NATURAL VENTILATION OPENINGS – A DISCUSSION OF DISCHARGE COEFFICIENTS

P. Karava1, T. Stathopoulos2, A. K. Athienitis3

Centre for Building Studies, Department of Building, Civil and Environmental Engineering,
Concordia University, Montreal, Quebec, Canada

ABSTRACT
This paper, in addition to reviewing the current literature in discharge coefficients of natural ventilation
openings, presents comparisons among different studies in an attempt to reconcile some of the existing
approaches and identify possible discrepancies. The existing information regarding the parameters
affecting the discharge coefficient of openings has been summarized and analyzed separately for single-
sided and cross ventilation configurations. The discharge coefficient is a result of a multivariable impact
and the use of a constant value such as that given in textbooks or works of reference is a simplification,
which may cause considerable errors. Large discrepancies in discharge coefficient values were found
among the different studies especially at small opening porosities where the fully developed turbulent flow
assumption (orifice equation) is incorrect and the impact of building envelope leakage might be important.
A significant variation of the discharge coefficient with the wind incidence angle has been found; however,
more research is required to clarify this issue so that it can be considered in natural ventilation system
design. Several problems with the application of the orifice equation to the calculation of wind-driven
cross ventilation rate are pointed out. However, there is disagreement among the different literature
sources regarding the proposed solution to the problem. Finally, the review of the literature for buoyancy-
driven single-sided ventilation shows considerable variation of the discharge coefficient with temperature
difference and lower discharge coefficient values compared to those for cross ventilation.

INTRODUCTION
Integrated multi-zone airflow and thermal models are increasingly popular for detailed simulation of yearly
performance of natural ventilation systems and their impact on energy consumption of buildings (Li and
Heiselberg 2002). It is therefore necessary to ensure that the fundamental governing equations are
correct and that the ventilation rate can be predicted accurately. The most common equation describing
the airflow through an opening is the orifice equation, which is based on Bernoulli’s equation for steady
incompressible flow. This equation can be used for a relatively large opening area (typical dimension
larger than 10 mm). In that case, the flow tends to be turbulent under normal pressures and the flow rate,
Q (m3/s), is proportional to the square root of the pressure difference across the opening:
2 ⋅ ∆P 0.5
Q = CD ⋅ A ⋅ ( ) (1)
ρ
where
CD = discharge coefficient of the opening
A = flow area (m2)
∆P = pressure difference across the opening (Pa)
ρ = air density (kg/m3).

The discharge coefficient is a characteristic parameter for a specific opening; it depends on the geometry
of the opening, the Reynolds number of the flow, and includes the influence of contraction and friction.
Andersen (1996) provides a thorough discussion of inlet and outlet coefficients. For turbulent flow, CD is
constant at a fixed Reynolds number and therefore, the flow is proportional to √∆P. For sharp-edge orifice
flow the discharge coefficient is almost independent of the Reynolds number. However, in reality the flow
through building envelope openings is not fully developed and CD appears to be variable due to geometry
of the openings and the variation in pressure difference with the environmental conditions inside and
outside the building. Therefore, CD is the result of a multivariable impact and it is difficult to analyze
simultaneously all the parameters involved. Usually, only limited cases are examined, hence the

1
P. Karava is Graduate Research Assistant (e-mail: [email protected])
2
T. Stathopoulos is Professor and Associate Dean (e-mail: [email protected])
3
A. K. Athienitis is Professor and Undergraduate Program Director (e-email: [email protected])

1
discharge coefficient values should be used within the limits of their applicability. Typical discharge
coefficients given in textbooks vary between 0.6 and 0.65 (ASHRAE Fundamentals, 2001) for small
square-edged openings and between 0.9 and 0.95 for round edge openings (Andersen 2002). Hence,
textbooks and other references provide very limited information considering the variety of typical building
openings and their combinations while complex cases such as orientation of openings relative to wind
direction, relative shape and sequence of aligned openings are not considered.

