Building and Environment 37 (2002) 13291342

www.elsevier.com/locate/buildenv
Theoretical and experimental investigation of impinging jet ventilation
and comparison with wall displacement ventilation
T. Karimipanah
a
, H.B. Awbi
b,
a
Air Innovation AB, Sweden
b
Indoor Environment and Energy Research Group, Department of Construction Management and Engineering, The University of Reading,
Reading RG6 6AW, UK
Received 2 August 2001; received in revised form 1 October 2001; accepted 18 October 2001
Abstract
This paper focuses on evaluating the performance of a new impinging jet ventilation system and compares its performance with a wall
displacement ventilation system. Experimental data for an impinging jet in a room are presented and non-dimensional expressions for the
decay of maximum velocity over the oor are derived. In addition, the ventilation eciency, local mean age of air and other characteristic
parameters were experimentally and numerically obtained for a mock-up classroom ventilated with the two systems. The internal heat
loads from 25 person-simulators and lighting were used in the measurements and simulations to provide a severe test for the two types of
ventilation systems. In addition to a large number of experimental data CFD simulations were used to study certain parameters in more
detail. The results presented here are part of a larger research programme to develop alternative and ecient systems for room ventilation.
? 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Impinging jet ventilation; Displacement ventilation; Classroom ventilation; Air jets; Room air movement; CFD
1. Introduction
Although the traditional mixing systems have poor
ventilation eciency and are less energy ecient, they still
occupy a large portion of the market. When displacement
ventilation (DV) was rst introduced almost three decades
ago, it seemed at the time to be a promising ventilation
concept due to its high ventilation eciency and strati-
cation principle. In DV systems, cool air supplied at low
level is entrained by plumes rising from heat sources in
the room. To create an eective ventilation system in the
occupied zone, there should be a balance between the mo-
mentum in the supply air and thermal (buoyancy) forces
due to heat sources. In this low momentum displacement
ow, the buoyancy forces created by heat sources have a
tendency to take over and thus often causing poor venti-
lation eciency in some zones of the room [1]. Another
disadvantage of a displacement system is that it can only be
used for cooling and is not suitable for winter heating. To
overcome this problem new systems like ceiling mounted

Corresponding author. Tel.: +44-118-931-6786; fax: 44-118-931-

3856.
E-mail address: h.b.awbi@reading.ac.uk (H.B. Awbi).
textile bag supply and down-to-oor impinging supply have
been developed, see Ref. [2].
A new method of air distribution known as the Air Queen
(AQ) has been developed in Sweden, which is based on im-
pinging jet ventilation (IJV) [3]. This method is based on
the principle of supplying a jet of air with high momentum
downwards onto the oor. As the jet impinges onto the oor
it spreads over a large area causing the jet momentum to re-
cede but still has a sucient force to reach long distances.
Unlike the DV system which oods the oor with supply
air, the resulting ow from an IJV is a very thin layer of air
over the oor. This method has the advantages of both the
mixing and displacement ventilation systems without known
disadvantages. Impinging jet ventilation has lower momen-
tum than mixing and higher momentum than wall displace-
ment ventilation (WDV), i.e. when the air terminal unit is
wall mounted at low level. Although higher momentum than
WDV, IJV produces a similar ow eld and has, therefore,
promising applications [2,4].
2. Properties of impinging jet
The impingement of a turbulent jet on a at plate (wall)
has been widely studied with dierent congurations. In all
0360-1323/02/$ - see front matter ? 2002 Elsevier Science Ltd. All rights reserved.
PII: S0360- 1323( 01) 00117- 2
1330 T. Karimipanah, H.B. Awbi / Building and Environment 37 (2002) 13291342
Nomenclature
(The International system of units (SI) is used through-
out this paper)
A
out
Supply outlet area (m
2
)
C
i
Inlet contaminant concentration (ppm)
C
o
Outlet contaminant concentration (ppm)
IJV Impinging jet ventilation (dimensionless)
1
o
Outlet temperature (

