Content uploaded by Zhaleh Hedayat
Author content
All content in this area was uploaded by Zhaleh Hedayat on Sep 07, 2016
Content may be subject to copyright.
The 3rd International Conference and Exhibition on Solar Energy
ICESE-2016
5-6 September, 2016, University of Tehran, Tehran, Iran
___________________________________________________
1
ICESE2016-1128
Energy performance analysis of solar-wind catchers under hot and dry climatic condition in
Iran-Yazd
Zhaleh.HEDAYAT *1 , Nima.SAMKHANIANI 2, Bert.BELMANS1 , M.Hossein.AYATOLLAHI 3,
Ine.WOUTERS 1, Filip.DESCAMPS 1
1. Dept. of Architectural Engineering , Vrije Universiteit Brussel, Belgium , Zhaleh.Hedayat@vub.ac.be ,
bbelmans@vub.ac.be , Ine.Wouters@vub.ac.be ,Filip.Descamps@vub.ac.be
2. Dept. of Mechanical Engineering, Tarbiat modares University, Iran , Nima.samkhaniani@modares.ac.ir
3. Dept. of Architecture,Yazd University, Iran , hayatollahi@yazd.ac.ir
Abstract
Ancient wind catchers of Yazd city as net zero energy
building components request the detailed numerical and
experimental studies to renovate and reuse in old part of
the city and for developing the new design strategies in
future. In this ingenious natural ventilation strategy no
other energy sources are required except wind and sun.
In this research, the effect of both wind and sun on the
flow and thermal behaviour of the solar-wind catchers
of Yazd city - a city with very high solar radiation - in
Iran, are investigated numerically and experimentally.
This paper represents a systematic evaluation of
isothermal 3D CFD modelling for predicting mean air
velocity distributions over and through the wind catcher
at three reference wind directions of 330°, 90° and 150°,
using Open source CFD package (OpenFOAM). The
evaluation is based on a grid-sensitivity analysis and on
validation with full scaled long term measurement data
set. The validated CFD simulation results show that
steady RANS with Standard k−ɛ model, in spite of its
limitations, can accurately predict the flow behaviour of
wind catchers at wind direction of 90° with the average
deviations of 19 %, , which is considered a good
agreement against the long term experimental results.
Moreover the thermal behaviour of wind catcher during
a summer day is discussed. The study reveals that the
solar wind catchers of Yazd city, is quite effective in
lowering air temperature and enhancing air circulation
through the buildings. The result of this study will be
useful for designing solar-wind catchers under hot and
dry climate conditions.
Keywords: Solar-Wind catcher, long-term experimental
dataset, Yazd city, Iran.
Introduction
Wind catchers of Yazd city in Iran so called Baud-Geers
in Persian, extract and supply air into the buildings
using ventilation principles of stack ventilation and
wind-induced ventilation, respectively. In this ingenious
natural ventilation strategy no other energy sources are
required except wind and sun. In hot and dry regions,
the impact of solar radiation on wind catcher
performance is higher comparing to the other regions.
Many researchers have been studied the functional
behavior of the wind towers in Iran. Bahadori [2] ,
performed full analysis of the design of wind towers in
several locations such as Yazd city and presented two
new designs of wind towers. Yaghubi M.A, 1991
studied a thermal performance of three naturally
ventilated building linked to the wind tower.A short-
term experimental study is conducted to analyze the
wind towers performance [13]. Montazeri [8] also
investigated the effects of the numbers of openings by
modeling a circular cross section wind tower that has
several openings at equal angels. In this research, the
wind and sun energy performance analysis of the solar-
wind catchers of Yazd city - a city with very high solar
radiation - in Iran, are investigated experimentally and
numerically. Hence, in present study the effect of both
sun and wind simultaneously is considered when
evaluating the performance of the wind catcher system
using numerical simulation , long term and short term
experimental dataset.
