Numerical and experimental performance analysis of a four-sided wind tower adjoining parlor and courtyard at different wind incident angles

Accepted Manuscript

Numerical and experimental performance analysis of a four-sided wind tower adjoining parlor and courtyard at different wind incident angles Hossein Dehghani Mohamadabadi , Ali Akbar Dehghan , Abdul Hamid Ghanbaran , Alireza Movahedi , Abolfazl Dehghani Mohamadabadi PII: DOI: Reference:

S0378-7788(18)30049-5 10.1016/j.enbuild.2018.05.006 ENB 8547

To appear in:

Energy & Buildings

Received date: Revised date: Accepted date:

8 January 2018 27 April 2018 4 May 2018

Please cite this article as: Hossein Dehghani Mohamadabadi , Ali Akbar Dehghan , Abdul Hamid Ghanbaran , Alireza Movahedi , Abolfazl Dehghani Mohamadabadi , Numerical and experimental performance analysis of a four-sided wind tower adjoining parlor and courtyard at different wind incident angles, Energy & Buildings (2018), doi: 10.1016/j.enbuild.2018.05.006

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Highlights

AC

CE

PT

ED

M

AN US

CR IP T

 " Four-sided wind tower performance combined with parlor and courtyard is analyzed experimentally and numerically.  " Combination of wind tower, parlor and courtyard reflects typical vernacular houses in hot-arid regions of Iran.  " At 61.5% of wind angles, the wind tower dominantly performs as an air extracting device.  " At other wind angles, the wind tower acts as an air exchanging device (supply=extract).  " "Air extracting" function provides passive ventilation using conditioned air from microclimates such as courtyards.  " It is emphasized on investigating wind tower performance in combination with the whole building layout.

1

ACCEPTED MANUSCRIPT

Numerical and experimental performance analysis of a four-sided wind tower adjoining parlor and courtyard at different wind incident angles

CR IP T

Hossein Dehghani Mohamadabadi1, Ali Akbar Dehghan2*, Abdul Hamid Ghanbaran3, Alireza Movahedi4, Abolfazl Dehghani Mohamadabadi5

1, 3- School of Architecture and Urban Design, Shahid Rajaee University, Tehran, Iran 2, 4- School of Mechanical Engineering, Yazd University, Yazd, Iran

AN US

5- School of Architecture and Urban Design, Yazd University, Yazd, Iran

Abstract:

Ventilation performance of a scaled four-sided wind tower connected to parlor and courtyard at

M

different wind incident angles was studied both numerically and experimentally. Parlor (Talar in Persian) is a large room with one open side that connects the wind tower to the courtyard. The

ED

combination of wind tower, parlor, and courtyard forms a typical layout of summer section of vernacular houses in hot-arid regions of Iran. This study aims to uncover ventilation strategies

PT

considered in the design of wind tower and its interaction with surrounding architectural

CE

elements. The numerical study was conducted on a 1:25 scaled model for 13 wind incident angles with 15 degrees intervals. Interested parameters were the flow rate and the ratio of total

AC

extract to supply flow rates at each wind incident angle. A structured grid was generated, and ANSYS Fluent software was employed for simulating the flow field. Wind tunnel experiment was carried out on the same scaled model and a semi-empirical approach was adopted for predicting the air flow rate at each wind angle. Numerical simulation results were validated with the wind tunnel measurements and good agreement was observed. Results indicate that for

2

ACCEPTED MANUSCRIPT

61.5% of wind incident angles, the dominant function of four-sided wind tower is extracting air out of the building and at other wind angles, with approximately equal values of supply and extract flow rates, it acts as an air exchanger device. Accordingly, it can be concluded that neglecting stack effect, four-sided wind towers in hot-arid regions of Iran are primarily employed

CR IP T

for heat dissipation rather than inducing outdoor breezes to the interior spaces.

Keywords: Wind tower/catcher, Courtyard, Parlor, Natural ventilation, Incident angle, Wind

Nomenclature

CE

PT

ED

M

Velocity magnitude (m/s) Cartesian co-ordinates (m) Reynolds number Air density (kg/m3) Kinematic viscosity (m2/s) Volume flow rate (m3/s) Supply flow rate (m3/s) Extract flow rate (m3/s) Gravitational acceleration (m/s2) Turbulence intensity Cross-sectional area (m2) Pressure (Pa) Total pressure (Pa) Static pressure (Pa) Length (m) Width (m) Height (m)

AC

V X, Y, Z Re ρ k Q Qs Qe g I A P Pt Ps L W H

AN US

tunnel.

3

ACCEPTED MANUSCRIPT

1. Introduction Energy consumption is directly related to the global warming which is the most serious threat to the future of all communities [1]. The challenging fact is that buildings are responsible for about

CR IP T

40% of world energy consumption. The energy used for heating, ventilation and air conditioning (HVAC) systems accounts for about 50-60% of the building annual energy consumption [2]. A considerable number of studies have focused on the concept of applying natural ventilation strategies to decrease energy consumption as well as improving indoor air quality (IAQ) and

AN US

thermal comfort [3].

