0
Research Papers: Fundamental Issues and Canonical Flows

# Effect of Channel Inlet Blockage on the Wake Structure of a Rotationally Oscillating Cylinder

[+] Author and Article Information
S. Kumar

Associate Professor
Department of Aerospace Engineering,
Indian Institute of Technology,
e-mail: skmr@iitk.ac.in

Contributed by the Fluids Engineering Division of ASME for publication in the JOURNAL OF FLUIDS ENGINEERING. Manuscript received September 15, 2015; final manuscript received July 6, 2016; published online September 12, 2016. Assoc. Editor: Peter Vorobieff.

J. Fluids Eng 138(12), 121203 (Sep 12, 2016) (18 pages) Paper No: FE-15-1667; doi: 10.1115/1.4034193 History: Received September 15, 2015; Revised July 06, 2016

## Abstract

This paper investigates, experimentally for the first time, the effect of channel inlet blockage induced by bringing the channel inlet walls closer together on the wake structure of a rotationally oscillating cylinder. The cylinder is placed symmetrically inside the channel inlet. The Reynolds number (based on constant upstream channel inlet freestream velocity) is 185, and three channel wall spacings of two, four, and eight cylinder diameters are used. Cylinder oscillation amplitudes vary from $π/8$ to π, and normalized forcing frequencies vary from 0 to 5. The diagnostics is done using hydrogen-bubble flow visualization, hot-wire anemometry, and particle image velocimetry (PIV). It is found that rotational oscillations induce inverted-vortex-street formation at channel width of two cylinder diameter where there is no shedding in unforced case. The channel wall boundary layers at this spacing undergo vortex-induced instability due to vortex shedding from cylinders and influence the mechanism of inverted-vortex-street formation near the cylinder. At channel width of four cylinder diameter, the inverted-vortex-street is still present but the mode shape change seen at normalized forcing frequency of 1.0 in the absence of channel walls is delayed due to the presence of nearby walls. The wake structure is observed to resemble the wake structure in unbounded domain case at channel width of eight cylinder diameter with some effect of channel walls on forcing parameters where mode shape change occurs. The lock-on diagram is influenced by the closeness of the channel walls, with low-frequency boundary moving to lower frequencies at smallest channel width.

<>

## Figures

Fig. 1

Schematic of the problem

Fig. 2

Schematic of the experimental setup for flow visualization and PIV experiments. (a) Top view, (b) side view, and (c) end view (looking upstream). For PIV studies, platinum wire and graphite anode were absent.

Fig. 3

Side view of the vortex shedding column ∼20D around the visualization plane in the absence of any walls. The flow is from left to right, and the vertical cylinder is visible on the left of the image.

Fig. 4

Boundary layer profile on one channel side wall flat plate at the location of cylinder placement in the absence of second channel side wall. Square symbol, measured and solid line, theoretical Blasius boundary layer profile.

Fig. 5

Velocity profiles measured in the channel at three channel widths with no cylinder: square, T/D=8; △, T/D=4; and ○, T/D=2

Fig. 6

Effect of forcing frequency on the wake structure at T/D=2.0 and forcing amplitude of π/2. The numbers on each frame show the value of FR, and the top and bottom of each frame represent the channel boundary. Each frame is at the same phase of cylinder oscillation with the cylinder about to turn clockwise.

Fig. 7

Effect of amplitude on the wake structure at T/D=2.0: (a) FR = 0.6 and (b) FR = 5.0. The numbers on each frame show the value of forcing amplitude, and the top and bottom of each frame represent the channel boundary. Each frame is at the same phase of cylinder oscillation with the cylinder about to turn clockwise.

Fig. 8

Inverted-vortex-street formation process at T/D=2.0, θo=π/2, and FR = 0.8. The time stamp on each frame is in terms of the cylinder oscillation period, τ, and the instant t = 0 represents when the cylinder is about to turn clockwise.

Fig. 9

Inverted-vortex-street formation process at T/D=2.0, θo=π/2, and FR = 0.8 with no guidelines for clarity purposes. The time stamp on each frame is in terms of the cylinder oscillation period, τ, and the instant t = 0 represents when the cylinder is about to turn clockwise.

Fig. 10

Cylinder oscillation frequency and amplitude effect on the mean velocity profiles at x/D=1 and T/D=2.0. The velocity profile with circle symbols shows the profile in the channel in the absence of the cylinder. (a) Frequency effect: θo=π/2; solid black, no oscillation; solid green, FR = 0.4; solid blue, FR = 0.6; solid red, FR = 1.0; solid brown, FR = 2.0; dashed, FR = 2.5; dotted, FR = 3.0; and dashed–dotted, FR = 5.0; and (b) amplitudeeffect: FR = 5.0; solid, θo=π/8; dashed, θo=π/4; dashed–dotted, θo=π/2; dotted, θo=3π/4; and dashed–dotted–dotted, θo=π.

