You are on page 1of 19

Architectural Science Review

ISSN: 0003-8628 (Print) 1758-9622 (Online) Journal homepage: http://www.tandfonline.com/loi/tasr20

A numerical investigation into the feasibility of


integrating green building technologies into row
houses in the Middle East

John Kaiser Calautit, Ben Richard Hughes & Saud Abdul Ghani

To cite this article: John Kaiser Calautit, Ben Richard Hughes & Saud Abdul Ghani (2013)
A numerical investigation into the feasibility of integrating green building technologies
into row houses in the Middle East, Architectural Science Review, 56:4, 279-296, DOI:
10.1080/00038628.2012.686433

To link to this article: http://dx.doi.org/10.1080/00038628.2012.686433

Published online: 28 May 2012.

Submit your article to this journal

Article views: 356

View related articles

Citing articles: 23 View citing articles

Full Terms & Conditions of access and use can be found at


http://www.tandfonline.com/action/journalInformation?journalCode=tasr20

Download by: [University of Strathclyde] Date: 29 June 2017, At: 08:45


Architectural Science Review, 2013
Vol. 56, No. 4, 279296, http://dx.doi.org/10.1080/00038628.2012.686433

A numerical investigation into the feasibility of integrating green building technologies


into row houses in the Middle East
John Kaiser Calautit1 , Ben Richard Hughes1 and Saud Abdul Ghani2
1 School of Civil Engineering, University of Leeds, Leeds, LS2 9JT, UK; 2 School of Mechanical Engineer,
Qatar University, Doha, Qatar

A traditional row house was re-designed to be adapted to the hot and arid climate of the Middle East. The vernacular design fea-
tures include a number of cooling devices such as an open courtyard, wind towers and heat-storing building materials to reduce
overheating during the summer months. This paper reports on the results of a computational uid dynamics (CFD) modelling
study of the performance of a wind tower incorporated into the row house to replace the traditional ventilation devices. The
study investigated the ways in which the resulting natural air ows in the house operated using the ANSYS Fluent CFD tool to
develop a numerical model of an optimized wind tower system. Achieved ventilation rates and temperature distribution inside
the structure were investigated. The results demonstrated that the proposed wind tower conguration was able to increase the
average indoor air velocity by 63%. An improved airow distribution is observed inside the modied row housing model.
Keywords: CFD; wind tower; green building; vernacular architecture; natural ventilation; row house

Introduction particularly in Doha, and rural areas, mainly in the north of


In recent years, the architectural design of buildings high- the country. The wind tower is a well-known element of the
lighted the potential advantage of natural ventilation sys- traditional architecture of many parts of the Middle East.
tems. Natural ventilation has become an attractive solution Used to providing natural ventilation and passive cooling,
not only for reducing the energy usage and operational costs the actual extant wind towers in the Gulf are generally ele-
but also for providing a good indoor air environment while vated structures ranging in height between 3 and 5 m above
sustaining a comfortable, healthy and productive internal the parapets of houses. Their vertical openings at their head
climate. An example of one such low-carbon ventilation direct air ow into the interior living spaces of the building
device is the wind tower. Green building technologies are below. Previous studies have highlighted the use of pools
widely promoted in line with policy drivers to reduce the and pots to incorporate evaporative cooling into their sys-
negative impacts of buildings on the environment and to tems and the passage of wind tower air through underground
increase the sustainability of developments (USGBC 2011). cooling tunnels in some buildings is also use to assist the
Approximately 40% of the principal energy consumed in associated cooling of the air streams (Boustani 2009).
buildings can be traced from the use of air-conditioning Figure 1 illustrates the arrangement of a traditional mud-
devices. Therefore, an alternative method to the more pre- walled vernacular row house in the Middle East. The house
vailing approach of using mechanical ventilation needs to features a number of adaptive devices to ease the stress
be developed in order to alleviate the associated problems caused by the hot and arid climate of the region.
with energy usage in buildings (WBCSD 2009). This study investigated a number of ways in which the
The traditional vernacular style of architecture in Qatar design and operation of traditional wind towers could be
was inuenced mainly by the local climate. Qatar has a improved by changing their materials and related thermal
yearly mean temperature of between 22 and 33 C and has mass, eliminating dust, reducing internal air short-circuiting
high daytime temperatures reaching up to 40 C or more and between windward and leeward ducts to optimize the
relative humidity of 34.580% throughout the year (WWIS systems. A number of previous studies have been used
2012). Houses were constructed close to each other, with to inform and improve modern ambient ventilation sys-
high walls, creating narrow alleys in between the struc- tems (Montazeri and Azizian 2007). Modern wind tower
tures which provided shading for the inhabitants throughout examples have been developed that incorporate louvers to
the day (Hughes et al. 2012). Traditional row house with prevent dust, direct sunlight and noise from entering the
central courtyards are scattered around the urban regions, device while allowing the external air into the enclosed

Corresponding author. Email: B.R.Hughes@leeds.ac.uk

2013 Taylor & Francis


280 J. K. Calautit et al.

Four sided wind towers


catches wind at higher
velocity

Tall building structure


minimizes direct sunlight

Constructed from local Windows allow good


mud-based materials ventilation and efficient
lighting
Raised platform for
shading of the courtyard

Open central courtyard


help to cool the house in
warm weather

Figure 1. Traditional vernacular style of architecture in the Middle East.

