Content uploaded by T. Boulard
Author content
All content in this area was uploaded by T. Boulard on Feb 27, 2019
Content may be subject to copyright.
Biosystems Engineering (2006) 93 (4), 427–436
doi:10.1016/j.biosystemseng.2005.01.001
SE—Structures and Environment
Effect of Vent Openings and Insect Screens on Greenhouse Ventilation
N. Katsoulas
1
; T. Bartzanas
1
; T. Boulard
2
; M. Mermier
3
; C. Kittas
1
1
Department of Agriculture Crop Production and Rural Environment, School of Agricultural Sciences, University of Thessaly,
Fytokou St. N. Ionia Magnisias, GR-38446, Greece; e-mail of corresponding author: ckittas@uth.gr
2
INRA-URIH 400, Route des Chappes, BP 167, 06903 Sophia Antipolis, France
3
INRA d’Avignon; Unite
´de Bioclimatologie, Site Agroparc, 84914, Avignon Cedex 9, France
(Received 13 April 2004; accepted in revised form 9 January 2005; published online 10 March 2006)
The objective of this work was to experimentally investigate the influence of vent type (side, roof or both) and
of an anti-aphid insect screen used to prevent insect intrusion on the ventilation rate of a round arch with
vertical side walls, polyethylene covered greenhouse. The greenhouse was equipped with two side roll-up vents
and a flap roof vent located at the University of Thessaly near Velestino in the continental area of Eastern
Greece. Microclimate variables as well as the airflow rate were measured during summer. Two measuring
methods were used for the determination of ventilation rate: (a) the decay rate ‘tracer gas’ method, using
nitrous oxide N
2
O as tracer gas, and (b) the greenhouse ‘energy balance’ method. In order to study the effect
of vent type on ventilation rate, in a greenhouse with an anti-aphid insect screen in the vent openings, airflow
was determined during periods with ventilation being performed by: (i) roof, (ii) side or (iii) both roof and side
vents. Furthermore, in order to study the effect of insect proof screen on airflow, measurements were also
carried out during periods that ventilation was performed by side vents without a screen in the openings. A
good correlation was found between the air exchange rate values calculated using the two methods, with the
values obtained by the tracer gas method being slightly lower than those obtained by the energy balance
method. Furthermore, the data of ventilation rate obtained by the tracer gas method fitted better to the model
used for the prediction of ventilation rate. In addition, the use of anti-aphid screen in vent openings caused a
33% reduction in greenhouse ventilation rate. From greenhouse ventilation performance point of view, it was
found that the most effective vent configuration was the combination of roof and side vents, followed by side
vents only (46% reduction in ventilation), while the least effective was roof vent (71% reduction ventilation).
r2006 Silsoe Research Institute. All rights reserved
Published by Elsevier Ltd
1. Introduction
The use of insect screens in Mediterranean green-
houses reduces insect migration on the crop and
subsequent crop damage (Teitel et al., 1999), thus
reducing the need for pesticide application; on the
opposite the screens impede ventilation, intensifying the
problem of high internal temperatures (Kittas et al.,
2002;Soni et al., 2005).
Natural ventilation is the most frequently used
method for cooling the greenhouse atmosphere and
maintaining more suitable conditions for plant growth
(Boulard et al., 1997;Kittas et al., 1997;Luo et al.,
2005). The efficiency of natural ventilation depends on
wind speed and the inside to outside air temperature
difference.
Methods which have been applied to measure green-
house ventilation performance include measurement of
flow and pressure differences between inside and
outside, with the ‘tracer gas’ and ‘energy balance’
methods being the most widely used (De Jong, 1990;
Fernandez & Bailey, 1992;Kittas et al., 1996;Wang &
Deltour, 1996;Kittas et al., 2002;Demrati et al., 2001;
Shilo et al., 2004).
Sophisticated techniques have recently been devel-
oped for visualisation and quantification of airflows
created by natural ventilation, including the use of sonic
anemometers (Wang et al., 1999;Boulard et al., 2000;
ARTICLE IN PRESS
1537-5110/$32.00 427 r2006 Silsoe Research Institute. All rights reserved
Published by Elsevier Ltd
Teitel et al., 2005) and greenhouse models in wind
tunnels (Lee et al., 2003). Advanced models to study and
characterise natural ventilation in greenhouses include
elaborate computer software tools, such as computa-
tional fluid dynamics (CFD) programs (Mistriotis et al.,
1997;Boulard & Wang, 2002;Bartzanas et al., 2004).
One of the issues that has received little attention is
the influence of vent configuration on greenhouse
ventilation performance. Montero et al. (2001) investi-
gated and compared four different vent configurations
on greenhouse ventilation rates using wind tunnel
facilities. Kittas et al. (1997) examined the ventilation
performance of a tunnel bi-span greenhouse equipped
with side and roof openings and found that, for low
wind velocities (o25ms
1
), the combination of roof
and side vents was more efficient than roof vents only.
The reduction of ventilation caused by different kinds
of screens (anti-thrip, anti-aphid and shade screens) was
quantified by Montero et al. (1997),Fatnassi et al.
