H e a t T r a n s f e r f r o m I m p i n g i n g Slot J e t s o f Air
P a r t 1: A e r o d y n a m i c C h a r a c t e r i s t i c s o f F r e e J e t s
Abstract. Aerodynamic data on the axial velocity and turbulence development of free slot jets is
presented. Jets varying from 0.0635 to 0.375 inches (1.59 to 9.53 mm) in width and with air
velocities between 2,000 and 10,000 ft/min (t0.2 to 50.8 m/sec) were examined. This study
was the first phase of a research program aimed at developing empirical data describing imping-
ing slot jet heat transfer commonly used in veneer drying.
Introduction
The use of impinging air jets for heating and cooling surfaces has steadily increased
over the past two and one-half decades. The relatively high heat transfer rates and
uniformity of heat delivery, compared with nonimpinging techniques, are largely
responsible for the popularity of this heat transfer method.
Impinging air jets are used for a wide variety of applications, ranging from the
tempering of glass sheets to the cooling of gas turbine blades. Impingement drying
systems are quite common to the forest products industry, being used in the drying
of veneer and paper. Although they are more efficient than parallel air flow systems,
there is still need for improved and more efficient design and operation of this type
of drier. This is due to ever-increasing fuel costs, desires for greater drier through-
put in shorter times, and the rather large capital investment costs involved.
Adequate information to predict heat transfer under various operating conditions
such as with multipl e slot jets at low jet exit to surface distances is not presently
available. Relatively few studies have been made for multiple slot jets with nozzle
to surface distances less than 10 slot widths. In one [Schuh, Pettersson 1966],
measurements were fnade at a single distance from the jet exit. Kerscher et al. [1968]
give predictive equations for multiple slot jets over a wide range of vertical distances,
but the equations for distances less than 5 slot widths are based on only 58 meas-
urements. In light of the large number of feasible slot widths, exit velocities, and
vertical distances, this number of measurements appears much too small.
0043-7719/78/0012/0237/$ 2.60
238 D.G. Arganbright, H. Resch and J. R. Olson
Purpose
The purpose of this study was to investigate the heat transfer produced by arrays of
normally impinging two-dimensional air jets and the functional relationships of a
number of different design and operating parameters. The desired end goal was to
develop empirically derived predictive equations describing this form of heat transfer,
which could be used for the design ~vcI operation of driers at optimum efficiency.
The study was divided into the following two main problem areas:
1. Aerodynamic characterization of two-dimensional jets
2. Measurement of average and local heat transfer coefficients.
Only the first of these problem areas is discussed in the present paper, the other
is considered separately in a second part. In regard to jet aerodynamics, data were
collected for each of four jet widths, in order to determine whether jets of different
sizes differ aerodynamically. In addition to centerline velocity decay, inforlaation
on jet turbulence levels was desired in order to explain variations in heat transfer.
Background information
No attempt has been made here to review the current literature pertinent to such a
study, as a general review is already available [,~ganbright, Resch 1971]. It seems
desirable, however, to make a few brief comments on the nature of air jets for those
not directly working in the field and to clarify the terminology used.
Two-dimensional jets are commonly referred to as slot jets. Jets can be classified
on the basis of their proximity to a surface. A free jet occurs when the jet fluid
discharges from the jet orifice into a large open space without any surrounding sur-
face impeding its flow. If a jet strikes a surface at some distance from the jet exit
it is termed an impinging jet. Further, if the jet travels in such a fashion that a
plane drawn through its centerline strikes a plane surface exactly at right angles, it is
known as a normally impinging jet.
A normally impinging slot jet of air is depicted in Fig. 1, and lines representing
lateral spread and velocity distributions at various distances from the jet exit are
shown.
Immediately upon lea~ing the jet orifice, the outermost edges of the emerging jet
begin to entrain air from the surrounding environment. This has two effects, one
that the jet increases in size laterally, and secondly the entrained air diffuses toward
the centerline of the jet slowly decreasing the velocity of the disturbed portion of
the jet. At a distance of from 3 to 5 slot widths away from the jet exit, the en-
trained air finally reaches the centerline of the jet and the entire jet has a lower
velocity than it did upon initially emerging. This reduction in jet velocity is exceed-
ingly important as the heat transfer capabilities of impinging jets are strongly depen-
dent upon the jet velocity.
