Table of

Contents

FORECAST OF THE PROPERTIES OF A COLD VERTICAL WALL JET COMING

FROM A FAN COIL

D. Marchal, P.J. Vialle

Centre Technique des Industries Aérauliques et Thermiques

27-29 Boulevard du 11 Novembre 1918
69603 Vileurbanne Cedex, FRANCE

ABSTRACT V : air blowing velocity (m/s)

The study of the flow in a room
cooled by a fan-coil pointed out how the
form of air flow and comfort could be
influenced by the characteristics of the cold
jet blowing out. It is based both on practical
experiment and on numerical simulation
using CFD code. Combining these methods
allowed a large number of configurations to
be studied, in association with different
conditions for the appliance.
Using the results in combination
enabled a relation to be established between
the problem data, the device characteristics
and the comfort conditions obtained. A
simple rule was derived from this, which can
be used in air-conditionning premises, in
order to make the right choice or scaling of
the air-conditionning appliance depending on
its conditions of use.
KEYWORDS
Comfort
Mixing ventilation
Numerical simulation
Full-scale experiments
SYMBOLS
f: width of the fan-coil unit supply grille (m)
l: length of the fan-coil unit supply grille (m)
g : gravity acceleration (ms-2)
h' : distance between the blowing plane and
the ceiling (m)
β: air volume coefficient of expansion (K-1)
∆ t : temperature difference between intake
and blowing (K)
ν : kinematic viscosity (m2/s)
Re : Reynolds number
Fr : Froude number
INTRODUCTION
To cool inhabited premises that are
subject to heat loads, cold-air ventilation is
generally used. Cold air is denser than
ambient air and tends to fall, making it
difficult to distibute to every places where it
is needed. To remedy this difficulty, the
property of air jet to adhere to the walls,
called Coanda effect, is resorted-to. The
good performance of the cooling system is
so, strongly dependent of the characteristics
of the jet blowed out by the fan-coil. This is
again more significant if the fan-coil is
located at the basement of a window. Due to
the low position of the blowing plane in this
case, the jet has first to go upward what is
opposite to its inclination to fall down.
The work discussed here is designed
to define the conditions that has to fulfill the
flow blowed out of a fan-coil located at the
basement of a window, in order to assume a
good comfort. It is based on both
experiment and numerical simulation, the
latter being used as a means of extending the
scope of the experimental results. The study
achieves a non-dimensional presentation of
the results, which can be used in many Variable (unity) Minimum Maximum
situations for practical design of appliances. value value
room length (m) 3.0 6.0
METHODS room width (m) 2.5 4.5
First part of this work [1] consisted in room height (m) 2.5 3.5
experimentations performed in a test
chamber with scope for varying the heat thermal load (W) 650 1600
loads and the operating conditions of the fan
coil installed. Two types of measurements l (m) 0.5 1.2
were taken : measurements relating to the f (m) 0.03 0.1
flow of the cold air jet from the device, and
the measurements of the velocities and α jet angle (°) -30 30
temperatures in the occupied area of the
room. The latter measurements enabled to Table 1 Variation range for the parameters
characterize the usual criteria to be used to studied
describe the comfort of the occupants,
particularly the ADPI criterion [2], which In each of these cases, several simulations
takes into account both the were performed in which the blowing
homogeneousness of temperature and the velocity of the fan coil (assumed to be
lack of excessively ventilated areas. The vertical and uniform) was varied
experimental work covered an important simultaneously with the corresponding
number of situations, albeit limited by the temperature, so as to compensate the heat
dimensions of the test chamber and the air- load in order to achieve always the same air-
conditionning appliance chosen. It has also intake temperature (25°C). This exploration
enabled the accuracy of the numerical of each configuration was so conducted as
simulations performed with a CFD code to highlight the variations of the comfort
(FLUENT) to be verified for a number of criterion (ADPI).
cases, as regards both the velocity and
temperature profiles in the air jet and the RESULTS
velocity and temperature fields in the
occupied zone. Numerical part
Three dimensionnal numerical The synthesis of the results obtained
simulation was then used to explore a large after the parametric study pointed out the
number of situations which may arise in ability to link the comfort criterion ADPI to a
practice. Every cases studied concerned non dimentional number X defined in the
empty, oblong rooms with heat loads following way :
applied, on the one hand by conduction
through the upper part of the building-front 1
against which the fan coil is installed X= (1)
Re( f ). Fr 2 ( h' )
(glazing), and on the other hand to the
greatest extent, by internally-generated heat this is the reciprocal of the product of
assumed to be uniformly distributed in the two usual non-dimentional numbers :
occupied zone. The study consisted of
systematically varying the parameters listed - Reynolds number relating to the width
in table 1, which shows the interval f of the supply grille of the device :
explored, representing some twenty f .V
configurations. Re = (2)
ν
 in which V and ν refer to the blowing velocity, is designed to account satisfactorily
velocity and kinematic viscosity of air for the effects of inertia, diffusion and gravity,
respectively; and which involve forces, the relationships of
which taken in pairs are represented by the
- the Froude number relating to the Reynolds and Froude numbers. The rapid
heigt h' to which the jet has to rise : change of the ADPI criterion from 0.9 to 0.3
V around a value for X of 1.9.10-5 corresponds
Fr ( h' ) = (3)
h' . g. β. ∆t to the transition between the two flow
situations described in the figures 3 and 4.
Figures 1 and 2 diagrams, relating to a
vertical uniform jet (α = 0) show how the
ADPI indice varies versus the X number, for a
constant heat load and different values of the
dimensional parameters of the case regarded :
width f of the supply grille (figure 1) or h'
height (figure 2).

