure 2.Sampling positions from the cylinder near field to its far field.

Figure 2

Figure 3.A uniform flow over a circular cylinder: $r$ is the radial distance from the origin and $\theta$ is the angle with the upstream centreline.

Figure 3

Figure 4.Instantaneous flow fields of a circular cylinder at $Re=3900$ and $Ma=0.4$: (a–c) vorticity magnitude with contour levels between $2 c_{\infty }/D$ and $50 c_{\infty }/D$; (d–f) dilatation with contour levels between $-0.01 c_{\infty }/D$ and $0.01 c_{\infty }/D$ at $tc_{\infty }/D=210$, $581$ and $983$, respectively. The arrows point to the negative-dilatation region around the flow separation point.

Figure 4

Figure 5.Statistics of the flow field at three stations in the wake of the cylinder. (a) Mean velocity in the streamwise direction, (b) mean velocity in the cross-flow direction, (c) variance of velocity in the streamwise direction, (d) variance of velocity in the cross-flow direction. Solid line, present DNS at $Ma=0.2$; dash-dotted line, LES of Mani (2009); $\diamond$, hot-wire measurements of Ong & Wallace (1996).

Figure 5

Figure 6.Streamwise velocity on the centreline in the wake of the cylinder. Solid line, present DNS at $Ma=0.2$; dash-dotted line, LES of Kravchenko & Moin (2000); $\diamond$, hot-wire measurements of Ong & Wallace (1996).

Figure 6

Figure 7.Pressure coefficient on the cylinder surface. Solid line, present DNS at $Ma=0.2$; dash-dotted line, LES of Kravchenko & Moin (2000); $\diamond$, previous experimental result referenced in Kravchenko & Moin (2000).

Figure 7

Figure 8.The PSD of density fluctuation normalized by the free stream flow density and the diameter of the cylinder at location $x=2D$ on the wake centreline. Solid line, present DNS at $Ma=0.4$; line with circles, LES of Mani (2009) at $Ma=0.4$.

Figure 8

Table 1.Comparison of flow statistics with existing references. The statistics from left to right are the drag coefficient, base pressure coefficient, Strouhal number of vortex shedding and minimum averaged streamwise velocity. The errors shown in the experimental and LES cases are based on a $95\,\%$ confidence interval. Experiments are from different incompressible studies referenced in Mani (2009); the LESs are results of Mani (2009); the DNSs are present results. For all cases, the Reynolds numbers are $3900$ except that the Reynolds number of $-C_{pbase}$ from the experiments is $4020$.

Figure 9

Figure 9.Contour of the SPSL in decibel at the vortex shedding frequency with a close-up around the cylinder. Here (a) $Ma=0.2$ and (b) $Ma=0.4$. For the two contours, different colour scales are used.

Figure 10

Figure 10.Near-field directivity plot of the normalized root mean square pressure fluctuation $p'_{rms}/\rho _{\infty }c_{\infty }^2$ in the cylinder midspan plane. Dotted line, along cylinder surface for $Ma=0.2$; dash-dotted line, along cylinder surface for $Ma=0.4$; solid line, along a circle centred at $(0.338D, 180^{\circ })$ with a radius of $2.19D$ for $Ma=0.2$; dashed line, along a circle centred at $(0.338D, 180^{\circ })$ with a radius of $2.19D$ for $Ma=0.4$.

Figure 11

Figure 11.Cross-correlation coefficient peak between the near-field separated acoustic (or hydrodynamic) component at $(2.50D, 160^{\circ })$ and the far-field pressure fluctuation at different polar angles. (a) Acoustic and hydrodynamic cross-correlation coefficient peak with the far-field pressure fluctuation selected at a polar angle of $50^{\circ }$ for $Ma=0.2$. Here $\circ$, acoustic component; $\vartriangle$, hydrodynamic component. (b) Acoustic cross-correlation coefficient peak with the far-field pressure fluctuation selected at different polar angles for $Ma=0.2$. Here $\diamond \ \theta =130^{\circ }$; $\cdot \ \theta =120^{\circ }$; $\times \ \theta =110^{\circ }$; $\circ \ \theta =100^{\circ }$; $\triangledown \ \theta =90^{\circ }$; $\square \ \theta =80^{\circ }$; $+\ \theta =70^{\circ }$; $\vartriangle \ \theta =60^{\circ }$; $\ast \ \theta =50^{\circ }$. (c) The same as panel (a) but at a polar angle of $100^{\circ }$ for $Ma=0.4$. (d) The same as panel (b) but for $Ma=0.4$. The iteration at which the threshold is selected for the wavelet separation procedure is highlighted with a vertical dashed line.