Once the jet strikes the surface below it (i. e., the exchange surface), the jet
divides and begins to flow roughly parallel to the surface. The point on the surface
Heat transfer from impinging slot jets of air 239
immediately below the centerline of the jet is the stagnation point. In conforming
with prevailing conventions vertical distances are denoted as Z values and lateral
distances as X (Fig. 1). Vertical and lateral distances are also commonly expressed
dimensionally by dividing them by the slot width, as Z/B's, or X/B's, or expressed
in another fashion, in terms of the number of slot widths in either direction.
d,
I - F----vELoc,TY PROFILE
?=
X
\\\\\\\\\\\\\\\\\\ -,~ \ \ x \ \ \ \ - , ~ ~ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \
EXCHANGE SURFACE WALL JET FLOW
STAGNATION
POINT
Variations in the velocity of jets are also extremely important, and jets can be
further classified according to the nature of their velocity fluctuations. Any fluid
in motion has a mean velocity and there are invariably slight variations in the flow
around this mean. Note that variations in mean velocity should not be confused
with these instantaneous or turbulent fluctuations. Any jet exhibiting these instanta-
neous fluctuations is referred to as a turbulent jet while jets lacking these are classi-
fied as being laminar. The importance of this classification lies in the fact that
turbulent flows yield higher heat transfer rates than do corresponding laminar flows.
At any point within the body of a jet there are three mean velacity components.
Two are transverse components (V and W), while the third is the axial component
(U). In a turbulent jet, the instantaneous velocity components corresponding to
each mean velocity component are denoted as v', w', and u' respectively. Because
fluctuations of this nature may be positive or negative they are represented by their
root-mean-square values x / ~ , x / ~ , and x/-~, and are called the intensity of turbu-
lence. If these values are divided by their corresponding mean velocity value they
are then said to be normalized turbulence intensity values (~7'/V, ~'/W, and fi'/U).
240 D.G. Arganbright, H. Resch and J. R. Olson
Methods
Experimental variables
Aerodynamic data were collected for free jets at the following slot widths-0.0625
(1.59m), 0.1250 (3.18 ram), 0.2500 (6.36mm), and 0.3750 (9.53mm) inches. The
following measurements were made on single jets of each width:
a. Mean centerline velocities (Uz) at 1 slot-width intervals from Z/B values of
1 to 16 for exit velocities of 2,000 (10.2 m/s), 4,000 (20.4m/s), 6,000 (30.6m/s),
8,000 (40.7 m/s), and 10,000 ft/min (50.8 m/s).
b. Mean lateral velocities (Ux) were measured at appropriate lateral distances,
(X/B), for vertical distances away from the jet (Z/B) equal to 4, 8, 12, and 16 slot
widths.
c. Axial components (u') of turbulence intensity were measured at the jet center-
line at 1.0 slot width intervals from Z/B values of 1 to 16.
d. Axial components of turbulence (u') were also measured at appropriate dis-
tances away from the jet centerline (X) for vertical distances of 4, 8, and 16 slot
widths.
Experimental procedures
rounded inlet edges (Fig. 2). Uniform edges were machined on each jet section
using a 0.2500 inch radius cutter as shown in the figure.
~ JET PLATE
l,
~.// I
/,/ 3/16 in.
/
/
!
\\ ,/
\ /
\ /
The centerline velocity at various distances from the jet exit (Uz), obtained for the
varying slot widths (B) and exit velocities (Uexit) tested, were normalized by dividing
each value by its corresponding original exit velocity. These normalized values were
then plotted against dimensionless distance from the jet exit (Z/B) (Fig. 3). As ex-
pected, centerline velocity remains constant for a few slot widths, after which it be-
gins a rapid decline. In all, 320 measurements were used in making this plot, many
of which are superimposed on the graph.
1.0 t
0.6
0.2 R2 = 0 . 8 2
0.0 ~ I ~ I ~ I i I i I i I ~ I ~ i
0 2 4 6 8 IO 12 14. 16
Z/B
Fig. Variation in normalized centerline velocity (Uz/Uexit) with dimensionless distance
3.
from the jet exit (Z/B)
jet exit (Z/B) was the only variable found to be related to the rate of velocity change.
Thus changes in slot width and exit velocity do not affect the rate of normalized
velocity decay of free slot jets.
A simple linear regression analysis was then made between (Uz/Uexit) and (Z/B)
for only that portion of the data in which Uz/Uexit ( 1.0. The resulting equation
was used to calculate the average (Z/B) at which (Uz/Uexit) was first less than 1.0.
These results permit a complete description of the centefline velocity distribution as
follows:
Uz/Uexit = 1.0976 - 0.02986 (Z/B) for 3.3 < Z/B ~< 16 (2)
These two relationships are represented in Fig. 3 by the two straight lines. Some
82 percent of the variation in velocity was directly attributable to changes in Z/B,
with the remaining variation unexplained.
In an attempt to clarify certain aspects of the slot jet centerline velocity decay,
the data are also presented in several other forms.