1.000 f=0.06 fan coil

f=0.045
unit
0.800
f=0.03

0.600 f=0.1
Figure 3 Sketch of the flow situation with
0.400 good comfort conditions
0.200

0.000
5.0E-06 1.0E-05 1.5E-05 2.0E-05 2.5E-05 3.0E-05 3.5E-05
1/[Re(f).Fr(h')^2]

room
Figure 1 Variation of the X parameter for
different values of the grille's width f.

1.000 h'=2.1

h'=1.8
0.800
h'=2.3

0.600 h'=2.8
Figure 4 Sketch of the flow situation with
bad comfort conditions
0.400

0.200
0.000
5.0E-06 1.0E-05 1.5E-05 2.0E-05 2.5E-05 3.0E-05 3.5E-05 4.0E-05
1/[Re(f).Fr(h')^2]
Figure 2 Variation of the X parameter for
different values of the h' height
The choice of this expression of X, of
which value increases when the temperature
difference is increased at the expense of
In figure 3 (X < 1.9.10-5) the jet
blowed out of the grille has a sufficient
energy to reach the ceiling and then surround
with cooled air the occupation zone, wich
implies a satisfying comfort; at the opposite
on the figure 4 (X > 1.9.10-5) the jet falls
down due to the effect of the negative
buoyancy forces and then only the vicinity of
the appliance is cooled.
A study concerning the length of the
room pointed out that this parameter hardly
affects the value of X for which the transition
between comfort and uncomfort conditions α α

appears, see table 2.

Room length X value for the

(m) transition
3 2.4.10-5
4.5 1.9.10-5
6 1.8.10-5

Table 2 X values for the transition between α α

comfort conditions for different values of the
room length.

Excepted for the angle of the jet α, the

results obtained show that the other
parameters listed in table 1 have only
secondary effects in the change of the flow
conditions.
Till now the results presented assumed
Figure 5 Sketchs of converging and
that the velocity profile at the exit of the grille
diverging jets
of the fan-coil was uniform, without any
horizontal component. This hypothesis is not
Though the value of X for which the
very representative of real conditions, but it is
transition appears between the flow
hardly avoidable in order to make a general
conditions is, as we saw before, 1.9.10-5 for
study. However in order to quantify the
the parallel jet, it increases to 2.2.10-5 for the
effects of such perturbations, some cases have
30° converging jet. The transition is so
been studied with the jet either converging or
obtained for a lower flow rate which tends to
diverging as shows the front views of the
show that the converging jet is more likely to
appliance in figure 5.
provide a good comfort.
Different values of α angle (angle
between the vertical direction and the
extremities of the jet) have been studied : 30° 1

