Professional Documents
Culture Documents
Ventilation Roofs
Ventilation Roofs
a r t i c l e
i n f o
a b s t r a c t
This paper aims at optimizing the performance of roong components equipped with a ventilated interspace
in order to reduce the heat load inside the animal houses. By means of a theoretical model and experimental
tests, the three-dimensional prole of the air temperature in the duct is determined. Consequently the air
temperature inside a typical house for livestock is calculated. The air ow is considered uniform inside the
rectangular duct and the thermal problem is solved by adopting H2 boundary conditions. The air temperature
in the house is then calculated and the optimal duct thickness minimizing it was determined.
2009 Elsevier Ltd. All rights reserved.
1. Introduction
In livestock housing the control of the climatic parameters is a
determinant factor for production. In fact the metabolic activity of
the animals varies remarkably with the hygro-thermal parameters, in such a way that as more they diverge from the optimal
values as more the productive performance worsens. While the
lower optimal thresholds (about 10 C for the dairy cows and a bit
higher for pigs and poultry) can easily be respected in winter time
with simple constructive solutions, the upper thresholds (around
24 C for all the species) are very often exceeded in the hot
climate conditions (being uneconomical the adoption of air
conditioning systems). Indeed the high density of animals
producing metabolic heat makes the inside temperatures higher
than the outside for nearly the whole day. In such extreme
conditions, with production losses of 20% or more, an even small
reduction of the internal temperature can lead to signicant
economical benets. Therefore it is interesting to nd out
constructive solutions capable of maintaining the inside temperature as low as possible during all the day, minimizing the solar
energy load and maximizing the heat discharge during the night.
To this purpose, in addition to a careful control of ventilation, the
adoption of shield components combining a high thermal
resistance to the solar radiation with a low resistance to the
heat transfer from inside to outside can be very helpful. This is
mainly relevant for the roof, the part of the building shield more
involved in the heat exchange, night and day. Considering the
requirements above mentioned both the insinuated and the
cW
AT
=
Af
!
A2 T
A2 T
A2 T
+
+
:
An2
A2
Af2
P. Liberati et al. / International Communications in Heat and Mass Transfer 36 (2009) 432437
433
Nomenclature
a
b
c
D
Gr
Gz
h
k
L
Nu(z)
Nufd
Pr
q
R
Re
s
T0
Ti
Tb(z)
T(.)
Tw
Qlamp
Qw
Qroof
W
x,y,z
Greek symbols
dimensionless temperature
y=
;
a
f
1
z= 2
;
=
Gz
a W
T T0
=
:
qa
X
X
m=0 n=0
y
:
Fmn z cosnx cos m
Introducing the expression (6) into the Eq. (4) and integrating, for
x varying from 0 to 1 and y varying from 0 to , thanks to the property
of the orthogonal complete system of cosines, an Eq. (2) is determined
permitting to obtain the terms of the function system Fmn(z).
Developing the calculations and solving the integrals [3], after
some algebra we obtain:
"
!#
2 2
z
2 X
1
n
y
+ 2
:
x; y; z =
1 exp 2 z
cos n
2
n=1 n
A
A
A
j
j
j
=
=
= 0:
Ax x = 0
Ax x = 1
Ay y =
2 X
1
n2 2
+
2
w y = 0; z =
exp 2 z
2
3
n=1 n
X
1
;
=
2
6
n
1
X
1n
;
=
2
12
n
1
X
1
:
=
4
90
n
1
8
9
10
being [3]:
L= 2
:
a W
Fig. 1. Cross section of the ventilated roof and the experimental facility.
434
P. Liberati et al. / International Communications in Heat and Mass Transfer 36 (2009) 432437
2
6
=
1 + w b
1+
1
1
6
2
P
n=1
1
n2
:
2 2
exp n2 z
11
b2 W
:
12
2 X
1
y
y2
=
y+
:
cos
n
2
2
3
2
n=1 n
13
The power transferred from the building to the air owing into the
R
cooling channel is aq = a 0 Ti R Tw d1, that is:
RL
=
a
i w dz:
14
Fig. 5. Length of the entrance region as function of the air velocity, for different
thickness of the rectangular duct.
Xs
ki
1
:
15
R
L
3
23 X
1
n2 2
i =
+
+
+ 4
exp 2 L :
a
2
3
45L
L n = 1 n4
The series t into the Eq. (16) is strongly convergent; the numerical
analysis permits to deduce a simple exponential function capable of
expressing the building internal air temperature as:
3
R
L
KL
+
+
+
exp 2
i =
a
2
3
45L
45L
!
17
being K = 1.0113.
The dimensionless temperature i is resulting as a function of and
L; therefore the air temperature of the building Ti is depending on , L
and q. The functional relationship between i and q is simply
proportional, whilst more complicated is the dependence of the
temperature on and L.
The aim of this work is to nd out the optimal cooling conditions,
that is the height of the ventilation duct able to minimize the
internal building temperature, given the duct length and the air inlet
velocity. To determine such height we have to nd out the value
minimizing the building internal temperature by analysing the
derivative:
di
L
6 X
1
n L
=
+ 4
exp
2
d
15L
2
2
L 1 n4
!
2 2
4 X
1
n L
1
exp
+ 2
+ :
2
3
1 n
2
2
16
2 2
!
18
The Figs. 6 and 7, where the graphs of the temperature i and of its
derivative versus are reported as a function of (following the L
variation), show that the internal air temperature i is a monotonous
function decreasing with (in the physical domain 0 b b 1), tending
to a constant value increasing with L.