Wind tower (also known as wind catcher) is one of the natural ventilation systems dating back to 1500 years ago, that exploits natural force of the wind to operate [4,5]. In central arid regions of

M

Iran, wind towers are manipulated to provide fresh air for the living spaces in compact urban textures [6]. They also used to provide natural air circulation for public cisterns which are long-

ED

term underground cold water reservoirs used to deliver cool potable water during summer time

PT

for the local communities [7].

Wind towers are highly functional devices that utilize clean and fresh air as well as higher wind

CE

speed at roof level [4]. Investigations conducted by Elmualim [8] showed that the flow rate induced by a wind tower is considerably greater than that of a window with equivalent aperture

AC

area. In general, the performance of a wind tower relies on the natural driving force of a blowing wind as the primary force and the buoyancy effect [4]. The wind-driven ventilation depends upon pressure coefficient distribution on building and wind tower faces [9], while the buoyancydriven ventilation depends on the difference between inside and outside temperatures [10]. Results of a CFD investigation conducted by Hughes and Cheuk-Ming [11] indicated that 4

ACCEPTED MANUSCRIPT

ventilation performance of a wind tower based on the first phenomenon is 76% more effective than the latter. They also realized that once there is no induced airflow from the outside, the stack effect is negligible. Results of two other studies indicated that the buoyancy force slightly

CR IP T

affects the flow rate of a wind tower at lower wind speeds (less than 2 m/s) and it is negligible at higher wind speeds (higher than 3m/s) [12,13].

Wind towers in the vernacular architecture of Iran are designed based on the harsh condition of climate and direction of seasonal winds. Depending on whether the desirable winds blow from

AN US

one or several directions, traditional wind towers are classified into five types, including one, two, four, six and eight-sided [5]. One-sided and two-sided wind towers are suitable for regions with one specific prevailing wind direction [14]. The use of multi-sided wind towers makes it

M

possible to benefit from various wind directions. Operation of multi-sided wind towers is reliant upon the vital role of cross dividers constructed inside the wind tower which can also cause a

ED

phenomenon called ―short circuiting‖ in which the induced air through windward channels is immediately drawn out by leeward ones [15,16].

PT

According to Fig. 1, a typical layout of houses in Yazd utilizing a wind tower includes the

CE

integration of wind tower with parlor and courtyard [5]. In hot-arid regions of Iran, climatic strategies such as compact urban texture, courtyard building layout, division of a house into

AC

summer and winter sections, use of underground and shaded spaces as well as local mud materials, have been adopted to create a tolerable environment [6]. Employing a courtyard as a passive technique provides better cooling, ventilation, and daylight for the living spaces [4]. Parlor (or Talar in Persian) is a semi-closed room with a high vaulted ceiling located on the southern side of the courtyard that generally functions as a gathering place in summer [6]. Except 5

ACCEPTED MANUSCRIPT

for a short time of receiving sunshine in the morning, the parlor is shaded during the entire day owing to its orientation toward northeast [17]. Most of the wind towers used in the vernacular architecture of Yazd city are the four-sided type with a rectangular cross-section. In most of the

CR IP T

houses, wind tower is placed behind the parlor on the symmetry axis crossing the parlor and courtyard [5]. The case under the present investigation, called ―Hazire-ei house‖, was chosen due to its conformity with the typical pattern of summer section of houses in the vernacular architecture of Yazd city.

AN US

The study of wind tower performance consists of two fields, including hydrodynamic and thermal analysis. Since the function of wind tower is primarily reliant upon wind-driven forces, this study only inspects the hydrodynamic analyses.

M

Some studies have focused on the ventilation performance of wind tower as a single device in order to explain its function and evaluate its ventilation performance. Calautit and Hughes [18]

ED

studied the efficiency of a modern four-sided wind tower mounted atop a classroom by calculating ventilation parameters such as the wind tower flow rate, mean age of air, air change

PT

rate and air change effectiveness. They found the minimum wind velocity at which the wind

CE

tower met the required ventilation standards. Results also showed that the highest efficiency of four-sided wind tower is achieved at the wind angle of 45º. Montazeri et al. [19] carried out an

AC

experimental and numerical study on a two-sided wind tower and observed that short-circuiting phenomenon appears in the wind tower as wind incident angle increases and it reaches the maximum value at the wind angle of 60º. Results also revealed that the suction capability of wind tower increases as wind angle increases and this trend continues until the wind tower turns into a total suction device at wind angle of 90º with the highest recorded efficiency. 6

ACCEPTED MANUSCRIPT

Some studies have assessed the impact of various modifications to the structure of wind tower on its ventilation performance. Elmualim and Awbi [20] compared the flow rate of two wind towers with square and circular cross-sections and concluded that the square wind tower provided a

CR IP T

higher efficiency due to the formation of larger flow separation region caused by sharp edges. Dehghan et al. [14] experimentally showed that the geometry of one-sided wind tower roof significantly influences its airflow rate. Montazeri [21] investigated the ventilation performance of wind towers with different number of openings. Tests were carried out on five scaled

AN US

cylindrical wind towers and the flow rate of wind tower quadrants was estimated. Results showed that the sensitivity of wind tower to wind incident angle decreases as the number of openings increases, while the induced airflow rate decreases. Studies conducted by Nejat et al.

