Experimental Thermal and Fluid Science 32 (2008) 1754–1758

Contents lists available at ScienceDirect

Experimental Thermal and Fluid Science

journal homepage: www.elsevier.com/locate/etfs

A flow visualization study of the development of vortex structures in a round

jet impinging on a flat plate and a cylindrical pedestal
John C. Duda, Francis D. Lagor, Amy S. Fleischer *
NovaTherm Laboratory, Department of Mechanical Engineering, Villanova University, 800 Lancaster Avenue, Villanova, PA 19085, USA

a r t i c l e i n f o a b s t r a c t

Article history: Smoke-wire flow visualization is employed to investigate the behavior of a round jet issuing from a
Received 28 April 2008 straight tube and impinging on a cylindrical pedestal mounted on a flat plate. Velocity and turbulence
Accepted 4 May 2008 intensity measurements at the jet exit show correlation with the formation of visualized structures near
the impingement surface. The effects of jet exit-to-surface spacing (H/d) and Reynolds number (Re) are
explored. It is found that as jet exit-surface spacing increases from 2 to 5, small but significant effects
Keywords: can be seen. These effects include a stabilization of the flow, a lengthening of the potential core and a
Jets
reduction in vortex formation prior to impingement. An increase in Re is found to increase the turbulence
Jet impingement
Flow visualization
in the shear layer and suppress vortex formation. The presence of the cylindrical pedestal is found to sta-
bilize the jet prior to impingement and also creates a recirculation region at the edge of the pedestal for
low Re.
Ó 2008 Elsevier Inc. All rights reserved.

1. Introduction investigate the behavior of a round jet impinging on both concave

and convex surfaces with high curvature values and Fleischer et al
Jets have been widely used to enhance heat transfer from a sur- [5] used the smoke-wire visualization technique to investigate the
face in a variety of applications, examples of which range from tex- behavior of a round jet impinging on a convex surface. In both
tile manufacturing to electronics cooling. Consequently, several cases, [4,5] the geometry of the impingement surface was found
studies have identified and investigated the effects of key parame- to have a significant effect on the formation of vortex structures
ters that can affect jet impingement cooling of flat plates. The in the jet shear layer.
parameters include but are not limited to jet diameter (d), orienta- Lee et al. [6] studied heat transfer from turbulent round jets on
tion, Reynolds number (Re), and jet-to-surface spacing which is inclined concave surfaces for 11,000 6 Re 6 50,000 and found that
typically nondimensionalized as the distance from the mouth of the stagnation point Nusselt number decreases with increasing
the jet to the surface divided by the diameter of the jet (H/d). Com- surface angle. Lim et al. [7] studied heat transfer from turbulent
prehensive reviews of two decades worth of heat transfer data round jets on inclined convex surfaces for Re = 23,000 and also
have been compiled by Jambunathan et al. [1] and Viskanta [2]. found that stagnation point Nusselt number decreases with
Despite this wealth of heat transfer data, the fluid dynamics increasing surface angle.
driving jet performance have received less attention and are still More recently, jet impingement on cylindrical pedestals has
not completely understood. Flow visualization permits a direct been of interest. Mesbah [8] measured the local heat transfer from
look at the fluid flow and developing vortex structures, and allows a 38.1 mm cylindrical pedestal to a 40.3 mm impinging round jet
an analysis of the parameters that affect jet formation. Popiel and using the preheated wall transient method. H/d varied from 1 to
Trass [3] employed the smoke-wire visualization technique to de- 6 jet diameters for Reynolds numbers of 23,000, 50,000, and
velop a greater physical understanding of the formation and move- 70,000. Parneix et al [9] considered a similar geometric configura-
ment of vortex structures in both free jets and jets impinging on tion and parameters for a numerical study in which they investi-
flat surfaces. In this study, vortex initiation and fluid entrainment gated the accuracy of the V2F turbulence model for this type of
processes were identified and correlated to the H/d spacing for jet. Fleischer and Nejad [10] investigated jet impingement cooling
Reynolds numbers (based on jet diameter) of 10,000 and 20,000. of a discretely heated portion of a cylindrical pedestal 50 mm in
Cornaro et al [4] used the smoke-wire visualization technique to diameter and 6.4 mm high. H/d ranged from 2 to 5 jet diameters
at Reynolds numbers of 10,000, 20,000, and 30,000. In all cases
the presence of the protruding pedestal was found to increase heat
* Corresponding author. Tel.: +1 610 519 4954; fax: +1 610 519 7354.
E-mail address: amy.fl[email protected] (A.S. Fleischer). transfer from the discretely heated region when compared to the

0894-1777/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.expthermflusci.2008.05.001
 J.C. Duda et al. / Experimental Thermal and Fluid Science 32 (2008) 1754–1758 1755