1. Material and method
For numerical simulation and to study the whole image
of the flow field and the airflow pattern around and
inside the wind-induced natural ventilation wind towers
of Yazd city, a reference case study is selected at the old
city center of Yazd.The model of experimental house
with wind tower is built and used in the open source
CFD package (OpenFOAM) for simulation. Figure1
shows the model of the experimental house linked to the
wind tower. The isothermal simulation results are used
to validate against the long-term empirical measurement
results.
For the experimental study and evaluation of thermal
behavior of wind catcher , a short term experimental
data of the warmest day of measuring period was
selected .All required data of wind and temperature was
stored in sub-hourly timescale. A mean average hourly
data was used to look at the daily wind and solar pattern
during a summer day .
The measurements data set and the template CFD model
of this research are intended to support and used in
future studies on performance analysis of the other wind
catchers of Yazd and, this way, to contribute to
improved solar -wind catcher design in future.
2
ICESE2016, 5-6 September, 2016
2. Experimental measurements A representative four-sided wind catcher in the historic
city center of Yazd was selected to equip with
temperature, wind, air velocity and solar sensors. The
monitoring was performed during 2014-2015.
Based on the literature review, on-site measurements
should be made to provide validation data for the CFD
simulations[6, 12] .The on-site measurements, in spite
of their disadvantages, represent the complex reality
without simplifications and are therefore the true
validation data for numerical models. It is however
important that these measurement are made during a
sufficiently long measurement period. To establish a
long-term dataset on a full scale wind tower the outdoor
wind speed and the air velocities in four tower shafts of
the experimental house ( Mortaz house) wind catcher in
city campus of Yazd university, were monitored during
a full year monitoring campaign in 2014-2015.
In this study a yearly and daily pattern of climate dataset
were used for the CFD validation purposes and thermal
analysis of the wind catcher performance .
2.1. O-site data measurements A local weather station was installed to obtain the
outdoor wind and temperature data of the site.Wind
measuring equipment at the weather station was located
at a height of 10m above ground level. For the duration
of 1 year, the data was filtered in case of sensor
malfunction, missing data and other typical errors [AWS
1997]. For CFD validation purposes the wind data were
carefully analyzed in view of selecting the yearly
prevailing wind directions over the year which can be
more effective to analyze the performance of wind
catcher. Three prevailing wind directions of east, south-
east and north-west were identified which the mean air
velocities were computed for the different tower shafts.
Apart from the experimental wind data measurement, it
is important to gain information on the type of wind
activity which is present at the location and from which
direction the wind is predominant, so that the reference
wind velocity could be specified in the CFD model at
the inlet boundaries. All these information were
obtained from wind data recorded at the Yazd
meteorological station, which is the nearest weather
station.
2.2. Indoor data measurements Indoor air velocity measurements in four tower shafts
have been performed using four air velocity sensors (A4
– A1 – A2 – A3 ) with the working range of 0-20 m/s
were installed in the middle of four tower shafts (shaft
A – B – D – E) in height of 2.00 m above the outlet of
the tower.A long-term measurement has been conducted
during 2014-2015.
3. CFD simulation :process and results In CFD simulations, a large number of choices have to
be made by the user. These choices can have a very
large impact on the simulation results. In a typical CFD
simulation, the user has to choose the approximate
equations describing the flow (steady RANS, unsteady
RANS (URANS), LES or hybrid URANS/LES), the
level of detail in the geometrical representation of the
buildings, the size of the computational domain, the
type and resolution of the computational grid, the
boundary conditions, the discretization schemes, the
initialization data and the iterative convergence criteria.