(2002) and Kittas et al. (2002), while the resistance of
insect screens was investigated using the Bernoulli’s
approach (Munoz et al. 1999;Teitel et al., 1999;
Fatnassi et al., 2003) or based on the flow through
porous media, using the Forchheimer equation (Miguel
et al., 1997).
The aim of the present study was:
(1) to compare the two most widely used methods for
greenhouse ventilation rate measurement, namely
the ‘tracer gas’ and the ‘energy balance’ methods;
(2) to determine the effect of an insect screen on wind
driven ventilation; and
(3) to examine the influence of vent configuration on
greenhouse ventilation performance.
2. Theory
2.1. Greenhouse energy balance
In a greenhouse, energy inputs equal the sum of the
energy losses and greenhouse transient energy content. The
energy balance equations used in the present study are
similar to those described by Demrati et al. (2001).Airflow
rate Gin m
3
s
1
was deduced by the following equation:
G¼AgRnet Fg
AcKDTioþChDTic
ðÞ
raCpDTioþlDHio
(1)
where: A
g
and A
c
are greenhouse soil and cover area,
respectively, in m
2
;R
net
is the net radiation measured
ARTICLE IN PRESS
Notation
A
c
greenhouse cover area, m
2
A
g
greenhouse ground area, m
2
A
R
roof opening surface area, m
2
A
S
side opening surface area, m
2
A
T
total opening surface area, m
2
C
d
discharge coefficient, dimensionless
C
h
convective heat exchange coefficient be-
tween the inside air and the cover,
Wm
2
1C
1
C
p
specific heat of air at constant pressure,
Jkg
1
[air] K
1
C
w
global wind-effect coefficient of venti-
lation, dimensionless
ewater vapour pressure, kPa
F
g
thermal flux in the soil, W m
2
Gventilation flow rate m
3
[air] s
1
G
0
leakage ventilation flow rate m
3
[air] s
1
ggravitational constant, m s
2
hvertical distance between the midpoint of
side and roof openings, m
Kglobal sensible heat loss coefficient of the
greenhouse through the cover, W m
2
1C
1
Nair exchange rate, h
1
N
energy
air exchange rate calculated by the energy
balance method, h
1
N
tracer gas
air exchange rate calculated by the tracer
gas method, h
1
R
2
coefficient of determination
R
net
net radiation, W m
2
R
s
total solar radiation inside the green-
house, W m
2
Ttemperature, 1C
T
c
cover temperature, 1C
T
i
mean air temperature, 1C
T
l
leaf temperature, 1C
T
o
outside air temperature, K
uoutside wind velocity, m s
1
V
g
greenhouse volume, m
3
DHi2oabsolute humidity difference between in-
side and outside air, kg [vapour] kg
1
[air]
DT
ic
temperature difference between inside air
and the plastic cover, 1C
DT
io
temperature difference between inside and
outside air, 1C
llatent heat of vaporisation of water,
Jkg
1
[vapour]
r
a
air density, kg [air] m
3
[air]
N. KATSOULAS ET AL.428
above the crop and below the plastic cover in W m
2
;F
g
is
the thermal flux in the soil in W m
2
;Kis the global
sensible heat loss coefficient of the greenhouse through the
cover in W m
2
1C
1
;DT
io
is the temperature difference
between inside and outside air in 1C; C
h
is the convective
heat exchange coefficient between inside air and the plastic
cover in W m
2
1C
1
;DT
ic
is the temperature difference
between inside air and the plastic cover in 1C; r
a
is the air
density in kg m
3
;C
p
is the air specific heat at constant
pressure in J kg
1
1C
1
;listhelatentheatofwater
vaporisation in J kg
1
;andDH
io
is the absolute humidity
difference between inside and outside air in kg kg
1
.
2.2. Ventilation models
Based on the application of Bernoulli’s equation, G
can be derived by taking into account the two main
driving forces of natural ventilation: the wind and stack
effects (Boulard & Baille, 1995;Baptista et al., 1999).
Kittas et al. (1997) proposed calculating Gby
G¼Cdffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ARAS
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
A2
RþA2
S
q
0
B
@1
C
A
2
2gDTio
To
h
þAT
2
2
Cwu2
v
u
u
u
u
t(2)
for a greenhouse equipped with side and roof openings
and by
G¼AT
2Cdffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2gDTio
To
h
4þCwu2
s(3)
for a greenhouse equipped with roof openings only,
where: A
R
,A
S
,A
T
are the roof, sides and total openings’
surface area in m
2
,gis the gravitational constant in
ms
2
,uis the outside wind speed in m s
1
,C
d
is the
discharge coefficient, C
w
is the wind effect coefficient, T
o
is the outside air temperature in K and his the vertical
distance between the midpoint of side and roof open-
ings, in m.