Natural ventilation through large openings is distinguished into single-sided and cross ventilation. For
single-sided ventilation, airflow through large openings is usually considered to be bi-directional. The total
volumetric flow through one-half of the opening for buoyancy-driven flow is given by the following formula:
A ∆T ⋅ g ⋅ H
Q = CD ⋅ ⋅ (2)
3 T
where
∆T = temperature difference across the opening (°C)
g = gravitational acceleration (m/s2)
H = height of the opening (m).

For wind-driven cross ventilation the so-called total pressure loss coefficient ζ is often used instead of the
discharge coefficient (i.e. Murakami et al. 1991; Vickery and Karakatsanis 1987). The following equation
given by Andersen. (1996) can be used to calculate CD values based on total pressure loss coefficient
values:
2
2 A  1
ζ = ∆Pt ⋅ ⋅  o  = 2 (3)
ρ  Q  CD
where
∆Pt = total pressure difference across the opening (Pa)
Ao = opening area (m2)
Q = airflow (m3/s)
ρ = air density (kg/m3).

There are several problems with the application of Equation 1 to the calculation of wind-driven cross
ventilation rate. Murakami et al. (1991) found that the conventional method based on the application of
Bernoulli’s law (approximation of the airflow through small openings, such as thin orifices of sufficiently
large Re) for cross ventilation can lead to unrealistic results. Similar studies have been carried out by He
et al. (1991) and Ishihara (1969). Sandberg (2002) pointed out the need for a more elaborate model than
the orifice equation. A better approach might be to regard the flow through openings as a flow catchment
problem.

Since the validity of the “traditional” orifice equation remains questionable for large openings, the
selection of an appropriate equation to represent the flow characteristics of openings is problematic. This
paper, in addition to reviewing the current literature in discharge coefficients of natural ventilation
openings, presents comparisons among different studies in an attempt to reconcile some of the existing
approaches and identify possible discrepancies. The existing information regarding the parameters
affecting the discharge coefficient of openings has been summarized and analyzed separately for
buoyancy-driven single-sided ventilation and cross ventilation configurations.

REVIEW OF CURRENT LITERATURE RESULTS

Table 1 summarizes existing information regarding the parameters affecting the discharge coefficient of
openings. The first column gives the bibliographical reference. Since all the experiments were not carried
out under the same conditions, it is necessary to define the experimental set up (i.e. wind tunnel, field
measurements). Columns 2 and 3 give the ventilation strategy (single or cross ventilation) and the
method (i.e. analytical, experimental or numerical) used. Column 4 gives the opening type and columns 5
to 12 give all the parameters affecting the pressure difference across an opening and therefore, the CD.

2
 TABLE 1. Literature sources regarding the evaluation of discharge coefficients.

Building
Ventilation Opening Opening Internal Envelope
Source Method A T V θ Pfan CD
Type Configuration Location Partition Leakage
Area
Aynsley et al.
cross vent. empirical rectangular
(1977)
Vickery and
rectangular and
Karakatsanis cross vent wind tunnel
holes
(1987)
−0.313
 ∆T 
Riffat (1989) single-sided
full-scale, internal CD = 0.0835⋅  
real building doorways  T

Kiel and Wilson full-scale, exterior

(1989)
single-sided
real building doorways CD = (0.4 + 0.0075⋅ ∆T)

Pelletret et al. full-scale test doorways CD = 0.21⋅ H

single-sided
(1991) building 1.5m≤h≤2m or CD about 0.43
Murakami et al. windows
cross vent. wind tunnel power balance model
(1991)
Flourenzou et cross and windows and
real building CD = 0.6 ± 0.1
al. (1998) single vent. doors
Heiselberg et
no constant value of
al. (1999, 2001, laboratory side and bottom
cross vent. CD
2002a and experiment hung windows
2002b)
0.7 at large porosities
Sandberg holes in circular or
cross vent. wind tunnel
(2002) disk or cylinder flow catchment
problem
Jensen et al. cross vent. holes in circular
CFD
(2002a) disk or cylinder
Jensen et al.
cross vent. CFD rectangular
(2002b)

