International Journal of Heat and Mass Transfer 47 (2004) 1827–1845

www.elsevier.com/locate/ijhmt

An experimental study of fluid mechanics and heat transfer

in an impinging slot jet flow
a,*
V. Narayanan , J. Seyed-Yagoobi b, R.H. Page c

a
Department of Mechanical Engineering, Oregon State University, 204 Rogers Hall, Corvallis, OR 97331-6001, USA
b
Department of Mechanical, Materials, and Aerospace Engineering, Illinois Institute of Technology, 10 W. 32nd Street,
Chicago, IL 60616, USA
c
Department of Mechanical Engineering, Texas A&M University, MS 3123, College Station, TX 77843-3123, USA
Received 4 September 2002; received in revised form 12 October 2003

Abstract
An experimental study of flow field, surface pressure, and heat transfer rates of a submerged, turbulent, slot jet
impinging normally on a flat plate is presented. Two nozzle-to-surface spacings of 3.5 and 0.5 nozzle exit hydraulic
diameters, which correspond to transitional and potential-core jet impingement, respectively, are considered. Fluid
mechanical data include measurements of mean flow field and variance of normal and cross velocity fluctuations, mean
surface pressure, and RMS surface pressure fluctuations along the nozzle minor axis. Local heat transfer coefficients are
calculated from detailed surface temperature measurements. The heat transfer data follow a trend similar to previous
studies, exhibiting high heat transfer rates in the impingement region for transitional jet impingement, and a non-
monotonic decay in heat transfer coefficient for potential-core jet impingement. The fluid flow results indicate that past
impingement, locations of high streamwise fluctuating velocity variance occur in the wall jet flow for both nozzle
spacings. The RMS surface pressure fluctuation profile exhibits a maximum at the impingement line for the transitional
jet impingement, and corresponds well with the rise in near-wall velocity fluctuation variance in the free jet prior to
impingement. For potential-core jet impingement, the streamwise location of peak RMS pressure fluctuations corre-
sponds to the highly correlated turbulence in the outer region of the wall-bounded flow. Further, there is a good
correspondence between the locations of secondary peak in heat transfer and near-wall streamwise fluctuating velocity
variance. The occurrence of the outer peak prior to the near-wall peak in streamwise velocity variance for the potential-
core jet impingement suggests that the interaction between correlated motion in the outer region and near-wall tur-
bulence causes the rise in heat transfer coefficient towards a secondary peak.

2003 Elsevier Ltd. All rights reserved.

1. Introduction are some of the key factors that have lead to the sustained
and widespread use of jet impingement technologies.
Impinging jets are used in a wide variety of applica- The most commonly used geometries are axisymmetric
tions such as cooling of electronics and turbine blades, (circular orifice or pipe) and slot (two-dimensional)
and in the heating, cooling, or drying of pulp, paper, nozzles.
textile, food, and chemicals. The ability to control heat The flow field of an impinging jet can be divided into
transfer from the surface by varying flow parameters three zones: (1) the free jet prior to impingement, (2) the
such as jet exit velocity and flow temperature, and geo- impingement region, and (3) the wall jet region. The free
metrical parameters such as jet exit opening, nozzle-to- jet region consists of one or more of the following zones
surface spacing, and nozzle-to-nozzle spacing in arrays, depending on the nozzle-to-surface spacing: (1) the po-
tential core, where the jet maintains its exit velocity at
the centerline while mixing and diffusing with the
* ambient fluid in its shear layers; (2) the transition zone,
Corresponding author.
E-mail address: [email protected] (V. Narayanan). where the core velocity has decayed completely, and the
0017-9310/$ - see front matter
2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijheatmasstransfer.2003.10.029
1828 V. Narayanan et al. / International Journal of Heat and Mass Transfer 47 (2004) 1827–1845

Nomenclature

bo free jet virtual origin (m) x streamwise coordinate along the minor axis
DPm
DPm;ref of the nozzle; x ¼ 0:0 corresponds to the
Cp pressure coefficient, Cp ¼
0:5qVmo 2 projected nozzle centerline on the impinge-
Dh hydraulic diameter of the nozzle ¼ 2Xws for ment surface (m)
slot jet (m) Yn nozzle-to-surface spacing along the y axis
h heat transfer coefficient (W/m2 K), hloc ¼ (m)
qelec
qCu y normal (vertical) co-ordinate; y ¼ 0:0 loca-

ðq00rad þ q00nc Þ
q00net Afoil tion corresponds to the impingement sur-
¼
Th
Tad Th
Tad face, unless otherwise mentioned (m)
k thermal conductivity (W/m K) z spanwise coordinate along the major axis of
L foil length (m) the nozzle; z ¼ 0:0 location corresponds to
m_ mass flow rate (kg/s) the projected nozzle centerline on the
hloc Dh
Nu Nusselt number, Nu ¼ impingement surface minor axis (m)
k
P pressure (Pa)
q heat energy rate (W) Greek symbols
q00 heat energy flux (W/m2 ) D differential
r radial coordinate for a circular (axisym- g non-dimensional geometrical scaling
x
metric) jet (m) parameter for a free jet, g ¼
Vmo Dh y
bo
Re Reynolds number, Re ¼ q density (kg/m3 )
m
T temperature (C)
U time-averaged (mean) velocity in the x- Subscripts
direction (m/s) ad adiabatic
uRMS RMS velocity fluctuation in the x-direction atm atmospheric
(m/s) Cu copper bus bars (heat loss)
huui variance of normal velocity fluctuation in elec electrical
x-direction (m/s)2 h heated
huvi variance of cross velocity fluctuation in the loc local
x–y plane (m/s)2 m mean value (volume or mass averaged)
V time-averaged (mean) velocity in the max maximum
y-direction (m/s) nc natural convection
Vmax time-averaged (mean) velocity along the net net value of quantity
nozzle centerline at a distance y from o nozzle exit
the nozzle exit (with reference to Fig. 3) rad radiation
(m/s) ref reference value of quantity
Vmo time-averaged (mean) nozzle exit velocity in RMS root mean square
the y-direction based on flow rate (m/s) u along u velocity direction
vRMS RMS velocity fluctuation in the y-direction v along v velocity direction
(m/s) w along w velocity direction
hvvi variance of normal velocity fluctuation in x x-direction
the y-direction (m/s)2 y y-direction
Xws slot width (m) z z-direction

shear layers interact at the jet centerline; and (3) the fully example, [2,3]) have determined that the maximum heat
developed zone, where the jet attains a self-similar transfer in the impingement region occurs when the
behavior. Fig. 1a shows a schematic of a slot jet with the transitional jet impinges on the surface.
transitional region of the free jet impinging normal to Table 1 summarizes the experimental studies in a
the surface, while Fig. 1b shows the case where the jet turbulent, subsonic, submerged, single slot jet heat
potential core impinges on the surface. Typically, the transfer and/or fluid mechanics. Several experimental
length of the potential core varies between 5 and 6 slot studies have focused on either the fluid mechanics or
widths from the jet exit, depending on the exit condi- heat transfer of impinging axisymmetric or slot jets, but
tions and nozzle geometry [1,2]. Previous studies (for very few have reported on both aspects. Donaldson and
 V. Narayanan et al. / International Journal of Heat and Mass Transfer 47 (2004) 1827–1845 1829

Fig. 1. Schematic of slot jet impingement on a flat surface (a) transitional jet impingement, and (b) potential-core jet impingement.