Important best practice guidelines have been developed
and/or published by many researchers [4,5, 6, 10, 11,
12] . In this section, the CFD simulation parameters
are briefly outlined :
3.1. Computational geometry and grid
A computational model was made of the full scale
building model linked to the wind tower used for the on-
site measurements. The dimensions of the
computational domain were chosen based on the best
practice guidelines by Franke et al. [7]. The upstream
domain length is 5H = 70 m. The resulting dimensions
of the domain were W×D×H=500×500×80 m3. The
computational grid was created using the blockMesh
and snappyHexMesh utilities in OpenFOAM resulting
in a hybrid grid with 4.294.876 polyhedral and
hexahedral cells. Two refinement boxes control the cells
located in the immediate surroundings of the building
model and the tower model. 17 and 10 cells are used
along the width and depth of the each tower shafts,
respectively, as shown in Figure.1 The minimum and
maximum cell volumes in the domain are approximately
3.15 e-06 m3 and 3683.49 m3, respectively.
The distance from the center point of the wall adjacent
cell to the wall, for the wind-catcher and ground plane is
23.19 m and 221 m respectively. This corresponds to y+
values between 20 and 350. As standard wall functions
are used in this study, these values ensure that the centre
point of the wall-adjacent cell is placed in the
logarithmic layer. The domain shape and size (figure.1)
allows modelling different wind directions330° , 90°
and 150° corresponding to the prevailing wind
directions from experimental measurement.
Figure 1: (top) Computational domain. (left) Grid at building
surfaces. (right) Detail of grid near and inside wind tower
shafts. 3.2. Boundary conditions
The atmospheric boundary layer inflow at the inlet of
the domain consists of the profiles of mean wind speed,
turbulent kinetic energy and turbulence dissipation rate
[10]. The mean wind speed profile is prescribed by the
logarithmic law with z0 according to the average
3
ICESE2016, 5-6 September, 2016
building height (h=7 m, z0 = 0.7 m). Based on the
meteorological data , the reference wind speed is
different in different wind directions e.g. Uref = 5.5 m/s
at 10 m height for wind direction of 90° . The ABL
profile is derived from the friction velocity, flow
direction and the direction of the parabolic coordinate.
(figure2)
(1)
where U* is the frictional velocity, K is Karman's
constant, z the vertical coordinate [m], z_0 is the
surface roughness length [m] and zg as minimum value
in z direction [m].
In the simulations the inlet boundary conditions (mean
velocity U, turbulent kinetic energy k and turbulence
dissipation rate were based on the measured incident
vertical profiles of mean wind speed U and longitudinal
turbulence intensity Iu (figure. 2) The turbulent kinetic
energy k was calculated from U and Iu using Eq. (1),
where a is a parameter in the range between 0.5 and 1.5
[2,3]., the turbulence kinetic energy with the a= 1.5 is
chosen .The turbulence dissipation rate ɛ with the von
Karman constant (= 0.41) are given by Eq. (2).
(2)
At the outlet, zero static pressure is specified. At the
sides and the top of the domain, symmetry boundary
conditions are imposed (i.e. zero normal velocity and
gradients). At the “walls”, the standard wall functions
are used.
Figure 2: Log law profile of ratio of mean wind speed U to
mean wind speed UH at building height.
3.4. CFD results and comparison with on-site
measurements The pressure and velocity distribution around the wind
catcher model are shown in figures 3 and 4. Figure 5
shows the air flow pattern through the tower shaft for
three reference wind directions 330, 90 and 150. The
down ward and upward air flow pattern (Uz) are
indicated for the leeward and wind ward shafts at each
wind direction.
Figure 3: pressure distribution for wind direction of 90 °.
For validation purposes, the measured mean air velocity
values are converted to velocity ratios by division by the
reference wind speed (Um/Uref). The simulated velocity
ratios are obtained by division of the simulated mean air
velocity of each shaft by the simulated reference wind
speed (UCFD/Uref).
Figure 4: Velocity distribution ( Umag ) at experimental house
position ( H= 10 m) for wind direction 90.