The stack effect can be significant in greenhouses
equipped with roof and side openings, while, for green-
houses equipped with roof or side openings only, the
potential is small because the vertical dimension of vents
is small. Generally, the stack effect should be considered
in ventilation models only at low wind speeds (De Jong,
1990;Kittas et al.,1996). When the contribution of stack
effect is negligible, the ventilation rate can be expressed
by the following equation (Kittas et al.,1996):
G¼AT
2Cdffiffiffiffiffiffiffi
Cw
pu(4)
If G
0
represents the airflow rate when the greenhouse is
closed (leakage ventilation, A
T
¼0), Eqn (4) becomes as
follows:
G¼AT
2Cdffiffiffiffiffiffiffi
Cw
puþG0(5)
The greenhouse air exchange rate per hour Ncan be
given by the following:
N¼3600G=Vg(6)
where V
g
in m
3
is greenhouse volume.
3. Materials and methods
3.1. Site and greenhouse description
The experiments were performed in a round arch
with vertical side walls, plastic covered greenhouse,
NE/SW oriented (361declination from north 01), located
at the University of Thessaly near Volos, (latitude 391440;
longitude 221790; altitude 85 m) on the continental area of
Eastern Greece, during June and July of 2000 (26–6–2000,
5–7–2000, 7–7–2000 and 10–7–2000) between 10:00 and
17:00. The geometrical characteristics of the greenhouse
were as follows [Fig. 1(a)]: eaves height of 24m; ridge
height of 41 m; total width of 8 m; total length of 20 m;
ground area A
g
of 160 m
2
,andvolumeV
g
of 572 m
3
.T
he greenhouse was next to a similar greenhouse 45m
away from its east side; and another one 10 m away
from its north-front side. It was equipped with two
side roll-up windows located at 06 m above the ground,
with 27 m
2
maximum opening area (2 vents of 15 m
length by 09 m opening height) for both vents (about
17% of the greenhouse ground area). A flap roof window
[Fig. 1(a)] was also located longitudinally on the whole
greenhouse roof (20 m long) with 09 m maximum opening
height (18 m
2
opening area, about 112% of greenhouse
ground area). The prevailing wind in the region has
north–south direction. Greenhouse soil was totally covered
by a double-sided (black downwards–white upwards)
plastic film.
In order to study the effect of vent type on G, three
different cases examined, where greenhouse ventilation
was achieved by:
(1) side vents;
(2) roof vents; or
(3) both side and roof vents.
Furthermore, in order to study the effect of an anti-
aphid insect screen (55 by 27 mesh size and 50%
porosity) on G, two different cases examined where the
greenhouse ventilation was conducted by side vents:
(1) with a screen installed with in the openings; and
(2) without screen in the openings.
ARTICLE IN PRESS
EFFECT OF VENT OPENINGS AND INSECT SCREENS 429
3.2. Crop
The tomato crop present in the greenhouse during the
period of measurements, had about 18 m average
height. Tomato plants (cv. Condesa) were transplanted
on January 2000 and plant density was 24 plants m
2
.
The plants were 033 m apart in four double rows, with
an intra-row distance of 075 m and an inter-row
distance of 08 m. Water and fertilisers were supplied
by a drip system, which was automatically controlled by
a fertigation computer. The plants were grown following
local agricultural practices. Crop leaf area index was
about 25m
2
[leaf] m
2
[ground] during the period of
measurements.
3.3. Climatic measurements
The following climatic data were recorded:
(1) air temperature T
i
in 1C and water vapour pressure e
in kPa, with ventilated psychrometers (wet and dry
bulb, model VP1, Delta-T Devices, Cambridge,
UK), located at the centre of greenhouse and 18m
above ground and outside 15 m away from the
greenhouse on a mast 4 m above ground;
(2) total solar radiation R
s
in W m
2
,withsolar
pyranometers (model Middleton EP08-E, Brunswick,
Victoria, Australia), placed inside the greenhouse at
the centre and 2 m above ground and outside of the
greenhouse in the same mast as the ventilated
psychrometer;
(3) net radiation R
net
in W m
2
, with a net radiometer
(model Q-7, Radiation and Energy Systems, Seattle,
Wash.) located between the top of the tomato plants
and the cover of the roof;
(4) conductive soil heat flux inside and outside the
greenhouse in W m
2
, with flux meters (EKO Heat
Flow Meter CN-140, EKO Instruments Co., Ltd.)
situated 001 m below the soil surface;
(5) leaf temperature T
l
in 1C, with copper-constantan
thermocouples (type T, 01 mm wire diameter,
Omega, UK), with thermocouple junctions glued
to leaf lower surface and canopy temperature
calculated as mean value of temperature measure-
ments on six healthy and mature leaves distributed
randomly along the canopy layers;
(6) greenhouse plastic cover temperature T
c
in 1C, with
temperature measurements on eight positions dis-
tributed along the greenhouse sides and roof of the
greenhouse using copper-constantan thermocouples
(type T, wire diameter of 1 mm, Omega, UK), with
thermocouple junctions secured to the cover with
transparent adhesive tape;
(7) air velocity near vent openings, with a three-
dimensional (3-D) sonic anemometer (model R3-50
Ultrasonic Research Anemometer, Gill Instruments
ARTICLE IN PRESS
2.4 m
2.5 m 1.5 m
4.1 m
0.6 m
20 m
N2O analyser
20
m
8 m
8 m
(a) (b)
Fig. 1. (a) Experimental greenhouse and (b) distribution of pipes used for air sampling in the greenhouse
N. KATSOULAS ET AL.430
Ltd., UK). The three components of air velocity
were measured at six positions equally distributed
along the length of ventilation opening, at 11m
above ground, which coincided with middle height
of the full opening, with a sampling frequency of
5 Hz and 5 min acquisition period for each position.
Data were logged to a personal computer with a
special data acquisition card.