Snedeker [4], Donaldson et al. [5], Hoogendoorn [6], and simultaneous flow field and temperature measurements
Lytle and Webb [7] studied flow field and heat transfer in of a slot jet impinging at a height of 7 slot widths from
axisymmetric jets. Donaldson et al. provided fluid the surface using particle image velocimetry and laser
mechanical and heat transfer data only for relatively induced fluorescence techniques.
large nozzle-to-surface spacings (3:5 < Dh < 15), while Many prior studies have reported a non-monotonic
Lytle and Webb studied impingement at Yn =D < 0:5. trend in local heat transfer coefficient distribution in
Gardon and Akfirat [2] reported streamwise turbulence both axisymmetric and two-dimensional jets at low
intensity in the absence of the impingement surface, nozzle-to-surface spacings, (Yn =Dh < 2:5) when the po-
along with extensive heat transfer data for impinging tential core of the free jet impinges on the surface. In
slot jets. Yokobori et al. [3,8] studied the behavior of circular jet impingement flow at very close nozzle-to-
turbulence at the impingement point of two-dimensional surface spacings, Hoogendoorn [6] and Lytle and Webb
and axisymmetric jets, respectively. They confirmed, for [7] showed that the heat transfer coefficient distribution
transitional jet impingement, the existence of pairs of exhibited two peaks away from the stagnation region.
counter-rotating, large-scale, streamwise vortices at The primary peak in the heat transfer distribution oc-
stable nodal locations along the impingement line and curred near the stagnation point, at r=Dh ¼ 0:50, and
discussed their relation with the behavior of upstream was attributed to local thinning of the accelerating
shear layer vortices. These coherent structures, they laminar boundary layer. Hoogendoorn [6] showed that
emphasized, were predominant in the transport of heat the radial distribution of local streamwise turbulence
in the impingement region. Sakakibara et al. [9] reported intensity, measured at the location of maximum radial
 1830
Table 1
Summary of experimental fluid mechanic and heat transfer studies on submerged turbulent single slot jets
Investigation Nozzle geometry Impingement surface, Yn =Dh Reynolds Measured/calculated Techniques used
(aspect ratio) jet incidence number quantities
Becko [31] Literature survey––heat
transfer, emphasis on
turbine blade cooling

V. Narayanan et al. / International Journal of Heat and Mass Transfer 47 (2004) 1827–1845
Martin [1] Literature survey––mainly
heat transfer
Hrycak [32] Literature survey
Looney and Walsh [33] Literature survey––fluid
mechanics, numerical study
Downs and James [34] Literature survey––heat
transfer
Polat [35] Literature survey––process
applications
Viskanta [13] Literature survey
Gardon and Akfirat [2,36] 96, 48, 24 Flat, normal 0.25–30 450–50,000 Heat transfer Heat flux sensor
Korger and Krizek [14] Xws ¼ 5–40 mm Flat, normal 0.13–20 6040–37,800 Mass transfer Naphthalene sublimation
Schlunder et al. (Martin [1]) Flat, normal 2.5–10 22,000–93,600 Mass transfer Evaporation of water
Belatos and Rajaratnam [37] 65 Flat, normal 14.04–67.5 5270–9400 U, V , wall pressure Preston probe
Gutmark et al. [38] 38.46 Flat, normal 50 30,000 U, V , W , normal and cross Cross wire anemometer
Reynolds stress, and
turbulence triple products
Yokobori et al. [3] 10, 30 Flat, normal 1–7.5 2000–20,000 Surface temperature, Thermocouple, hydrogen
stagnation vortex pair bubble flow viz.
wavelength and frequency
Striegl and Diller [39] 56.7, 34, 17, 11.3 Flat, normal 3–12.5 3000–40,000 Heat transfer Heat flux gage
Kataoka et al. [40] 25 Flat, normal 3000–6000 Impingement region flow Hydrogen bubble flow
visualization, heat transfer, viz., liquid crystal, two
and turbulence intensity single film sensor
anemometry
Kataoka [41] Flat, normal Summary of work in flow Hydrogen bubble flow
structures, turbulence, and viz., thermocouple,
heat transfer in the anemometer
impingement region
Gau and Chung [42] 7.5, 10, 15 Convex and concave, 1–8 6000–350,000 Flow visualization, heat Smoke-wire viz.,
normal transfer thermocouple
Gau and Lee [21] 7.5, 12.5, 25, 43 Ribs, normal 1–8 2500–11,000 Flow visualization, heat Smoke-wire viz.,
transfer thermocouple
Tu and Wood [27] 61–400 Flat, normal 1–20.6 3040–11,000 Mean wall surface pressure Stanton tube, Preston tube
and shear stress
Sakakibara et al. [9] 10.35 Flat, normal 4 4000 Simultaneous, instantaneous DPIV, LIF
U , V , Tflow -focus on stagna-
tion region flow structure
 V. Narayanan et al. / International Journal of Heat and Mass Transfer 47 (2004) 1827–1845 1831

1-D LDA, thermocouple, velocity, and Nusselt number attained a minimum value

Four holed fast response

1-D LDA, piezoresistive

at approximately the same radial location for nozzle-to-
surface spacings of 2, 4, and 5.9 hydraulic diameters.

Near-wall hot-wire
The minimum in heat transfer coefficient, observed for

2-D LDA, PIV

pressure probe

transducer, IR
Thermocouple

Thermocouple

thermography
Yn =Dh ¼ 2 and 4, was attributed to a transition from a
anemometry laminar to turbulent boundary layer. The secondary
peak following this minimum was observed only for
DPIV

cases when a low-turbulence contoured nozzle was used

as opposed to a pipe nozzle. Lytle and Webb [7] reported
instantaneous, statistical and
surface averaged heat trans-

stresses, mean and fluctuat- the existence of a secondary peak for a pipe nozzle with

U, V , huui, huvi, huvi, P , p,

ing pressure, hupi correla-
instantaneous planar flow

Stagnation region normal

a jet turbulence level of around 10% near the impinge-

tion; mean shear stress
phase averaged planar

U, V , huui, mean skin

U, V , huui, huvi, huvi,

ment region. The location and magnitude of the sec-

U, huui, heat transfer

ondary peak in heat transfer depended on the Reynolds

heat transfer, flow

number, nozzle-to-surface spacing, and nozzle geome-
Heat transfer

visualization
try. Goldstein et al. [10] attributed the local minimum in
the recovery factor distributions at low nozzle spacings
velocity

friction

profile

to energy separation in the vortex flow. Popiel and Trass

field
fer

[11] performed flow visualization studies on a circular jet

impinging at close nozzle-to-surface spacings to quali-
1780, 2960, 4740
600–8000 (based

tatively document the flow behavior. Fox et al. [12]

20,000–60,000

6700–110,000
on diameter)

studied the influence of vortex structures on the adia-

4000–12,000

batic wall temperature distribution in an impinging jet.

23,000
88,000

They showed, by means of an analytical model and

4000

experiments, that the wall temperature was modified by

the presence of secondary vortex structures that were
induced near the plate surface by primary vortices in the
outer region. They attributed the region of lower wall
1, 2, 2.5
4, 6, 10

0.5, 3.5
0.5–5.5

temperature at low nozzle-to-surface spacing to second-

1–4.6
4–12

5–25

ary vortices. Summarizing results from prior investiga-

4

tions, Viskanta [13] noted that there was considerable

speculation and disagreement among researchers as to
the explanation behind the secondary peak in heat
transfer coefficient. However, in general, the non-
Concave, normal
Cylinder, normal

monotonic trend in radial distribution of heat transfer

Flat, oblique

Flat, normal

Flat, normal

Flat, normal

Flat, normal

Flat, normal

coefficients in circular jet impingement at close nozzle-

to-surface spacing was attributed to an accelerated
laminar boundary layer in the vicinity of the impinge-
ment point (primary peak) and the interaction of the
large-scale turbulence generated in the mixing layer, and
transition to developed turbulent radial wall jet (sec-
(fluted nozzle)

ondary peak).
Gardon and Akfirat [2] and Korger and Krizek [14]
8, 12, 24

17.5–70

studied the local heat and mass transfer distributions

10.35

5.77

respectively, for potential-core slot impingement at close

9.1
30

15

20

nozzle-to-surface spacings. Gardon and Akfirat ob-

served a secondary peak in heat transfer coefficient dis-
tribution for slot jet impingement at a nozzle-to-surface
McDaniel and Webb [43]

spacing between 1 and 3 hydraulic diameters, and

Maurel and Solliec [46]

attributed the peak to a transition from a laminar to a

Sakakibara et al. [28]

Guo and Wood [24]

Bietelmal et al. [44]

Zhe and Modi [15]

turbulent boundary layer. They also reported a local

minimum at the stagnation line, surrounded by primary
Choi et al. [45]