C)
1
i
Inlet temperature (

C)
U Air velocity in radial direction (m}s)
U
o
Air velocity at supply outlet (m}s)
U
c
Jet centreline velocity (m}s)
U
max
Maximumair velocity at dierent positions from
oor (m}s)
WDV Wall displacement ventilation (dimensionless)
d Supply outlet diameter (m)
h Impingement height from oor level (m)
q
v
Supply airow rate (m
3
}s)
x Radial distance from the jet impingement point
to measuring point (m)
x
0
Radial distance from the jet impingement point
to the virtual origin (m)
, Vertical distance from oor (m)
,
1}2
The height from oor corresponding to half of
the maximum air velocity (m)
c
c
Ventilation eectiveness for contaminant re-
moval (%)
c
t
Ventilation eectiveness for heat distribution}
removal (%)
DV Displacement ventilation
PMV Predicted mean vote
PPD Predicted percentage of dissatised (%)
r Radius of supply outlet (m)
early investigations, the prediction of the momentum ux
follows the same procedure as in a classical boundary layer.
The impinging jet is an interesting test case due to its dier-
ent ow regions (i.e. the potential core region, the free jet
region, the impinging or ow deection region and the wall
jet region).
Some early and recent investigators have also studied the
free and impinging jet phenomena. Beltaos and Rajaratnam
[5] provided detailed results for the impingement region and
enlarged the scope of the results which were available in
the free jet and wall jet regions. Gutmark et al. [6] pre-
sented an experimental study of the turbulent structure on
the centre-line of a two-dimensional impinging jet. They
pointed out that the impingement of the jet is not aected
by the presence of the plate over 75% of the distance be-
tween the nozzle and the plate and also the turbulent prop-
erties of the jet change from their equilibrium level close
to the impingement region. The impinging jet is somewhat
similar to a stagnation ow in which an innite stream im-
pinges on a nite body. In our investigation we used the
impinging jet conguration shown in Fig. 1. The ow eld
is divided into three regions, see also Ref. [5]: (i) a free
jet region, (ii) the impingement region and (iii) the wall
jet region. There are also transitional zones between these
regions.
The numerical analysis and development of impinging
ows have received large attention, not because of their
simplicity, but due to the presence of dierent ow regions.
However, the result from the analysis may depend on the
turbulence model used.
A jet approaching a plate will at some distance from the
plate begin to feel the presence of the plate. For example
the mean velocity on the centre line is similar to a free jet
up to some distance and then decreases (faster) to the zero
value at the impingement point, see Fig. 2. The distance
from the plate where the centre line velocity starts to deviate
from the free jet curve is taken as the location of the end
of the free jet region and the beginning of the impingement
region. From the gure one can see that the eect of the
plate is felt when the distance from the plate is 0.14h,
which is in agreement with the results obtained by Bealtos
and Rajaratnam [5].
Fig. 1. Impinging jet conguration.
T. Karimipanah, H.B. Awbi / Building and Environment 37 (2002) 13291342 1331
Fig. 2. Decay of centre line velocity with dierent impingement heights
(d = 25 mm).
3. Impinging jet ventilation system
A test method has been developed for impinging jet sup-
ply devices, which is based on the radial impinging jet
theory, see Ref. [7]. It was based on the assumption that
the velocity distributions on the impinging surface have the
same proles as the wall jet region illustrated in Fig. 3.
Independent tests show that the radial velocity (U) in the
y-direction with dierent radial distances from the origin
give a similar velocity prole described in dimensionless
form, see Ref. [8].
Using these similarity criteria, the following proportion-
ality relation is obtained:
log

U
max
U
o

log

A
out

, (1)
where U
max
is the maximum velocity in radial direction for
a specic point located at horizontal distance x from origin
(impingement point), U
o
is supply velocity and A
out
is the
supply outlet area.
Fig. 3. Radial jet impinging on a at plate.
The resulting empirical relation based on the measured
U
max
at dierent distance (x) in the radial direction, can give
a basis for use of impinging jet supply system.
3.1. Methodology and test conditions
Velocity eld measurements were carried out in a test
room (classroom) at the Centre f or Built Environ-
ment in G avle, Sweden. The supply device Air Queen
model AQ-D54}24 was used in the tests. Two impinge-
ment heights of 300 and 950 mm were studied, see
Fig. 4(a).
The aim of velocity measurements was to establish a ba-
sis of dimensioning physical parameters for design purposes
of the impinging jet supply. This was to investigate the vari-
ation of the maximum air velocity on the oor with the
jet outlet velocity and to establish the vertical height above
the oor where this maximum occurs as this will be impor-
tant for draught considerations. The tests were also aimed
to show the inuence of impingement height on the ow
pattern over the oor.
The supply airow rate was regulated by a variable
speed fan and the supply temperature was set according to
the cooling load in the test room. The room temperature
was kept constant by cooling the room walls. The airow
was measured by an orice plate. During the tests an air-
ow of 150 l}s, supply temperature of 18