M

[22] showed that equipping wind tower with wing wall enhances its flow rate by 50%. They also suggested an optimum angle for the wing wall. Hosseini et al. [23] numerically examined various

ED

designs of wind tower with modifications to their height and width and concluded that the width of the wind tower considerably affects its performance.

PT

Limited studies have focused on the impact of architectural features, surrounding elements and

CE

building apertures on the ventilation performance of wind tower. Karakatsanis et al. [9] experimentally estimated the flow rate of a wind tower by calculating pressure coefficient on its

AC

openings. Tests were conducted on three models, including a single wind tower, a wind tower connected to a room, and adding a courtyard to the second case. Results indicated that pressure coefficient on surfaces of architectural features such as a courtyard, wind-shading elements, and building apertures significantly influences the wind tower performance. In their experiment, the house model was mounted in the test section with its sidewalls exposed to the wind, resulting in 7

ACCEPTED MANUSCRIPT

flow separation at edges of outer walls, which led to different flow pattern compared to an urban courtyard house surrounded by other buildings. Besides, estimating the pressure coefficient at few external points resulted in a provisional assessment of flow behavior inside the wind tower.

CR IP T

Calautit et al. [24] studied the optimum arrangement and spacing of multiple wind towers and concluded that unlike parallel arrangement, the staggered arrangement of wind towers improves their ventilation performance. Su et al. [12] in a numerical-experimental study concluded that upstream flow condition considerably affects the performance of cylindrical wind tower. They

AN US

also observed that the flow rate of wind tower is greatly influenced by room pressure, which is dependent on the room ceiling configuration. Ameer et al. [25] investigated the performance of a wind tower integrated with five different roof configurations and stated that roof design is a key

M

element in optimizing the wind tower function. Cruz-Salas et al. [26] experimentally examined different configurations of wind tower mounted on a room with a window on the windward side.

ED

They concluded that wind tower efficiency is mainly dependent on the orientation of its openings. Results also showed that the presence of wind tower improves the induced flow rate to

PT

the room up to 420 % for specific orientations. Montazeri and Montazeri [27] evaluated the

CE

impact of size and position of a window on flow characteristic inside a room ventilated by a wind tower. They realized that increasing the size of opening enhances airflow rate and air

AC

change efficiency. They also examined other configurations by changing the type of wind tower (one or two-sided) and its position on the roof. To the best of our knowledge, in most of the studies conducted on the ventilation performance of wind towers, the impact of adjoining building layout has rarely been taken into account. Many studies have examined the wind tower as a single device or at best connected to an isolated 8

ACCEPTED MANUSCRIPT

hypothetical room without being affected by outside pressure field. Furthermore, wind tower was generally regarded as a single device for capturing cool breezes. However, this concept at best suits the performance of one-sided wind towers when oriented toward prevailing winds

CR IP T

[4,28,29], while for two-sided and four-sided wind towers, the induced flow through windward and air suction through leeward quadrants happens simultaneously [16]. Investigations conducted on two-sided wind tower emphasized on its suction capability that leads to its maximum efficiency [19]. The mechanism of four-sided wind tower is more complicated than

AN US

the two-sided one, hence requiring more in-depth assessment. The novelty of the present study is considering the performance of four-sided wind tower combined with its surrounding architectural elements (particularly parlor and courtyard), which has not yet been addressed by

M

previous studies. It will be shown that the mentioned building layout, including parlor and courtyard, makes the wind tower perform dominantly as a suction device, which alters general

ED

perception about wind towers as devices for catching cool winds.

PT

1.1. Extract and supply flow rate of four-sided wind tower Figure 2 illustrates the full CFD assessment methodology flowchart considered for performance

CE

analysis of the four-sided wind tower integrated with parlor and courtyard at various wind

AC

angles. CFD modeling can be a useful method for evaluating the wind tower performance as well as assessing the accuracy of measurement procedure and uncertainty of experimental results [19]. In the present study, a combination of CFD simulation and semi-empirical modeling based on wind tunnel experiment is adapted to provide a validated and comprehensive prediction of wind tower performance.