Present study

peaks on either side for a very close spacing of

Yn =Dh ¼ 0:33. They proposed that the transition to
turbulent flow was triggered by the disappearance of the
favorable pressure gradient which existed in the accel-
erating flow past impingement. The transition resulted in
1832 V. Narayanan et al. / International Journal of Heat and Mass Transfer 47 (2004) 1827–1845

a rise in local heat transfer coefficients at a x=Dh ¼ 2. was attached to the central slot in the plenum, and all
The rising trend in heat transfer coefficient persisted other slots were sealed. The geometry of the nozzle used
until the boundary layer thickened to a turbulent profile, in this study is provided in Fig. 2b. The nozzle width and
along with a decrease in mean flow velocity due to jet length were 0.0127 m and 0.102 m, respectively. The
spread. These two opposing effects, they speculated, nozzle aspect ratio was 20:1 excluding the unconfined,
resulted in a secondary peak in heat transfer at a semi-circular ends. A shroud of width 1:20Dh on either
x=Dh ¼ 3:5, beyond which heat transfer decreased side from the nozzle centerline was located flush with the
monotonically. They observed that the non-monotonic nozzle exit, as shown in Figs. 1 and 2b. Preliminary
variation of heat transfer coefficient was suppressed in experiments, performed in the absence of the impinge-
the case of high exit jet turbulence intensity. More re- ment plate, indicated that the exit mean velocity along
cently, Zhe and Modi [15] measured mean velocity and the major axis of the nozzle, x ¼ 0:0, was uniform up to a
turbulence intensity in the near-wall jet flow, and esti- distance of z=Dh ¼ 3:5. In this paper, all flow
mated mean wall shear stress on the impingement sur- field, surface pressure, and heat transfer results are re-
face. They observed a secondary peak in mean skin ported along the z=Dh ¼ 0:0 plane. Two interchangeable
friction around x=Dh ¼ 2:5 for three nozzle-to-surface impingement plates were used in these experiments: (a) a
spacings of Yn =Dh ¼ 1, 1.5, and 2. Their study did not Plexiglas surface of dimensions 63.5 cm · 73.2 cm for
include local heat transfer measurements, thereby pre- flow field and surface pressure measurements, and (b)
cluding a discussion on the effect of the wall shear stress a heated inconel impingement surface of dimensions
on heat transfer distribution. 38.1 cm  38.1 cm for heat transfer measurements.
The primary aim of this paper is to present results Mean flow field and turbulence intensities were
of a combined flow field and heat transfer study of a measured using a 1-D fiber optic laser Doppler ane-
turbulent, submerged, impinging slot jet at the two mometer (LDA) system (TSI Inc.), with collection optics
nozzle-to-surface spacings of interest as highlighted in arranged in a backscatter configuration. The probe was
literature: a spacing of 3:5Dh (7.0 jet exit widths), which traversed in three dimensions by means of an accurate
corresponds to transitional jet impingement, and a servo motor actuated traverse system with position
spacing of 0:50Dh (1.0 jet exit widths), which corre- encoders. A 514.5 nm, 150 mW Argon-ion laser pro-
sponds to near-wall potential-core impingement on the vided the coherent light source. The optic train consisted
surface. The exit Reynolds number was 23,400 for flow of a beam splitter, 40 MHz Bragg cell, fiber optic beam
field measurements, and 22,500 for heat transfer mea- launchers, and the transmitting lens within the fiber
surements for both nozzle spacings. The paper provides optic probe. The detected light was focused onto a
an insight into the distinctly different flow field and heat photo-detector at the end of the fiber optic. The elec-
transport mechanisms that occur during impingement of tronics consisted of a frequency downmixer and a signal
the jet transitional and potential-core regions. The non- processor (TSI Inc., model IFA550). Data acquisition
monotonic trend in heat transfer at low spacings is ex- and probe traverse was controlled by means of a com-
plained based on the flow field, surface pressure, and puter located outside the test chamber. Light-scattering
transient surface temperature results. Further details on particles for velocity measurement were generated by a
transient surface temperature measurements are pre- mixture of glycol and water, drawn into the air stream
sented elsewhere [16]. due to suction created by an ejector nozzle located in the
piping upstream of the plenum entrance. The average
size of the particles at the nozzle exit was estimated to be
2. Experimental apparatus and instrumentation between 3 and 5 lm in diameter.
Local surface temperatures were measured non-
Fig. 2a shows a schematic of the jet impingement set- intrusively with an 8–13 lm wavelength infrared (IR)
up used in this study. Two dedicated screw compressors camera (Mikron Inc., model 6T62 thermotracer). The
located in a separate room supplied compressed air for temperature resolution of the camera in the normal
the jet flow. The air was dried, filtered, and directed into mode of operation was 0.1 C, and as high as 0.025 C in
a plenum through two settling tanks and pressure reg- the high signal-to-noise improvement mode. The camera
ulators. The mass flow rate was determined using a sonic optics in front of the single cooled detector scanned the
nozzle at the exit of the last settling tank. The plenum field of view for temperatures over 256 · 206 spatial
consisted of a honeycomb flow straightener, followed by locations. The camera was placed in a vented enclosure
a series of screens and a convergent section. The jet directly below the foil (see Fig. 2a). The temperature rise
impingement set-up, shown in Fig. 2a, was located in a in the enclosure did not exceed 1 C during the test
large test chamber in order to prevent stray light and duration of approximately 5 h. The heat transfer
other disturbances from affecting the measurements. The impingement surface consisted of a central constant heat
flow exited the plenum through an end plate consisting flux surface surrounded by Plexiglas end plates on all
of seven contoured slots of aspect ratio 20:1. The nozzle sides. The entire test section was 63.5 cm long  63.5 cm
 V. Narayanan et al. / International Journal of Heat and Mass Transfer 47 (2004) 1827–1845 1833

Flow inlet
coarse screen plenum section 1

honeycomb plenum section 2

fine screen
fine screen
plenum
convergence
fine screen section

plenum section 4

interchangeable
nozzle plate
impingement
slot nozzle plate

IR camera IR camera
enclosure

vertical
adjustment
jack
(a)

(b)

Fig. 2. Schematics of the (a) jet impingement facility and (b) nozzle geometry. All dimensions are in mm.

wide. The central heat transfer test section consisted of a calibration of the IR camera with respect to a calibrated
0.0254 mm thick Inconel 600 foil 38.1 cm long · 38.1 cm thermistor located on the foil surface. This value was not
wide, stretched tightly between two solid copper bus expected to change considerably with temperature, since
bars. The foil was electrically heated by means of a high the foil was heated to less than 15 C above ambient
current DC power supply to provide a constant heat flux temperature.
surface. Solid copper rods were used to transmit power Mean and fluctuating wall pressures were measured
to the bus bars. Voltage difference was measured at the using a piezoresistive differential pressure transducer
bus bars, and the electrical contact resistance between (Kulite Semiconductor Products, Inc., model XCS062).
the foil and bus bars was considered while determining The transducer had a pressure sensitive diameter of 0.71
the net voltage drop across the foil. Current through the mm, and was mounted flush with the surface. The
foil was measured at the power supply readout. The transducer output signal was amplified and low-pass
underside of the foil and the surrounding Plexiglas sur- filtered at 20 kHz, using a strain gage amplifier module
face was coated with flat black paint to provide a con- (Encore Electronics, model 619M). For mean pressure
stant high emissivity, diffuse surface for radiation measurements, the output voltage from the amplifier
detection. The emissivity of the painted foil surface was was directly read by a high-resolution digital multimeter
determined to be 0.96, based on ambient temperature (Hewlett Packard, model 3478A) and recorded through
1834 V. Narayanan et al. / International Journal of Heat and Mass Transfer 47 (2004) 1827–1845