Figure 12

Figure 12.The PSD of the near-field pressure fluctuation and its separated acoustic and hydrodynamic components for sampling time series of 100 and 160 shedding cycles, respectively, for $Ma=0.4$ at two different positions: position A $(2.05D, 70^{\circ })$ and B $(2.50D, 160^{\circ })$. (a) Near-field pressure fluctuation at position A, (b) near-field pressure fluctuation at position B, (c) acoustic pressure fluctuation at position A, (d) acoustic pressure fluctuation at position B, (e) hydrodynamic pressure fluctuation at position A, (f) hydrodynamic pressure fluctuation at position B. Dashed lines, 160 shedding cycles; dotted lines, 100 shedding cycle. The spectra are bin-averaged over 1/6 octave. A Hann window is applied to avoid spectral leakage.

Figure 13

Figure 13.The p.d.f. of the near-field pressure fluctuation at $(2.50D, 160^{\circ })$ and its separated acoustic and hydrodynamic components for sampling time series over 100 and 160 shedding cycles, respectively, at $Ma=0.4$. Red lines, 160 shedding cycles; blue lines, 100 shedding cycles. Solid lines, near-field pressure fluctuation; dash-dotted lines, acoustic pressure fluctuation; dashed lines, hydrodynamic pressure fluctuation.

Figure 14

Figure 14.Contours of the cross-correlation coefficient peak between the near-field separated acoustic (or hydrodynamic) component and the near-field pressure fluctuation for $Ma=0.2$ and $Ma=0.4$. (a) Hydrodynamic cross-correlation coefficient peak, $Ma=0.2$, (b) acoustic cross-correlation coefficient peak, $Ma=0.2$, (c) hydrodynamic cross-correlation coefficient peak, $Ma=0.4$, (d) acoustic cross-correlation coefficient peak, $Ma=0.4$.

Figure 15

Figure 15.The PSD of the near-field pressure fluctuation and its separated acoustic and hydrodynamic components for $Ma=0.2$ and $Ma=0.4$ at two different positions. (a) For $Ma=0.2$ at $(1.96D, 50^{\circ })$, (b) for $Ma=0.2$ at $(2.50D, 160^{\circ })$, (c) for $Ma=0.4$ at $(1.96D, 50^{\circ })$, (d) for $Ma=0.4$ at $(2.50D, 160^{\circ })$. Dashed line: near-field pressure fluctuation, dotted lines: acoustic component, solid line: hydrodynamic component. The spectra are bin-averaged over 1/6 octave. A Hann window is applied to avoid spectral leakage.

Figure 16

Figure 16.The p.d.f. of the near-field pressure fluctuation at $(2.50D, 160^{\circ })$ and its separated acoustic and hydrodynamic components for sampling time series over 100 shedding cycles for both $Ma=0.2$ and $Ma=0.4$. Red lines, $Ma=0.2$; blue lines, $Ma=0.4$. Solid lines, near-field pressure fluctuation; dash-dotted lines, acoustic pressure fluctuation; dashed lines, hydrodynamic pressure fluctuation.

Figure 17

Figure 17.The PSD of the near-field pressure fluctuation and its separated acoustic and hydrodynamic components for $Ma=0.4$ at four near-field positions of different polar angles: (a) $(1.86D, 10^{\circ })$; (b) $(2.05D, 70^{\circ })$; (c) $(2.28D, 110^{\circ })$; (d) $(2.48D, 150^{\circ })$. Dashed line, near-field pressure fluctuation; dotted lines, acoustic component; solid line, hydrodynamic component. The spectra are bin-averaged over 1/6 octave. A Hann window is applied to avoid spectral leakage.

Figure 18

Figure 18.Contour of the SPSL of the separated hydrodynamic and acoustic components at the vortex shedding frequency in the midspan cylinder near field for $Ma=0.4$: (a) hydrodynamic component; (b) acoustic component. The solid lines separate the cylinder near field into three sectorial zones. The vertical dashed lines show the locations where the acoustic and hydrodynamic SPSLs are extracted to study how they decay, they are along $x=0$ and $x=2.48D$, respectively. For the two contours, different colour scales are used.