Kurabuchi, et local similarity model

cross vent. wind tunnel rectangular *
al. (2002) PR

Ohba et al. rectangular local similarity model

cross vent. wind tunnel *
(2002) and circular PR
full scale window
Sawachi (2002) cross vent.
wind tunnel openings
Andersen vents with
empirical
(2002) movable flaps

3
These variables are the opening area (A), temperature (T), wind speed (V), wind incidence angle (θ),
mechanical pressure (Pfan), location of the opening on the façade, internal partitions and building
envelope leakage. In most of the studies the opening area is referred with respect to area of the wall
(porosity) and is calculated by the following equation:
A opening
porosity = (4)
A wall
Column 14 gives the proposed formula for CD or outcome whenever applicable.
Examination of Table 1 reveals the following:
• No study has considered all the parameters involved.
• Although wind is the most important driving force for single-sided ventilation in low rise buildings or for
small temperature difference, no experimental study has been carried out for the evaluation of discharge
coefficient in that case, to the authors’ best knowledge.
• The impact of the porosity of the building envelope has only been considered by Heiselberg et al.
(2001; 2002a; 2002b). It was found that the absolute value of discharge coefficient is uncertain at small
opening areas and measured values above 1 can be caused by incorrect estimation of the geometrical
opening and infiltration/exfiltration. However, the relationship between discharge coefficient and porosity
of the building envelope has not yet been examined and more research is required to clarify this issue.
• The case of more than one opening in the same façade has not been considered.

COMPARISON AMONG DIFFERENT STUDIES

In order to identify the possible discrepancies between the various proposed discharge coefficient values,
a comparison among different studies was performed. However, the opening type and the conditions
under which the experiments were carried out are different among the various studies making such
comparisons problematic. The authors attempted to compare “similar” cases by using specific (partial)
data from various literature sources. For more information the reader should refer to these literature
sources where additional cases are discussed. Cross and single-sided ventilation are considered
separately.

Cross ventilation. In this section a comparison among different studies for cross ventilation is carried
out. Figure 1 illustrates the building models. For more information regarding the experimental set up, the
building models, opening type and dimensions the reader should refer to the corresponding literature
sources. Figures 2 and 3 present the discharge coefficient as a function of the porosity for circular and
rectangular (windows) openings respectively. Cases with equal inlet and outlet opening area, thin opening
walls, and normal wind incidence angle are considered.

Analysis of Figure 2 shows the following:

• Discharge coefficient values vary from 0.6 to 1.
• The experimental results by Sandberg (2002) and Jensen et al. (2002a) give similar CD values only for
porosities between 10 and 20%, hence there is lees dependence on the opening configuration for larger
porosity.
• For small porosities (less than 10%), large differences are observed and discharge coefficient values
even more than 1, which does not have any physical meaning, have been reported. This is may be due to
the fact that the orifice equation was used for very small openings, and the assumption of fully developed
turbulent flow is not valid. Note that small porosities are more important for natural ventilation.

Analysis of Figure 3 reveals the following:

• Large discrepancies between the empirical discharge coefficient values proposed by Aynsley et al.
(1977) and the experimental studies. In fact, Aynsley et al. (1977) results show an increase of the
discharge coefficient with porosity, which could be expected based on the traditional resistance approach,
while Heiselberg et al. (1999 and 2002b) found that the discharge coefficient decreases with increase of
the porosity (side hung window and the configuration W2 - inner window. For configuration W1 (corner

4
window) CD is almost constant. However, for the configuration W1-4 an increase of the discharge
coefficient with the porosity was observed. Similar observations were reported by He et al. (1995); In this
study, wind tunnel experiments with a cubical building model with and without a partition wall having an
opening of the same dimension as each pair of openings on the exterior wall showed that the airflow rate
became larger with a partition wall, although the total airflow resistance was increased. This is probably
due to the fact that the resistance approach does not consider the Reynolds number dependence of the
discharge coefficient. However, experimentally deduced discharge coefficient values could be
overestimated at low porosities due to the fully developed turbulent flow assumption (orifice equation) or
inaccurate opening area estimation.

a. Sandberg (2002) and Jensen et al. (2002a).

b. Murakami et al. (1991) model 1 and 2 respectively. c. Sawachi (2002).

d. Heiselberg et al. (1999) side hung window. e. Heiselberg et al. (2002b), bottom hung window.