9

ACCEPTED MANUSCRIPT

Wind tunnel experiments were carried out on a 1:25 scaled model at different wind incident angles. A semi-empirical approach was adapted to predict the performance of wind tower in which inputs were the measured static and total pressure at particular points inside the wind tower channels, and outputs were the estimated flow rates of each channel. Finally, experimental

CR IP T

and semi-empirical modeling outputs were used to validate CFD results at 5 out of 13 wind angles.

For CFD modeling, the geometry of the wind tower and the house model was created using a commercial CAD software, and then it was imported to ICEM CFD for generating the structured

AN US

grid. The CFD simulations were conducted using the commercial software ANSYS Fluent. The same boundary conditions as the wind tunnel experiments were applied to make a valid comparison between CFD and experimental results.

M

In order to evaluate the performance of wind tower, the flow rate of all wind tower channels at different wind angles was extracted from the software. Afterward, total summation of extract and

ED

supply flow rates were calculated at each wind incident angle. The ratio of total suction to supply

PT

flow rate (α) was defined as an index of wind tower ventilation performance.

CE

2. Methodology

2.1. Computational analysis

AC

In the present study, CFD modeling was conducted assuming steady-state incompressible fully turbulent flow. A Finite Volume Method (FVM) was used to discretize the Navier–Stokes equations and the flow fields were estimated using the 3D Reynolds Averaged Navier–Stokes (RANS) in combination with the SST k-ω turbulence model. For discretization of the convection

10

ACCEPTED MANUSCRIPT

terms, a second-order upwind scheme was adopted, and for Pressure-Velocity Coupling, the Semi-Implicit Method for Pressure-Linked Equation (SIMPLE) was applied. The SST k-ω turbulence model was chosen for its better prediction of flow at low Reynolds numbers inside and outside of the building where details of wall boundary layers characteristics

CR IP T

affect the wind tower flow rate [30]. Studies have proven that choosing a turbulence model depends highly on the target parameters. The accuracy of SST k-ω turbulence model was proved to be acceptable in the prediction of airflow in buildings [31–33]. The governing equations are as follows:

AN US

Mass conservation:    (  )  0

(1)



Equation (1) is valid for both compressible and incompressible flows, where ρ is density and 

M

refers to the velocity vector. Momentum conservation:      (   )  p    ( )  g  F

ED



(2)



PT

Here, p is the pressure, g and F are the gravitational and external body forces, respectively, and

 stands for the stress tensor described as: 





2



CE

   (   T )    I  3  

(3)

AC

Where µ is molecular dynamic viscosity, I is the unit tensor, and the second term on the righthand side represents the effect of volume dilation. Transport equations for the SST k-ω turbulence model are as follows:   kui     k k xi x j  x j

 ~   Gk  Yk  S k  

(4)

11

ACCEPTED MANUSCRIPT

 ui    xi x j

   ~     G  Y  D  S  x  j  

In these equations,

k

and

??

(5)

are the sources of turbulent kinetic energy due to average velocity

CR IP T

gradient and generation of ω, respectively. k and  represent the effective diffusivity of k and ω. Yk and Y are terms that represent dissipation of k and ω due to turbulence. D is the cross-

AN US

dissipation, and Sk and S  are source terms that are defined by the user.

2.2. Computational geometry

Figure 3a shows the computational domain created for the simulation of the external wind, including a cuboid domain of the length, width, and height of 300, 300 and 100 cm, respectively.

M

Figure 3b illustrates the geometry of the wind tower integrated with parlor and courtyard. As

ED

shown, the wind tower is located behind the parlor. They are connected through a small room called the wind tower room located beneath the wind tower. The parlor adjoins the courtyard

PT

with a completely open side and connects the wind tower to the courtyard. Table 1 presents different parts of the model and their dimensions. In order to simulate the house in its surrounded

CE

environment of urban texture in which the flow passes over the buildings and blows above the

AC

courtyards, the house geometry was positioned under the bottom level of the computational domain making the courtyard seem like a cuboid hole in it, hence, only the wind tower was directly exposed to the wind.

2.3. Computational grid and its verification

12

ACCEPTED MANUSCRIPT

In CFD simulations, the accuracy of the results, as well as acceptable convergence behaviors, are highly dependent on the grid quality [29]. As the geometry contained many perpendicular corners, to better capture those corners and hence providing a high-quality grid with a fewer

CR IP T

number of cells, a structured grid was generated (Fig. 3c). In order to provide a validated computational model, grid sensitivity analysis was conducted by solving the governing equations for various gradually refined mesh sizes of 3.3, 5.4, 7.3, and 10 million cells. The grid was concentrated around critical areas, where the flow parameters were

AN US

subjected to high gradients. The grid refinement process continued until a negligible difference was observed between the values of vertical velocity profile extracted from a line drawn along the center of supply channel (C2) at 0º angle. The difference between the results of two

M

consecutive grids of 7.3 and 10 million cells was observed to be 0.1%, and hence the grid of 7.3

ED

million cells was chosen for the rest of the numerical simulations.