Fig. 11

Effect of forcing frequency on the wake structure at T/D=4.0 and forcing amplitude of π/2. The numbers on each frame show the value of dimensionless frequency, and the top and bottom of each frame represent the channel boundary. Each frame is at the same phase of cylinder oscillation with the cylinder about to turn clockwise.

Fig. 12

Effect of amplitude on the wake structure at T/D=4.0 : (a) FR = 0.6 and (b) FR = 5.0. The arrowheads show some instances of the separated wall boundary layer vortex. The numbers on each frame show the value of forcing amplitude, and the top and bottom of each frame represent the channel boundary. Each frame is at the same phase of cylinder oscillation with the cylinder about to turn clockwise.

Fig. 13

Inverted-vortex-street formation process at T/D=4.0, θo=π/2, and FR = 1.0. The time stamp on each frame is in terms of the cylinder oscillation period, τ, and the instant t = 0 represents when the cylinder is about to turn clockwise. The top and bottom of each frame represent the channel boundary.

Fig. 14

Cylinder oscillation frequency and amplitude effect on the mean velocity profiles at x/D=1 and T/D=4.0. The velocity profile with triangle symbols shows the profile in the channel in the absence of the cylinder. (a) Frequency effect: θo=π/2; solid black, no oscillation; solid red, FR = 0.5; solid green, FR = 1.0; solid blue, FR = 1.5, dashed–dotted, FR = 2.0; dotted, FR = 3.0; dashed, FR = 4.0; and dashed–dotted–dotted, FR = 5.0; and (b) amplitude effect: FR = 5.0; solid, θo=π/8; dashed, θo=π/4; dashed–dotted, θo=π/2; dotted, θo=3π/4; and dashed–dotted–dotted, θo=π.

Fig. 15

Effect of forcing frequency on the wake structure at T/D=8.0 and forcing amplitude of π/2. The numbers on each frame show the value of dimensionless frequency, and the top and bottom of each frame represent the channel boundary. Each frame is at the same phase of cylinder oscillation with the cylinder about to turn clockwise.

Fig. 16

Effect of amplitude on the wake structure at T/D=8.0: (a) FR = 0.6 and (b) FR = 5.0.The numbers on each frame show the value of forcing amplitude, and the top and bottom of each frame represent the channel boundary. Each frame is at the same phase of cylinder oscillation with the cylinder about to turn clockwise.

Fig. 17

Cylinder oscillation frequency and amplitude effect on the mean velocity profiles at x/D=1 and T/D=8.0. The velocity profile with square symbols shows the profile in the channel in the absence of the cylinder. (a) Frequency effect: θo=π/2; solid, FR = 0.5; dashed, FR = 1.0, dashed–dotted, FR = 2.0; dotted, FR = 3.0; long-dash, FR = 4.0; and dashed–dotted–dotted, FR = 5.0; and (b) amplitude effect: FR = 5.0; solid, θo=π/8; dashed, θo=π/4; dashed–dotted, θo=π/2; dotted, θo=3π/4; and dashed–dotted–dotted, θo=π.

Fig. 18

Frequency spectrum showing the effect of confining walls on vortex shedding frequency from a nonoscillating cylinder

Fig. 19

Spectrum evolution with forcing frequency at T/D=2.0 and θo=π/2 showing lock-on: (a) FR = 0.1; (b) FR = 0.5; (c) FR = 0.8; (d) FR = 1.0; (e) FR = 1.4; (f) FR = 1.8; (g) FR = 2.0; (h) FR = 2.5; (i) FR = 3.0; (j) FR = 3.5; (k) FR = 4.5; and (l) FR = 5.0

Fig. 20

Spectrum evolution with forcing frequency at T/D=4.0 and θo=π/2 showing lock-on: (a) FR = 0.1; (b) FR = 0.4; (c) FR = 0.7; (d) FR = 1.0; (e) FR = 1.4; (f) FR = 1.8; (g) FR = 2.0; (h) FR = 2.5; (i) FR = 2.6; (j) FR = 3.0; (k) FR = 4.5; and (l) FR = 5.0

Fig. 21

Spectrum evolution with forcing frequency at T/D=8.0 and θo=π/2 showing lock-on: (a) FR = 0.1; (b) FR = 0.4; (c) FR = 0.6; (d) FR = 0.8; (e) FR = 1.0; (f) FR = 1.5; (g) FR = 2.0; (h) FR = 2.4; (i) FR = 2.5; (j) FR = 3.0; (k) FR = 4.0; and (l) FR = 4.5

Fig. 22

Wake lock-on diagram: regions showing wake lock-on in the forcing amplitude–forcing frequency plane at three channel wall spacings and x/D=6. The data of Kumar et al. [14] are also shown for the case of unbounded domain at x/D=4.

## Discussions

Some tools below are only available to our subscribers or users with an online account.

### Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related Proceedings Articles
Related eBook Content
Topic Collections