space, adjustable control dampers and ceiling diusers lower level to replace the air that is released (Bansal and
to optimize the fresh air ow rate and indoor conditions Bhandari 1994).
(El-Shorbagy 2001). In hot and arid regions such as the Middle East, there
A wind tower provides natural ventilation by taking is a huge dependency on electricity to run mechanical
advantage of the pressure dierences surrounding the struc- ventilation systems and provide thermal comfort. In these
ture. The slight change in air pressure can create sucient areas, using wind towers to achieve comfort is a well-
airow to improve the thermal comfort of the inhabitants. known traditional technique. The ventilation of a building
The tower can act both as a wind catcher and an exhaust. is directly related to thermal comfort. The simplest strat-
The upper part of the tower is open on four sides and a egy for improving comfort when the indoor temperature
diagonal internal partition divides the tower into four equal is felt as too warm is by daytime ventilation: providing
quarters through its height. The entrained prevailing wind comfort through higher indoor airspeeds, thus enhancing
from the windward direction is forced down the tower under the cooling sensation of the inhabitants (Givoni 1998). The
pressure to the spaces below. Warm, stale air inside the occupants of the building will feel more comfortable, even
building can simultaneously rise up the leeward ducts of the though the building is not actually being cooled. Accept-
tower to be exhausted as shown in Figure 2 (Battle 1999, able air temperature will increase as the air speed increases
Su et al. 2008). (De Dear and Brager 2002). A eld study (Hwang et al.
In the absence of wind, the driving force in the tower 2009) conducted in residents in a hot and humid region was
is the stack eect, resulting from indoor and outdoor air carried out to investigate the thermal comfort perceptions
temperature dierence. The subsequent variation of air den- of occupants and general adaptation methods. The survey
sity and pressure gradient of the indoor and outdoor air results have shown that a higher percentage of the sub-
masses forces the warm air (less dense) up to escape through jects in private homes opted for natural ventilation methods
the head vents of the tower. Fresh air is drawn in at the over using air-conditioner and adjusting clothing level. It
was found that lowering the temperature and increasing air
velocity are the most eective methods of enhancing ther-
mal comfort. A similar study (Candido et al. 2008) assessed
- Leeward
the air movement acceptability inside a naturally ventilated
+ Windward
building using a questionnaire-based survey which focused
on thermal comfort inside naturally ventilated spaces. The
(Macro Climate)
Stale Air Out work concluded that the air movement inside the room was
Fresh Air In
(Micro Climate) found to be rather adequate at speeds well excess of the
values suggested in the literature. The study also suggested
that higher air speeds are advantageous in naturally venti-
lated buildings situated in hot and humid regions in order
to improve the thermal comfort of the occupants. The study
also highlighted the importance of the ability of the inhab-
Figure 2. Cross-section of a modern conventional wind tower itants to control the airow inside the buildings according
device. to their requirements.
Architectural Science Review 281

Previous studies of the wind tower is inuenced mainly by the external


Aboulnaga and Elsheshtawy (2011) evaluated the environ- wind velocities and wind direction with respect to the wind
mental sustainability and performance of buildings situated towers quadrants. The wind tower system was modelled
in the hot and arid region of the Middle East using a for external wind speeds of the range 0.56 m/s and dier-
dynamic simulation programme was used to estimate the ent wind directions. The achieved numerical results were
carbon dioxide emission level in commercial and residen- compared with published experimental data of Elmualim
tial buildings. The study concluded that retrotting existing and indicated a good correlation between both methods of
buildings is a more eective solution in reducing the gas analysis. The work concluded that the wind tower system is
emissions than improving the energy performance of new an ecient way of delivering fresh air into a room at higher
buildings and eciency of mechanical ventilation systems. air ow rates.
Experimental, theoretical and numerical investigations of Hughes and Cheuk-Ming (2011) used experimental and
the performance of dierent types of wind tower systems CFD modelling to study the wind pressure and buoyancy-
have been reported in the literature. However, the study of driven ows through a four-sided wind tower. The work
wind tower performance has generally been limited to con- examined the relationship between the two driving forces
trolled experiments. Hence, it is dicult to compare the for the natural ventilation device. The experimental and
performance of the wind tower systems. Few of the papers CFD results showed that wind-driven forces is the main
reviewed, investigated the detailed thermal conditions and driving force for the wind tower, providing 76% more
indoor air quality of rooms using a wind tower system. indoor ventilation than buoyancy-driven forces. The study
Bouchahm et al. (2010) investigated the thermal per- also found that the eect of buoyancy forces is insignicant
formance of a one-sided wind tower system integrated in wind tower systems without exterior airow openings
to a climate adaptable house design located in a hot and other than the window itself. However, it was found that
arid region. The wind tower is combined with an evapora- the eective addition of external windows and vents in
tive cooling system. System analysis and site measurement combination with buoyancy-driven ows has the poten-
provided data which assisted the validation of the numeri- tial of increasing the internal ventilation and overcoming
cal model. The analytical model was validated against the the unavailability of wind-driven forces in dense and urban
experimental measurements and a good agreement between estates. The numerical results were validated against exper-
the results was observed. The results conrmed that the imental data and a good correlation between the dierent
airow induced by the 0.75 0.70 m tower has a direct methods of analysis was observed. To provide a steady-
eect on the reduction of internal temperature. Increasing state equivalent, time-averaged results of the internal and
the number of conduits (small-sized partitions) was found external velocities were used.
to be more ecient than increasing the height of the wetted Montazeri et al. (2010) performed a similar experimen-
column of the cooling tower. Bouchahms research con- tal analysis but chose to evaluate the ventilation perfor-
cluded that wind towers can provide a fresh supply of air and mance of a two-sided wind tower system. The analysis
improve the thermal comfort of the inhabitants regardless of was conducted at steady state and with a three-dimensional
the extreme external conditions. The work demonstrated the model. The study demonstrated the eect of the pres-
signicance of passive cooling towers and their potential as sure coecient distributions in the wind catcher openings
an alternative to the more prevalent mechanical ventilation on the natural ventilation performance of the wind tower
systems. device. The experimental results conrmed that the pres-
Kalantar (2008) also used computational uid dynam- sure coecient varies sharply with the air incidence angle.
ics (CFD) modelling and experimental studies to assess the The pressure coecient experiences its maximum value
cooling performance of a similar wind tower system in Iran. at zero air incidence angle at the wind ward side. For
The investigation was based on the weather conditions and the leeward side of the wind tower a uniform pressure
comfort criteria for the city of Yazd. A CFD tool, ANSYS distribution is observed. Increasing the incidence angle,
Fluent was used to analyse the air ow for steady-state con- average pressure coecient reduces rapidly to reach its min-
ditions and three-dimensional ow elds. The eects of the imum value at 90 . The work demonstrated the potential
parameters including the height of the tower, the indoor of multi-directional wind catcher for enhancing the natural
and outdoor air temperatures, the wind velocity and the ventilation inside residential buildings. Furthermore, ana-
relative humidity were investigated. The simulated stream- lytical and CFD modelling was used to assess the accuracy
lines were used to analyse the airow distribution inside the of the experimental results. Good agreement between the
structure. This approach showed a good correlation between dierent methods of analysis was observed.
the CFD results and the published experimental data. The Elmualim (2006) evaluated the performance of a wind
results conrmed that the wind tower system reduced the tower system integrated to a test room using numerical and
indoor air temperature by 1015 C at its peak performance. experimental analysis. The air ow achieved through the
Li and Mak (2006) used CFD to evaluate the ventilation wind tower depends on the speed and direction of the wind.
capabilities of a similar wind tower device integrated to a The study showed that the CFD calculation correlates with
test room. The study showed that the thermal performance the experimental results of the wind tunnel testing. The
282 J. K. Calautit et al.