C and a room
temperature of 21.5

C were used, see Ref. [7] for further

details.
For supply heights of h = 950 and 300 mm, velocities
were conducted at 12 measuring points with dierent dis-
tances (x) and 3 angles from the impingement point, see
Fig. 4(b). At each measuring point the traversing for ve-
locity was carried out from the oor up to 300 mm with
10 mm between traversing intervals. At each point in the
traversing interval, 480 samples were recorded for a sam-
pling interval of 0.5 s to give the mean velocity at that
point.
1332 T. Karimipanah, H.B. Awbi / Building and Environment 37 (2002) 13291342
Fig. 4. A plane view of test facility with measuring points. All dimensions in metres.
The following measuring instruments were used in the
tests:
Air velocity measuring instrument of type Swema Air 300
with hotwire anemometer of type CTDA.
Micromanometer of type Alnor MP3 KDS (pressure mea-
surement for supply duct).
Temperature measuring instrument of type Mitec AT30
with thermistor.
The measuring errors were estimated to be:
Supply airow (orice) 4%.
Room temperature 0.5 K.
Supply temperature 0.5 K.
Air temperature in the jet 0.2 K.
Air velocity 0.05 m}s.
The measured parameters for the two cases considered
were:
Supply height 950 mm
Supply airow rate, q
v
= 0.15 m
3
}s.
Supply outlet area, A
out
= 0.094 m
2
characteristic value

A
out
0.30 m.
Supply air velocity at outlet, U
o
= 1.60 m}s.
Supply height 300 mm
Supply airow rate, q
v
= 0.15 m
3
}s.
Supply outlet area, A
out
= 0.117 m
2
characteristic value

A
out
0.34 m.
Supply air velocity at outlet, U
o
= 1.29 m}s.
The reason for varying the supply outlet area and the
velocity in the two chosen heights was because the extended
noise damping cover of supply duct in the case with supply
height of 300 mm was missing.
3.2. Test results
The results from these tests are given in Figs. 510.
The equation in Fig. 6 shows a best line t through the
measured values. This is used to calculate the jet spread rate,
i.e. 0.081 in the equation and also the virtual origin which is
0.34. The same procedure is repeated for other angles, see
Figs. 8 and 10, which show the dierence in the jet spread
and the location of virtual origin for the two angles. The
dierences obtained are due to the vicinity of the wall and
the entrained air into the jet boundaries for the measuring
cases with 30

and 90

which leads to a higher spread rate,

0.081 and 0.074, respectively. These values are very close
to the wall jet spread rate obtained by Karimipanah [3]. For
the 60

case the jet spread rate is as low as 0.067, see Fig.

8 and the conguration in Fig. 4. For the later case, the
connement eects are less important and the jet seems to
be almost unaected by the walls.
3.3. Empirical equations
The empirical equations for each angle given below show
how the maximum velocity varies with distance from the
supply device as a function of supply area and outlet veloc-
ity. The equations have been derived from the data in Figs.
510.
Angle 30

U
max
= 3.38U
o

A
out

1.52
, (2)
Angle 60

U
max
= 2.66U
o

A
out

1.19
, (3)
Angle 90

U
max
= 2.45U
o

A
out

1.10
. (4)
T. Karimipanah, H.B. Awbi / Building and Environment 37 (2002) 13291342 1333
0
0.5
1
1.5
2
2.5
3
1 11 13 15
out
x
A
-1.5202
0
3.3784
out
U x
U A
=
max
0
U
U
max
3 5 7 9
Fig. 5. Decay of maximum velocity for angle 30

and supply height of 950 mm.

-0.10
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00
out
x
A
out
out
y
1/2
A x
A
= 0.081
1/2
y
x
Aout
x
+ 0.0278
Fig. 6. The rate of spread of the jet for angle 30

and supply height of 950 mm.