a GPIB interface onto a PC. For RMS pressure mea- foil was measured through the voltage and current
surements, the voltage data was high-pass filtered at 0.01 readings, and the known heated foil area. Heat losses
Hz. A 12-bit data acquisition board (National Instru- due to radiation and natural convection from both the
ments, model 6024E) was used to further amplify and upper and lower surfaces of the foil, and end conduction
digitize the signal. In order to obtain pressure profiles, to the copper bus bars were taken into consideration
the surface (and correspondingly, the sensor) was tra- while calculating the net heat flux at each pixel location.
versed along the x-direction by means of a computer For the radiation loss calculation from the upper side of
controlled traverse system. the foil, a constant emissivity of 0.10 was used from
literature [19], based on the normal total emissivity value
for a 0.0254 mm polished Inconel foil at 300 K. The
3. Data collection and analysis procedure local heat transfer coefficient was determined from the
net heat flux and the local temperature difference be-
LDA measurements were obtained with three differ- tween the heated and unheated foil.
ent rotary orientations of the probe in order to estimate The surface pressure experiments consisted of tra-
the mean velocity and streamwise, normal, and cross versing the impingement surface with the piezoresistive
fluctuating velocity variance at each location using a transducer along the minor axis of the nozzle,
component resolution technique [17,18]. Preliminary z=Dh ¼ 0:0. At each location, dc voltage readings for the
experiments indicated that the flow measurements mean pressure, followed by ac voltage readings of the
remained unchanged between experimental runs. An pressure fluctuations were obtained. This procedure
estimated repeatability in the measurement volume eliminated any position-related uncertainty between the
coincidence to within ±0.1, ±0.1, and ±0.25 mm, in the mean and RMS fluctuating pressure data. An average of
x, y, and z directions respectively, between the three 300 samples of dc voltage readings from the voltmeter
probe rotary orientations was achieved by consistent was used to determine the mean pressure differential.
alignment procedures. Five thousand data points were For the fluctuating component, two million data points
recorded at each location in the flow field, and analyzed were recorded through the data acquisition card at a
to obtain mean flow and turbulence quantities. A tran- sampling rate of 60 kHz. To prevent aliasing effects, the
sit-time-weighting technique was used to eliminate signal was low-pass filtered at 20 kHz in the amplifier
velocity bias. module prior to digitization of the signal.
During a typical heat transfer experiment, eight Uncertainties in reported data were calculated based
consecutive temperature maps of the unheated foil with on propagation of errors method [20] and are presented
jet impingement were obtained at steady state, and in Table 2. The velocity profile at the slot exit was
averaged in order to estimate the mean local adiabatic integrated to obtain a mass balance. The mass flow rate
foil temperatures. The foil was then heated and the was found to agree to within 3% of the flow rate cal-
temperature distribution measured after steady state culated by the flowmeters. Table 3 provides a compari-
temperature was attained. Again, eight consecutive son of local Nusselt numbers, Nuloc , determined in the
frames were averaged to obtain a mean local heated present study with data from literature for a slot jet
temperature distribution on the foil. The heat flux to the under similar Reynolds numbers and nozzle spacings.

Table 2
Uncertainty estimates for fluid mechanics and heat transfer data
Variable Percent uncertainty
Mass flow rate, m_ (kg/s) 1.5%
Mean velocity, U (m/s) (calculated) 2.7% of local velocity
Mean velocity, V (m/s) (measured) 2.7% of local velocity
Fluctuating velocity variance, huui (m2 /s2 ) (calculated) Maximum of 15% of local huui, typically less than 10% in near-
wall regions, and less than 5% at y=Dh > 0:15
Fluctuating velocity variance, hvvi (m2 /s2 ) (measured) Maximum of 13.7% of local hvvi, typically <10% near the wall in
the high-velocity regions, and less than 2% at y=Dh > 0:15
Fluctuating velocity variance,
huvi (m2 /s2 ) (calculated) Typically less than 20% of local huvi for near-wall measurements,
and less than 10% at y=Dh > 0:15
Pressure coefficient 8.7% for Cp;max
RMS surface fluctuating pressure (Pa) 10% of DPRMS;max
Heat transfer coefficient, Nuloc maximum of 7.1% of Nuloc , with an average of 5.0% of Nuloc
Pixel location uncertainty in temperature measurements 1.5% of 1 pixel value
 V. Narayanan et al. / International Journal of Heat and Mass Transfer 47 (2004) 1827–1845 1835

Table 3
Comparison of Nuloc between present experiments and literature
x=Dh 0 1 2 3 4 5
a
1. Present data 145 101.2 76.3 69.5 64.2 57
2. Schlunder et al. (Martin [1])b 145 89 72.5 62 56 51.5
% difference (1)2)/[(1+2)/2] 0.0 12.9 5.2 11.4 13.7 10.1
a
Re ¼ 22; 500, Yn =Dh ¼ 3:5.
b
Re ¼ 22; 000, Yn =Dh ¼ 4:0.

The present data are in fairly good agreement with those 4. Results and discussion
of Schlunder et al. (reported in [1]).
Uncertainties in mean flow velocity and fluctuat- 4.1. Flow field
ing velocity variance were calculated at each measure-
ment location. Typical uncertainties in the measured Fig. 3a and b present results from free jet experiments
mean and fluctuating velocity variance for the slot at an exit Reynolds number of 35,000. In these figures,
jet at Yn =Dh ¼ 0:5 are provided in Table 2. Uncer- the x- and y-coordinates represent the streamwise and
tainties in the measured mean values were less than
3% in all cases, while that in the measured fluctuating
1.2
velocity variance were as large as 13.7% at near-wall Goertler(in[22])
locations with a large mean velocity component. The y/Dh=10.5
y/Dh = 12.5
velocity uncertainty estimate includes errors due to the 1.0 y/Dh = 14.5
y/Dh = 16.5
uncertainty in fringe spacing and the clock resolution
of the processor. The uncertainties in the calculated
0.8
quantities, U , huui, huvi were based on the component
resolution equations, and were typically larger than
V/Vm

those in quantities that were directly measured (V and 0.6

hvvi).
For surface pressure measurements, the transducer,
0.4
amplifier, and data acquisition board were calibrated
as a single unit in the range of measurement using a
micromanometer as the reference. The electronics were 0.2
kept on for the entire duration of the experiments to
prevent any errors due to thermal drift in the elec-
tronics for both flow field and surface pressure mea- -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3
surements. Although the response of the pressure (a) η= x/(y-bo)
transducer was linear, base readings (with no air flow)
taken prior to the experimental run and after the 1
experimental run exhibited a slight zero drift. This
Vo/Vo,max , vrms/v rms,o,max

V o/Vo,max
introduced an uncertainty in the mean pressure mea- 0.8
surements alone, since the ac values were high-pass
filtered prior to data acquisition. To account for the 0.6 vrms/vrms,o,max
zero drift, for each experiment, two (base) pressure
0.4
measurements were obtained with no flow––the first
prior to the experiment and the second on completion 0.2
of the experiment. The pressure measurements at each
location were subtracted from these base readings, and 0
the mean of the two readings was taken to be indic- -0.5 -0.4 -0.3 -0.2 -0.1 0
ative of the pressure differential at that location. This (b) x/Dh
method of taking a difference in voltages between the
Fig. 3. Preliminary free jet results. (a) Mean non-dimensional
measured location and the base also eliminated any velocity profiles of the slot jet in the similarity region, and (b)
zero offset of the transducer recorded under no flow mean non-dimensional exit velocity and turbulence intensity
condition. Typical resolution for the mean pressure profiles at z ¼ 0. An x=Dh ¼ 0 corresponds to the jet centerline,
measurements with the multimeter was as high as and a vRMS =vRMS;o;max ¼ 0:2 corresponds to 3.5% turbulence
0.08% of the maximum measured pressure. intensity at the jet centerline.
1836 V. Narayanan et al. / International Journal of Heat and Mass Transfer 47 (2004) 1827–1845

3.5 nozzle
14.4 m/s
3.0

2.5

2.0
y/Dh

1.5

1.0

0.5

0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
(a) x/Dh

1.0 14.4 m/s

0.8
0.6
y/Dh

0.4
0.2

1.5 2 2.5 3 3.5 4 4.5

(b) x/Dh

Fig. 4. Plots of the mean velocity field in a slot jet flow. (a) Yn =Dh ¼ 3:0 and (b) Yn =Dh ¼ 0:50.

lateral direction, respectively. The origin of the coordi- the mean exit velocity. The exit Reynolds number, based
nate system is at the nozzle exit. Fig. 3a shows profiles of on the nozzle hydraulic diameter is 23,400. Flow field in
the dimensionless mean velocity, V , normalized by the regions of the free jet near the nozzle exit could not be
local jet centerline velocity, Vmax , for the free jet at four determined due to obstruction of the beam by the nozzle
streamwise locations. The mean velocity profiles exhib- walls.
ited a self-similar behavior, collapsing onto a single Fig. 4a show the mean flow velocity vectors of the
curve past y=Dh ¼ 10:5 from the nozzle exit. The solid slot jet flow field for the nozzle spacing of Yn =Dh ¼ 3:50.
line represents Goertler’s semi-analytical solution for a The potential core of the jet extended between 5 and 5.5
free jet (presented in [22]), with a spread parameter of slot widths from the nozzle exit. The mean flow kinetic
7.67. Fig. 3b shows the dimensionless mean velocity and energy was high in the potential core up to around
turbulence intensity profile at the jet exit. The average y=Dh ¼ 0:8, after which the flow rapidly decelerated to
exit RMS velocity fluctuation in the streamwise direc- zero normal velocity at the impingement point. Beyond
tion was 3.5% of the mean jet exit velocity. x=Dh ¼ 1:00, the developing wall jet flow showed
Figs. 4–7 present flow field results of slot jet deceleration due to a spread of momentum normal to
impingement at the two nozzle spacings of Yn =Dh ¼ 3:5 the surface. Fig. 4b shows the mean flow velocity vectors
and 0.50. The axis representation and relevant geomet- of the slot jet flow field for the nozzle spacing of
rical parameters are indicated in Fig. 1a and b. Mean Yn =Dh ¼ 0:50. The kinetic energy of the mean flow at the
velocity components were made non-dimensional with first detailed measurement location of x=Dh ¼ 1:75 was
the mean exit velocity while the fluctuating velocity concentrated in y=Dh locations below 0.15, indicating
variance was made non-dimensional with the square of that the high-velocity region of the wall-bounded flow
 V. Narayanan et al. / International Journal of Heat and Mass Transfer 47 (2004) 1827–1845 1837