The simulation results have been compared with on-site
wind speed measurements for the wind directions 330°,
90° and 150° (Figure 6) . The average deviations for
wind direction 90° in shafts B , D , E and A are 12.5 %
, 9.5 % , 29.5 % and 25.5 % , respectively. A possible
reason for this observation is that shaft B and D with
low wind speed ratios are generally characterized by
lower turbulence intensities and lower wind direction
fluctuations, and can therefore be better described by
the statistically steady RANS approach for the turbulent
flow. The overall average deviation between simulated
and measured wind speed was 19% for Standard k−ɛ
model for wind direction 90, which is considered a good
agreement. Note that for the wind direction of 90 ° the
wind ward shafts are shaft E and B providing downward
flow behavior and Shaft A , shaft D considering the
leeward shafts with upward flow pattern .
The overall average deviations in wind directions 330
and 150 are 48% and 45% respectively which at least
partly caused by the deficiencies of steady RANS
modelling and by neglecting the effect of buoyancy
driven force. It should be mentioned that in this paper,
an isothermal simulations are justified, in which
mechanical turbulence generation dominates and
thermal effects are absent.
4
ICESE2016, 5-6 September, 2016
4. CFD simulations: sensitivity analysis To analyze the sensitivity of the results to geometrical
and computational parameters [11], systematic changes
are made to the reference wind tower case in one of the
simulated wind direction = 90.
4.1. Impact of grid resolution Performing a grid-sensitivity analysis is important to
reduce the discretization errors and the computational
time. In this study, a grid-sensitivity analysis was
performed based on three different grid resolutions near
and inside the wind tower model. Coarsening and
refining was performed in refinement box surrounding
the wind tower shafts.
The results for mean velocity ratio (Umag/Uref) profiles
along shaft A from the three different grid resolution ,
indicating only a very limited dependence of the results
on the grid resolution. It can be indicated that the
differences between grid 2 and grid 3 are significant,
Figure 7: Comparison of mean velocity ratio (U mag/U ref)
profiles along shaft A from the three different grid resolutions
for wind direction 90°.
whereas the differences between the grid 2 and grid 1
(final grid) are small. This seems to indicate that the
grid 1 is a good compromise between computational
accuracy and computational cost. This uniform grid is
therefore selected for further analysis.
4.3. Impact of turbulence model
Isothermal 3D steady RANS simulations were made in
combination with two turbulence models: the standard
k−ɛ model (Sk−ɛ) and the realizable k−ɛ model (Rk−ɛ).
The average and total deviations in four equipped tower
shafts for each turbulence model , at the wind direction
of 90° are given in “Table 1”.
The differences between the standard k−ɛ model (Sk−ɛ)
and the realizable k−ɛ model (Rk−ɛ) are most
pronounced in shaft D, where the Rk−ɛ model tends to
Figure 6: Comparison of mean velocity ratios by
CFD simulation results and experimental
measurements data in four tower shafts , for three
prevailing wind directions 330 , 90 , 150
a b c
Figure 5: Air flow behavior inside four tower shafts ( H= 6.35 m) for three prevailing wind directions of 330 (a) , 90 (b) , 150 (c)
5
ICESE2016, 5-6 September, 2016
overestimate the velocity variations. The overall
average deviation between simulated and measured
wind speed was 19% for standard k−ɛ model and 30 %
for realizable k−ɛ model at wind direction of 90° .
Hence the standard k−ɛ model generally provides an
accurate results in four tower shafts at wind direction of
90°.
Table 1: The average absolute deviations for the equipped
tower shafts for each turbulence model , for the wind direction
of 90 .
Turbulence
model
Shaft
B
Shaft
D
Shaft
E
Shaft
A
Total
deviation
(percentage)
Standard
k−ɛ
0.12
0.09
0.29
0.25
19%
Realizable
k−ɛ
0.25
0.13
0.45
0.40
30%
5. Thermal behaviour of wind catcher
To monitor the thermal behaviour of the reference wind
catcher case during a hot summer day , the data from the
temperature sensors was averaged over the warmest day
of measuring period in shaft B (wind ward shaft ) at
wind direction 90°. Figure 8 shows the wind catcher
plan and section and the position of the temperature
sensors in Shaft B . Where the T out is the ambient air
temperature of 26th June 2015 .T1and T2 are coding as
the surface temperature of the external and internal
tower walls at the height of 10 m respectively . and T3as
the surface temperature of the internal wall at height of
3m.Ta1 is the air temperature (H=10m) and Ta2 the air
temperature at height 10m ( the same height as the air
velocity sensor height ) . The temperature data was
recorded during the summer day at the warmest hour of
day has shown in “Table 2”.