Nomenclature

d jet tube diameter, m U local velocity of jet, m/s

D pedestal diameter, m UP velocity fluctuation of jet, m/s
h pedestal height, m UC velocity at jet exit centerline, m/s
H jet-to-surface distance, m x axial coordinate of jet measured from jet centerline at
L tube length, m tube exit, m
r radial coordinate from stagnation point, m m Kinematic viscosity, m2/s
Re Reynolds number based on jet diameter (Re = Ud/m),
dimensionless

flat-plate control arrangement. Chung [11] measured local heat adjustment of the flow rate. The tubing connects to the plenum
transfer coefficients across cylindrical fins 21.5 mm in diameter where layers of mesh and honeycomb in the plenum break down
with height-to-diameter aspect ratios 0.5, 1.0, and 1.5. Smoke-wire any turbulent structures induced upstream.
visualization was employed for a Reynolds number 2300 to inves- At the bottom of the plenum is an aluminum tube of 21 mm ID
tigate the development of the recirculation regions along the sides and length (L) of 10 diameters (210 mm) which forms the jet. The
of the cylinder for different cylinder height-to-diameter aspect tube entrance is packed with a honeycomb material and a series of
ratios. screens which straighten the flow within the tube. The jet issues
The present study is designed to provide physical insight into from the end of the tube and impinges on either a flat plate or a
the fluid dynamics of a round jet impinging on a cylindrical pedes- cylindrical pedestal. The cylindrical pedestal is the same as that
tal with low height-to-diameter aspect ratios for Reynolds used in Fleischer and Nejad [10]. This pedestal impingement sur-
numbers below 2000. This fulfills a unique role in the field as face is formed by a 19.9 mm  19.9 mm  6.4 mm square pedestal
previous visualization studies have considered only jets at Rey- machined into a solid copper cylinder of 31.8 mm in diameter
nolds number 2300 or greater which often feature high jet veloci- insulated at the pedestal edges by a snugly fitting insulating ring
ties and increased acoustic noise, both of which can be undesirable fabricated from a 6.4 mm thick, 50 mm diameter Teflon ring (see
in certain applications. Fig. 1). This forms a cylindrical pedestal height (h) of 6.4 mm and
diameter (D) of 50 mm with a center copper section which can
2. Experimental apparatus and data collection be discretely heated. In this study, the pedestal is not heated.
Flow visualization was completed using the smoke-wire tech-
This experiment was conducted using a jet issuing from a tube nique. The smoke is generated from a 0.1 mm diameter Nickel–
21 mm in diameter at Reynolds numbers based on jet diameter of Chromium wire with resistance of 21.3 X/m strung tightly across
400 and 1400 for H/d spacing of 2–5. A schematic of the experi- the mouth of the jet. An oil mixture was applied to the wire gener-
mental apparatus is shown in Fig. 1. Air from the building supply ating evenly spaced oil droplets across the mouth of the jet. When
enters through a pressure regulator. At the exit of the regulator power is supplied to the wire, the oil droplets heat via Joule heat-
is a system of valves that directs the flow through one of three flow ing and begin to smoke. This fine smoke is entrained in the jet flow,
meters depending on the flow rate required. Downstream of the tracing streaklines. The smoke streaklines were recorded using
flow meters, the piping converges and enters a length of both a Canon Rebel XT digital still camera and a Sony DCR-
19.1 mm ID reinforced vinyl tubing. A needle valve allows precise VX2000 high speed digital video camera. All images were taken

Fig. 1. Schematic of jet arrangement and pedestal surface.

