Bhattacharya, S. (2010)

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

Experimental Thermal and Fluid Science 34 (2010) 633–637

Contents lists available at ScienceDirect

Experimental Thermal and Fluid Science


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

A note on unsteady impinging jet heat transfer


S. Bhattacharya *, A. Ahmed
Aerospace Engineering Department, Auburn University, Auburn, AL 36849, USA

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

Article history: Reported are the results of experiments on a regular in-line-jet that was modified prior to its impinge-
Received 14 July 2009 ment to improve heat transfer on the plate. Modifications consisted of insertion of two bluff bodies
Received in revised form 9 November 2009 and a streamlined body in the jet. The improvement in transport coefficients was attributed to the peri-
Accepted 14 December 2009
odicity in the flow which caused the attachment point to sweep the reattachment region and the turbu-
lence created was distributed over a large area compared to simple attachment.
Ó 2009 Elsevier Inc. All rights reserved.
Keywords:
Impinging jet
In line jet
Impinging jet heat transfer

1. Introduction the vicinity of the attachment point [11]. Popiel et al. [17] recom-
mended placement of the flat plate at the end of the potential core
Impinging jets have been studied over the years for their impor- region of the jet in order to attain maximum heat fluxes at the stag-
tance in industrial applications, mainly in cooling, heating or dry- nation point. Guo and Wood [8] made measurements along the
ing. The high levels of convective heat and mass transfer that can stagnation streamline of a plane jet impinging on a flat plate and
be achieved are the prime motivation for their continued use maintained the spacing in such a way so that the stagnation
[15,10,9,3]. Impingement of jets from a number of different nozzle streamline remained in the potential core of the jet. They found
geometries and configurations such as planar, single and multiple that in the region close to the plate the attenuation of streamwise
round and elliptical nozzles have been investigated to date velocity caused a large velocity gradient which contributed to in-
[4,1,6]. However single round jets are preferred due to symmetric creased level of normal stress. The spectral decomposition of the
boundary conditions and simplicity. Martin [16], Down and James velocity and pressure fluctuation showed that the increase in the
[5] and more recently Jambunathan et al. [12] provide an excellent level of energy production was mainly due to the low frequency
review of the heat transfer characteristics of circular impinging motion.
jets. A typical streamline pattern of an impinging jet without excita-
From the previous work it is evident that the important factors tion or flow modification is shown in Fig. 1. Since the heat transfer
that play a significant role in the enhancement of the transport is high in the vicinity of the stagnation point, two techniques have
properties of impinging jet are the Reynolds number, nozzle diam- been attempted in experiments to improve the performance of
eter, impingement distance i.e. spacing between the nozzle and the in-line jets. First by increasing the turbulence levels by pulsation
wall, and turbulence levels in the jet. Lee et al. [13] showed that lo- [2], excitation [14], or by introducing intermittency in the flow
cal Nusselt number increased in the stagnation point region with [18], and secondly by physically displacing the stagnation stream-
an increase in the nozzle diameter and attributed it to the increase line over a wider region [7].
in the jet momentum and turbulent intensity levels attained with This paper describes experimental investigation of an inexpen-
large nozzle diameter. In the case of annular jets it has been shown sive and simple method to improve convective heat transfer of an
that with decreasing distance between the nozzle and the plate, impinging jet. Unsteadiness in the jet was introduced by placing an
the velocity approaching the impingement surface increased eccentrically mounted round cylinder and an airfoil near the nozzle
resulting in higher heat transfer along the flow direction and in exit. As a result of the fluid–structure interactions both oscillated
freely about the longitudinal axis and imparted multiple scales to
the jet in addition to oscillatory dislocation of the stagnation
streamline from its mean trajectory. In addition a stationary round
cylinder was also tested. The heat transfer coefficients and the sur-
* Corresponding author. Tel.: +1 334 444 1674; fax: +1 334 844 6803.
E-mail addresses: bhattacharya.30@buckeyemail.osu.edu (S. Bhattacharya), face pressure distributions of all three were compared with the
aahmed@eng.auburn.edu (A. Ahmed). unperturbed in-line jet.