It should be pointed out that almost every jet examined showed a slight "vena
contracta" effect, or a contracting in width of the jet immediately upon leaving the
1.0
0.9
~O
lo
0,8
0.7
0.6
^
N (15
=.
oA
o3
0.2
r
o.i
^8o~
~ - ~ ~
B
o
o.o
6 5 4 3 2 I r I 2 3 4 5 6
XIB
Fig. 4. Variation in normalized lateral velocity (Ux/Uz) at horizontal distances from the jet
centerline for vertical positions of Z/B = 4, 8, and 16
244 D . G . Arganbright, H. Resch and J. R. O1son
The centerline axial turbulence levels (u'z) and mean centerline velocities (Uz), at in-
creasing distances from the jet exit are given in Fig. 5 for a 0.2500 inch wide slot jet
with an exit velocity of 6,000 ft/min. The results of this particular test were highly
representative of the other tests made. Observing the centerline velocity data first,
one sees that they are essentially constant until a Z/B = 3 is reached after which it
700
I 0'z -'/"
'~" ~ i
2 600 ' 00o
m
Z
50C
la.l
// ~'-~
"% Uz
-, ._,I 5
z_
W
4OO 'b--a..
9 I
_14oo0
"<
N 300
m
~- 2 0 0 o.
1
x
,~ I00 I
I
O J ' ' I I I i ; , , , i f , , , 12000
O I 2 3 4 5 6 7 8 9 I0 II 12 13 14 15 16 17
Z/B
Fig. 5. Variation in centerline axial turbulence (U~) and velocity (Uz) with increasing distance
(Z/B) from t h e jet exit
Heat transfer from impinging slot jets o f air 245
begins to steadily decline. The axial turbulence, on the other hand, is almost
negligible when the jet first emerges from the orifice, and then rapidly increases
with increasing distance until it reaches its maximum value at a Z/B = 8. From this
point on, it begins a continuous decrease at approximately the same rate as the mean
centerline velocity.
The above turbulence data, normalized in several different ways, are replotted in
Fig. 6. The lower plot (Uz/Uexit) SHOWSturbulence at different Z/B's as a ratio of
the turbulence at the jet exit (actually at a Z/B = 1.0). This shows that, at its
maximum value, the turbulence has increased almost 6 times over its value at the
jet exit. Centerline turbulence is also plotted as a percentage of the exit velocity
(U'z/Uexit); note that the maximum turbulence is approximately equal to 13 percent
of the exit velocity. Lastly, turbulence is given as a percentage of the jet velocity
(U'z/Uz) at the same vertical distance. In this, the uppermost line, turbulence is
seen to increase until it is about 17 percent of the jet velocity after which it re-
mains constant, showing again that once the turbulence begins to decrease it does
so at a rate equal to that of the jet velocity.
One of the major reasons in measuring turbulence values was to determine
whether jet width affects turbulence development. This was assessed in two ways.
17 ~ ~ - ~ - - ~ o ~.~ (i00 )
16 z
15
14
13
w 12
i
nn
n-- 9
~ 8
7
5 u~
423 ~ "u~xlt
I
I I I L I L I I I I I I I I 1 I
I 2 3 4 5 6 7 8 9 I0 II 12 13 14 15 16
ZIB
Fig. 6. Variation in turbulence, normalized in different ways, with increasing distance from the
jet exit
246 D.G. Arganbright, H. Resch and J. R. Olson
First graphs of turbulence as a ratio of jet exit turbulence versus Z/B were made
for each of the jet widths tested, in order to see if differences existed in either the
rate of turbulence development or the Z/B at which maximum turbulence occurred.
These data, although showing rather large variations, failed to reveal any consistent
effect of slot width on either the rate of turbulence development or the maximum
value. In a second test, turbulence data as a percentage of the jet exit velocity,
were averaged over a Z/B range of 1 to 16, to test whether different slot widths
generated different average turbulence levels. These results also failed to reveal any
differences between jets of varying width. It was concluded that the magnitude of
turbulence development is not affected by slot width over the range tested.
Jet exit velocity was found, however, to have a consistent although relatively
unimportant effect on turbulence development. At very short distances from the
jet exit (Z/B ~< 3) the turbulence level of jets at low exit velocities was appreciably
lower (about 89 as large) than that at higher exit velocities. This difference is, how-
ever, very short-lived and after a distance near Z/B = 3 is reached the degree of
turbulence development becomes independent of jet exit velocity.