Figure 9 shows the temperature difference between
external and internal tower walls during a hot summer
day on 26th June 2015. The temperature difference
between the hot external surface and internal wall of the
wind tower at 15 pm can explained by the high thermal
inertia of the adobe walls.The amount of coolness which
can be stored in the tower mass , due to the specific heat
of the energy-storing material of the wind catcher walls
to meet the cooling needs of the building during a hot
summer day.
Figure 10 shows the temperature variation inside and
through the tower shaft B comparing to the ambient air
temperature. The results show the key role of tower
walls to reduce the uncomfortable temperature
fluctuations inside and through the wind catcher shafts.
This phenomena can be seen in passive solar
architecture by the use of massive walls which is called
thermal flywheel effect. Based on the literature review ,
to achieve the desired time lag in the temperature at the
inside surface of the massive wall requires a very thick
wall (typically 12-18"),and there is considerable
attenuation in the temperature wave as it passes through
the wall.
In present study the homogeneous adobe walls of the
reference wind catcher case with the thickness of 14"
and time lag of 24 hours , can significantly effect on
lowering temperature of outlet air (Ta2) from the wind
tower .
Figure 11 shows the wind and air velocity variations
during a hot summer day which can provide a full
analysis of the supply air from the tower during a hot
summer day.The wind speed , wind direction and air
velocity at 3 P.M. can be indicated from the daily data
recorded on 26th June 2015.(Table 3)
Figure 8: Wind catcher plan and the tower shafts coding
(right) , wind catcher section and the position of the
temperature sensors ( Tx) and air velocity sensor ( A1)
in Shaft B ( left)
Table 2: Temperature data from temperature sensors on
a hot summer day (26th June 2015) at 15 pm.
Date/Time
T out
(°C)
T wind
(°C)
T1
(°C)
T2, T3
(°C)
Ta1
(°C)
Ta2
(°C)
26th June
2015
45
43.9
55.6
40.1
39.8
39.2
Figure 9: Temperature difference between external (T1 )
and internal (T2) tower walls during a hot summer day.
Table 3: wind and air velocity data on a hot summer day
(26th June 2015) at 15 pm.
Date/Time
Wind
speed
(m/s)
Wind
direction
(°)
Air
velocity
(m/s)
26th June
2015
1.5
90
0.85
Tables 2and 3 show the supply air flow velocity of 0.85
m/s with the temperature of 39.2 °C in shaft B at 3
P.M.on 26th June 2016.
6
ICESE2016, 5-6 September, 2016
Figure 10: the temperature difference between the inlet
air , out let air and the ambient air during a hot summer
day .