author also concluded that CFD calculations are faster and Traditional row housing model
easier to use in obtaining results compared with wind tunnel A three-dimensional model of a standard row housing
testing which is expensive and time consuming. model with of the dimension 10 m 9 m with a vertical
height of 6 m was used for the CFD modelling. The house
surrounds the courtyard with high walls to provide shad-
Aims of this study ing for the inhabitants. The rooms with large windows
Building on the ndings of previous studies this study aimed were built around the open courtyard, allowing the wind
to optimize the performance of the parameters and compo- to circulate freely throughout the room and provide day-
nents using the developed computational model. A novel light. However, the row house eliminates the option of side
wind tower system design was developed which can be window as shown in Figure 3.
retrotted into existing row housing models. The numer- The unique feature of house is the 1.5 1.5 m wind
ical simulation and experimental results demonstrated the tower, rising 2.5 m above the roof level as shown in Figure 3.
importance of the use of computational uid dynamics in The upper part of the tower is open on four sides and a pair
assessing the performance of a wind tower system. The of divider is placed diagonally across each other along its
good correlation between both methods of analysis sug- length. Unlike conventional wind towers, air is supplied
gests that the CFD techniques in use were suitable for this to the enclosed space through the diusers located near
type of device and such has been used for the purpose of the oor, rather than from the ceiling. This is to decrease
this research. the air velocity at the base of the tower allowing the dust
The study evaluates the performance of the proposed particles to settle on the dust collectors. Hence, the total
model by simulating and analysing the air movement in area of the house and the free space available for ventila-
the structure using a CFD model. The level of performance tion is reduced. However, oor ventilation systems can be
of the traditional and modied design is assessed based on integrated with evaporative cooling systems. In traditional
the results of the internal air velocity, pressure and tem- structures the airow is passed through wetted columns or
peratures. This would allow nding an alternative device to fountains located at the lower oors to cool the air before
the traditional ventilation techniques, which would be more inducing it to the living space. Therefore it is important to
energy ecient, less expensive and provide the best results have openings and vents between the tower and the upper
of a decreased indoor air temperature oors, routing all the air to the ground oor and carry the
air to the oors above. As air ows out from the basement,
the dust drops on the oor, and then moves up through the
Design geometry external vents to the rest of the structure. In this case,
The geometry was created using Solid Edge ST software the upper oor is linked to the ground oor via the stair
and imported into ANSYS Workbench Model Designer case. The performance of the traditional house was inves-
to generate a numerical model of the design. The geom- tigated under the wind speed of 4 m/s and a constant wind
etry of the standard row housing model (microclimate) direction.
with the height, width and length of 6, 10 and 9 m was
used for the computational analysis. An enclosure repre-
senting the macroclimate was created to obtain a sucient Modied row housing model
distance between the boundary conditions (Mehta 1991). Figure 4 displays the row housing model highlighting the
The enclosure with a height, width and length of 23, 40 and integrated green technologies such as the solar panels,
39 m creates a direct interface through the geometry. shading devices and energy-ecient building materials.

Figure 3. Schematic of a traditional row housing model with two four-sided wind towers.
Architectural Science Review 283

The four-side wind tower geometry of dimension hence maintenance cost is reduced. The open courtyard is
1.5 m 1.5 m and height of 2 m is positioned above the also connected to a carport to enhance the airow rate.
house replacing the ancient cooling tower, as illustrated The wind tower system supplies air to the enclosed space
in Figures 5 and 6. Similarly, the ventilation device is through the diusers located at ceiling level. Hence, more
divided into four segments in order to induce air ow into free space is available for ventilation on the ceiling than on
the enclosed space regardless of direction of the prevailing the corresponding oor of equal area. The lower oors are
wind. A one-sided wind tower of the dimension 1.5 m 2 m connected to the ventilated space through a stair case which
and height of 1.8 m is located opposite the multi-directional is located just below the multi-directional wind tower. This
tower to maintain the building cross-ventilation. In the maximizes the cross-ow ventilation within the structure.
absence of wind, the one-sided wind tower functions as The performance of the modied row housing model was
a solar chimney. During the day, the tower is heated and investigated under the wind speed of 4 m/s and a constant
the air within, creating a pressure gradient which allows wind direction.
less dense hot air to rise up and exit out the top. External
vents are positioned on the other end of the structure to
assist the wind towers. Similarly, the rooms facing the cen- Computational model
tral courtyard are equipped with operable windows. Large The CFD analysis was carried out using the ANSYS
overhangs are positioned on top of the building to provide 12.1 Fluent software. The CAD model was imported to
shading to the courtyard and direct the air ow down to the Workbench: Model Designer to generate a computational
open space and improve the air circulation. In addition, solar model. An enclosure was created to represent the external
panels are mounted on the inclined overhangs to provide an wind environment (macroclimate). The air streamline had a
alternative energy source. Tilted solar panels collect less constant velocity of 4 m/s through the enclosure to simulate
dust and debris compared with horizontal-oriented panels, the actual wind environment. The enclosure was set with