The absolute values for the distance of the virtual origin
are also determined from the gures for each angle. The
following relations show the spread of jet with distance
from the supply jet centreline and the distance of the virtual
origin for angles of 30

, 60

and 90

. It is shown in Figs.
6, 8 and 10 that the virtual origin is always downstream of
the jet centreline.
Angle 30

x
0

A
out

=|0.34|, (5)
Angle 60

x
0

A
out

=|1.90|, (6)
Angle 90

x
0

A
out

=|0.95|. (7)
3.4. Discussion of results
The distance (x) form the impingement point up to the
actual measuring point gives the best parameter for dimen-
sionless representation of the experimental data. The results
showthat velocities for dierent angles considered are some-
what dierent. Although the largest velocities occur for an-
gles of 60

and 90

, the velocities for these two angles at

the same measuring points are almost the same. The small
variations are due to measuring errors. The lower velocities
for angle 30

are due to the damping eect of the adjacent

wall.
Since the aimis to produce dimensionless correlations that
can be used for selecting impinging jet devices the results
1334 T. Karimipanah, H.B. Awbi / Building and Environment 37 (2002) 13291342
2.6571
0.00
0.50
1.00
1.50
2.00
2.50
3.00
1.00 3.00 5.00 7.00 9.00 11.00 13.00 15.00
out
x
A
max
0
U
U
max
0 out
U A
-1.1881
=
U x
Fig. 7. Decay of maximum velocity for angle 60

and supply height of 950 mm.

-0.10
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
-2.00 -1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00
1/2 0.0671 0.1273
out
out
y
A x
A
=
1/2
y
x
A
out
x
A
+
x
out
Fig. 8. The rate of spread of the jet for angle 60

and supply height of 950 mm.

for angle 90

, which gives somewhat higher values, will be

used for this purpose. Thus, if one considers a maximum
velocity at a specic distance from the supply device, it is
clear that the value at the same distance for other directions
may be lower than that for 90

.
Based on this assumption, the equation for the maximum
velocity at a distance x from the jet centreline is
U
max
= 2.45U
o

A
out

1.10
. (8)
The exponent for an ideal case should be 1.0. However,
due to friction losses and entrainment of the room air into the
jet and also because the jet momentum cannot be conserved
in a conned enclosure, a value of 1.10 is obtained from
measurements.
Concerning the supply device height above the oor, the
results show that the velocity proles for both 300 and
950 mm supply heights are almost the same. The marginal
dierence in the maximum velocities is probably due to the
smaller supply area in the case with h=950 mm which gives
somewhat higher outlet velocity. A higher supply outlet ve-
locity results in a higher momentum and can aect the jet
velocity spread over the oor. Therefore the height of the
supply outlet has no signicant inuence on the velocity pro-
les over the oor, but small inuences on the ventilation
eciency [4].
T. Karimipanah, H.B. Awbi / Building and Environment 37 (2002) 13291342 1335
0
0.5
1
1.5
2
2.5
3
1 5 9 11 13 15
out
x
A
max
0
U
U
max
0
2.4476
U x
U A
=
3 7
out
-1.0977
Fig. 9. Decay of maximum velocity for angle 90

and supply height of 950 mm.

-0.2
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
- 2 -1 0 4 9 10
1/2
y
x
A
1/2 0.0744 0.0705
out
out
y
A x
A
=
+
out
x
A
1 2 3 5 6 7 8
out
Fig. 10. The rate of spread of the jet for angle 90

and supply height of 950 mm.

4. Application to a classroom
A test room of 8.47.23 m
3
and 25 person-simulators,
placed in the room to represent a teacher and 24 students,
was used for evaluating the performance of the IJV system.
A climate chamber attached to the room was used to sim-
ulate extreme winter and summer conditions, see Fig. 11.
The heat output from the person-simulators were 2375 and
525 W was considered for lighting giving a total cooling
load of 2900 W (48 W}m
2
). In all tests the air ow rate
was 10 l}s per person and the supply air temperature was
kept constant at 15