Figure 1. Illustration of the models used by (a): Sandberg (2002) and Jensen et al. (2002a); (b): Murakami
et al. (1991); (c): Sawachi (2002); (d): Heiselberg et al. (1999); (e): Heiselberg et al. (2002b).

5
 1.4

1.2

0.8
CD

0.6
Saldberg (2002), case 1a, experimental data
Sandberg (2002), case 1a, regression
0.4 Sandberg (2002), case 1a, CFD
Sandberg (2002), case 1a, CFD, regression
Sandberg (2002), case 1b, experimental data
Sandberg (2002), case 1b, regression
0.2 Jensen et al. (2002a), case 3, experimental data and CFD
Jensen et al. (2002a), case 3, regression
Andersen (2002)
0
0.001 0.01 Porosity 0.1 1

Figure 2. Discharge coefficient as a function of the porosity of the opening (circular openings).

1.4

1.2

0.8
CD

0.6

0.4 Sawachi (2002), total

Muracami et al. (1991), inlet
Aynsley et al. (1977), outlet
Heiselberg et al. (1999)
Heiselberg et al. (2002b), W1
0.2 Heiselberg et al (2002b), W2
Heiselberg et al. (2002b), W1-4, inlet
Heiselberg et al., (2002b), W1-4, outlet
ASHRAE (2001)
0
0.001 0.01 Porosity 0.1 1

Figure 3. Discharge coefficient as a function of the porosity of the opening (rectangular openings,
windows).

• Disagreement between wind tunnel results by Sawachi (2002) and the other studies. This study
showed no variation of the discharge coefficient with porosity. Moreover, the proposed CD value of 0.48
for porosities of 1% and 9% is much lower than the values found from the other experimental studies with
the same porosity.
• The experimentally deduced discharge coefficient values given by Murakami et al. (1991) and
Heiselberg et al. (1999 and 2002b) show larger variability with porosity especially at low porosity values
compared to Aynsley et al. (1977).

6
• Higher inlet discharge coefficients values compared to outlets (Heiselberg et al. 2002b). However,
larger outlet CD values have been reported by Vickery and Karakatsanis (1987), Murakami et al. (1991)
and Jensen et al. (2002a). In fact, discharge coefficient values higher than 1 for the outlet reported by
Vickery and Karakatsanis (1987) and Jensen et al. (2002). Unrealistic results for outlet discharge
coefficients (pressure loss coefficient) also reported by Murakami et al. 1991 which was explained by the
fact that the approximation of the airflow through small openings is no longer valid in the case of cross
ventilation. The application of power balance model (energy conservation law) instead showed very
reasonable results.

Figure 4 indicates the discharge coefficient variation with the wind incidence angle. The discharge
coefficient decreases with the increase of the wind angle as it was expected. More specifically, the
discharge coefficient varies between 0.58 and 0.1 for wind incidence between 0 and 90°. This is a
considerable variation, which should be considered in natural ventilation design. Unfortunately, discharge
coefficient values for wind incidence other than 0° are not available in textbooks and more research is
required to clarify this issue. It appears also that CFD predictions (Jensen et al, 2002b) show good
agreement with the experimental data. Only for small angle of incidence a slight overestimation of the
CFD values is observed. However, the proposed CD values for θ = 0° are much lower compared to those
proposed by Murakami et al. (1991) and Heiselberg et al. (1999 and 2002b) for the same porosity (see
Figure 3).

1.4
Wind tunnel, Saw achi (2002)
1.2 A = 1.5 sq.m
A = 0.16 sq.m
CFD, Jensen et al. (2002b), A = 0.25 sq.m
1
Opening

0.8
CD

0.6 θ

0.4

0.2

0
0 50 100 150
Wind incidence, θ (degrees)

Figure 4. Discharge coefficient as a function of wind incidence angle.