Insulated window shades


Shading device
incorporates tilted solar
panels for alternative Modern wind catcher with
energy source external louvers used to deflect
dust, direct sunlight and noise.
Overhangs directs
airflow down to the Operable external vents
open courtyard assist in cooling by
circulating air better
Energy efficient
building materials Open courtyard

Car port linked to the open


courtyard improves airflow
circulation

Figure 4. Modied row housing model arrangement with a four- and one-sided wind towers.

Figure 5. Schematic of a modied row housing model integrated with modern wind towers.
284 J. K. Calautit et al.

Top Hat
Cross Divider

See Detail A
45 Louvre

Adjustable
Dampers
Roof Level

Ceiling Diffuser
Figure 6. Schematic of a 1.5 m 1.5 m four-sided wind tower system.

boundary walls of the macroclimate (the ow domain) at Flow Face In Flow Face Through Flow Face Out
a distance from the geometry to avoid reversed ow in the Enclosure Fluid Indoor- Fluid Enclosure Fluid
region (Mehta 1991). (Macro Climate) (Micro Climate) (Macro Climate)

Mesh generation
The quality of the mesh has important implication on the
convergence and the level of accuracy of the achieved
results (Chung 2002). A non-uniform mesh was applied 2
to the volumes of the computational model. The mesh 1
arrangement consisted of 30,153 hybrid non-uniform mesh
elements with 54,315 nodes, as shown in Figure 7. The
size of the mesh element was extended smoothly to resolve
the sections with high gradient mesh and to improve the
accuracy of the results.
A similar approach was used when applying mesh to the
modied design with a modern wind catcher device. The Figure 8. Flow domain representation of the physical geome-
mesh arrangement consisted of 49,920 hybrid non-uniform try of the modied design under investigation and location of set
boundary conditions.
mesh elements with 90,645 nodes, as illustrated in Figure 8.

Boundary conditions
Flow Face In Flow Face Through Flow Face Out The k-epsilon (2eqn) viscous model with standard wall
Enclosure Fluid Indoor- Fluid Enclosure Fluid
(Macro Climate) (Micro Climate) (Macro Climate)
functions was used for the numerical simulation as this is the
most appropriate viscous model for an inlet wind velocity
of 4 m/s and lower (Ayad 1999). The two equation model
includes two extra transport equations to represent the tur-
bulent properties of the ow. The rst transported variable
is turbulent kinetic energy, k, it determines the energy in
the turbulence. While the second transported variable in
2 this case is the turbulent dissipation. It is the variable that
determines the scale of the turbulence. The k-epsilon model
1
is used for free-shear layer ows with relatively small pres-
sure gradients. The geometry (microclimate) was modelled
as an open structure with openings on the windward and lee-
ward side which allows the incoming air to pass through it,
in order to simulate and analyse the air ow pattern inside
the structure and determine the average interior velocity.
Figure 7. Flow domain representation of the physical geometry The geometry faces were set as solid walls with a thickness
of the traditional design under investigation and location of set of 0.04 and 0.02 m to create a boundary between the interior
boundary conditions. and exterior.
Architectural Science Review 285

Table 1. CFD model boundary conditions. wind ward and leeward side of the four-sided wind towers.
The measurement vertices were based on the entry location
Traditional Modern
of the air and the exit location of air at the rst and second
Geometry Solid Solid tower. The vertex points were positioned 350-mm away
Inlet face Fluid (air) Fluid (air) from the inlet and outlet openings of the geometry. Vertex
Outlet face Fluid (air) Fluid (air) points were also located inside the courtyard to measure the
Ground Wall Wall average velocity and temperature as shown in Figure 9.
Velocity inlet (m/s) 4 4
Pressure outlet Atmospheric Atmospheric Similarly, points were created at the wind ward and lee-
Viscous model k-epsilon k-epsilon ward side of the four-sided wind tower and leeward side of
(equation (2)) (equation (2)) the exhaust tower. The measurement vertices were based on
Near-wall treatment Standard function Standard function the entry location of the air and the exit location of air. The
Velocity formulation Absolute Absolute vertex points were positioned 350-mm away from the inlet
Time Steady Steady
Gravity 9.81 (Z-axis) 9.81 (Z-axis)
and outlet openings of the geometry as shown in Figure 10.
Wall material Clay Concrete
Wall thickness (m) 0.04 0.02
Free stream 315 315 Grid adaptation
temperature Grid adaptation was used to validate the programming
External 315 315 and computational operation of the computational model.
temperature (K)
The numerical grid was rened and locally enriched using
Thermal conductivity 1.8 1.5
Heat transfer 8.9 8.9 the hp-method grid adaptation technique. This procedure
coecient of evaluation requires the use of dierent mesh sizes or
(W/m2 K) higher-order approximations. The accuracy of the results is
improved by using successively smaller cell sizes for the
computation (Blocken et al. 2009).
The enclosure (macroclimate) was created to simulate The grid adaptation procedure increases the number of
the external wind velocity. The enclosure consists of a nodes and elements of the numerical model. The grid was
velocity inlet (operating velocity) on the left hand side of evaluated and rened until the error becomes insignicant
the enclosure, and a pressure outlet (atmospheric pressure) between the number of elements and computational iter-
on the opposing boundary wall of the enclosure as shown in ations (Alauzet et al. 2007). The grid adaptation method
Figures 6 and 7. The rest of the macroclimate sides were set starts with a coarse mesh and gradually renes it until the
as symmetry faces to alleviate the return velocities which variation observed between the results are smaller than the
may alter the results (Hughes and Cheuk-Ming 2011). The predened acceptable error as shown in Tables 2 and 3. The
incoming and free ow velocities were set as a constant applied boundary conditions were remained xed through-
4 m/s. The base of the structure is set as ground (wall) out the simulation process to ascertain precise comparison
and the remaining faces were set as symmetry to allevi- of the numerical results. Also, owing to the complexity of
ate any return velocities which could aect the precision of the geometry, the mesh was generated using hybrid mesh.
the results. The boundary conditions for the CFD model are Figures 11 and 12 show the solved variables at the
shown in Table 1. points that are specied in Figure 9 versus the number of
To measure the velocity entering and exiting the stan- elements. At 48,449 elements the error between the grid
dard row housing model, vertex points were created at the renements was at its lowest in the last two steps. An