C. The outdoor temperature was kept

at 21

C to simulate winter conditions and at +25

C to
simulate summer conditions. However, only results for the
summer condition are presented here. Results from tests un-
der winter conditions can be found in Ref. [4]. Two types
of air supply devices have been evaluated; one is based on
the IJV principle and the other on the WDV concept. The
devices were tested by measuring the air temperature, air
velocity and air quality (local mean age of air) in the occu-
pied zone. The room area was divided into 12 zones and a
stand placed in the middle of each zone was used to measure
1336 T. Karimipanah, H.B. Awbi / Building and Environment 37 (2002) 13291342
Fig. 11. Plan view of the classroom with supply positions. All dimensions in metres.
Fig. 12. Velocity vectors in a horizontal plane 10 cm from oor and age of air contours in a vertical plane: (a) wall displacement ventilation;
(b) impinging jet ventilation.
the required quantities at dierent heights. The local mean
age of air was measured by using the tracer gas decay tech-
nique at 1.2 m above oor. Further details can be found in
Refs. [2,4].
To assess the eectiveness of the two ventilation systems
for measurements and CFD simulations, well-known param-
eters given below are used [9].
Ventilation eectiveness for heat distribution or removal
(c
t
): This is similar to a heat exchanger eectiveness and is
dened by
c
t
=
1
o
1
i

1 1
i
. (9)
Ventilation eectiveness for contaminant removal (c
c
):
This is a measure of how eective the ventilation system is
in removing internally produced contamination. It is dened
by
c
c
=
C
o
C
i

C C
i
. (10)
In Eqs. (9) and (10), 1 is temperature (

C), C is the con-

taminant concentration in parts per million (ppm), the sub-
scripts i and o refer to inlet and outlet, respectively, and
() represents the mean value for the occupied zone (to a
height of 1.8 m). The values of c
t
and c
c
is dependent on the
method of room air distribution, room characteristics, heat
and contaminant sources, etc.
T. Karimipanah, H.B. Awbi / Building and Environment 37 (2002) 13291342 1337
18
18.5
19
19.5
20
20.5
21
21.5
22
22.5
0.01 0.09 0.42 0.74 0.89 1.19 1.50 2.07 2.38 2.46 2.61 2.85
Height [m]
T
e
m
p
r
a
t
u
r
e

[
o
C
]
IJV CFD
IJV measured
WDV CFD
WDV measured
Fig. 13. Mean temperature proles for two ventilation systems.
0
0.05
0.1
0.15
0.2
0.25
0.3
0.01 0.09 0.42 0.74 0.89 1.19 1.5 2.07 2.38 2.46 2.61 2.85
Height [m]
M
e
a
n

v
e
l
o
c
i
t
y

[
m
/
s
]
IJV CFD
IJV measured
WDV CFD
WDV measured
Fig. 14. Mean velocity gradients for two ventilation systems.
In addition, expressions for the predicted percentage of
dissatised (PPD) and predicted mean vote (PMV) are used
here and these are dened in Ref. [10].
5. CFD calculations
The CFD program VORTEX [11] has been used to pre-
dict the airow properties in the classroom. This program
has been developed for the simulation of airow, heat trans-
fer, mean age of air distribution, PPD and PMV in enclo-
sures. The code uses the standard c turbulence model
and has been developed for ventilation research, which may
be more suitable for ventilation simulations than the more
general-purpose codes. In the simulations, the measured
mean temperatures of the six room surfaces have been used
as boundary conditions. The number of nodes used were
49 39 37 giving a resolution of 0.17 m in the horizon-
tal plane and 0.081 m in the vertical direction. The distance
from the oor of the impinging jet outlet was 0.90 m in both
the measurements and the CFD simulations.
6. Results from CFD and measurements in classroom
Fig. 12(a) and (b) shows the air ow patterns from the
CFD results for the two supply methods at a height of
10 cm from the oor. Although, the air supply velocity of
1338 T. Karimipanah, H.B. Awbi / Building and Environment 37 (2002) 13291342
0
0.05
0.1
0.15
0.2
0.25
0.00 1.05 3.15 5.25 7.35 8.00
Distance from window, x[m]
M
e
a
n

v
e
l
o
c
i
t
y

[
m
/
s
]
.
IJV measured
IJV CFD
WDV measured
WDV CFD
Fig. 15. Predicted and measured mean velocity distribution in a horizontal plane 10 cm above the oor for two ventilation systems.
0
5
10
15
20
25
30
0.01 0.09 0.42 0.74 0.89 1.19 1.5 2.07 2.38 2.46 2.61 2.85
Height [m]
P
P
D