Additionally, measurements of wind speed, with a cup
anemometer (model AN1-UM-3, Delta-T Devices,
Cambridge, UK) and wind direction, with a wind vane
(model WD1-UM-3, Delta-T Devices, Cambridge, UK)
were carried out on a mast 4 m above the ground and
15 m away from the greenhouse.
All measurements were recorded in a data logger
(model DL3000, Delta-T devices, Cambridge, UK) with
1 Hz measuring frequency.
3.4. Ventilation rate measurements
The air exchange rate measurements were performed
with the impulse peak method using nitrous oxide (N
2
O)
as tracer gas and Gwas calculated as described by
several authors (Baptista et al., 1999;Kittas et al., 2002).
The tracer gas was distributed up to 200 p.p.m while
the vents were closed. After gas injection, 5min were left
to pass before the vents were opened to the desired
position, in order to obtain uniform gas distribution
inside the greenhouse. Air samples were continuously
taken at six points in the greenhouse, by means of six
equally distributed plastic pipes of the same length,
located at a height of approximately 18 m above the
ground [Fig. 1(b)]. The air from six positions was then
mixed and pumped to an infrared gas analyser (model
7000, ADC gas analyser, analysis up to 200 p.p.m,
accuracy at 75 p.p.m, Analytical Development Com-
pany, Hoddenson, UK) and N
2
O concentration was
recorded with a data logger every one second. Experi-
ments lasted between 5 and 20min, depending on
environmental conditions and ventilation opening. Dur-
ing the experiments, wind speed varied between 1 and
5ms
1
and ventilation opening ranged from 0 to 27 m
2
for side windows and from 0 to 18m
2
for roof windows.
3.5. Data processing and analysis
During data processing, the measurements obtained
during the experimental period were screened and
measurements performed during periods with a rela-
tively stable wind direction (north to south, between 01
and 451,i.e. almost parallel to the greenhouse ridge) are
shown. From a total of 68 observations that had been
obtained during the experimental period, 54 fulfilled the
above criterion.
The screened experimental data were fitted to Eqn (5),
using Marquardt’s algorithm (Marquardt, 1963).
4. Results and discussion
Values of outside climate characteristics during the
period of measurements are presented in Table 1.
Climatic conditions outside the greenhouse were similar
between the periods of measurements with screens in vent
openings, but during the period of measurements without
insect screens in the openings, outside solar radiation was
relatively low and wind speed was relatively high.
4.1. Methods comparison
Greenhouse ventilation rate was calculated using the
‘tracer gas’ and ‘energy balance’ methods. In Fig. 2, air
exchange rate calculated by the tracer gas method N
tracer
gas
is plotted versus the air exchange rate calculated by
the energy balance method N
energy
for a greenhouse
ventilated by roof, side or roof and side vents and insect
screens in the openings. A good correlation was found
between air exchange rate values calculated using the
two methods (tracer gas and energy balance), which for
54 observations, was:
Ntracer gas ¼081 Nenergy (7)
ARTICLE IN PRESS
Table 1
Average values (period 10:00–17:00 local time) of outside climatic parameters during the period of measurements. Standard deviation
values are given in parenthesis
Date Vent Screen Air
temperature, 1C
Relative
humidity, %
Solar radiation,
Wm
2
Wind speed,
ms
1
Wind direction,
deg
10/7/2000 Side No 252(05) 533(15) 284 (112) 456 (128) 35 (15)
26/6/2000 Side Yes 352(14) 261(36) 708 (178) 177 (066) 33 (12)
7/7/2000 Roof Yes 322(14) 344(38) 714 (167) 294 (058) 26 (14)
5/7/2000 Side & roof Yes 371(25) 295(34) 850 (69) 220 (034) 10 (10)
EFFECT OF VENT OPENINGS AND INSECT SCREENS 431
with a coefficient of determination R
2
of 076. The
results show that the calculated values of air exchange
rate using the tracer gas method were lower than those
calculated using the energy balance method. The model
used for the determination of ventilation flux by the
energy balance method does not include a term
representing the latent heat storage, thus causing an
error in calculating Gfor small time steps. Similar
results were also found by Fernandez and Bailey (1992)
and Shilo et al. (2004), who also found a good
correlation between the two methods of calculating
ventilation flux and the average value of ventilation rate
obtained by the energy balance method to be slightly
higher than that obtained by the tracer gas method.
Furthermore, Fernandez and Bailey (1992) noted that
measurement precision increased with the length of the
measurement time scale and vent opening. Moreover,
the data of Gobtained by the tracer gas and the energy
balance methods were used for the calibration of Eqns
(3) and (5); and it was found that the Gvalues obtained
by the tracer gas method fitted better to the models.