0894-1777/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.expthermflusci.2009.12.004
634 S. Bhattacharya, A. Ahmed / Experimental Thermal and Fluid Science 34 (2010) 633–637

2.1. Pressure measurement

An acrylic flat plate with a single pressure tap in the middle


was used for pressure measurement. The plate was mounted on
a computer controlled single degree of freedom traversing system
that traversed the plate±10 cm from the centerline of the jet. The
vertical spacing between the nozzle and the plate was 10 cm and
was held constant for all tests. The pressure tap was connected to
a calibrated differential pressure transducer. Analog signals from
the pressure transducer and traversing system were digitized
using an A/D board and sampled at 1500 Hz. At each position
5 s of data was recorded with a 10seconds of settling time. Pres-
sure transducer calibration was checked before and after each
test condition. From the datasheet provided by the manufacturer,
Fig. 1. Streamlines of an impinging jet. the accuracy of the pressure transducer was found to be 0.5% of
full scale output. The total relative standard uncertainty due to
transducer calibration and A/D board resolution error was below
2. Experimental setup 3%.

The test setup is presented in Fig. 2. The nozzle consisted of a 2.2. Temperature measurement
straight tube of 50 mm inside diameter and 15 cm in length con-
nected to a reservoir with flow conditioning screens and honey- Test setup for the heat transfer measurement is presented in
comb. A separate attachment was clamped near the exit of the Fig. 3. An Inconel 600 nickel alloy foil of 1 mil (25.4 lm) thickness
nozzle to hold the cylinder and airfoil without disturbing the jet. was held securely on a porous acrylic plate mounted on a plenum
Cylinder models were made from solid aluminum stock of nominal connected to a vacuum pump. With adequate seals provided at the
diameter of 1 cm. One cylinder had a mounting pin at the center edges a slight vacuum ensured flatness of the foil. Two linear roller
and the second cylinder had the pin located at 2.5 mm from the type electrodes were connected to a power supply for heating the
center allowing the cylinder to oscillate when positioned in the foil at a constant voltage and for energy input calculations. A ther-
jet. The streamlined airfoil consisted of a NACA 23012 airfoil sec- mal image of the unheated foil (with the airflow coming out of the
tion and had a chord length of 12 mm. It was equipped with the nozzle and impinging the foil) was taken to establish a base tem-
mounting pin that was located at 3 mm from the leading edge. Be- perature distribution. A power of 50 W was supplied from a power
cause of the camber of the airfoil the flow was not symmetric and source (HP model 6023A) with an average of 2 V and 25 ampere
as a result the airfoil oscillated when placed in jet. current. After the voltage was applied across the foil, it was al-
Tests consisted of measurement of mean surface pressures and lowed to reach steady state in approximately 20 min. Then a sec-
heat transfer on a flat plate during jet impingement for four cases ond thermal image of the heated foil with impinging jet was
consisting of (1) in-line jet, (2) in-line jet with stationary cylinder, recorded. Both the images were stored in the computer digitally
(3) in-line jet with eccentrically mounted oscillating cylinder, and and then the image of the unheated foil was subtracted from the
(4) in-line jet with oscillating cambered airfoil. A calibrated in-line image of the heated foil. Considering the foil material to be homog-
flow meter was used to determine the flow rate through the noz- enous (without any imperfection which promote formation of hot
zle. The mass flow rate was held constant at 0.061 kg/s that re- spots), the heat input per unit area can be taken as constant. Since
sulted in a jet exit velocity of 25 m/s and a corresponding the temperature change and energy per heated area is known, local
Reynolds number of 83,000 based on nozzle diameter. Exit velocity heat transfer coefficient can be calculated using the following
was also calculated from the difference between the stagnation relationship:
pressure and the atmospheric pressure and was found to be in
good agreement with the velocity calculated from the flow meter. q ¼ hðT h  T c Þ
Here h is the convection heat transfer coefficient and Th and Tc
refers to hot and cold temperatures of the foil.

Fig. 2. Setup. Fig. 3. Setup for temperature measurement.