Figure 9. Location of vertex measurement points positioned 350-mm away from the inlet and outlet of the wind towers.
286 J. K. Calautit et al.

Figure 10. Location of vertex measurement points positioned 0.35-m away from the inlet and outlet of the wind towers.

Table 2. Grid adaption for verication of the CFD model (traditional design).

Number of nodes 54,315 55,844 55,909 86,706 87,381


Number of elements 30,153 31,328 31,388 47,920 48,449
Relevance centre Coarse Coarse Coarse Medium Medium
Smoothing High Medium Low High Medium
Mesh matrix Tet4, Hex8, Wed6 Tet4, Hex8, Wed6 Tet4, Hex8, Wed6 Tet4, Hex8, Wed6 Tet4, Hex8, Wed6
Average velocity inlet (m/s) 1.571 1.418 1.618 1.609 1.612
Average velocity outletT2 (m/s) 0.852 0.819 0.951 0.943 0.945
Average indoor velocity (m/s) 0.350 0.339 0.362 0.363 0.363
Average pressure inlet (Pa) 0.903 0.875 0.947 0.961 0.962
Avg. pressure OutletT2 (Pa) 0.431 0.240 0.451 0.481 0.483
Average indoor pressure (Pa) 0.114 0.106 0.115 0.116 0.116

Table 3. Grid adaption for verication of the CFD model (modied design).

Number of nodes 90,645 91,089 92,795 101,909 101,231


Number of elements 49,920 50,191 51,487 58,426 57,859
Relevance centre Coarse Coarse Coarse Medium Medium
Smoothing High Medium Low High Medium
Mesh matrix Tet4, Hex8, Wed6 Tet4, Hex8, Wed6 Tet4, Hex8, Wed6 Tet4, Hex8, Wed6 Tet4, Hex8, Wed6
Average velocity inlet (m/s) 1.681 1.741 1.775 1.783 1.785
Average velocity outletT2 (m/s) 1.422 1.438 1.449 1.458 1.459
Average indoor velocity (m/s) 0.531 0.557 0.560 0.563 0.563
Avg. pressure inlet (Pa) 2.435 2.437 2.439 2.441 2.442
Avg. pressure outletT2 (Pa) 1.386 1.385 1.448 1.465 1.468
Avg. indoor pressure (Pa) 0.272 0.279 0.282 0.283 0.283

acceptable compromise was reached between the accuracy it is seen that the air ow is entering from the left side of
of results and computational iterations. the enclosure (velocity inlet). The airow splits at the wind
Figures 13 and 14 show the solved variables at the points ward side of the structure with the air entering the wind
that are specied in Figure 10 versus the number of ele- tower openings and the remaining ow shearing across the
ments. At 57,859 elements the error between renements structure and exiting to right side of the enclosure (pres-
was at its lowest in the last two points. An acceptable com- sure outlet). The air ow passing the wind tower channel is
promise was reached between the accuracy of results and accelerated as it enters the interior room, reaching a maxi-
computational iterations. mum velocity of 1.8 m/s. The air ow is directed down to
lower oors and ows up to the rst oor via the staircase.
Air-short circuiting is observed for both wind towers, as
CFD results
most of the air ow exits the exhaust vent without owing
Vectors of velocity magnitude into the structure. The air that is released from exhaust of
Figure 15 illustrates the velocity vector of a middle plane tower 1 re-enters the structure through the inlet of tower 2.
in the traditional row house geometry. From the diagram Hence, the stale air is re-circulated back inside the enclosed
Architectural Science Review 287

Figure 11. The solutions using various grid adaptation methods. The variables are the average velocity at the inlet and outlet of the wind
towers and interior of the structure (traditional design).

Figure 12. The solutions using various grid adaptation methods. The variables are the average dynamics pressure at the inlet and outlet
of the wind towers and interior of the structure (traditional design).

space which has reduced the fresh air supply capabilities of is directed down to the enclosed space through the ceiling
this conguration. diuser. The airow is accelerated as it shears down the
Figure 16 shows the velocity vector of the cross- walls and oor of the structure, reaching a maximum veloc-
sectional plane in the modied row housing model. From ity of 2.1 m/s. Similarly, air-short circuiting is observed,
the illustration it is seen that the air ow enters from the some of the air ow exits the leeward quadrant without
left side of the enclosure (velocity inlet). The airow splits entering the ventilated space, this is due to the pressure dif-
at the wind ward side of the four-sided wind tower with ferentials surrounding the device. The warm and stale air
the air entering the ventilation device and the remaining rises up and exits the one-sided wind catcher located at the
ow passing around the sides of the device and exiting other end of the structure. However, air-short circuiting can
to right side of the enclosure (pressure outlet). The air be reduced by adjusting the dampers of the modern wind
ow entering the windward quadrants of the wind tower tower device.
288 J. K. Calautit et al.