1756 J.C. Duda et al. / Experimental Thermal and Fluid Science 32 (2008) 1754–1758

in black and white to increase streakline contrast and visibility. The velocity at a given location, U, normalized by the center line veloc-
images underwent no post-processing or refinement. ity, Uc. The profiles are measured radially across the mouth of the
The experiment was conducted at Reynolds number based on jet and are presented nondimensionally as the radial distance, r,
jet diameter of 400 and 1400 for H/d spacing of 2–5 jet diameters. from the centerline of the jet divided by the jet diameter, d. The ax-
The desired flow rate was set using the flowmeters and the result- ial profile is measured by positioning the hot-film probe at the jet
ing velocity and turbulence intensity profiles for the jet were exit centerline and measuring the velocity at various distances
checked radially and axially. A TSI IFA-300 constant temperature away from the jet exit, x. Both radial and axial profiles were re-
anemometer system was used to measure the hot film voltage data corded for the free jet.
and TSI ThermalPro software was used to calculate the velocity and The radial distributions of velocity and turbulence intensity for
turbulence intensity profiles. The probe was calibrated by TSI at Re = 400 (Fig. 2a) depict characteristics of laminar flow with low
their facility over the velocity range 0–10 m/s and the calibration turbulence intensity values. The radial distribution of velocity
features a mean error of less than 2% on voltage measurement and turbulence intensity at the jet exit for Re = 1400 depict the
and correlation to velocity/turbulence. The IFA-300 contains a characteristics of the beginnings of transitional flow as seen in
thermocouple circuit that compensates for any temperature drift Fig. 2b. The slight skewness seen in the normalized velocity profile
over the course of the study. In the high turbulence regions veloc- is within the accuracy of the velocity measurements. The velocity
ity measurements are fluctuating 8–13%, thus the reported results is fluctuating with a turbulence intensity of 8–10%. The higher tur-
are an average of 10–20 individual measurements. These high tur- bulence intensity leads to more data point scattering and the influ-
bulent fluctuations can affect the overall accuracy. The hot film ence of this increased turbulence will be clearly seen in the flow
probe position is controlled using a bislide driven by a stepper mo- visualization.
tor. The probe was positioned at even intervals of approximately
0.9 mm, yielding 24 data points over the 21 mm jet diameter.
The turbulence intensity is presented as the percentage of the 3. Results and discussion
ratio of the velocity fluctuations at a given location, UP, over the
velocity at that point, U. The velocity profiles are presented as 3.1. Flow visualization of the jet impinging on a flat plate

Several previous studies have found that toroidal vortices form

in the shear layer of an impinging jet prior to impacting a flat sur-
face. When these structures reach the impingement surface, the
vortices follow the jet flow, moving radially outward from
the impingement point along the surface [3–5]. The location of
the impingement surface from the jet exit has been found to have
an effect on the formation of these structures. For small H/d values
the length of the free jet region is not sufficient for these structures
to form prior to impingement, however, vortices may still form in
the wall jet region as identified by Cornaro et al. [4].
This study features significantly smaller Reynolds numbers than
all previous works and yet confirms the formation of vortices in the
wall jet region even for these low Re conditions. A dependence on
H/d spacing is also confirmed. As can be seen in Fig. 3a for Re = 400,
H/d = 3, no vortices form in the shear layer of the jet prior to
impingement. This impingement location is very close to the end
of the potential core which is at H/d = 3.5 for this jet. The break-
down of the potential core features an increase in turbulence and
the effects of this can be seen. While no vortices form prior to
impingement, the vortices form rapidly in the wall jet region and
are spaced more closely together than for lower H/d values. The
close spacing of the vortices leads to interesting behavior as the
wall jet develops. The vortices tend to collide resulting in vortex
combination and dissipation as seen in Fig. 3a. This behavior was
not observed for H/d = 2 (Fig. 3b).
A Reynolds number effect is also seen. As Reynolds number in-
creases, the turbulence in the shear layer increases and the coher-
ent vortex structures are less discernible. Those few vortex
structures that do form are significantly smaller than those at low-
er Reynolds numbers. In the wall jet, the vortices form closer to the
stagnation region, typically between x/d = 0.75 and 1. Additionally,
significant oscillation of the stagnation point is observed in con-
trast to the more stable stagnation point for lower Re as seen in
Fig. 3a, b. This stagnation point oscillation leads to instabilities in
the wall jet and faster transition to a turbulent wall jet leading to
early dissipation of the vortex structures.

3.2. Flow visualization of the jet impinging on a cylindrical pedestal

Fig. 2. Free jet velocity and turbulence intensity distributions for (a) Re = 400 and The addition of the 50 mm diameter, 6.4 mm high cylindrical
(b) Re = 1400. pedestal to the impingement surface is found to have a significant
 J.C. Duda et al. / Experimental Thermal and Fluid Science 32 (2008) 1754–1758 1757

Fig. 3. Jet impingement on a flat plate for Re = 400, (a) H/d = 3 and (b) H/d = 2.