S. Bhattacharya, A. Ahmed / Experimental Thermal and Fluid Science 34 (2010) 633–637 635

This process was repeated for each test configuration. The enough the stationary and oscillating cylinders show similar trends
uncertainty in the thermal imaging experiment was well within but that was expected as the amplitude of oscillations was ob-
5% for 15 image average. served to be well within the range of shear layer excursions. How-
ever the pressures registered for the oscillating cylinder are higher
2.3. Flow visualization and is attributed to the narrowing of the wake. In comparison the
cambered airfoil produced less velocity deficit because of the
Flow visualization experiments were conducted in a water tun- streamlining and hence the jet traversing the airfoil imparted high-
nel to understand the kinematics of the flow. An Argon Ion laser er momentum to the wall. Asymmetry in the pressure distribution
with a combination of cylindrical and spherical lenses was used is due to the camber of the airfoil. The lowest centerline pressure
to produce light sheet that illuminated the jet impinging on the was exhibited by the stationary cylinder. The oscillating cylinder,
tunnel floor. A Panasonic CCD camera and a VCR were used to ob- being free to rotate, allowed more fluid to be driven down toward
tain and record the data. Video records were played back frame by the plate directly below it than the stationary cylinder.
frame for qualitative analysis. The entire nozzle assembly was in- The image presented in Fig. 5 shows the temperature differen-
stalled in the water tunnel on a three axis traversing system. Water tial after the thermal image of the unheated foil with impinging
was supplied to the nozzles by a flexible hose and metered with jet was subtracted from the thermal image of the heated plate with
the help of a King Instruments rotameter. Average flow rate was impinging jet, taken in quiescent environment. Thus a lower tem-
maintained at 120 L/min that provided jet exit velocity of 1 m/s perature indicates higher heat transfer rate. Although the experi-
and a Re of 50,000. Because of the high flow rate, hydrogen bubbles mental setup prevented viewing of the region directly, it did not
and injection of fluorescent dye were not successful. However it significantly affect the extraction of important thermal parameters.
was noted that an adequate amount of air was present in the oscil- It is evident that the region of highest heat transfer exists directly
lating jet system for flow visualization. below the jet surrounded by a ring of lower heat transfer as the
flow moves outwards. Because of the high velocity, jet cannot turn
smoothly after impingement and consequently it separates and
3. Results and discussion reattaches resulting in the formation of secondary ring around 1
nozzle diameter from the centerline. This ring also signifies the
Surface pressures measured along the plane of symmetry of the presence of a separation bubble or recirculation region consisting
jet for four cases are presented in coefficient form in Fig. 4. Pres- of trapped eddies. Trapped eddies are a common feature however
sure coefficients were calculated using: their residence time in a ‘‘trapped position” varies for different
p  patm flows. Eventually they are ejected and replaced by other eddies.
Cp ¼ 1 In an instantaneous snap shot, these eddies do appear as trapped.
q U2
2 jet jet A separation bubble is a characteristic of impinging jets. They are
Here p  patm is the differential pressure referenced to atmo- formed as a consequence of impingement, separation and reattach-
spheric pressure and qjet and Ujet are density and exit velocity of ment of shear layer vortices on the plate .This feature was observed
the jet respectively. for all cases and confirmed during flow visualization tests.
As can be seen in Fig. 4 the spread of jet after impingement is In calculating the Nusselt number Nu ¼ hd k
, the local convective
confined to ±2 jet diameters from the centerline and becomes a heat transfer coefficient was determined for a given element of
wall jet that loses momentum rapidly as it moves outwards. In-line area A from:
jet produces a single high pressure peak directly underneath the jet
Q conv
and is symmetric. With the cylinder in place the pressure coeffi- hloc ¼ :
AðT plate  T jet Þ
cient plots shows two distinct peaks that are due to the impinge-
ment of shear layers from either sides of the cylinder with one For the average Nusselt number, the average heat transfer coef-
side dominating and is typical of bluff body wakes. Interestingly ficient then written as
R
R R0f    
2p 0 hloc Rr0 d Rr0
havg ¼  2
Rf
p R0

In which r/R0 is a dimensionless radius measured from the jet


centerline and Rf/R0 is the dimensionless radius over which the
integration is performed. The value of havg at a given radial location
Rf is the effective heat transfer coefficient for the entire disk from
r = 0 to r = Rf.