3.5 nozzle

3.0

2.5 0.02
0.00 0.01 0.02 0.03
0.00
2.0
y/Dh

1.5 0.00
0.01
1.0
0.02
0.01
0.5
0.03

0.02
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
(a) x/Dh

0.01 0.02 0.03 0.04

1.0
0.01
0.8 0.01
0.6 0.02
y/Dh

0.03
0.4
0.04
0.2
0.0
1.5 2 2.5 3 3.5 4 4.5
(b) x/Dh

2
Fig. 5. Contour plot of the non-dimensional x-direction fluctuating velocity variance, huui=Vmo , in a slot jet flow. (a) Yn =Dh ¼ 3:50
and (b) Yn =Dh ¼ 0:50.

was not in contact with the shroud. Flow visualization above the impingement line, the stress exhibited a local
studies of Popiel and Trass [11] demonstrated that for a high value at the jet centerline, before being damped by
2
jet impinging at Yn =Dh ¼ 1:2, the presence of a shroud, the presence of the wall. Past impingement, huui=Vmo
extending to 33 cm around the circular nozzle exit, did reached a maximum value of 0.037 in the outer region of
not significantly impact the vortex filament pattern in the developing boundary layer at x=Dh ¼ 2:0 and
2
the outer region of the wall boundary layer. Near-wall y=Dh ¼ 0:30. The nearest-wall huui=Vmo attained a
flow acceleration persisted up to about x=Dh ¼ 1:50, maximum value of 0.030 at x=Dh  1:5. Fig. 5b presents
2
where a peak U =Vmo ¼ 1:2 was measured at y=Dh ¼0:03. huui=Vmo contours for the slot jet at Yn =Dh ¼ 0:50. Re-
Fig. 5a presents a contour plot of the non-dimen- gions of high-turbulence intensity were located in the
sional velocity fluctuation variance parallel to the outer shear layer between the wall jet and the ambient
2
impingement surface, huui=Vmo , for the slot jet at surrounding fluid. Past impingement, huui in the outer
Yn =Dh ¼ 3:50. The variance was high in the shear layers mixing layer was higher compared the larger nozzle
of the free jet close to the nozzle exit, and in the outer spacing. The nearest-wall measurement of huui peaked
mixing region of the wall-bounded flow past impinge- at x=Dh  3:0.
ment. In the free jet region prior to impingement, the The contour of the non-dimensional y-direction
shear layers spread with progression downstream, component of the fluctuating velocity variance,
2
reducing the extent of the potential core. At y=Dh  0:09 hvvi=Vmo , for the slot jet at Yn =Dh ¼ 3:50 is shown in
1838 V. Narayanan et al. / International Journal of Heat and Mass Transfer 47 (2004) 1827–1845

3.5 nozzle

3.0
0.01 0.02 0.02 0.03
2.5
0.03

2.0
y/Dh

0.01
1.5

1.0 0.01
0.02

0.5
0.02
0.03
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
(a) x/Dh

0.005 0.010 0.015 0.020

1.0
0.8 0.005
0.6 0.010
y/Dh

0.4
0.020 0.015
0.2
0.0
1.5 2 2.5 3 3.5 4 4.5
(b) x/Dh
2
Fig. 6. Contour plot of the non-dimensional y-direction fluctuating velocity variance, hvvi=Vmo , in a slot jet flow. (a) Yn =Dh ¼ 3:50 and
(b) Yn =Dh ¼ 0:50.

2
Fig. 6a. High values of hvvi=Vmo in the free jet shear flow, a value of huui approximately twice that of hvvi is
layer, with a lateral spread with increasing distance expected, since production of turbulence kinetic energy
downstream from the nozzle exit were observed. In from the mean flow occurs solely in the transport
2
the measured region of the flow, a maximum hvvi=Vmo equation for huui. The normal velocity fluctuation re-
of 0.038 occurred at y=Dh ¼ 2:75 and x=Dh ¼ ceives its energy through a redistribution of turbulent
2
0:30. Past impingement, hvvi=Vmo was high in the initial kinetic energy by fluctuating pressure and velocity gra-
part of the developing wall-bounded flow between dient interactions [23].
y=Dh ¼ 0:15 to 0.25 and x=Dh ¼ 0:25 to 1.0, where a Fig. 7a shows a plot of the non-dimensional x–y
large streamwise mean flow velocity gradient existed. In (cross) component of the fluctuating velocity variance,
2 2
contrast, as seen in Fig. 5a, huui=Vmo attained a maxi-
huvi=Vmo , for the slot jet at Yn =Dh ¼ 3:50. A high po-
mum value in the outer wall-bounded flow significantly sitive value in the shear layer of the developing free jet
later, at x=Dh  2:0. In contrast, for the Yn =Dh ¼ 0:50 near the nozzle exit was observed, while the outer region
2
spacing shown in Fig. 6b, hvvi=Vmo in the outer region of the developing boundary layer showed an equally high
exhibited high values at the same locations as the negative value. The opposing signs of mean velocity
2
streamwise component, huui=Vmo (see Fig. 5b). The gradient and fluctuating cross velocity variances indi-
magnitude of peak hvvi was around 45% of the peak huui cated a net generation of turbulence at both locations.
2
in the outer shear layer of the wall jet. For a free shear Fig. 7b shows a plot of
huvi=Vmo for the slot jet at
 V. Narayanan et al. / International Journal of Heat and Mass Transfer 47 (2004) 1827–1845 1839

3.5 nozzle

3.0
-0.008 -0.003 0.002 0.008
2.5

2.0
y/Dh

1.5
0.008
-0.003
1.0 0.002

0.5
-0.008

0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
(a) x/Dh

-0.017 -0.012 -0.006 0.000

1.0
0.8
0.000
0.6 -0.006
y/Dh

0.4 -0.012
0.2 -0.017
0.0
1.5 2 2.5 3 3.5 4 4.5
(b) x/Dh
2
Fig. 7. Contour plot of the non-dimensional x–y component of fluctuating velocity variance,
huvi=Vmo , in a slot jet flow.
(a) Yn =Dh ¼ 3:50 and (b) Yn =Dh ¼ 0:50.

Yn =Dh ¼ 0:50. Large negative value of fluctuating cross around the centerline, z=Dh ¼ 0:0. The heat transfer re-
velocity variance was observed in the outer mixing layer, sults are typical of those reported in jet literature [1,2].
similar to the transitional jet impingement (Fig. 7a). Figs. 8a and 9a present dimensionless plots of local heat
High values of mean streamwise velocity gradient normal transfer coefficient, Nuloc , for Yn =Dh ¼ 3:50 and 0.50,
to the wall, oU =oy, and fluctuating cross velocity vari- respectively, as a function of streamwise distance from
ance indicated a production of turbulence in this region. the impingement line. For Yn =Dh ¼ 3:5, the heat transfer
coefficient decreased monotonically with increasing dis-
4.2. Heat transfer tance from the jet centerline, with a change (reduction)
in negative gradient at x=Dh ¼ 1:5. For spacings of
Two-dimensional contours of local heat transfer Yn =Dh < 2:5, a non-monotonic decay in Nuloc distribu-
coefficient for the slot jet nozzle at a turbulent exit tion was observed with a primary peak at the impinge-
Reynolds number of 22,500 were determined from local ment line, followed by a secondary peak that occurred
surface temperature distribution for nozzle-to-surface closer to the nozzle centerline with reduced nozzle-to-
spacings between Yn =Dh of 0.5 and 5.0. Although the slot surface spacing. In particular, for the Yn =Dh ¼ 0:5 slot
jet nozzle used in this study was not confined at its ends, jet impingement (see Fig. 9a), a region of low heat
the contour plots of local heat transfer for largest transfer was observed at x=Dh ¼ 1:6, followed by a sec-
spacing of Yn =Dh ¼ 5:0 indicated two-dimensionality ondary peak region further downstream at x=Dh ¼ 3:2.
1840 V. Narayanan et al. / International Journal of Heat and Mass Transfer 47 (2004) 1827–1845

Fig. 8. (a) Combined plot of dimensionless heat transfer coefficient, mean and RMS pressure, and near-wall fluctuating velocity
variance distributions on impingement surface for the Yn =Dh ¼ 3:5 slot jet nozzle along the minor axis of the nozzle, at centerline,
z=Dh ¼ 0:00. (b) Dimensionless fluctuating velocity variance profiles at four locations along the wall-bounded flow of the slot jet at
Yn =Dh ¼ 3:50.