Figure 13. The solutions using various grid adaptation techniques. The variables are the average velocity at the inlet and outlet of the
wind towers and interior of the structure (modied design).

Figure 14. The solutions using various grid adaptation techniques. The variables are the average dynamics pressure at the inlet and outlet
of the wind towers and interior of the structure (modied design).

Contours of velocity magnitude outlet of the cooling device. The airow entering the device
is directed towards the interior rooms via the wind tower
Figure 17 illustrates the airow distribution inside the stan- channels. The air velocity is reduced as it ows up to the sec-
dard row housing model with two four-sided wind towers. ond oor and increases as it exits the exhaust openings. The
From the velocity contour plot, a distinct area of higher volume ow rate of the air supplied by the traditional wind
velocity is seen directly below the wind towers. This is tower system is 730 L/s (8.11 L/s/m2 ) at external velocity
a result of the pressure dierences between the inlet and of 4 m/s.
Architectural Science Review 289

Figure 15. Velocity vectors of a cross-sectional plane in the traditional row housing model with an inlet velocity of 4 m/s.

Figure 16. Velocity vectors of a cross-sectional plane in the modied row house model with an inlet velocity of 4 m/s.

Figure 17. Velocity contour lines of a cross-sectional plane in the traditional row housing model with an inlet velocity of 4 m/s.
290 J. K. Calautit et al.

Figure 18. Velocity contour lines of a cross-sectional plane in the modied house model with an inlet velocity of 4 m/s.

Figure 19. Contours of dynamic pressure (traditional design).

Figure 18 illustrates the airow distribution at a middle device. From the illustration, it is observed that the areas of
plane passing through the centre of the wind tower geom- high pressure are observed inside the wind tower channel
etry. Similarly, higher velocity is observed directly below and the room below it. Although low-pressure areas are seen
the four-sided wind tower. This was due to the displace- inside the microclimate and at the outlet of the wind tower.
ment eect between the inlet and outlet of the device. The Figure 20 shows the pressure distribution around a stan-
air ow entering from the windward side of the structure dard row housing model integrated with a conventional
creates a positive pressure and an equivalent negative pres- wind tower. Similarly, areas of high pressure are observed
sure is formed at the leeward side drawing the warm air at the windward side of the wind tower device and areas
out of the house. The internal air velocity is decreased as it of low pressure on the leeward side, this was expected.
ows down to the lower oor and increases as it exits the It is observed that pressure is signicantly reduced inside
exhaust tower. The ow is circulated throughout the struc- the microclimate. Also, a pressure build-up is seen directly
ture and an improved air velocity distribution is observed below the wind tower which is a result of the pressure
for this conguration. The volume ow rate of the air sup- dierence between the inlet and outlet of the device.
plied by the roof-mounted wind tower system is 1064 L/s
(11.82 L/s/m2 ) at an external velocity of 4 m/s.
Flow visualization
Figure 21 illustrates the velocity streamlines for the tradi-
Contours of dynamic pressure tional geometry representing the ow distributions within
Figure 19 shows the pressure distribution at a cross- the micromacro climate. The ow enters uniformly from
sectional plane passing through the centre of the ventilation the velocity inlet before it splits with some of the air entering
Architectural Science Review 291

Figure 20. Contours of dynamic pressure (modied design).

Figure 21. (a) Velocity streamlines on the ground oor with top view. (b) Velocity streamlines on the rst oor with top view. (c) Velocity
streamline with isometric view.

the wind towers and some shearing across the sides of the Figure 22 shows the velocity streamline pattern inside
external walls. Uneven airow distribution is observed on the macromicro climate of the modied row housing
the lower oors. model integrated with a modern wind tower device. From
292 J. K. Calautit et al.

Figure 22. (a) Velocity streamlines on the ground oor with top view. (b) Velocity streamlines on the rst oor with top view. (c) Velocity
streamline with isometric view.

the diagram, improved airow distribution is observed on and external conditions with the ambient external air tem-
the ground and rst oor. The air ow is circulated inside the perate set at 315 K and the free stream temperature or the
microclimate reaching the walls and oors of the structure. reference temperature set at 315 K. Internal heat sources
Figure 23 shows the airow distribution inside the court- and solar loading were not considered in the numerical
yard of the modied row housing model incorporating large model.
overhangs. Unlike the rst conguration, only one opening Figure 24 illustrates the contour of static temperatures
is used for the analysis and the raised platforms are removed inside the row housing model incorporating two four-sided
to prevent airow disruption. It is observed that air ow is wind towers. It is observed that the temperature follows
accelerated as it enters the lower areas of the courtyard, the same route as the velocity. Hence, lower temperature is
reaching a maximum velocity of 3 m/s. observed inside and below the ventilation devices. The aver-
age temperature inside the microclimate is 313 K with the
macroclimate temperature set at 315 K. This was expected,
Temperature distribution when a building is cross-ventilated during daytime hours
The row housing model was used in the CFD analysis the temperature of the indoor air and surface follow closely
to investigate the inuence of the wind tower ventilation the ambient temperature (Givoni 1998).
rates on the thermal performance of the structure. The Figure 25 illustrates temperature distribution inside the
computational models were run for a xed external wind modied row housing model integrated with a modern wind
speed of 4 m/s with an average internal velocity mea- tower device. Similarly, it is observed that the tempera-
sured at 0.34 and 0.55 m/s. Similar wall material properties ture follows the same path as the velocity. Hence, lower
were used for the structure of the row house (Table 1). A temperature is observed inside the towers column and
convective thermal model was used to simulate the internal across the walls of the structure. The average temperature
Architectural Science Review 293

Figure 23. Air ow distribution inside the courtyard with large overhangs.

Figure 24. Temperature distributions within a building integrated with two four-sided wind towers.

Figure 25. Temperature distributions within a structure integrated with a modern four-side wind tower and up-draught tower.
294 J. K. Calautit et al.