for the converging jet, and 15° and 30° for

the diverging one, other parameters being 0.8

kept constant. The results presented in a

relationship form between the X value and the 0.6

ADPI comfort criterion in figure 6 show how

0.4 Parallel (aplha=0)
this comfort indice is sensitive to this
Convergent 30°
perturbation.
0.2 Divergent 15°

Divergent 30°

0
5.00E-06 1.00E-05 1.50E-05 2.00E-05 2.50E-05 3.00E-05
1/[Re(f).Fr(h')^2)

Figure 6 Variations of the ADPI indice

versus the X parameter depending on
the convergence of the jet.
 This is the opposite for the diverging jet
for which X value has to be lower than
0.7.10-5 in order to obtain a satisfying value
of the comfort criterion. Noting this reduce
the scope of the results listed before ; but
neverthless it allows to realize how the
divergence of the jet observed for some fan-
coils is negative for a good working.
Actually, it induces to make them run with
higher flow rates what is unfavorable for a
good acoustical comfort.
Experimental verification
Previous paragraph conclusions are completly
depending on numerical results. They have
been verified using the test facility and the
fan-coil already used for the first part of the
study [1]. Only the width f of the blowing
grille has difered, some of the experiments
have been performed with a rating. This law derived from combining
0.06 m f value and others with 0.09 m. For
each case, blowing temperature and volume
flow rate were both modified in order to
balance the global thermal load which was
kept constant (1000 W). The average rating
velocity profiles at the exit of the blowing
grille have been measured by laser Doppler
anemometry ; ADPI's method has been used
to determine the comfort indice in the
occupied zone (80 points). The results are
shown in the diagram of the figure 7.
1.000
f=0.06
0.800
f=0.09
0.600
0.400
0.200
0.000
1.00E-06 1.00E-05
X=1/[Re(f).Fr(h')^2]
Figure 7 Experimental results
It can be seen that the two curves
representing the variation of the criterion
comfort ADPI for two different values of the
blowing grille's width f are relatively close.
However, they differ from the curves
obtained by the numerical simulation with an
homogeneous vertical jet (figures 1 and 2) but
they look like the curve of the diagram of the
figure 7 corresponding to a diverging jet with
a maximum value of the ADPI index for
about X = 0.5.10-5. This result is so to be
attributed to the diverging aspect of the jet
blowed out by the fan-coil used for these
experiments.
CONCLUSION
The study presented here leads to a
non-dimensional relationship which allows
foreknowledge of the flow behaviour caused
by a fan-coil unit used for air-conditioning an
inhabited room and the resulting comfort
numerical results obtained in the case of
strictly vertical and homogeneous velocity
profile for the jet at the exit of the blowing
grille, can be used in practical cases
respecting this condition in order to make the
correct choice or scaling of the air-
conditionning appliance, allowing for the
conditions of its use.
The reliability of the formulation
presented here obviously depends on the
following assumptions, basically regarding the
homogeneousness heat loads in the occupied
area of the room and that of the air flow from
the appliance blower nozzle. Concerning that
latter hypothesis, a study to determine the
sensitivity of the comfort indice to the jet's
potential lack of parallelism showed how a
diverging jet is a significant factor of lower
comfort. This is a point to take into account
for the fan-coil's manufacturing.
1.00E-04
 REFERENCES

[1] A. Meslem, C. Isnard, P. Barles.

Experimental study on the air flow
characteristics of air-conditioned office
rooms.
ROOMVENT 96, 17th-19th July 1996,
Nagoya, Japan

[2] Method of testing for room air diffusion.

ANSI/ASHRAE 113-1990, American
National Standards Institute, 1990.