[
%
]
IJV CFD
WDV CFD
Fig. 16. Predicted PPD proles for two ventilation systems.
1.56 ms
1
in the case of IJV was much higher than in the
displacement ventilation case the velocity close to the oor
decays very rapidly and it is only quite large at a short dis-
tance from the jet impinging point on the oor. The spread
of the impinging jet over the oor produces a velocity near
the oor which is similar in magnitude to that of the dis-
placement ventilation system. However, the oor layer in
this case is much thinner than that in the case of the WDV.
The temperature gradients for the two cases are shown
in Fig. 13. The agreement between the measurements and
CFD simulations are good. The recommended temperature
gradient of 3 K or less is achieved.
The mean velocity gradients are shown in Fig. 14. Ex-
cept for heights up to 15 cm above the oor the velocities
are much lower than 0.15 m}s for both cases, indicating a
comfortable environment. The velocity distribution over a
horizontal plane 10 cm above the oor (see Fig. 11) is plot-
ted in Fig. 15. Only the wall displacement system shows a
tendency for a velocity higher than the recommended max-
imum of 0.15 m}s closed to the walls.
Fig. 16 shows the PPD. One can see from the gure
that in a large portion of the room PPD is below 10%
for both systems and this is acceptable for such large heat
loads [10].
T. Karimipanah, H.B. Awbi / Building and Environment 37 (2002) 13291342 1339
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
1 2 3 4 5 6 7 8 9 10 11 12
Measuring point number
N
o
r
m
a
l
i
s
e
d

m
e
a
n

a
g
e

i

[
-
]
WDV measured
WDV CFD
IJV measured
IJV CFD
Fig. 17. Predicted and measured normalised local mean age of air proles at the breathing zone (1.2 m above oor) for two ventilation systems.
19
20
21
22
23
24
25
1 2 3 4 5 6 7 8 9 10 11 12
Measuring point number
T
e
m
p
r
a
t
u
r
e

[
o
C
]
WDV measured
WDV CFD
IJV measured
IJV CFD
Fig. 18. Predicted and measured local mean temperature proles at the breathing zone (1.2 m above the oor) for two ventilation systems.
The normalised mean age of air (i.e. the local age divided
by that at the exhaust) is shown in Fig. 17 for all the mea-
suring points at a height of 1.2 m above the oor (breathing
zone level). There is some agreement between the WDV
and the IJV results at some of the points but not at others.
This may be due to the dierence between the momentum
in the two systems and its interaction with the local
buoyancy forces. The CFD results show similar trends
to measurements.
The temperature and velocity proles at all measuring
points 1.2 m above the oor (breathing zone) are shown in
Figs. 18 and 19. Both systems show good temperature dis-
tributions but the WDV gives higher values. Considering
the velocity eld for both cases one can see that the ve-
locities are very low and there is little tendency to draught
risk.
Ageneral consideration of all the parameters studied show
that both systems can handle the extreme situation they
have been exposed to with some small dierences between
them.
The predicted contaminant removal eectiveness (c
c
),
was 140% for WDV and 130% for IJV indicating the ability
1340 T. Karimipanah, H.B. Awbi / Building and Environment 37 (2002) 13291342
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
1 2 3 4 5 6 7 8 9 10 11 12
Measuring point number
M
e
a
n