2.4. Boundary conditions

PT

Boundary conditions for CFD simulation were chosen to be the same as the wind tunnel

CE

experiment. Table 2 shows the boundary conditions assigned to different parts of the geometry. Since the simulation of different wind angles was conducted using a single structured grid, to

AC

change the free stream direction, inlet and outlet boundary conditions for side faces of the computational domain were modified. To simulate a velocity flow field for perpendicular wind incident angles, e.g. 0º, the windward face of the domain was set as velocity inlet, and the leeward side faces were set as pressure outlet. For non-perpendicular wind angles, e.g. 45º, the two windward faces were set as velocity inlet, and the other two faces were set as pressure outlet. 13

ACCEPTED MANUSCRIPT

To change velocity direction at the inlet boundary conditions, velocity components along X and Y coordinate axes were modified accordingly. 2.5. Experimental set-up

CR IP T

The experimental investigation was carried out in an open circuit subsonic wind tunnel located in the aerospace engineering department of Amirkabir University of Technology (located in Tehran, Iran). The wind tunnel has a test section with length, width, and height of 180 ×100 ×100 cm, respectively. The turbulence intensity of the free stream is 0.28%, and maximum wind

AN US

speed in the test section can reach to 60 m/s using a suction fan (Diagram shown in Fig. 4).

For wind tunnel testing, the wind tower model was scaled down by the factor of 25 so that it could be assembled and tested within the permissible range of blockage ratio. Choosing this

M

scale also enabled us to respect the considerations of data recording as well as blockage ratio of pitot and static tubes inside wind tower channels [19,21,24]. Since the study aimed to predict the

ED

wind tower performance at different wind angles, the house model was mounted in a way to be rotated easily inside the test section in the range of 0º to 180º. In order to make the wind blow

PT

over the courtyard, a board parallel to the bottom surface of the test section was mounted at the

CE

same height of the courtyard that inevitably reduced the height of the test section (Fig. 4). Based on the formula of blockage ratio calculation [34], the maximum and minimum blockage percent

AC

of the wind tower in the reduced section area was calculated at wind incident angles of 45º (8.2%) and 90º (3.1%), respectively. Moreover, to keep the Reynolds number of the model identical to the real size building, for a 1:25 scaled model, a velocity of 25 times greater than the urban wind speed should be generated. Therefore, for wind tunnel experiments, the highest possible wind speed (25.5 m/s) was set for 14

ACCEPTED MANUSCRIPT

all of the tests. Such a speed not only made the experimental results more recognizable to read and record but also made the flow behave as if the speed of 1 m/s (3.6 km/h) was set for the fullscale model. However, studies have proven that due to the forced flow separation at the sharp

CR IP T

corners of cuboid geometries, the hydrodynamic behavior and the flow pattern are independent of the Reynolds number [28].

2.6. Measurements procedures

AN US

The semi-empirical approach adopted for the prediction of wind tower performance includes calculation of velocity through pressure measurement and accordingly estimation of flow rate inside the wind tower’s channels. For this purpose, the horizontal cross-sectional area of each

M

channel was divided into three portions. Then, the velocity for each portion was estimated. Finally, the flow rate of the channel (Q) was estimated using Equation (6). n

Q   AiVi

(6)

ED

i 1

respectively.

PT

In this equation, Ai and Vi represent the area and velocity of the ith portion of each channel,

CE

In order to estimate the flow velocity, several tiny and sensitive pitot and static tubes were mounted in the upper and lower parts of each channel to measure the total and static pressures.

AC

Pitots were made of the stainless steel tubes with outer and inner diameters of 1.2 mm and 1 mm, respectively. Figure 4 shows the location of pitot and static tubes in each channel. The upper and lower pitot tubes were rotated toward the negative and positive direction of Z-axis, respectively. Such an arrangement helped to capture any downward airflow by the lower sensors and any upward flow by the upper ones after the flow became rather uniform in the channel. Finally, to calculate the flow velocity, equation (7) was used. 15

ACCEPTED MANUSCRIPT

pt  p s 

1 V 2 2

(7)

Where V is the flow velocity, ρ is the density, and pt and ps are the measured total and static pressures, respectively.

CR IP T

The uncertainty analysis of the experiments was conducted adopting the approach provided in ref. [35]. The uncertainty regarding total and static pressure measurements was found to be around 7%. Related parameters affecting the uncertainty were the bias and precision errors in the calibration of the transducers, resolution limits of the A/D converter, bias error in the A/D

AN US

converter circuit, pitot tube misalignments and total error of the transducer. The uncertainties of velocity and hence the flow rate estimations in the wind tower channels were determined to be approximately 8%.