Ignoring the slight effect of exit velocity at very low values of Z/B, permits one
to conclude that in this experimentation the only factor affecting turbulence is
distance from the jet exit (Z/B). The turbulence data, in the form of uPz/Uexit,
R 2= 0 . 7 8 8
16
15
14
13 08 ~ 8 8 ~ . o o
12
II
I0 g; ~ oo ; t o
0 ~ 0 o 0 o
9 8 ,~ 8 ~ ~ ~ - o o ~ ~ o
_o 8
o g/8 8
5
4
Jo:i ~00
3
2
I
l l l [ ( l l I l l l l l l l
I 2 3 4 5 6 7 8 9 I0 II 12 13 14 15 ~
Z/B
Fig. 7. Axialturbulencedata(u~/UexitxlOO) p o o l e d o v e r ~ o t w i d t h a n d e x i t v e l o c i t y versus
z/B
Heat transfer from impinging slot jets o f air 247
obtained from testing of different slot widths and exit velocities were, therefore,
combined and are shown in Fig. 7. The turbulence data were regressed against (Z/B)
using a polynomial equation with the following result:
(u'z/Uexit) 1~176
= - 0.9355 + 2.9317 (Z/B) - 0.1872 (Z/B) 2 + 0.00306 (Z/B) 3 (3)
In light of the observed variation within the data and since the slight effect of
Uexit at low Z/B was ignored the correlation is excellent-R 2 = 0.79. As shown be-
fore with the 0.2500 inch jet, turbulence is seen to be small at low Z/B. With in-
creasing distance from the jet exit it increases rapidly reaching a maximum value at
a Z/B --- 10.5, after which it begins to decrease. At its maximum point the turbu-
lence is, on the average, equal to 12.7 percent of the original velocity at the jet exit.
The axial turbulence data at lateral positions away from the jet centerline were
analyzed in a similar manner. As with the centerline turbulence data, neither slot
width nor exit velocity was found to significantly influence turbulence. The data
were, therefore, combined and average values of lateral axial turbulence calculated
for the different Z/B tested (Fig. 8). An interesting feature of these data was the
0 /
54 3 2 ~ 2
I
34 5I
14
12 ~ ~B
I0
Ux
(100)
i
Uz 6 8
4
2
0
5 4-3 2 0 I 2 3 4i 5I
I [ I I I I I I
18
16141210684 i/B=I6
X/B
F~. 8. Lateral tttrbulence (ux/Uz) from the jet centerline line (X/B = 0) for vertical distances
of Z/B = 4, 8, and 16
248 D.G. Arganbright, H. Resch and J. R. Olson
occurrence of maxima values of turbulence at some distance away from the center-
line. A maximum was found at each of the Z/B tested. The plot also shows that
the jet is highly turbulent across the greatest part of its width.
Aerodynamic data on axial velocity (U z and Ux) and turbulence (u') were obtained
on free jets with widths between 0.0625 to 0.375 inches (1.59 to 9.53 ram). Meas-
urements were made at vertical distances from 1 to 16 slot widths over air velocities
between 2,000 to 10,000 ft/min (10.2 to 50.8 in/sec). The major findings were as
follows:
1. The velocity of a free slot jet is constant up to a distance of around 3.3 jet
widths away from the jet exit.
2. From this point on it decreases linearly at least for distances less than 16 slot
widths. Thus, the greater the distance from the jet exit the lower the air velocity.
This decrease is considerable, amounting to 26 percent of the original velocity at a
Z/B equal to 12.
3. This decay in velocity was found to be independent of slot width and exit
velocity and was only affected by distance from the jet exit (Z/B). An equation
relating normalized velocity (Uz/Uexit) and dimensionless distance from the exit
when applied to all of the data pooled accounted for 82 percent of the variation.
4. As expected the jets expanded as they moved away from the jet exit. The
rate expansion, which is accompanied by a decrease in the jet velocity was also
independent of dimensionless jet width and original velocity and was only depen-
dent on distance. The maximum velocity was at the jet centerline with a minimum
at the outer edges.
5. Centerline axial turbulence was found to be essentially zero as the jet emerges;
then it rapidly increases with increasing distance until it reaches a maximum at a
Z/B of approximately 10.5. Afterward it declines continuously at a rate about equal
to the decrease in mean jet velocity.
6. At its maximum value the turbulence is approximately equal to 12.5 to 13 per-
cent of the original exit velocity.
7. Turbulence was not found to be influenced by slot width nor by exit velocity
except slightly for Z/B's less than 3.0.
8. The jets examined were highly turbulent across their entire widths. The
turbulence levet at lateral points other than the jet centerline was also unaffected by
dimensionless jet width and the velocity employed.
References
Arganbright, D. G.; Resch, H. 1971. A review of basic aspects of heat transfer under impinging
air jets. Wood Sci. Technol. 5 (2): 73-94
Gordon, R.; Cobonpue, J. 1962. Heat transfer between a flat plate and jets of air impinging on
it. Inter. Develop. Heat Transfer 454-460, ASME, New York
Heat transfer from impinging slot jets of air 249