Therefore, the tracer gas method was selected for
calculating greenhouse ventilation rate Gin the follow-
ing experiments.
4.2. Effect of screen on ventilation rate
For a greenhouse equipped with roof and side
openings, Kittas et al. (1997) considered that the
stack effect is important if the ratio uDT05
iois o1.
Considering the greenhouse with roof and side openings,
it was found that the above mentioned ratio uDT05
io
was equal to 178 m s
1
1C
1
with standard deviation
(SD) of 05ms
1
1C
1
, indicating that the stack effect
can be neglected. Furthermore, in greenhouses equipped
with side only (Kittas et al., 1996) or roof only openings
(De Jong 1990) the stack effect can be neglected when
wind speed exceeds 1–15sm
1
. In our case, during the
period of measurements with side only or roof only
openings, mean outside wind speed was higher than
15ms
1
(Table 1). Thus, for all cases studied, the stack
effect was neglected and Eqn (5) was used for airflow
rate Gcalculations.
In Fig. 3, the airflow rate measured, using the tracer
gas method, in a greenhouse without or with screens in
the openings, is plotted versus the half product of the
effective opening area and wind velocity (05A
S
u). The
experimental data were fitted to Eqn (5), using
Marquardt’s algorithm, and a regression, which is
represented by straight lines in Fig. 3, was obtained
for each case. The calculated wind-related coefficients,
Cdffiffiffiffiffiffiffi
Cw
pand leakage airflow rate G
0
along with their
standard errors and the determination coefficients, are
shown for each case in Table 2. The calculated wind-
related coefficient Cdffiffiffiffiffiffiffi
Cw
pwas found to be equal to
0078700057 and 0052700017 for the greenhouse
without and with screen in the vent openings, respec-
tively. Use of the anti-aphid insect screen in the openings
caused a 33% reduction in the value of the wind-related
coefficient Cdffiffiffiffiffiffiffi
Cw
pand therefore, according to Eqn (5),
in order to obtain the same G, as for the case without
ARTICLE IN PRESS
0.0
1.0
2.0
3.0
4.0
5.0
0 1020304050
0.5 ASu, m3 s−1
Ventilation rate G, m3 s−1
Fig. 3. Measured ventilation rate (G) versus the half product of
opening surface area and wind velocity (05A
S
u), in a green-
house with side vents (J) without screen in the openings and
(K) with screen in the openings; the straight lines were obtained
by linear regression
Ntracer gas = 0.81 Nenergy
R2 = 0.76
0
10
20
30
40
50
01020304050
Air exchange rate Ntracer gas, h−1
Air exchange rate Nenergy, h−1
Fig. 2. Measured ventilation rate using the tracer gas method
N
tracer gas
versus measured ventilation rate using the energy
balance method N
energy
, during the period with screens in vent
openings; the straight line was obtained by linear regression; R
2
,
coefficient of determination
N. KATSOULAS ET AL.432
screen, a 50% increase in the vent opening area is
needed to prevent greenhouse overheating.
From the results shown in Table 2,at-test was used to
determine if the two calculated wind-related coefficients
for the above cases are statistically different. Motulsky
and Christopoulos (2003) identified the calculation of t
value as
t¼0078 0052
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
00057ðÞ
2þ00017ðÞ
2
q¼437 (8)
This value is higher than 238, which is the
corresponding tvalue for 95% of confidence and 26
degrees of freedom (the sum of degrees of freedom for
each fit) and accordingly, the two Cdffiffiffiffiffiffiffi
Cw
pcoefficients
calculated are statistically different.
The estimated values of leakage air flow rate G
0
were
013570149 and 031370022 m
3
s
1
(Table 2) for the
greenhouse without and with screen in the vent open-
ings, respectively, values that are similar to those usually
reported (Boulard et al., 1997) and not statistically
different as indicated by the 118 value of t(with
corresponding tvalue for 95% of confidence and 26
degrees of freedom equal to 238). However, the higher
leakage ventilation rates observed in the greenhouse
with insect screens could be considered that ‘hides’ the
stack effect which indeed must be greater for the case
with an insect screen because the DTis of course greater
and that has not been taken into account in Eqn (5).