M

2.7. CFD Validation

Figure 5 compares estimated flow rate of wind tower through semi-empirical modeling with

ED

CFD results at wind angles ranging from 0º to 180º with 45º intervals. All positive values show

PT

the flow rate of extract channels (upward flows) while negative values indicate the flow rate of supply ones (downward flows). According to the column chart, the maximum difference between

CE

semi-empirical modeling and CFD results is 11.8%, indicating a good agreement between the numerical prediction and experimental results. The main reason for the difference observed is the

AC

error caused by pitot and static tubes in the case of flow rotation and inclination in the wind tower channels. In such a condition, the pitot suffers an error due to its axial sensitivity to the flow direction [21].

3. Results and discussions 16

ACCEPTED MANUSCRIPT

To have a better understanding of the wind tower performance, a 3D view of 13 wind incident angles and the house model is depicted in Fig. 6a. In addition, wind tower components including edges, faces, channels, and openings are named so that they could be easily referred to while discussing the results.

CR IP T

Figure 6b presents the CFD results of the supply and extract flow rate at different wind angles ranging from 0º to 180º with 15º intervals. The horizontal and vertical axes represent the wind angle and the flow rate of channels, respectively. For a particular wind angle, six columns are presented each of which shows the flow rate of a particular channel. Supply and extract channels

AN US

can be distinguished considering their positive or negative values. Positive values indicate the flow rate of suction channels and negative values show the flow rate of supply ones.

M

3.1. Ventilation performance at different wind angles

Analysis of the column chart data, presented in Fig. 6b results in the classification of the wind

ED

tower performance into five categories based on the wind incident angle. Figure 7 shows the simulation results of the velocity distribution inside and outside the wind

PT

tower at different air incident angles. The last row shows the location of planes from which the

CE

corresponding contour plots in the column are extracted. The left column illustrates wind incident angles. The middle column shows velocity contour plots overlaid with flow streamlines

AC

around the wind tower on a plane located 5 cm above the ground level (bottom surface of the domain). The streamlines show the flow path around the wind tower, separation points, shear layers and formation of recirculation zones on the wake side of the wind tower. The right column shows Z velocity contour plot of the horizontal cross-sectional plane inside the wind tower located 1.5 cm below the lower edge of wind tower’s openings. As was mentioned before,

17

ACCEPTED MANUSCRIPT

whether a channel acts as air supplier or air extractor, can be distinguished by its negative or positive velocity value according to the legend. In the following sub-sections, five categories of the wind tower performance, based on the wind incident angles, are discussed.

CR IP T

3.1.1. Ventilation performance at 0º, 15º and 30º wind angles According to Fig. 7, at 0º wind angle, external flow separation is observed at E2 and E3 edges of the windward face (F3). This phenomenon leads to the formation of shear layer and a lowpressure wake region that surrounds the entire leeward side of the wind tower resulting in air

AN US

suction in four channels at sides and one channel (C5) at leeward face. At these wind angles, wind tower supplies the air only through channel C2.

As pressure coefficient is the main factor influencing the flow direction from room to wind tower

M

or vice versa [28], and since the parlor’s mouth is oriented leeward to the wind (Fig. 6a (0º)), the flow direction is supposed to be from the tower to the parlor. However, results reveal that the

ED

flow moves mostly toward the opposite direction due to the function of five extract channels that eventually keep the total exhaust flow rate more than the supplied value. Therefore, as the

PT

direction of flow from the parlor to the wind tower shows (Fig. 8), the wind tower overall

CE

function is extracting the air out of the building. Similar to what was observed in previous studies [9,16], as the wind incident angle increases

AC

from 0º to 30º, due to a reduction in average pressure coefficient on windward faces and also strengthening of flow rotation inside the wind tower, a reduction in the flow rate is observed for the supply channel (Fig. 6b (0º-30º)). According to Fig. 8, at 0º wind angle, due to the presence of positive pressure on the mouth of the windward opening (I2), flow separation is observed near the lower edge of the opening that

18

ACCEPTED MANUSCRIPT

causes the formation of an inactive zone and consequently increasing the flow velocity downward the channel due to a reduction in air passage area. This phenomenon which is observed near the lower edge of windward openings at all wind incident angles (Fig. 8 and 10),

3.1.2. Ventilation performance at 45º and 60º wind angles

CR IP T

was previously reported in the literature [14,19].

According to Fig. 7 (Also Fig. 6b), as wind incident angle changes from 30º to 45º, the function

AN US

of C1 and C6 channels suddenly changes from suction to supply. This change is due to a shift at the separation edge (From E2 to E1), causing two faces of F2 and F3 receive positive pressure, and two faces of F1 and F4 receive negative pressure. Due to the angle between free stream

M

direction and wind tower’s faces, the flow enters the supply channels with rotating movements, which causes lower flow speed compared to perpendicular wind angles. However, in this state,

ED

the wind tower provides the maximum value of supply flow rate due to exposing the largest area (blockage surface area) for capturing the wind. Thus, the pressure difference between wind tower

PT

openings and the room increases, causing air supply through three channels (C1, C2, and C6) and

CE

air suction by the rest of them (C3, C4, and C5). Only at the wind angle of 45º, the supply flow rate of wind tower slightly exceeds the extract flow rate (by 6%). The capability of the four-sided

AC

wind tower in providing the highest value of supply flow rate at the wind angle of 45º was also reported by Calautit and Hughes [16]. At wind angles of 45º and 60º, parlor’s mouth is still in the negative pressure area, but the flow rate of exhaust and supply channels are almost equal. Thus, the wind tower acts as an air exchanger device.