The variation of the discharge coefficient with the pressure difference and the opening area for a side and
a bottom hung window was studied by Heiselberg et al. (2001). The study found that especially in the
case of the side hung window the discharge coefficient varies at small pressure differences across the
opening, while it becomes constant at large pressure differences. This indicates a Re dependency that
might be important to consider since natural ventilation systems most of the time operate at small
pressure differences. Kurabuchi et al. (2002) and Ohba et al. (2002) investigate the ratio of ventilation
driving pressure of crossflow at the opening. A new model was proposed to estimate cross ventilation
flow rate and inflow angle at an opening. This model named as local similarity model of cross ventilation
is based on the fact that dynamic similarity of airflow can be approximately determined under the
condition where the ratio of ventilation driving pressure to the interfering crossflow dynamic pressure in
the vicinity of the opening is consistent. The interfering flow is defined as the airflow tangential to the
opening, which might disturb the approach flow entering the building through the opening. However, the
proposed discharge coefficient values for a window opening are much lower (between 0.2 and 0.4)
compared to those found by Heiselberg et al. (2001) (between 0.65 and 1).

7
Single-sided ventilation. Several studies have been carried out to investigate the heat and mass
transfer through doorways but there are limited studies that provide discharge coefficient values for
natural ventilation configurations such as windows in the case of single-sided ventilation; this might be
due to additional complexities caused by the bi-directional character of the flow for stack-driven ventilation
or the uncertainties due to the wind turbulence for wind-driven ventilation.

A simulation study for the prediction of air flow through natural ventilation configurations using large-eddy
simulation (Chen, 2003) found that it appears more difficult to predict the performance of single ventilation
openings compared to cross ventilation. In a study carried out by Dascalaki et al. (1995), totally 52 single-
sided ventilation configurations were tested and the measured flow rates were compared with those
predicted by using several airflow network tools. Similar values were predicted by all tools for each
experiment, since all tools use the same technique and equations to predict the air flow, but the predicted
flow rates were far from the mean measured values due to the use of the same discharge coefficient for
any opening configuration (CD = 1). Haghighat et al. (1991) studied the mechanism of single-sided
ventilation due to wind induced pressure. The study found that the turbulent wind pressure has a direct
impact on the ventilation performance in buildings. This effect is more significant when the mean pressure
difference across openings is low and its turbulent component is high, which usually occurs in single-
sided ventilation situations. A review of the existing information on heat and mass transfer through large
openings was presented by Santamouris et al. (1995). A sensitivity analysis performed to identify the
discrepancies between the various proposed formulas showed differences up to 40% in the case of
buoyancy dominated flow.

Figure 5 shows the variation of the discharge coefficient with the temperature difference for different
single-sided ventilation configurations. In this figure the results of full-scale experiments in four different
test buildings regarding the discharge coefficient of internal doorways published by Pelletret et al. (1991)
are compared with empirical formulas proposed by Riffat (1989). The empirical formula proposed by Riffat
(1989) is the outcome of an experimental study carried out in a real house aiming to determine the heat
and mass transfer through an internal doorway separating the lower and upper floors of the house. The
empirical formula proposed by Kiel and Wilson (1989) gives the discharge coefficient of an exterior
doorway and is the outcome of an experimental study carried out in a full-scale test house.
1
CSTB Sophia Antipolis, H = 1.56 m, Pelletret et al. (1991)
0.9
INSA Lyon, H = 1.85 m, Pelletret et al. (1991)
0.8 CSTB Sophia Antipolis, H = 2.08 m, Pelletret et al. (1991)
University of Liege, H = 2.16 m Pelletret et al. (1991)
0.7 CD = 0.4 + 0.0075.(∆T), Kiel and Wilson (1989)
CD = 0.0835.(∆T/T)-0.313, Riffat (1989)
0.6
CD

0.5

0.4

0.3

0.2

0.1

0
0 1 2 3 4 5 6
o
Temperature difference ( C)

Figure 5. Discharge coefficient as a function of the temperature difference.