effect on the flow patterns observed at both Re = 400 and 1400, de- height in length and one pedestal height high and is independent
spite the pedestal’s low height-to-diameter aspect ratio of 0.128. of jet-to-surface spacing for the range examined in this study.
At Re = 400 and H/d = 2 the formation of vortices in the wall jet is The Reynolds number effect for the pedestal configuration is the
strongly affected by the presence of the pedestal and this influence same as for the flat plate. The turbulence in the shear layer in-
can be clearly seen by comparing Fig. 3b to Fig. 4a. For impinge- creases with Reynolds number and vortex structures are less likely
ment on the flat plate (Fig. 3b), the impingement is smooth and to form. Significant oscillation of the stagnation point is again seen
uniform and the wall jet develops slowly along the length of the leading to instabilities in the wall jet and faster transition to a tur-
impingement plate. Vortices form in the wall jet at x/d = 2–4. In bulent wall jet. The turbulent flow does not clearly separate over
contrast, with the cylindrical pedestal in place (Fig. 4a), the flow the pedestal edge, but instead rapidly mixes and dissipates.
quickly separates over the pedestal edge. This leads to three dis-
tinct regions of the wall jet. There is a recirculation region at the
base of the pedestal, very similar to flow seen over a backwards 4. Conclusions
facing step, a separation zone where flow detaches from the pedes-
tal surface and does not reattach downstream, and a region of sep- Smoke-wire visualization is used to provide insight into the
arated flow which reattaches after the pedestal boundary and fluid dynamics of low Reynolds number jets impinging on flat
forms the wall jet. It is in this reattachment region that the vortices plates and cylindrical pedestals. The presence of the cylindrical
begin to form. However, the vortices form much closer to the pedestal on the impingement surface is found to have a significant
impingement point than for the flat plate, forming at x/d = 1 2. effect on the developed flow patterns. For impingement on the flat
As H/d increases some small but significant effects can be seen plate, the impingement is smooth and uniform and the wall jet
(Fig. 4b). First is the loss of the separation zone. As H/d increases, develops slowly along the length of the impingement plate with
the flow impacts the pedestal with less energy and the flow tends vortices forming in the wall jet between x/d = 2 4. With the cylin-
to reattach. In Fig. 4b, it can be seen that vortices do form in the drical pedestal in place however, the flow quickly separates over
wall jet region at x/d = 1 2, but the wall jet is much more turbu- the pedestal edge leading to three distinct regions of the wall jet.
lent than that seen in Fig. 3a for impingement on the flat plate. There is a recirculation region at the base of the pedestal, a separa-
The recirculation region tends to be roughly twice the pedestal tion zone where flow detaches from the pedestal surface and does
1758 J.C. Duda et al. / Experimental Thermal and Fluid Science 32 (2008) 1754–1758

Fig. 4. Jet impingement on a cylindrical pedestal for Re = 400, (a) H/d = 2 (b) H/d = 3.

not reattach downstream (for low H/d spacing), and a region of [5] A.S. Fleischer, K. Kramer, R.J. Goldstein, Dynamics of the vortex structure of a
jet impinging on a convex surface, Experimental Thermal and Fluid Science 24
separated flow which reattaches after the pedestal boundary and
(2001) 169–175.
forms the wall jet. It is in this reattachment region that the vortices [6] C.H. Lee, K.B. Lim, S.H. Lee, Y.J. Yoon, N.W. Sung, A study of the heat transfer
begin to form in the wall jet but much closer to the impingement characteristics of turbulent round jet impingement on an inclined concave
point that seen for the flat plate, forming at x/d = 1 2. With in- surface using liquid crystal transient method, Experimental Thermal and Fluid
Science 31 (2007) 559–565.
creases in H/d spacing, all the flow tends to reattach indicating a [7] K.B. Lim, C.H. Lee, N.W. Sung, S.H. Lee, An experimental study on the
stabilization of the flow with an increase in spacing. characteristics of heat transfer characteristics on the turbulent round jet
impingement according to the inclined angle of convex surface using the liquid
crystal transient method, Experimental Thermal and Fluid Science 31 (2007)
References 711–719.
[8] M. Mesbah, An experimental study of local heat transfer to an impinging jet on
[1] K. Jambunathan, E. Lai, M.A. Moss, B.L. Button, A review of heat transfer data non-flat surfaces: a cylindrical pedestal and a hemispherically concave surface,
for singular jet impingement, International Journal of Heat and Fluid Flow 13 Ph.D. Thesis, University of California, Davis, 1996.
(1992) 106–115. [9] S. Parneix, M. Behnia, P.A. Durbin, Predictions of turbulent heat transfer in an
[2] R. Viskanta, Heat transfer to impinging isothermal gas and flame jets, axisymmetric jet impinging on a heated pedestal, Journal of Heat Transfer 121
Experimental Thermal and Fluid Science 6 (1993) 111–134. (1999) 43–49.
[3] C.O. Popiel, O. Trass, Visualization of a free and impinging round jet, [10] A.S. Fleischer, S.R. Nejad, Jet impingement cooling of a discretely heated
Experimental Thermal and Fluid Science (1991) 253–264. portion of a protruding pedestal with a single round air jet, Experimental
[4] C. Cornaro, A.S. Fleischer, R.J. Goldstein, Flow visualization of a round jet Thermal and Fluid Science 8 (2004) 893–901.
impinging on cylindrical surfaces, Experimental Thermal and Fluid Science 20 [11] Y.S. Chung, Jet impingement cooling of chips equipped with cylindrical
(1999) 66–78. pedestal profile fins, Journal of Heat Transfer 127 (2005) 106–112.