The variation of impingement region heat transfer spacing has been attributed to opposing effects of
coefficient with nozzle spacing observed in the present increased turbulence with increase in Yn =Dh , and a
study was very similar to that observed by Gardon declining centerline velocity beyond the potential core
and Akfirat [2]. They reported a non-monotonic trend of the jet [2]. The impingement point Nusselt number
in the impingement line heat transfer coefficient at was greater by 39% for Yn =Dh ¼ 3:50, and by 6.3%
Re ¼ 22; 000 with an increase for spacings lower than for Yn =Dh ¼ 0:50, when compared with the analytical
Yn =Dh < 1. The present study indicated that the laminar stagnation flow heat transfer solution [26].
inflexion occurred at a slightly higher spacing between
Yn =Dh ¼ 1:0 and 1.5 [25]. The peak impingement line 4.3. Discussion
Nusselt number occurred between nozzle spacings of
Yn =Dh ¼ 3:5 and 4.0. The non-monotonic trend in Figs. 8a and 9a represent combined non-dimensional
impingement line heat transfer coefficient with nozzle plots of the local heat transfer coefficient, mean surface
 V. Narayanan et al. / International Journal of Heat and Mass Transfer 47 (2004) 1827–1845 1841

Fig. 9. (a) Combined plot of dimensionless heat transfer coefficient, mean and RMS pressure, and near-wall fluctuating velocity
variance distributions on impingement surface for the Yn =Dh ¼ 0:5 slot jet nozzle along the minor axis of the nozzle, at centerline,
z=Dh ¼ 0:00. (b) Dimensionless fluctuating velocity variance profiles at four locations along the wall-bounded flow of the slot jet at
Yn =Dh ¼ 0:50.

pressure, RMS surface pressure fluctuation, and near- For Yn =Dh ¼ 3:5, the mean pressure exhibited a
wall fluctuating velocity variances along the minor axis Gaussian distribution, as reported in several prior
for Yn =Dh ¼ 3:5 and 0.5, respectively. The near-wall studies (for example, [13,27]). The peak values of Nuloc
fluctuating velocity variances reported in this figure and Cp occurred at the impingement location. The peak
corresponded to a physical location of y=Dh ¼ 0:02. The RMS pressure fluctuation occurred in the impingement
mean surface pressure at each location was subtracted region and extended to approximately 0:4Dh on either
from a reference pressure measured at a downstream side of the jet centerline, followed by a monotonic de-
reference location of x=Dh ¼ 4:5, and is represented as a crease downstream. High RMS pressure fluctuation
pressure coefficient, Cp . The RMS surface pressure around the impingement region indicated significant
fluctuation was also subtracted from its reference value, unsteadiness in the jet flow approaching the impinge-
and normalized with the peak mean pressure differential. ment region, and can be attributed to the high near-wall
1842 V. Narayanan et al. / International Journal of Heat and Mass Transfer 47 (2004) 1827–1845

hvvi and huui of the jet prior to impingement. High huui impinging jet. They found high surface pressure fluctu-
and hvvi at a location of y=Dh  0:1 in the free jet above ations at the impingement point and attributed it to the
the impingement location were attributed primarily to velocity turbulence of the oncoming jet. They indicated
turbulent transport by convection from the free jet that, in turn, the surface pressure fluctuations produced
mixing region to the centerline, and to turbulence pro- radial velocity disturbances in the wall region of the
duction caused by the high mean streamwise velocity momentum boundary layer.
gradient, oV =oy, near the impingement region. For Yn =Dh ¼ 0:50 jet impingement, the combined
Prior studies (for example, [3,9,28]) have observed, fluid mechanics and heat transfer plot shown in Fig. 9a
by flow visualization and detailed PIV measurements, indicates that the mean surface pressure and local heat
the presence of stable streamwise vortex pairs along the transfer coefficient peaked in the impingement region.
stagnation line for transitional jet impingement, and The RMS pressure fluctuation in the impingement re-
have attributed the enhanced heat transfer in this region gion was lower than at the reference location down-
to their presence. In this study, the presence of, and stream, as indicated by the slight negative value of the
interaction between the vortex filaments was deduced non-dimensional RMS pressure profile. Although tur-
through temporal wall temperature measurements at dis- bulence in the flow field was not measured directly be-
crete x=Dh ¼ constant lines, and are detailed elsewhere neath the nozzle, the low values of RMS pressure
[16]. For a transitional jet impingement (Yn =Dh ¼ 5:0), fluctuation indicated that a low-turbulence potential
results indicated the presence of warm and cold thermal core of the jet impinged on the surface. As mentioned
streaks that meandered in time about discrete spatial before, the heat transfer coefficient at the impingement
z-locations at the impingement line, x=Dh ¼ 0:0. These line differed by only 6.3% from the analytical laminar
streaks persisted beyond the impingement region, and flow solution. The high mean streamwise velocity gra-
existed at a downstream location of x=Dh ¼ 1:19. A dient, oV =oy, of the free jet prior to impingement was
conceptual sketch of flow patterns based on flow field not accompanied by high near-wall turbulence (deduced
studies of other researchers (for example, [3,9]) sug- from the measured RMS pressure fluctuations), sug-
gested that the hot streaks corresponded to the location gesting an absence of significant levels of turbulence
of the counter-rotating vortex pair, while the cold production in the near-wall impingement region at this
streaks corresponded to the region between, where the nozzle spacing. Time trace of wall surface temperature
jet impinged directly on the wall. RMS temperature and RMS temperature fluctuations [16] indicated little
fluctuation data exhibited high values at locations be- unsteadiness at the impingement line, suggesting the
tween the hot and cold streaks, and were attributed to absence of streamwise counter-rotating vortex pairs at
the unsteady motion of the vortex pairs. this spacing.
Past impingement, hvvi peaked in the outer mixing In Fig. 5b, for the range of flow measurements, the
region, between x=Dh ¼ 0:25 and 1.00, (see Fig. 6a) prior maximum huui in the outer region occurred at
to the huui peak that occurred around x=Dh ¼ 2:0 (see x=Dh ¼ 1:75. Since this was the first downstream loca-
Fig. 5a). The fluctuating cross component of velocity tion of detailed measurements, it was unclear if a higher
variance, huvi, also peaked in the outer region at around huui existed prior to this location. The peak hvvi and huvi
x=Dh ¼ 2:0 (Fig. 7a). The region of high huui in the outer in the outer shear layer also occurred in this region (see
layer persisted up to x=Dh  3:0, followed by a decline, Figs. 6b and 7b), and there was a high correlation be-
primarily due to a reduction in mean flow kinetic energy. tween the turbulent fluctuations.
In the inner region of the developing flow past reat- From Fig. 9a, the favorable gradient of mean surface
tachment, huui peaked between x=Dh ¼ 1:25 and 1.50, pressure was larger than for transitional jet impinge-
prior to the location of the outer region peak further ment, indicating stronger flow acceleration up to
downstream at x=Dh ¼ 2:0. Fluctuating velocity vari- x=Dh  0:8. The RMS pressure fluctuation increased
ance profiles at four locations between x=Dh ¼ 1:50 and past x=Dh ¼ 0:8, and peaked around x=Dh ¼ 1:50–1:75
2.50, shown in Fig. 8b, indicated the development of a in the region of high turbulence in the outer mixing re-
region of increased huui in the outer mixing layer, along gion. Note that the peak RMS pressure fluctuation was
with a diminution of its value in the near-wall region three times lower than for transitional jet impingement.
with distance downstream of impingement. From Fig. Past impingement, the near-wall huui increased with
8a, the near-wall huui gradient exhibited a positive slope distance downstream, and attained a maximum at
from x=Dh  0:5 to 1.5, while the RMS fluctuating x=Dh ¼ 3:0. This location corresponded well with the
pressure gradient showed a negative slope in the same reduction in the rising slope in Nuloc towards the sec-
region, suggesting an inverse correlation between the ondary peak x=Dh  3:2. As with the transitional jet
streamwise gradients of RMS pressure fluctuation and impingement, the downstream locations of gradients of
near-wall huui. Kataoka et al. [29] reported local mass near-wall huui and RMS surface pressure fluctuation
transfer rate, velocity gradient fluctuations, and surface were inversely related. Non-dimensional fluctuating
pressure fluctuations on a turbulent axisymmetric velocity variance profiles at four locations between
 V. Narayanan et al. / International Journal of Heat and Mass Transfer 47 (2004) 1827–1845 1843