Figure 19

Figure 19.The difference between the acoustic and hydrodynamic OASPLs at sampling probes on two radial lines. Here $\circ$, $150^{\circ }$; $\vartriangle$, $160^{\circ }$.

Figure 20

Figure 20.The SPSL decay of the acoustic and hydrodynamic pressure fluctuations for $Ma=0.4$ in the cylinder midspan plane. Red dashed line, original pressure fluctuation along $x=0$; black solid line, acoustic pressure fluctuation along $x=0$; black dotted line, original pressure fluctuation along $x=2.48D$; black dashed line, acoustic pressure fluctuation along $x=2.48D$; black dash-dotted line, hydrodynamic pressure fluctuation along $x=2.48D$.

Figure 21

Figure 21.The OASPL decay of the acoustic and hydrodynamic pressure fluctuations for $Ma=0.4$ in the cylinder midspan plane. Same line styles as figure 20.

Figure 22

Figure 22.Cross-correlation coefficient peak between the near-field separated acoustic (or hydrodynamic) component at $(2.50D, 160^{\circ })$ and the selected near-field pressure fluctuation at different polar angles or the selected far-field pressure fluctuation at $(23.06D, 50^{\circ })$ for $Ma=0.2$ and $(27.17D,100^{\circ })$ for $Ma=0.4$. Here (a) $Ma=0.2$: $\circ$, acoustic component when far-field position is selected; $\vartriangle$, hydrodynamic component when far-field position is selected; $+$, acoustic component when near-field position at $(2.16D, 90^{\circ })$ is selected; $\ast$, hydrodynamic component when near-field position at $(2.16D, 90^{\circ })$ is selected. Here (b) $Ma=0.2$: $\circ$, acoustic component when far-field position is selected; $\cdot$, acoustic component when near-field position at $(2.05D, 70^{\circ })$ is selected; $\triangledown$, acoustic component when near-field position at $(2.16D, 90^{\circ })$ is selected; $\times$, acoustic component when near-field position at $(2.28D,110^{\circ })$ is selected. Here (c) $Ma=0.4$, with same symbol styles as panel (a). Here (d) $Ma=0.4$, with same symbol styles as panel (b). The iteration at which the threshold is selected for the wavelet separation procedure is highlighted with a vertical dashed line.

Figure 23

Figure 23.Contour of the cross-correlation coefficient peak between the near-field separated acoustic (or hydrodynamic) component and the pressure fluctuation at a far-field position $(27.17D,100^{\circ })$ or near-field position $(2.16D, 90^{\circ })$ for $Ma=0.4$. (a) Hydrodynamic cross-correlation coefficient peak, using the far-field position, (b) acoustic cross-correlation coefficient peak, using the far-field position, (c) hydrodynamic cross-correlation coefficient peak, using the near-field position, (d) acoustic cross-correlation coefficient peak, using the near-field position.

Figure 24

Figure 24.The PSD of the separated acoustic and hydrodynamic pressure fluctuation at $(2.50D, 160^{\circ })$ for $Ma=0.4$ using a far-field position and three near-field positions at different polar angles. Here $\vartriangle$, acoustic pressure fluctuation using a far-field position at $(27.17D,100^{\circ })$; $\circ$, hydrodynamic pressure fluctuation using a far-field position at $(27.17D,100^{\circ })$. Black solid line, acoustic pressure fluctuation using a near-field position at $(2.05D, 70^{\circ })$; black dashed line, hydrodynamic pressure fluctuation using a near-field position at $(2.05D, 70^{\circ })$; red solid line, acoustic pressure fluctuation using a near-field position at $(2.16D, 90^{\circ })$; red dashed line, hydrodynamic pressure fluctuation using a near-field position at $(2.16D, 90^{\circ })$; blue solid line, acoustic pressure fluctuation using a near-field position at $(2.28D,110^{\circ })$; blue dashed line, hydrodynamic pressure fluctuation using a near-field position at $(2.28D,110^{\circ })$. The spectra are bin-averaged over 1/6 octave. A Hann window is applied to avoid spectral leakage.