The values of Cdffiffiffiffiffiffiffi
Cw
pfound for the greenhouse
without and with screens in the side vents (Table 2) are
relatively low (0078 and 0052 for greenhouse vent
openings without and with screen, respectively). Simi-
larly, low Cdffiffiffiffiffiffiffi
Cw
pvalues have also been reported by
Boulard et al. (1997),Fatnassi et al. (2003) and Perez
Parra et al. (2004). The above results concerning the low
Cdffiffiffiffiffiffiffi
Cw
pvalues could be partly attributed to the fact that
the mature tomato plants 18 m high created a barrier
between the two greenhouse side vents; and accordingly,
significantly reduced greenhouse ventilation rate. Re-
duction of ventilation rates caused by crop presence was
also reported by Boulard et al. (1997). The low Cdffiffiffiffiffiffiffi
Cw
p
values are also influenced by the proximity of the
experimental greenhouse to a similar greenhouse 45m
away from its east side; and another one 10 m away
from its north-front side, as the pressure field of the
airflow around it was affected by the adjacent green-
houses (ASHRAE, 1993) reducing the air velocity near
vent openings. Figure 4 presents the normalised air
velocity at a height of 11 m along the length of side vent
openings, a height which coincides to the middle height
of full opening. The normalised air velocity was
obtained by the ratio of interior air velocity measured
(by the 3-D sonic anemometer) near the openings, to the
mean external wind speed. Mean value of normalised air
velocity near the west side opening (ventilation openings
without obstacles) was 024, whereas the respective
value near the east side opening (ventilation opening
facing the adjacent greenhouse) was 007, a value 67%
lower than air velocity in the west side ventilation
opening.
Using the calculated value of Cdffiffiffiffiffiffiffi
Cw
pcoefficient
(Table 2) and the theoretical model for the prediction of
ventilation rate [Eqn (5)], the ventilation flux Gcan be
calculated. For example, with maximum side opening
(27 m
2
) and 3 m s
1
wind speed u,Gis equal to 33m
3
s
1
(034 air changes min
1
) for a greenhouse without screen
and 24m
3
s
1
(025 air changes min
1
) for a greenhouse
ARTICLE IN PRESS
Table 2
Estimated values and standard error (95% confidence) of wind-related coefficient Cdffiffiffiffiffiffiffi
Cw
p
and leakage ventilation rate (G
0
), from
Eqn (5), for the cases studied; R
2
indicates the determination coefficient of the calibrated model [Eqn (5)]
Date Openings Screen Number of measurements Cdffiffiffiffiffiffiffi
Cw
pG
0
m
3
s
1
R
2
10/7/2000 Side No 14 0078700057 013570149 094
26/6/2000 Side Yes 12 0052700017 031370022 099
7/7/2000 Roof Yes 24 0028700037 022070058 072
5/7/2000 Side & Roof Yes 16 0096700096 034070225 088
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0 5 10 15
Side vent len
g
th, m
Normalised air velocity
Similar
greenhouse
West
vent
East
vent
Experimental
greenhouse
Fig. 4. Normalised air velocity values near vent openings of a
greenhouse with side vents (J)on east side and (K) west side
EFFECT OF VENT OPENINGS AND INSECT SCREENS 433
with screen. Consequently, with wind direction parallel
to greenhouse ridge, an optimum ventilation rate (075
air changes min
1
), as recommended by ASAE Stan-
dards (1999), requires an outside wind speed of 67ms
1
for openings without screen, and 98ms
1
for openings
with screen. However, wind speeds higher than about
55ms
1
are rare in the region, where the present study
was carried out. For the greenhouse with screens in the
side vent openings, recommended ventilation rates, with
outside average wind speed of 3 m s
1
, can be achieved
by a triplication of side vent opening surface area
(almost completely open greenhouse side walls).
4.3. Effect of vent type on ventilation rate
In Fig. 5, the airflow rates measured during periods with
(a) side, (b) roof or (c) both side and roof openings were
plotted versus thehalfproductoftheeffectiveopening
areaandwindvelocity(05A
T
u). The experimental data
were fitted to Eqn (5), and the calculated wind-related
coefficients, Cdffiffiffiffiffiffiffi
Cw
pand leakage airflow rates G
0
,along
with their standard errors and the determination coeffi-
cients, are shown for each case in Table 2.
Following the same statistical analysis presented above
and using the results presented in Table 2, it was found
that the coefficients calculated for cases (a), (b) and (c) for
the greenhouse with different vent openings, are statisti-
cally different between them for all cases.
The most effective vent configuration was the
combination of roof and side vents, followed by side
vents only, while the least effective was roof vent
(Fig. 5). In comparison with the most effective vent
configuration (side and roof vents), the ventilation rate
obtained by side vents only was 46% lower, while that
obtained by roof vents only was 71% lower (Table 2).
Bartzanas et al. (2004) used a computational fluid
dynamic code to investigate the influence of vent
configuration on airflow and temperature patterns.
The ratio of overall ventilation rate to opening surface
area was reduced from 016 m s
1
for a tunnel green-
house equipped with roof and side vent openings to
015 m s
1
for side only and to 007 m s
1
when the
tunnel greenhouse was ventilated with roof openings
only. Considering the efficiency of each vent configura-
tion on temperature difference between inside and
outside, the ratio of temperature difference to opening
surface area was found to be 0016 1Cm
2
when roof
and side openings were used for ventilation. This value
was increased to 0022 and to 0090 1Cm
2
when the
tunnel greenhouse was ventilated with only side or roof
openings, respectively.
However, Papadakis et al. (1996) examined the
ventilation performance of a two-span greenhouse
equipped with continuous roof vents and pivoting door-
type side openings and found that the most effective vent
configuration was that with roof openings only while the
least effective vent configuration was that with side
openings only. Boulard et al. (1997) also examined the
ventilation performance of some tunnel greenhouses
equipped with side and roof openings but without crop
inside. They found that side vents only were more
efficient than the combination of roof and side vents.