x=Dh ¼ 1:75 and 3.50 are shown in Fig. 9b. A compar- wise thermal streaks, indicating the presence of stream-
ison of Figs. 8b and 9b illustrates the differences between wise vortices, is unclear. Recently, Meola et al. [30]
the development of turbulence in the inner and outer proposed that a separation and reattachment mecha-
regions at the two nozzle spacings. It is clear that the nism caused the non-monotonic trend in heat transfer
peak huui in the near-wall region occurred at a location coefficient at low nozzle-to-surface spacing in circular jet
following the peak huui in the outer region for the po- impingement. Although an intermittent separation and
tential-core impingement, in contrast to observations for reattachment mechanism is also strongly suggested by
the transitional jet impingement. The profiles of surface data presented here, further studies are needed to verify
pressure fluctuation, outer- and near-wall huui strongly this hypothesis.
suggest that the enhancement resulting in the secondary Popiel and Trass [11] performed smoke-wire flow
peak in heat transfer is a consequence of the interaction visualization on circular jet impingement at Yn =Dh ¼ 1:2,
of correlated turbulent motion in the outer region and in which they observed that toroidal vortices, convected
streamwise turbulence in the near-wall region. from the free jet shear region to the outer region of the
Lytle and Webb [7] presented turbulence intensity wall-bounded flow, impinged on the wall (past the near-
and heat transfer measurements for circular air jets at wall accelerated flow region). By placing the smoke-wire
Yn =Dh < 0:5. They observed that the location of peak very close to the wall, they discovered the presence of
near wall huui (recorded at y=Dh ¼ 0:05 for Re ¼ 7800, ring-shaped wall eddies that rolled up on the plate sur-
and at y=Dh ¼ 0:025 for Re ¼ 13; 000) corresponded well face between the large-scale toroidal vortices rotating
with that of secondary peak in heat transfer. They counter to them. They hypothesized that these wall ed-
concluded that this secondary peak in local heat transfer dies, induced by the large-scale vortices impinging on the
coefficient at low spacings was a result of significantly surface, caused the enhancement in heat and mass
higher turbulence in the boundary layer, which resulted transfer rates leading to the secondary peak.
from the intense shear in the outer region between the
radially exiting jet and the stagnant ambient. While the
turbulence data reported by Lytle and Webb corre- 5. Conclusions
sponded to low nozzle-to-surface spacings, where a
secondary peak in heat transfer coefficient was observed, Flow field, surface pressure, and heat transfer rates
in the present study, a similar peak in near-wall turbu- associated with transitional and potential-core slot jet
lence also occurred in the developing flow past impingement was studied experimentally. Mean velocity
impingement for transitional slot jet impingement. At and fluctuating velocity variance were determined using
such spacing, however, the origin of high near-wall 1-D LDA, mean and RMS surface pressure fluctuations
turbulence intensity could not be related to turbulence in using a piezoresistive transducer, and surface tempera-
the outer region, and to a corresponding secondary heat tures using IR thermography.
transfer peak. It is important to note that for potential- For the transitional jet impingement, mean and
core jet impingement, the location of peak near-wall RMS-averaged fluctuating surface pressure, and local
turbulence intensity in both Lytle and Webb’s and the heat transfer coefficient peaked in the impingement re-
present study corresponded well with the location of the gion and decreased monotonically in the wall-bounded
reduction in slope of heat transfer coefficient towards flow past impingement. Generation of turbulence near
the secondary peak, which suggested that the peak near- the surface prior to impingement and the presence of
wall turbulence intensity marked a transition to a zero spanwise vortices in the stagnation region [16] indicated
pressure gradient turbulent boundary layer. an increase in near-wall turbulence, and could be the
Temporal wall temperature fluctuations, recorded at main factors that enhanced heat transfer rate from the
x=Dh ¼ 1:67 and 3.2 for Yn =Dh ¼ 0:5 jet impingement, surface in the transitional jet compared to the potential-
corresponding closely to the locations of the minimum core jet impingement.
and secondary peak in Nuloc , respectively, are presented Past impingement, the RMS surface pressure fluctu-
in Narayanan [16]. The time traces indicated a warm ation decreased monotonically and was uncorrelated
(compared to the mean surface temperature) thermal with the high streamwise fluctuating velocity variance in
streak at both locations. As mentioned previously, such the outer region at x=Dh ¼ 2:0. Peak streamwise fluctu-
streaks were not observed at the impingement region. ating velocity variance in the near-wall locations oc-
The observation of thermal streaks in the spanwise curred prior to that in the outer shear region, suggesting
direction at x=Dh ¼ 1:67 (and corresponding high RMS that turbulence in the outer region did not contribute to
surface pressure and temperature fluctuations) and the an increase in near-wall turbulence. However, the gradi-
rise in Nuloc indicated that the correlated turbulence in ents of RMS surface pressure fluctuation and near-wall
the outer region caused an unsteadiness in the thermal streamwise fluctuating velocity variance were inversely
boundary layer upon disappearance of strong mean related. A change in the slope of the heat transfer coef-
favorable pressure gradient. The origin of the stream- ficient profile at around x=Dh ¼ 1:5 corresponded to the
1844 V. Narayanan et al. / International Journal of Heat and Mass Transfer 47 (2004) 1827–1845