Figure 11:The wind and air velocity variations during a
hot summer day(26th June 2015)
Conclusions
To analyze the wind and temperature behaviour of the
wind catchers of Yazd city , a reference typical wind
catcher case in city campus of Yazd university has been
selected . The CFD simulation has been conducted to
predict the air flow pattern around and through the wind
catcher for three prevailing wind direction of 330 , 90 ,
150 and an experimental analysis has been done to find
the thermal behavior of tower in shaft B (wind ward
shaft ) at wind direction 90°.Based on the CFD
simulation , this paper has presented a systematic
evaluation of an isothermal 3D CFD for prediction of
the air flow behavior of wind catchers of Yazd. The
evaluation is based on a grid-sensitivity analysis and on
validation with long term full-scaled empirical
measurements .The present study has shown that 3D
steady RANS CFD, in spite of its limitations, is a
suitable tool to predict the wind driven flow behavior in
wind towers. The impact of the turbulence model and
suitable grid refinement has been investigated and it has
been shown that a careful selection of these parameters
is very important for accurate and reliable results. The
basic OpenFoam CFD simulation template of wind-
catcher reference case from the present paper easily
allow future parametric studies to evaluate alternative
design configurations of the wind catchers of Yazd city ,
especially when the different configurations , are all a
priori embedded within the same computational domain
and grid .Considering that in this paper, isothermal CFD
simulations are justified, in which mechanical
turbulence generation dominates and thermal effects are
small or absent. The deviation between the CFD and
experimental results at windward and leeward tower
shafts can be explained due to the buoyancy and
chimney effect and to the possibility of neglecting the
effect of thermal modelling .Hence the impact of
buoyancy driven force in wind catcher CFD model
should be considered and evaluated for future study. On
the other hand the thermal behavior of the wind catcher
reference case has been analyzed experimentally during
a hot summer day .The results reveals that the tower
massive walls has a significant impact on the thermal
behaviour of the wind catcher due to their thermal
properties in lowering air temperature ( around 5.8°c )
leaves the tower during a hot summer day.At the final
stage the effect of both sun and wind simultaneously
was evaluated on 26th June 2016 .The study results show
that the solar-wind catchers of Yazd city is quiet effect
in lowering air temperature and provide a comfort air
velocity during a hot summer day without any energy
cost .
References
[1]AWS, 1997. Wind Resource Assessment Handbook:
Fundamentals for Conducting a Successful Monitoring
Program. Prepared for National Renewable Energy
Laboratory.
[2] BahadoriMN. 1985.” An improved design of wind
towers for natural ventilation and passive cooling”.
Solar Energy;35 (2):119–129
[3]Blocken, B., Carmeliet, J. 2002. “Spatial and
temporal distribution of driving rain on a low-rise
building”. Wind Struct.; 5 (5): 441-462.
[4]Blocken, B., Carmeliet, J. 2004b. “A review of wind
driven rain research in building science”. J. Wind Eng.
Ind. Aerodyn.; 92 (13): 1079-1130.
[5]Blocken, B., Carmeliet, J., Stathopoulos, T. 2007.
“CFD evaluation of wind speed conditions in passages
between parallel buildings-effect of wall-function
roughness modifications for the atmospheric boundary
layer flow”. J. Wind Eng..; 95 (9-11): 941-962.
[6]Blocken, B., Carmeliet, J., Moonen, P., Stathopoulos,
T. 2008. “A numerical study on the existence of the
Venturi-effect in passages between perpendicular
buildings”. Journal of Engineering Mechanics-ASCE;
134 (12): 1021-1028.
[7]Franke, J., Hellsten, A., Schlünzen, H., Carissimo, B.
2007. “Best practice guideline for the CFD simulation
of flows in the urban environment”. COST Action 732;
COST Office: Brussels (Belgium).
[8] Montazeri H. 2011.”Experimental and numerical
study on natural ventilation performance of various
multi-opening wind catchers”,Building and
Environment;370–378.
[9]OpenFOAM project web site, URL:
http://www.OpenFOAM.org, OpenCFD Ltd.
[10]Richards, P.J., Hoxey, R.P. 1993. “Appropriate
boundary conditions for computational wind
engineering models using the k-ε turbulence model”. J.
Wind Eng. Ind. Aerodyn.; 46-47: 145-153.
[11]Roache, P.J. 1998. Verification of codes and
calculations. AIAA J.; 36: 696-702.
[12] Schatzmann, M., Leitl, B., 2011. Issues with
validation of urban flow and dispersion CFD models. J.
Wind Eng. Ind. Aerodyn. 99 (4), 169e186.
[13] Yaghoubi MA , Sabzevari A , Golneshan AA.
1991. “Wind towers: measurement and performance”.
SolarEergy;47(2):97–106.
The 3rd International Conference and Exhibition on Solar Energy
ICESE-2016
5-6 September, 2016, University of Tehran, Tehran, Iran
___________________________________________________
7