Analysis of Figure 5 shows the following:

8
• The experimental results show slight decrease of the discharge coefficient with the temperature
difference (except for the case of CSTB Sophia Antipolis, H = 2.08 m). The proposed formula by Riffat
(1989) shows the same trend. This may be due to the increase in interfacial mixing as a result of the
direct transfer of some cold air. In addition, the increase in density difference can cause an increase in
turbulence which will affect the coefficient of discharge (Riffat 1989). However, the formula for steady
buoyancy-driven flow proposed by Kiel and Wilson (1989) shows a slight increase of the CD with ∆T.
• There is significant impact of the opening height on discharge coefficient. The proposed formula by
Pelletret et al. (1991) is:

CD = 0.21⋅ H , 1.5 m ≤ H ≤ 2 m (6)

The difference in CD values (for the same ∆T and H) derived from the different experiments confirms that
the discharge coefficient is dependent also on a number of parameters such as Re, zone geometries and
experimental conditions.
• Relatively good agreement between the experimental data (Pelletret et al. 1991) and the discharge
coefficient values calculated using the formula proposed by Kiel and Wilson 1989, for temperature
difference up to 3 °C. On the contrary, the CD values calculated using the proposed formula by Riffat
(1989) do not seem to agree with the experimental results. This is most probably due to the fact that the
proposed formula by Riffat (1989) is for vertical and not horizontal airflow.
• Lower discharge coefficient values compared to those for cross ventilation (Figure 3). This is probably
because the airflow through an open window into a room still preserves a large part of its mean kinetic
energy when it remains inside the room. A major part of its preserved energy is directed outside the room
through the leeward opening without interior dissipation; this is reflected in the decreasing values of ζ (or
increasing values of CD - see Equation 3) in the case of cross ventilation (Murakami et al. 1991).

CONCLUSION
The review of current literature has shown that the discharge coefficient is the result of a multivariable
impact and the use by designers of a constant value such as that given in textbooks or works of reference
is an invalid simplification. More specifically, considerable variation of the discharge coefficient with
porosity, opening configuration, Re, wind incidence angle, local geometrical and airflow conditions,
opening height, and temperature difference has been found even in simple cases such as low-rise
buildings with one opening in the façade. Comparison of different studies showed large differences in
cross ventilation. The differences are larger (up to 50%) at small porosities (less than 15%). Several
problems with the application of the orifice equation to the calculation of wind-driven cross ventilation rate
have been pointed out. Generally, lower discharge coefficient values are observed for buoyancy-driven
single-sided ventilation compared to cross ventilation.

ACKNOWLEDGEMENTS
This study has been carried out with partial support from Natural Recourses Canada through a University
Research Network Grant, for which the authors are grateful.

REFERENCES
Andersen, K.T. 1996. Inlet and outlet coefficients: a theoretical analysis. Proceedings of Roomvent 1996,
Vol. 1, pp. 379-390.
Andersen, K.T. 2002. Friction and contraction by ventilation openings with movable flaps. Proceedings of
Roomvent 2002, Copenhagen, Denmark.
ASHRAE Fundamentals Handbook. 2001. American Society of Heating Refrigeration and Air-conditioning
Engineers, Atlanta, GA.
Aynsley, R.M., W. Melbourne, B. J. Vickery. 1977. Architectural Aerodynamics. Applied Science
Publishers LTD, London.