end of the relation between surface pressure fluctuations [4] C.D. Donaldson, R.S. Snedeker, A study of free jet
and near-wall huui, and indicated a transition from a impingement. Part 1. Mean properties of free and imping-
favorable to a zero pressure gradient hydrodynamic and ing jets, J. Fluid Mech. 45 (2) (1971) 281–319.
thermal boundary layer. [5] C.D. Donaldson, R.S. Snedeker, D.P. Margolis, A study of
free jet impingement. Part 2. Free jet turbulent structure
For potential-core jet impingement, the primary peak
and impingement heat transfer, J. Fluid Mech. 45 (3)
in heat transfer, which occurred in the impingement (1971) 477–512.
region, was followed by a region of local minimum and a [6] C.J. Hoogendoorn, The effect of turbulence on heat
secondary peak that occurred at around 1.5 and 3.2 transfer at a stagnation point, Int. J. Heat Mass Transfer
hydraulic diameters from the jet centerline, respectively. 20 (1977) 1333–1338.
The peak mean surface pressure occurred in the [7] D. Lytle, B.W. Webb, Air jet impingement heat transfer at
impingement region; however, the RMS-averaged sur- low nozzle-plate spacing, Int. J. Heat Mass Transfer 37 (12)
face pressure fluctuation profile exhibited a peak at a (1994) 1687–1697.
location of high outer region turbulence. There was a [8] S. Yokobori, N. Kasagi, M. Hirata, M. Nakamura, Y.
good correlation between the locations of the secondary Haramura, Characteristic behaviour of turbulence and
transport phenomena at the stagnation region of an axi-
peak in heat transfer and peak near-wall streamwise
symmetrical impinging jet, in: Proceedings of the 2nd
turbulence. However, local heat transfer coefficient Symposium on Turbulent Shear Flows, Imperial College,
could not be related solely to the near-wall streamwise London, 1979, pp. 4.12–4.17.
turbulence, and was strongly dependent on the relation [9] J. Sakakibara, K. Hishida, M. Maeda, Vortex structure
between outer shear layer and near-wall turbulence. and heat transfer in the stagnation region of an impinging
Similar to the transitional jet impingement, the gradients plane jet (Simultaneous measurements of velocity and
of RMS surface pressure fluctuation and near-wall temperature fields by digital particle image velocimetry and
streamwise fluctuating velocity variance were inversely laser-induced fluorescence), Int. J. Heat Mass Transfer 40
related. While the origin of high RMS pressure fluctu- (13) (1997) 3163–3176.
ation for transitional jet impingement was attributed to [10] R.J. Goldstein, A.I. Behbahani, K.K. Heppelmann,
Streamwise distribution of recovery factor and the local
the high turbulence near the stagnation region of the
heat transfer coefficient to an impinging circular air jet, Int.
impinging jet, for the potential-core jet impingement, it J. Heat Mass Transfer 29 (8) (1986) 1227–1235.
resulted from correlated turbulent motion in the outer [11] C.O. Popiel, O. Trass, Visualization of a free and imping-
region of the wall-bounded flow past impingement. Past ing round jet, Exp. Thermal Fluid Sci. 4 (3) (1991) 253–261.
x=Dh ¼ 1:5, the observation of thermal streaks (and [12] M.D. Fox, M. Kurosaka, L. Hedges, K. Hirano, The
corresponding high RMS surface temperature fluctua- influence of vortical structures on the thermal fields of jets,
tion) [16] and an increase in Nuloc indicated that the J. Fluid Mech. 255 (1993) 447–472.
correlated outer region turbulence caused an unsteadi- [13] R. Viskanta, Heat transfer to impinging isothermal gas and
ness in the thermal boundary layer upon disappearance flame jets, Exp. Thermal Fluid Sci. 6 (1993) 111–134.
of strong mean favorable pressure gradient. [14] M. Korger, F. Krizek, Mass-transfer coefficient in imping-
ing flow from slotted nozzles, Int. J. Heat Mass Transfer 9
(1966) 337–344.
Acknowledgements [15] J. Zhe, V. Modi, Near wall measurements for a
turbulent impinging slot jet, J. Fluids Eng. 123 (2001)
The authors acknowledge the financial support pro- 112–120.
vided by the Texas A&M University Drying Research [16] V. Narayanan, 2003, Time-resolved thermal flow structures
in impinging slot jet flows, HT2003-47493, in: Proceedings
Center. VN would like to thank Dr. Stephen Danczyk
of the 2003 Summer Heat Transfer Conference, Las Vegas,
and Prof. G.L. Morrison for insightful discussions and
Nevada.
help with the set-up of the LDA system. [17] S.E. Logan, A laser velocimeter for reynolds stress and
other turbulence measurements, AIAA J. 10 (7) (1972)
References 933–935.
[18] G.L. Morrison, private communication, May 2001.
[1] H. Martin, Heat and mass transfer between impinging gas [19] R. Siegel, J.R. Howell, Thermal Radiation Heat Transfer,
jets and solid surfaces, Adv. Heat Transfer 13 (1977) 1–60. third ed., Hemisphere Publishing Corporation, Washing-
[2] R. Gardon, J.C. Akfirat, The role of turbulence in ton, DC, 1992, p. 1040.
determining the heat-transfer characteristics of impinging [20] R.J. Moffat, Describing uncertainties in experimental
jets, Int. J. Heat Mass Transfer 8 (1965) 101–108. results, Exp. Thermal Fluid Sci. 1 (1988) 3–17.
[3] S. Yokobori, N. Kasagi, M. Hirata, N. Nishiwaki, Role of [21] C. Gau, C.C. Lee, Impingement cooling flow structure and
large-scale eddy structure on enhancement of heat transfer heat transfer along rib-roughened walls, Int. J. Heat Mass
in stagnation region of two-dimensional, submerged, Transfer 35 (11) (1992) 3009–3018.
impinging jet, in: Proceedings of the Sixth International [22] N. Rajaratnam, Turbulent Jets, in: Developments in Water
Heat Transfer Conference, Toronto, Canada, 1978, pp. Science, vol. 5, Elsevier Publishing Company, New York,
305–310. NY, 1976.
 V. Narayanan et al. / International Journal of Heat and Mass Transfer 47 (2004) 1827–1845 1845

[23] H. Tennekes, J.L. Lumley, A First Course in Turbulence, [34] S.J. Downs, E.H. James, Jet impingement heat transfer––a
first ed., MIT press, Cambridge, MA, 1972, p. 75 (sixteenth literature survey, in: Proceedings of the National Heat
printing). Transfer Conference, Pittsburgh, Pennsylvania, 1987.
[24] Y. Guo, D.H. Wood, Measurements in the vicinity of a [35] S. Polat, Heat and mass transfer in impingement drying,
stagnation point, Exp. Thermal Fluid Sci. 25 (2002) 605– Dry. Technol. 11 (6) (1993) 1147–1176.
614. [36] R. Gardon, J.C. Akfirat, Heat transfer characteristics of
[25] V. Narayanan, An experimental flow field and heat transfer impinging two-dimensional air jets, J. Heat Transfer 86
study of submerged double-sided reattachment and two- (1966) 101–108.
dimensional impingement slot jets, Ph.D. Dissertation, [37] S. Belatos, N. Rajaratnam, Plane turbulent impinging jets,
Texas A&M University, College Station, Texas, 2001. J. Hydraul. Res. 11 (1973) 29–59.
[26] W.M. Kays, M.E. Crawford, Convective Heat and Mass [38] E. Gutmark, M. Wolfshtein, I. Wygnanski, The plane
Transfer, third ed., McGraw-Hill, Inc., Hightstown, NJ, turbulent impinging jet, J. Fluid Mech. 88 (4) (1978) 737–
1993. 756.
[27] C.V. Tu, D.H. Wood, Wall pressure and shear stress [39] S.A. Striegl, T.E. Diller, The effect of entrainment temper-
measurements beneath an impinging jet, Exp. Thermal ature on jet impingement heat transfer, J. Heat Transfer
Fluid Sci. 13 (1996) 364–373. 106 (1984) 27–33.
[28] J. Sakakibara, K. Hishida, W.R.C. Phillips, On the vortical [40] K. Kataoka, M. Suguro, H. Degawa, K. Maruo, I. Mihata,
structure in a plane impinging jet, J. Fluid Mech. 434 The effect of surface renewal due to large-scale eddies on jet
(2001) 273–300. impingement heat transfer, Int. J. Heat Mass Transfer 30
[29] K. Kataoka, Y. Kamiyama, S. Hashimoto, T. Komai, (3) (1987) 559–567.
Mass transfer between a plane surface and an impinging [41] K. Kataoka, Impingement heat transfer augmentation due
turbulent jet: the influence of surface-pressure fluctuations, to large scale eddies, Heat Transfer 1 (1990) 255–273.
J. Fluid Mech. 119 (1982) 91–105. [42] C. Gau, C.M. Chung, Surface curvature effect on slot-air-
[30] C. Meola, G. Cardone, C. Carmicino, G.M. Carlamagno, jet impingement cooling flow and heat transfer process, J.
Fluid dynamics and heat transfer in an impinging air jet, in: Heat Transfer 113 (1991) 858–864.
G.M. Carlomagno, I. Grant (Eds.), Proceedings of the 9th [43] C.S. McDaniel, B.W. Webb, Slot jet impingement heat
International Symposium on Flow Visualization, Heriot- transfer from circular cylinders, Int. J. Heat Mass Transfer
Watt University, Edinburg, Scotland, paper number 429, 43 (2000) 1975–1985.
2000, CD-ROM. [44] A.H. Bietelmal, M.A. Saad, C. Patel, The effect of
[31] Y. Becko, Impingement Cooling––A Review, in: Von inclination on the heat transfer between a flat surface and
Karman Lecture Series 83, Brussels, Belgium, 1976, pp. an impinging two-dimensional air jet, Int. J. Heat Fluid
1–40. Flow 21 (2000) 156–163.
[32] P. Hrycak, Heat transfer from impinging jets, a literature [45] M. Choi, H.S. Yoo, G. Yang, J.S. Lee, D.K. Sohn,
review, Final report AFWAL-TR-81, Air Force Wright Measurements of impinging flow and heat transfer on a
Aeronautical Laboratories, Wright-Patterson Air-Force semi-circular concave surface, Int. J. Heat Mass Transfer
Base, Ohio, 1981. 43 (2000) 1811–1822.
[33] M.K. Looney, J.J. Walsh, Mean-flow and turbulent [46] S. Maurel, C. Solliec, A turbulent plane jet impinging
characteristics of free and impinging jet flows, J. Fluid nearby and far from a flat plate, Exp. Fluids 31 (6) (2001)
Mech. 147 (1984) 397–429. 687–696.