Table 4. Traditional design CFD results. contour plot. As observed from the diagram, the velocity
is reduced as it moves towards the structure and the ow is
signicantly decreased as it exits from the exhaust. Vertex
points were created at the wind ward and leeward side of
the four-sided wind towers to measure the average weighted
velocity and pressure at those areas.
With reference to Figure 15, computational air velo-
city of 1.58 m/s is obtained at the windward side of the
geometry. The measured average velocity inside the geom-
etry is 0.34 m/s which is signicantly lower than the initial
Average indoor velocity (m/s) 0.34 air velocity. This is due to the uneven air ow distribution
Average velocity at inlet (m/s) 1.57 inside the structure which is a result of air-short circuiting
Average velocity at outlet T1 (m/s) 0.85 of the two wind towers.
Average velocity at outlet T2 (m/s) 0.75 Figure 16 displays the velocity contours for the modied
Average courtyard velocity (m/s) 0.62
row housing model. As displayed, the windward air veloc-
Average indoor pressure (Pa) 0.11 ity obtained is 2.11 m/s which is 33% higher than the rst
Average pressure at inlet (Pa) 0.90 conguration. An average internal air velocity of 0.55 m/s
Average pressure at outlet T1 (Pa) 0.12
Average pressure at outlet T2 (Pa) 0.43 is obtained which is 63% higher than the standard tradi-
tional conguration. An improved air ow distribution is
Average indoor temperature (K) 313 observed inside the structure. In relation to air movement
Indoor temperature reduction (K) 1.7
Average courtyard temperature (K) 313
acceptability, it remains to be exceeding the recommended
comfort standards in the literature (Candido et al. 2008),
of 0.25 m/s.
Figure 17 shows the velocity contour plot obtained at
Table 5. Modied design CFD results
a free stream velocity of 4 m/s. A distinct area of higher
velocity is observed directly below the ventilation device.
This is due to the displacement eect between the inlet
and the outlet of the cooling tower. The air entering the
device generated a positive pressure at the wind ward
side and consequently negative pressure is generated on
the leeward side drawing the stale and warm air out of
the house. Uneven airow distribution is observed inside
the structure.
Figure 18 is based on the front view of the geometry
Average indoor velocity (m/s) 0.55
Average velocity at inlet (m/s) 2.10
displaying the velocity contour plot. The velocity contour
Average velocity at outletT1 (m/s) 0.56 of the cross-sectional planes in the basic house model repre-
Average velocity at outletT2 (m/s) 1.45 sents the air ow distribution within the structure. As seen
Average courtyard velocity (m/s) 0.80 from the illustration, the air is accelerated as it enters the
Average indoor pressure (Pa) 0.27 opening of the four-sided wind tower. The air stream is cir-
Average pressure at inlet (Pa) 2.55 culated inside the structure with some of the air rising up
Average pressure at outletT1 (Pa) 0.13 to the exhaust quadrant of the wind tower. Although the
Average pressure at outletT2 (Pa) 1.48 remaining air ow exits, the one-sided wind tower located
Average indoor temperature (K) 313 on the other end. An improved air velocity distribution is
Indoor temperature reduction (K) 2 observed inside the structure.
Average courtyard temperature (K) 313 As displayed in Figure 19, the dynamic pressure con-
tours for the standard row housing model are obtained with
an average macro-climate pressure of 10.67 Pa and a micro-
inside the microclimate is 313 K with the macroclimate climate pressure of 0.11 Pa. As seen from the diagram, the
temperature set at 315 K. dynamic pressure inside the structure is signicantly lower
Tables 4 and 5 compare the calculated CFD results of the compared with the inlet conditions.
traditional and modied row house computational models. Figure 20 displays the cross-sectional representation of
the dynamic pressure contours for the modied row housing
model. An average macroclimate pressure of 10.25 Pa is
Discussion obtained. From the gure, the average dynamic pressure
Figures 15 and 16 display velocity vectors with the ow inside the structure is 0.27 Pa which is relatively higher
direction coming in from left- to right-hand side of the than the traditional row house geometry.
Architectural Science Review 295