Such a result could be different whenever, as presently, a
high crop such as tomato plants can impede air
circulation at the side vent. It confirms the modifications
on air flow characteristics a crop such tomato plants, size
and planting arrangement can cause (Sase, 1989). It was
also found, in this study, that adding a screen in the side
vent openings the Cdffiffiffiffiffiffiffi
Cw
pcoefficient reduced from 0078
to 0052 and consequently decreased greenhouse ventila-
tion rate. In the current study, by adding a roof vent in
the greenhouse, the Cdffiffiffiffiffiffiffi
Cw
pwas increased from 0052 to
0096, resulting in about 46% higher ventilation rate than
for the case of side vents only without screen.
4. Conclusion
The aim of the present study was: (1) to compare the
two, most widely used, methods for greenhouse ventila-
tion rate measurement, namely the ‘tracer gas’ and the
‘energy balance’ methods; (2) to determine the effect of
an insect screen on wind driven ventilation; and (3) to
examine the influence of vent configuration on green-
house ventilation performance.
ARTICLE IN PRESS
0.0
1.0
2.0
3.0
4.0
5.0
0 1020304050
0.5 ATu, m3 s−1
Ventilation rate G, m3 s−1
Fig. 5. Measured ventilation rate (G) versus the half product of
opening surface area and wind velocity (05A
T
u), for a screened
greenhouse ventilated by roof (J), side (K) or roof and side
vents (n); the straight lines were obtained by linear regression
N. KATSOULAS ET AL.434
A good correlation was found between air exchange
rate values calculated using the two methods, with
values obtained by the tracer gas method being about
19% lower than those obtained by the energy balance
method. Furthermore, ventilation rate values obtained
by the tracer gas method provided better fit to the model
used for predicting ventilation rates.
When side vents only were used for greenhouse
ventilation, insect screens in the vent openings reduced
by 33% the Cdffiffiffiffiffiffiffi
Cw
pcoefficient (where C
d
is the
discharge coefficient and C
w
is the wind effect coeffi-
cient) and accordingly greenhouse ventilation rate.
However, in a greenhouse with side vents only, the
reduction in ventilation rate caused by installing anti-
aphid insect screens can be surpassed by installing of
roof vent, since, due to the drag effect of the crop, the
ventilation performance of greenhouse with side and
roof vents and screens in the openings was found to be
better than that greenhouse with side vents only without
screens in the openings.
Finally it was found that for a greenhouse with 18m
tall tomato plants which create a barrier between the
two greenhouse side vents, the most effective vent
configuration was the use of both roof and side vents,
followed by side vents only (46% less ventilation), while
the least effective was roof vent (71% less ventilation).
Acknowledgements
This work was funded by the Geothermiki (Greek
company), the General Secretary for Research and
Development of Greece and the Minister for Research &
Education of France, under the bilateral Greek-French
project PLATON.
References
ASAE (1999). Heating, ventilating and cooling greenhouses.
ANSI/ASAE standard EP406.3. ASAE Standards, ASAE,
St. Joseph, MI, pp 679–686
ASHRAE (1993). Airflow around buildings. In ASHRAE
Handbook of Fundamentals, vol 14, pp 14.1–14.3. Amer-
ican Society of Heating, Refrigerating, and Air-Condition-
ing Engineers, Atlanta, GA
Baptista F J; Bailey B J; Randall J M; Meneses J F (1999).
Greenhouse ventilation rate: theory and measurement with
tracer gas techniques. Journal of Agricultural Engineering
Research, 72(4), 363–374
Bartzanas T; Kittas C; Boulard T (2004). Effect of vent
arrangement on windward ventilation of a tunnel green-
house. Biosystems Engineering, 88(4), 479–490
Boulard T; Baille A (1995). Modeling of air exchange rate in a
greenhouse equipped with a continuous roof vents. Journal
of Agricultural Engineering Research, 61(1), 37–48
Boulard T; Feuilloley P; Kittas C (1997). Natural ventilation
performance of six greenhouse and tunnel types. Journal of
Agricultural Engineering Research, 67(4), 249–266
Boulard T; Wang S (2002). Experimental and numerical studies
on the heterogeneity of crop transpiration in a plastic
tunnel. Computers and Electronics in Agriculture, 34(1–3),
173–190
Boulard T; Wang S; Haxaire R (2000). Mean and turbulent air
flows and microclimate patterns in an empty greenhouse
tunnel. Agricultural and Forest Meteorology, 100(2–3),
169–181
De Jong T (1990). Natural ventilation of large multispan
greenhouse. PhD Thesis, Wageningen University, The
Netherlands 116pp
Demrati H; Boulard T; Bekkaoui A; Bouirden L (2001). Natural