The derivative versus of the internal air temperature is a
monotonous function with a negative value, tending to 0.
The constant value obtained by i as a function of L represents
the minimum possible value min of the air temperature inside the
P. Liberati et al. / International Communications in Heat and Mass Transfer 36 (2009) 432437
435
p
R
+ 0:7522 L:
a
19
X
opt
L
1
n
+
+
exp 2 L
3
45L
2opt
L n = 1 n4
opt
3opt
2 3opt
4
2 2
p
= 1:01 0:7522 L :
20
opt = 1:7735 L
0:5213
21
436
P. Liberati et al. / International Communications in Heat and Mass Transfer 36 (2009) 432437
Table 1
Inside air temperature, measured and calculated by the model, in day night (C).
Ti,
Measured
Model
Ti,
24.0
24.0
Ti, A Ti,
23.7
23.6
0.3
0.4
For the 95% of the testing time the air ow motion has revealed to
be a forced convection (Re2 NN Gr), and for the 65% the convection was
laminar too, especially in the case B. The condition of forced
convection was favoured by the northsouth orientation of the pitch
having the cross section of the ventilated channel perpendicular to the
dominating wind. The wind was mainly a breeze with peak persistent
values of about 1.0 m s 1 and average values of about 0.5 m s 1.
The validation of the model has been carried out selecting, among
the collected data, those periods where the steady conditions of the air
temperature, solar radiation and wind speed were respected. The
validation has been made in the nocturnal situation only because the
strong solar contribution didnt allow to fully satisfy the boundary
condition of adiabaticity assumed in the model.
The dimensionless temperature of the air in the room (i) was
calculated by means of the Eq. (16). The i values so obtained have
then been transformed by means of Eq. (2), made explicit in T. The q
parameter included in Eq. (2) was calculated using the experimental
data through the stationary thermal balance of the building as:
22
Qlamp = Qw + Qroof
23
Table 2
Heat removed by ventilation through the duct, total and per unit of the duct section, for
the two tested roofs.
Case A
Case B
Total (W)
Specic (W m 2)
Total (W)
Specic (W m 2)
51
1126
53
2266
Fig. 10. bopt as function of air velocity (W), for the tested roofs.
to the duct axis, we can see, again from the Eq. (22), that with
0.047 bopt 0.026 m all the anemological conditions of the experimental site inuencing the optimal dimensioning of b are covered
(Fig. 10).
5. Conclusions
The proposed method makes it possible to determine the prole of
the air temperature inside the ventilated duct and the air temperature
inside the room underneath when the thermal ow q is known, under
the hypothesis that the motion is laminar with forced convection, and
the upper roong sheet is perfectly insulated from the outer
conditions. In the case q is not known, its however possible to nd
out the optimal conditions in order to cool the room, by making
evident a relationship between the thickness of the ventilated
duct and an dimensionless parameter L, directly proportional to
the length of the channel and inversely proportional to the air velocity at the entrance of the channel (easy to be calculated as a function
of the wind speed and direction and of the discharge losses at
the inlet).
Once the width of the channel is xed, the optimal height is a
monotonous and increasing function of the parameter L; therefore, for
the same air velocity W, opt increases with the increase of the length
of the duct , whilst for an equal length the optimal height decreases
when W increases.
So for an animal house the ventilated duct should be designed
when the anemological situation of the site is known. If the wind
conditions are such to determine a laminar motion inside the duct, the
optimisation should be pursued taking into account that the air
temperature inside the room cannot decrease below the value:
Tmin = T0 + q R + 0:140
r!
:
W
24
0:5213
25
P. Liberati et al. / International Communications in Heat and Mass Transfer 36 (2009) 432437
The results show also how relevant is the presence of the thermal
entrance region that in many real cases can reduce to a negligible
quantity the length of the region of the thermally fully developed ow.
In such situations it would be rough to arrange numerical simulations
based on the temperature proles of the air inside the duct and on the
pertinent heat transfer coefcients (or Nusselt numbers) deducted by
the wide bibliography on the thermally fully developed ow.
The three-dimensional analysis, or at least the bi-dimensional one
but rigorously considering the axial coordinate z, becomes therefore
essential for a proper numerical simulation aimed at designing and
dimensioning the ventilated ducts.
Finally, it has been put into evidence that the optimal dimensions
of the duct cannot be calculated in abstract terms but, as shown by the
model and by the experiments, are depending on the anemological
characteristics of the site.
437
References
[1] R.K. Shah, A.L. London, Laminar Flow Forced Convection in Ducts, Academic Press,
New York, 1978, pp. 196208.
[2] M. Spiga, G.L. Morini, The developing Nusselt number for slug ow in rectangular
ducts, Int. J. Heat mass Transfer 41 (1998) 27992807.
[3] I.S. Gradshtein, I.M. Ryzhik, Table of Integrals, Series and Products, Academic Press,
New York, 1965.
[4] M. Spiga, G.L. Morini, The thermal entrance length problem for slug ow in
rectangular ducts, ASME J. Heat Transfer 118 (1996) 979982.
[5] S.X. Gao, J.P. Hartnett, Analytical Nusselt number predictions for slug ow in
rectangular ducts, Int. Commun. Heat Mass Transf. 20 (1993) 751760.
[6] M. Spiga, G.L. Morini, Nusselt numbers in laminar ow for H2 boundary conditions,
Int. J. Heat mass Transfer 39 (1996) 11651174.