Figures 21 and 22 display the air streamline pattern structure to provide shading to the inhabitants and improve
through the built geometry. Reduced circulated ow at the the airow inside the courtyard. An average air velocity
geometry interior is obtained in comparison to the free of 0.81 m/s is obtained from the modied conguration,
stream velocity of 4 m/s. it is noticeable that the air stream- which is 28% higher than the standard model.
lines in the second conguration is higher compared with its The computational model showed a good correlation
counterpart, thus highlighting the improved air circulation between the results, and the hp-grid adaptation technique
at the interior and the courtyard. is a well-recognized practice for achieving precise numer-
Figure 23 displays the air streamline pattern through ical results with the chosen software (Chung 2002). The
the open courtyard. An average air velocity 0.81 m/s is percentage of error between the grid renements is in the
obtained for the modied design which is 28% higher than range of 0.286.79%.
the standard model. This demonstrates the capabilities of the Traditional wind towers have been designed and built
overhangs in directing the air ow down to the courtyard. using old methodologies; therefore, they have design limi-
Figures 24 and 25 display the contour of static tempera- tations which can be eliminated with advanced technology.
tures inside the row housing model. Lower air temperature The achieved results demonstrate the capabilities of the new
is observed inside the microclimate. An average tempera- passive ventilation techniques in providing eective natural
ture of 313 K is obtained inside the modied row housing ventilation.
model, which is 13% lower than the traditional design. The conclusions of this work are based on the numerical
The results are based on steady-state CFD simulations results of CFD and further works will validate the results
and a full transient thermal model of a Qatari residence using wind tunnel experimentations. These ndings are part
can be designed to accurately replicate a dynamic ther- of a large Qatar sponsored project to develop a low-energy
mal response of a residential building to external climatic residential cooling device to cut heating, ventilation, and air
conditions. conditioning loads. Future work and results will be shared
as the device reaches eld trial stage.
Conclusions
This work compares the ventilation and thermal perfor-
mance of a traditional and a modied row housing model Acknowledgements
in terms of the achieved internal air velocity, pressure and This publication was made possible by a NPRP Grant from Qatar
temperatures. The investigation determined the feasibil- National Research Fund (A member of the Qatar Foundation). The
ity of retrotting traditional vernacular architecture with statements made herein are solely the responsibility of the authors,
NPRP 09-138-2-059.
green technologies such as wind tower devices. From the
obtained numerical results, it is concluded that the pro-
posed wind tower conguration is a sustainable alternative
to the traditional cooling methods for providing natural air References
ventilation. AboulNaga, M. and Elsheshtawy, Y., 2011. Environmental sus-
The results showed that the proposed wind tower cong- tainability assessment of buildings in hot climate: the case of
uration was able to increase the average indoor air velocity the UAE. Renewable Energy, 24, 553.
by 63%. An average temperature of 313 K is obtained inside Alauzet, F., et al., 2007. 3D transient xed point mesh adaptation
for time-dependent problems: application to CFD simula-
the modied row housing model, which is 13% greater tions. Journal of Computational Physics, 222, 592623.
than the traditional design. Improved internal air veloc- Ayad, S., 1999. Computational study of natural ventilation.
ities, airow movement and temperature distribution is Journal of Wind Engineering, 82, 4968.
observed in the proposed structure, based on the velocity Bansal, N.K. and Bhandari, M.S., 1994. A study of solar chim-
vectors and temperature contours obtained from compu- ney assisted wind tower system for natural ventilation in
buildings. Building and Environment, 29, 495500.
tational model. This was expected, the multi-directional Battle, G., 1999. Wind towers: detail in building. London:
roof-mounted wind tower system directly supplies air to Academy Press, 2527.
the enclosed space through the diusers located at ceiling Blocken, B., et al., 2009. High-resolution CFD simulations for
level. Hence, more free space is available for ventilation forced convective heat transfer coecients at the facade
on the ceiling than on the corresponding oor of equal of a low-rise building. Building and Environment, 44,
23962412.
area as shown in the CFD results. The lower oors are Bouchahm, Y., Bourbia, F., and Belhamri, A., 2010. Performance
connected to the ventilated space through a stair case analysis and improvement of the use of wind tower in hot dry
which is located just below the multi-directional wind climate. Renewable Energy, 36, 898906.
tower. This maximizes the cross-ow ventilation within the Boustani, F., 2009. Water utilization in Arid region of Iran by
building. Qanats international. Journal of Human and Social Sciences,
4, 505508.
The study also highlights the signicance of the open Candido, C., et al., 2008. Natural ventilation and thermal com-
courtyard as an integral part of the traditional row hous- fort: air movement acceptability inside naturally ventilated
ing design. Large overhangs are mounted on top of the buildings in Brazilian hot humid zone. Air Conditioning and
296 J. K. Calautit et al.

Low Carbon Cooling, 2729 July, Cumberland Lodge, Wind- Kalantar, V., 2008. Numerical simulation of cooling performance
sor, UK. London: Network for Comfort and Energy Use in of wind tower (Baud-Geer) in hot and arid region. Renewable
Buildings. Energy, 34, 246254.
Chung, T.J., 2002. Computational uid dynamics. Cambridge: Li, L. and Mak, C.M., 2006. The assessment of the performance
Cambridge University Press, 102125. of a windcatcher system using computational uid dynamics.
De Dear, R. and Brager, G., 2002. Thermal comfort in natu- Building and Environment, 42, 11351141.
rally ventilated buildings: revisions to ASHRAE Standard Mehta, U.B., 1991. Some aspects of uncertainty in computational
55. Energy and Buildings, 34, 549561. uid dynamics results. Journal of Fluids Engineering, 113,
Elmualim, A., 2006. Dynamic modelling of a wind catcher/tower 539543.
turret for natural ventilation. Building Services Engineering Montazeri, H. and Azizian, R., 2007. Experimental study on
Research & Technology, 27, 165182. natural ventilation performance of one-sided wind catcher.
El-Shorbagy, A., 2001. Design with nature: wind catcher as a Building and Environment, 43, 21932202.
paradigm of natural ventilation device in buildings. Inter- Montazeri, H., Azizian, R., and Mostafavi, S., 2010. Two-sided
national Journal of Civil & Environmental Engineering, 10, wind catcher performance evaluation using experimental,
2631. numerical and analytical modeling. Renewable Energy, 35,
Givoni, B., 1998. Climate considerations in buildings and urban 14241435.
design. New York: Wiley, 3839. Su, Y., et al., 2008. Experimental and CFD study of ventila-
Hughes, B.R., Calautit, J.K., and Ghani, S.A., 2012. The devel- tion ow rate of a Monodraught windcatcher. Energy and
opment of commercial wind towers for natural ventilation: a Buildings, 40, 11101116.
review. Applied Energy, 92, 606627. USGBC, 2011. U.S. green building council: green building facts
Hughes, B.R. and Cheuk-Ming, M., 2011. A study of wind and [online]. Available from: http://www.usgbc.org [Accessed
buoyancy driven ows through commercial wind towers. 20 June 2011].
Energy and Buildings, 43, 17841791. WBCSD, 2009. Energy eciency in buildings [online]. Available
Hwang, R., et al., 2009. Thermal perceptions, general adaptation from: http://www.wbcsd.org [Accessed 7 November 2011].
methods and occupants idea about the trade-o between WWIS, 2012. Weather information for Qatar [online]. Avail-
thermal comfort and energy saving in hothumid regions. able from: http://worldweather.wmo.int [Accessed 5 April
Building and Environment, 44, 11281134. 2012].

You might also like