v
e
l
o
c
i
t
y

[
m
/
s
]
WDV measured
WDV CFD
IJV measured
IJV CFD
Fig. 19. Predicted and measured local mean velocity proles in breathing zone (1.2 m above the oor) for the two ventilation systems.
Fig. 20. Smoke visualisation of impinging jet ventilation.
of both systems to remove the contaminants in an eective
way. The corresponding value of the predicted ventilation
eectiveness for heat distribution for the two cases was
1421 and 142 respectively, which again indicates a high
and very close value for the two systems. The PMVs were
0.63 and 0.61, respectively, which are slightly less than
0.5 that is suggested by Fanger [10] to be acceptable.
Fig. 20 shows four sequences of smoke visualisation for
an impinging jet supply device. One can see that when the jet
reaches the heat source (sitting person) a plume forms and
due to sucient momentum the jet continues along the oor.
This is the advantage of impinging jet ventilation compared
to a displacement system, in which case the ow sometimes
has insucient momentum to continue pass the heat source
and its totally consumed by the plume.
Similar types of impinging jet supply devices (Air Queen)
have been installed in dierent buildings, e.g. schools, of-
ces, industrial buildings, etc. The device is similar in looks
to a duct and is mounted onto a wall. It supplies air down-
ward to the oor and contaminated air is exhausted from
ceiling level. Four applications of impinging jet ventila-
tion in Sweden are shown in Fig. 21(a)(d). One of these
T. Karimipanah, H.B. Awbi / Building and Environment 37 (2002) 13291342 1341
Fig. 21. Practical applications of impinging jet ventilation: (a) System 3R international factory in Stockholm; (b) Heavy metal industries in Fagersta;
(c) Str omsbro school in G avle; (d) Swedish taxes oce in Stockholm.
represents an installation in a school, one in an oce build-
ing and two in industrial buildings. Fig. 21(a) shows an
installation of impinging jet supply in System 3R Interna-
tional, which was installed to replace displacement devices
in a building where precision instruments}tools are man-
ufactured. When displacement ventilation was used there
were many complains from the workers and after their re-
placement with impinging jet systems no compliment was
reported. Fig. 21(b) shows IJV installed in a factory for
heavy metal industry. Fig. 21(c) shows the application in the
Str omsbro School in G avle and Fig. 21(d) is in the Swedish
Taxes Oce in Stockholm. Since installing the impinging
jet ventilation system in all these buildings no complains
have yet been reported.
1342 T. Karimipanah, H.B. Awbi / Building and Environment 37 (2002) 13291342
Another advantage of impinging jet ventilation is that
the duct itself has a sound damping eect, which reduces
aerodynamic noise at the supply terminal.
7. Conclusions
The results presented in this paper have shown that a oor
level air distribution can handle a full room heat load in an
acceptable manner. Although both the WDV and the IJV
systems show similar tendencies, some small dierences in
their performance were observed. Because of the better bal-
ance between buoyancy and momentum forces the IJV sys-
tem shows slightly better mean age of air and velocity
distributions. The experiments show that after a short dis-
tance from the impinging region and due to the damping
eect of the oor, the jet velocities have decreased to an
acceptable level. Furthermore, the IJV system can also be
used for both heating and cooling purposes.
In this paper we have evaluated the IJV system un-
der certain conditions which are by no means exhaustive.
Therefore, this new system may need further studies by
ventilation researchers and designers to fully understand
its performance under various conditions experienced in
practice.
Acknowledgements
The authors gratefully acknowledge Prof. Mats Sandberg,
Leif Claesson (in BMG, G avle), Mr Bengt Svensson and
Lars Berthilson (VVESS Consulting AB, G avle), Per-Johan
Ohlsson, Johan Kostakis and

Orjan Josefsson for their
valuable contribution to this work.
References
[1] Etheridge D, Sandberg M. Building ventilation: theory and
measurement. UK: Wiley, 1996.
[2] Karimipanah T, Sandberg M, Awbi HB. A comparative study of
dierent air distribution systems in a classroom. In: Awbi HB, editor.
Air distribution in rooms: ventilation for health and sustainable
environment. Proceedings of the ROOMVENT 2000, Vol. 2. Oxford:
Elsevier, 2000. p. 10138.
[3] Karimipanah T. Turbulent jets in conned spaces. Ph.D. thesis, Royal
Institute of Technology, Sweden, 1996.
[4] Kostakis J, Josefsson

O. Thermal comfort and air quality in a
classroom. B.Sc. thesis, University College of G avle, 1999 (in
Swedish).
[5] Beltaos S, Rajaratnam N. Plane turbulent impinging jets. Journal of
Hydraulic Research 1973;11:2959.
[6] Gutmark E, Wolfshtein M, Wygnanski I. The plane turbulent
impinging jet. Journal of Fluid Mechanics 1978;88(Part 4):73756.
[7] Berthilson L, Ohlsson PJ. Down-to-oor Impinging Jet supply
deviceair exchange eciency. B.Sc. thesis, University College of
G avle, 1997 (in Swedish).
[8] Rajaratnam N. Turbulent jets. Amsterdam: Elsevier, 1976.
[9] Awbi HB. Energy ecient room air distribution. Renewable Energy
1998;15:2939.
[10] Fanger PO. Thermal comfort. New York: McGraw-Hill, 1972.
[11] Gan G, Awbi HB. Numerical simulation of the indoor environment.
Building and Environment 1994;29(4):44959.