19

ACCEPTED MANUSCRIPT

3.1.3. Ventilation performance at 75º, 90º, and 105º wind angles At wind angle of 90º, the narrower face of the wind tower (F2) is windward (Fig. 7 (90º)). The

CR IP T

flow separation occurs at E1 and E2 edges, resulting in the formation of negative pressure zone near three faces F1, F3, and F4. As Fig. 6b shows, the air is induced through two channels (C1 and C6) and is extracted from the rest of them (C2, C3, C4, and C5). Comparison of two perpendicular wind incident angles of 0º and 90º indicates that the four-sided wind tower

AN US

performance considerably differs at these wind angles. Although having a less effective exposed area to the wind, the flow rate of wind tower at 90º angle is 1.6 time more than that of the zero angle due to the presence of two openings on the windward face (one more opening compared to

M

the zero angle). At these wind angles, with equal values of total supply and extract flow rates, the wind tower acts as an air exchanger device, again.

ED

Figure 6b suggests that the wind incident angles of 75° and 105° are transition angles for C2 and C5 channels, respectively. The transition angle is the angle at which the direction of flow starts

CE

PT

to change inside the channel [14]. Thus, the flow rate of the mention channels is almost zero.

3.1.4. Ventilation performance at 120º and 135º wind angles

AC

According to Fig. 7, at wind angle of 135°, two faces (F1 and F2) are windward, flow separation occurs at E2 and E4 edges, and consequently, inactive areas appear behind the leeward faces (F3 and F4). Therefore, three channels (C1, C5, and C6) supply the air and the rest (C3, C4, and C2) extract the air from the building (Fig. 6b (135º)). Similar to the wind angle of 45º, flow rotation is observed in supply channels. Comparison between 45º and 135º wind angles offers that flow 20

ACCEPTED MANUSCRIPT

pattern and pressure distribution around wind tower itself looks similar for both wind angles, however, due to the presence of positive pressure on parlor’s mouth at 135° wind angle, the extract flow rate of the wind tower exceeds the supply flow rate.

CR IP T

Thus, at the wind angles of 120º and 135º, the dominant function of the wind tower is extracting the air out of the building.

3.1.5. Ventilation performance at 150º, 165º, and 180º wind angles

AN US

As wind angle changes from 135º to 150º, the function of C1 and C6 channels changes due to a shift at separation edge (from E2 to E1) that makes these channels extract the air from the parlor. At the angle of 180º, one face (F1) is windward. Flow separation occurs at E1 and E4 edges,

M

causing the formation of inactive zones in the vicinity of side faces F2 and F4 as well as the leeward face, F3. Therefore, the wind tower operates with one supply and five extract channels

ED

(Fig. 7 (180º)). The performance of wind tower at 150º, 165º, and 180º wind angles may be compared with corresponding symmetry wind angles of 30º, 15º and 0º, respectively (Fig. 6a).

PT

Obviously, if the wind tower was considered as a single device, its hydrodynamic behavior

CE

would expected to be the same for the corresponding symmetry angles. However, when it comes to observing the impact of the whole building layout, significant changes are seen in the wind

AC

tower performance.

Figure 9 gives a comparison of vertical velocity profile inside the wind tower channels at two selected wind angles of 0º and 180º. The figure demonstrates the Z velocity values extracted from a cross-line drawn at the same location of the pitot and static tubes (red dashed-lines). For extract and supply channels the line is located at the upper and lower side of the channels, 21

ACCEPTED MANUSCRIPT

respectively. According to the profiles, a significant augmentation in extracted air velocity is seen while slight reduction is observed in the velocity of supply channels. The main reason is the positive

CR IP T

pressure imposed on the mouth of the parlor at 180º wind angle. Another consequence is the short-circuiting phenomenon that occurs as the flow induced by C5 channel, is immediately extracted by five other channels. According to Fig. 10, wind tower induces no airflow to the parlor and the dominant direction of flow is from the parlor to the wind

AN US

tower. It can be inferred that the parlor acts as the air supplier for the living space at these wind angles.

M

3.2. Overall ventilation performance of the wind tower

Figure 11 compares the total supply and extract flow rates at each wind incident angle. The

ED

horizontal and vertical axes show the angle of attack and the total flow rate, respectively. The value for each column is obtained by summing the flow rate of all supply and all extract

PT

channels, separately. Equation (8) was used to calculate the ratio of total extract to supply flow

Q ( e) Q( s )

(8)

AC



CE

rate (α) as an index that denotes how the wind tower performs at each wind angle.