Fig. 4. Mean surface pressure distribution. Fig. 5. DT distribution for oscillating jet.
636 S. Bhattacharya, A. Ahmed / Experimental Thermal and Fluid Science 34 (2010) 633–637

selt number corresponding to the low heat transfer is discussed


earlier in conjunction with the Fig. 5.
Results of oscillating cylinder can be explained in terms of the
wake vorticity dynamics. Both configurations produced Karman
vortices however in the case of oscillating cylinder; the stagnation
streamlines moved a finite distance away from the centerline. The
overall heat transfer therefore is a result of higher turbulent mixing
due to a combination of shed vortices and excursions of attach-
ment flow in the stagnation region. Airfoil wakes have concen-
trated vorticity near the trailing edge and for low angle of
incidences there are no large vortices. Results however show that
the oscillating airfoil had the highest heat transfer. This is attrib-
uted to the curvature of streamlines in the wake due to the airfoil
camber. Consequently, the attachment streamlines were displaced
over a larger area and hence the higher heat transfer.
Limited flow visualization experiments were pursued to quali-
tatively determine the flow properties that governed the heat
transfer characteristics. The oscillating cylinder and the stationary
cylinder showed similar flow pattern however the Karman vortices
were more stable in the stationary cylinder wake. Because of the
presence of wall, shear layers on either sides of the cylinder were
Fig. 6. Distribution of local Nusselt number.
compressed in way that reduced the spacing between the Karman
vortices and shortened the vortex formation region. A topological
cartoon of the flow presented in Fig. 8 shows the presence of saddle
points between the vortices and the stagnation point in itself is a
Fig. 6 show the variation of local Nusselt number in the radial
half saddle of attachment. A considerable amount of recirculation
direction starting from the region near the geometric centerline
was also observed. As the flow turned outwards, vorticity in the
of the impinging jet. Note that a slight displacement of curves from
shear layer and the boundary layer being of opposite sign helped
r/d = 0 is primarily due to the experimental setup that blocked the
thermal imager view of the plate directly under the jet. The in-line
jets shows a rapid drop of Nusselt number away from the stagna-
tion point followed by another increase around r/d = 2 and tapering
off beyond that point. Similar characteristics of outer flow are also
visible for other cases as well and are attributed to the interaction
between the shear layer and the boundary layer on the plate as the
flow moves outwards as a decaying wall jet. In contrast, all other
impinging jets produced a gradual drop in Nusselt number and
higher heat transfer compared to the in-line jet. These trends also
apply to the average Nusselt number plot presented in Fig. 7. The
oscillating jets show improvements over the in-line jet by as much
as 20% for the configuration tested, with the oscillating airfoil
showing the highest heat transfer coefficients. Also, a drop in Nus-

Fig. 8. Flow pattern with stationary/oscillating cylinder.

Fig. 7. Distribution of average Nusselt number. Fig. 9. Flow pattern with oscillating airfoil.
S. Bhattacharya, A. Ahmed / Experimental Thermal and Fluid Science 34 (2010) 633–637 637