9
Dascalaki, E., M. Santamouris, A. Argiriou, C. Helmis, D.N. Asimakopoulos, K. Papadopoulos, A.
Soilemes. 1995. Predicting single-sided natural ventilation rates in buildings. Solar Energy, Vol. 55, pp.
327-341.
Etheridge, D., M. Sandberg.1996. Building Ventilation: Theory and Measurement. John Willey & Sons,
London.
Flourenzou, F., J. Van der Maas, C.A. Roulet. 1998. Natural ventilation of passive cooling: measurement
of discharge coefficients. Energy and Buildings, Vol. 27, pp. 283-292.
Haghighat, F., J. Rao, P. Fazio. 1991. The influence of turbulent wind on air change rates – a modelling
approach. Building and Environment, Vol. 26, pp. 95-109.
He, P., T. Katayama, T. Hayashi, J. Tsutsumi. 1995. Basic examination of cross ventilation rates of single
unit models by numerical simulation. Journal of Architectural Plan and Environmental Engineering, AIJ,
No 474, pp. 47-55.
Heiselberg, P., E. Bjorn, J.T. Jensen. 2002a. Window openings – air flow and thermal comfort.
Proceedings of 4th Annual Hybvent Conference, Concordia University, Montreal, Canada, May.
Heiselberg, P., E. Bjorn, P.V Nielsen. 2002b. Impact of open windows on room air flow and thermal
comfort. International Journal of Ventilation, Vol.1, No 2.
Heiselberg, P., H. Dam, L.C. Sorensen, P.V. Nielsen. 1999. Characteristics of air flow through windows.
First International One day Forum on Natural and Hybrid Ventilation, Sydney, Australia.
Heiselberg, P., K. Svidt, P.V. Nielsen. 2001. Characteristics of airflow from open windows. Building and
Environment, Vol. 36, pp. 859-869.
Ishihara, M. 1969. Building ventilation design. Asakura Publisher Co, Ltd.
Jensen, J.T., M. Sandberg, P. Heiselberg, P.V. Nielsen. 2002a. Wind driven cross-flow analyzed as a
catchment problem and as a pressure driven flow. International Journal of Ventilation, HybVent-Hybrid
Ventilation Special Edition, Vol. 1, pp. 88-101.
Jensen, J.T., P. Heiselberg, P.V Nielsen. 2002b. Numerical prediction of natural ventilation by means of
CFD. Proceedings of Roomvent 2002, Copenhagen, Denmark.
Jiang, Y., R. Alexander, H. Jenkins, R. Arthur, Q. Chen. 2003. Natural ventilation in buildings:
measurement in a wind tunnel and numerical simulation with large eddy simulation. Journal of Wind
Engineering and Industrial Aerodynamics, Vol. 91, pp. 331-353.
Kiel, D.E., D.J. Wilson. 1989. Combining door swing pumping with density driven flow. ASHRAE
Transactions, 95 (2), pp. 590-599.
Kurabuchi, T., M. Ohba, Y. Fugo, T. Endon. 2002. Local similarity model of cross ventilation: part 1
modelling and validation. Proceedings of Roomvent 2002, Copenhagen, Denmark.
Li, Y., P. Heiselberg. 2002. Analysis methods for natural ventilation and hybrid ventilation - a critical
literature review and recent developments. International Journal of Ventilation, HybVent-Hybrid
Ventilation Special Edition, Vol. 1, pp. 3-20.
Murakami, S., S. Kato, S. Akabashi, K. Mizutani, Y-D. Kim. 1991. Wind tunnel test on velocity-pressure
field of cross-ventilation with open windows. ASHRAE Transactions, 97 (1), pp. 525-538.
Ohba, M., T.Y. Kurabuchi, T. Endon. 2002. Local similarity model of cross ventilation: part 2 application.
Proceedings of Roomvent 2002, Copenhagen, Denmark.
Pelletret, R., F. Allard, F. Haghighat, G. Liebecq, J. Van der Maas. 1991. Modelling of large openings.
Proceedings of 12th AIVC Coference, Ottawa, Canada, Sept.
Riffat, S.B. 1989. A study of heat and mass transfer through doorway in traditional built house, ASHRAE
Transactions, 95 (2), pp.573-583.
Sandberg, M. 2002. Wind induced airflow through large openings: summary. International Energy Agency
Annex 35 project Technical Report, TR24 in Annex 35 CD, Ed. Per Heiselberg, Aalborg University,
Denmark.
Santamouris, M., A. Argiriou, D.N. Asimakopoulos, N. Klitsikas, A. Dounis. 1995. Heat and mass transfer
through large openings by natural convestion. Energy and Buildings, Vol. 23, pp. 1-8.
Sawachi, T. 2002. Detailed observation of cross ventilaton and air flow through large openings by full
scale building model in wind tunnel. Proceedings of Roomvent 2002, pp. 565-568, Copenhagen,
Denmark.
Vickery, B.J., C. Karakatstanis. 1987. External wind pressure distributions and induced internal ventilation
flow in low-rise industrial and domestic structures. ASHRAE Transactions, 93 (2), pp. 2198-2213.

10