ventilation and climatic performance of a large-scale banana
greenhouse. Journal of Agricultural Engineering Research,
80(3), 261–271
Fatnassi H; Boulard T; Bouirden L (2003). Simulation of
climatic conditions in full-scale greenhouse fitted with
insect-proof screens. Agricultural and Forest Meteorology,
118(1–2), 97–111
Fatnassi H; Boulard T; Demrati H; Bouirden L; Sappe G
(2002). Ventilation performance of a large canarian-type
greenhouse equipped with insect-proof nets. Biosystems
Engineering, 82(1), 97–105
Fernandez J E; Bailey B J (1992). Measurements and
prediction of greenhouse ventilation rates. Agricultural
and Forest Meteorology, 58(3–4), 229–245
Kittas C; Boulard T; Bartzanas T; Katsoulas N; Mermier M
(2002). Influence of an insect screen on greenhouse ventila-
tion. Transactions of the ASAE, 45(4), 1083–1090
Kittas C; Boulard T; Mermier M; Papadakis G (1996). Wind-
induced air exchange rates in a greenhouse tunnel with
continuous side openings. Journal of Agricultural Engineer-
ing Research, 65(1), 37–49
Kittas C; Boulard T; Papadakis G (1997). Natural ventilation
of a greenhouse with ridge and side openings: sensitivity to
temperature and wind effects. Transactions of the ASAE,
40(2), 415–425
Lee I; Okushima L; Sase S; Ikeguchi A (2003). Improvement of
particle image velocimeter (PIV) at large-sized wind tunnel
test for aerodynamic study of natural ventilation in multi-
span greenhouses. Transactions of the American Society of
Agricultural Engineers, 46(3), 763–777
Luo W; Stanghellini C; Dai J; Wang X; De Zwart H F; Bu C
(2005). Simulation of greenhouse management in the
subtropics, Part II: Scenario study for the summer season.
Biosystems Engineering, 90(4), 433–441
Marquardt D W (1963). An algorithm for least-squares
estimation of non-linear parameters. Journal Society of
Applied Mathematics, 2, 432–441
Miguel A F; Van de Braak N J; Bot G P A (1997). Analysis of
the airflow characteristics of greenhouse screening materials.
Journal of Agricultural Engineering Research, 67(2),
105–112
Mistriotis A; Bot G P A; Picuno P; Scarascia Mugnozza G
(1997). Analysis of the efficiency of greenhouse ventilation
with computational fluid dynamics. Agricultural and Forest
Meteorology, 85(3–4), 217–228
Montero J I; Hunt G R; Kamaruddin R; Anton A; Bailey B J
(2001). Effect of ventilator configuration on wind driven
ventilation in a crop protection structure for the tropics.
Journal of Agricultural Engineering Research, 80(1), 99–107
ARTICLE IN PRESS
EFFECT OF VENT OPENINGS AND INSECT SCREENS 435
Montero J I; Munoz P; Anton A (1997). Discharge coefficients
of greenhouse windows with insect-proof screens. Acta
Horticulturae, 443, 71–77
Motulsky H J; Christopoulos A (2003). Fitting models to
biological data using linear and nonlinear regression. A
practical guide to curve fitting. GraphPad Software Inc.,
San Diego, CA
Munoz P; Montero J I; Anton A; Giuffrida F (1999). Effect of
insect-proof screens and roof openings on greenhouse
ventilation. Journal of Agricultural Engineering Research,
73(2), 171–178
Papadakis G; Mermier M; Meneses J F; Boulard T
(1996). Measurements and analysis of air exchange rates in
a greenhouse with continuous roof and side opening.
Journal of Agricultural Engineering Research, 63(3),
219–227
Perez Parra J; Baeza E; Montero J I; Bailey B J (2004).
Natural ventilation of parral greenhouses. Biosystems
Engineering, 87(3), 355–366
Sase S (1989). The effect of plant arrangement on airflow
characteristics in a naturally ventilated greenhouse. Acta
Horticulturae, 245, 429–435
Shilo E; Teitel M; Mahrer Y; Boulard T (2004). Air-flow
patterns and heat fluxes in roof-ventilated multi-span
greenhouse with insect-proof screens. Agricultural and
Forest Meteorology, 122(1–2), 3–20
Soni P; Salokhe V M; Tantau H J (2005). Effect of screen mesh
size on vertical temperature distribution in naturally
ventilated tropical greenhouses. Biosystems Engineering,
92(4), 469–482
Teitel M; Barak M; Berlinger M J; Lebiush-Mordechai S
(1999). Insect proof screens: their effect on roof ventilation
and insect penetration. Acta Horticulturae, 507, 29–37
Teitel M; Tanny J; Ben-Yakir D; Barak M (2005). Airflow
patterns through roof openings of a naturally ventilated
greenhouse and their effect on insect penetration. Biosys-
tems Engineering, 92(3), 341–353
Wang S; Boulard T; Haxaire R (1999). Air speed profiles in a
naturally ventilated greenhouse with a tomato crop.
Agricultural and Forest Meteorology, 96(4), 181–188
Wang S; Deltour J (1996). An experimental ventilation
function for large greenhouses based on a dynamic energy
balance model. International Agricultural Engineering
Journal, 5(3–4), 103–112
ARTICLE IN PRESS
N. KATSOULAS ET AL.436