diffuse the flow more rapidly as evident from the relaxation of sur- [2] L.F.A. Azevedo, B.W. Webb, M. Queiroz, Pulsed air jet impingement heat
transfer, Experimental Thermal and Fluid Science 8 (1994) 206–213.
face pressures in Fig. 4.
[3] N. Celik, H. Eren, Heat transfer due to impinging co-axial jets and the jets’ fluid
In the case of the oscillating airfoil, a single stable vertical struc- flow characteristics, Experimental Thermal and Fluid Science 33 (2009) 715–
ture was observed (Fig. 9). This vortex was spatially not very stable 727.
and was driven outward by the flow soon after its formation. The [4] D.W. Colucci, R. Viskanta, Effect of nozzle geometry on local convective heat
transfer to a confined impinging air jet, Experimental Thermal and Fluid
stagnation point (again a half saddle point) was not stationary, Science 13 (1996) 71–80.
and no trapped recirculating eddy was observed. The last two [5] S.J. Down, E.H. James, Jet Impinging Heat Transfer – A Literature Survey. ASME
observations explain the high heat transfer coefficient of the Paper No. 87-H-35, 1987.
[6] J. Goldstein, A.I. Behbahani, K.K. Heppelmann, Streamwise distribution of the
impinging jet with airfoil induced oscillations. The fact that the air- recovery factor and the local heat transfer coefficient of an impinging circular
foil had only a small effect on the pressure distribution but higher air jet, International Journal of Heat and Mass Transfer 29 (8) (1986) 1227–
heat transfer characteristics after impingement indicate that this 1235.
[7] S. Goppert, T. Gurtler, H. Mocikat, H. Herwig, Heat transfer under a processing
was primarily due to the oscillatory dislocation of the stagnation jet: effects of unsteady jet impingement, International Journal of Heat and
streamline. Mass Transfer 47 (2003) 2795–2806.
[8] Y. Guo, D.H. Wood, Measurements in the vicinity of a stagnation point,
Experimental Thermal and Fluid Science 25 (2002) 605–614.
4. Conclusion [9] B.R. Hollworth, M. Durbin, Impingement cooling of electronics, Journal of Heat
Transfer 14 (1992) 607–613.
A simple method of improving the transport properties of in- [10] K. Ichimiya, K. Kobayashi, R. Echigo, Fundamental study on the flattening of
temperature distribution of a high temperature steel slab (3rd report, heat
line jets by inducing oscillations in the flow was tested. Three con- transfer characteristics due to a confined turbulent impinging jet flow),
figurations were used for this purpose, a stationary cylinder, an Transactions of JSME 53 (1987) 511–515.
eccentrically mounted cylinder and an airfoil subjected to self in- [11] K. Ichimiya, Heat Transfer characteristics of an annular turbulent impinging jet
with a confined wall measured by thermo sensitive liquid crystal, Heat and
duced oscillations. The ensuing oscillating jet configurations Mass Transfer 39 (2003) 545–551.
showed higher heat transfer rates than in-line jet tested under [12] K. Jambunathan, E. Lai, M.A. Moss, B.L. Button, A review of heat transfer data
the similar conditions. The improved heat transfer characteristics for single circular jet impingement, Journal of Heat and Fluid Flow 13 (2)
(1992) 106–114.
of the impinging jet were attributed to two mechanisms consisting [13] D.H. Lee, J. Song, M.C. Jo, The effects of nozzle diameter on impinging jet heat
of enhanced mixing/turbulence due to bluff body wakes vortex transfer and fluid flow, Journal of Heat Transfer 126 (4) (2004) 554–557.
dynamics and oscillatory dislocation of the stagnation streamline. [14] T. Liu, J.P. Sullivan, Heat transfer and flow structures in an excited circular
impinging jet, International Journal of Heat and Mass Transfer 39 (1996)
The oscillating airfoil showed the greatest heat transfer rates of 3695–3706.
the four configurations tested. The airfoil produced the greatest [15] E. Loo, A.S. Majumdar, A simulation model for combined impingement and
dislocation of the stagnation point on the plate, and it had a min- through drying using superheated steam as the drying medium Drying’84,
Hemisphere, Washington, 1984, pp. 264–280.
imal amount of recirculation directly under the airfoil. This simple
[16] H. Martin, Heat and mass transfer between impinging gas jets and solid
technique has promise in mixing and drying applications with surfaces, Advances in Heat Transfer, Academic Press, New York, 1977. pp. 1–
minimum external energy input. 60.
[17] C.O. Popiel, T.H.J. Van der Meer, C.J. Hoogendoorn, Convective heat
transfer on a plate in an impinging round hot gas jet of low Reynolds
References Number, International Journal of Heat and Mass Transfer 23 (1980) 1055–
1068.
[1] S.C. Arjocu, J.A. Liburdy, Identification of dominant heat transfer modes [18] D.A. Zumbrunnen, M. Aziz, Convective heat transfer enhancement due
associated with the impingement of an elliptical jet array, Journal of Heat to intermittency in an impinging jet, Journal of Heat Transfer 115 (2009)
Transfer 122 (2000) 240–247. 91–98.

You might also like