Where Q(e) and Q(s) are the total flow rate of extract and supply, respectively. Table 3 shows the overall performance of the wind tower integrated with parlor and courtyard at different wind incident angles. It is assumed that for values of α in the range of 0.8<α<1.2, the wind tower acts totally as an air exchanger unit, and for values smaller or greater than this range, wind tower performs as a supplier or extractor device, respectively. According to the data shown 22

ACCEPTED MANUSCRIPT

in table 3, the values of α exceed 1.2, for 8 out of 13 wind incident angles and for the rest, it is almost equal to 1. Thus, it can be concluded that at 61.5% of wind incident angles, the dominant function of four-sided wind tower is extracting air out of the building and at other incident

CR IP T

angles, with the approximately equal amount of supply and extract flow rates, it acts as an air exchanger device.

4. Conclusions

Detailed CFD analysis, as well as scaled wind tunnel testing, were conducted to assess the

AN US

ventilation performance of four-sided wind tower connected to a parlor and courtyard at different wind angles. The study aims to reflect ventilation strategies adopted for the most common layout of vernacular houses that benefit from wind tower in the central hot-arid regions of Iran.

M

Numerical simulation was conducted on a 1:25 scaled model for 13 wind incident angles with 15º intervals. Wind tunnel experiment was also conducted on the same scaled model and

ED

comprehensive pressure measurements were carried out for predicting the flow rate of the wind tower through a semi-empirical approach. Generally, a good agreement (0-11.8%) was observed

PT

between the semi-empirical modeling and CFD simulation results.

CE

Results show that at 61.5% of wind angles, the dominant function of four-sided wind tower combined with the parlor and courtyard is extracting air out of the building, which alters the

AC

general perception about wind towers as devices for catching cool winds. At other wind angles, with approximately equal values of extract and supply flow rate, it acts as an air exchanger device. The two functions of ―Air extracting‖ and ―Air exchanging‖ implies that neglecting the stack effects, four-sided wind towers in hot-arid regions of Iran are primarily responsible for heat dissipation from the structure rather than inducing outdoor breezes to the interior spaces. Air 23

ACCEPTED MANUSCRIPT

extracting function of the wind tower includes drawing out warm stale air and replacing it with cool fresh air from microclimate zones such as the courtyard and underground spaces. Such a smart strategy helps to provide natural cooling while avoiding harmful effects of

CR IP T

warm and polluted winds. Moreover, it is found that the ventilation performance of the wind tower integrated with parlor and courtyard is significantly influenced by wind incident angle. Generally, for wind angles ranging from 0º to 30º, the wind tower acts as an air extractor while within the range of 45º to

AN US

105º, with an equal amount of extract and supply flow rates, it acts as an air exchanger unit. For wind angle changing from 120º to 180º, with the positive pressure imposed on the mouth of the parlor, the wind tower acts as an air extractor.

M

This study emphasizes on considering wind tower integrated with the whole building layout, highlighting the fact that pressure distribution on the surrounding environment significantly

ED

influences the performance of wind tower. Parlor and wind tower are two complementary elements for maximizing the flow movement in the vicinity of the courtyard. The parlor is a

PT

large semi-closed architectural element that can provide fresh air for the living space by itself.

CE

For wind angle within the range of 120º-180º, by increasing the wind angle, airflow induced through the mouth of the parlor increases. For the range of 150º-180º, the wind tower induces no

AC

airflow to the building due to the positive pressure imposed on the mouth of the parlor. Since the internal pressure of the parlor is easily affected by the outside pressure field, it can provide significant air suction, air supply, and air circulation, depending on its orientation.

Acknowledgments 24

ACCEPTED MANUSCRIPT

The authors would like to thank the Aerospace Engineering Department of the Amirkabir University of Technology for providing experimental facilities. Also, efforts made by Marzieh Zeynali Fahadan for acquiring detailed information of the historic house case are greatly

CR IP T

appreciated.

References

F. Jomehzadeh, P. Nejat, J.K. Calautit, M.B.M. Yusof, S.A. Zaki, B.R. Hughes, M.N.A.W.M. Yazid, A review on windcatcher for passive cooling and natural ventilation in buildings, Part 1: Indoor air quality and thermal comfort assessment, Renew. Sustain. Energy Rev. 70 (2017) 736–756. doi:http://doi.org/10.1016/j.rser.2016.11.254.

[2]

L. Pe´rez-Lombard, J. Ortiz, C. Pout, A review on buildings energy consumption information, Energy and Buildings, Energy Build. 40 (2008) 394–398.

[3]

O. Saadatian, L.C. Haw, K. Sopian, M.Y. Sulaiman, Review of windcatcher technologies, Renew. Sustain. Energy Rev. 16 (2012) 1477–1495. doi:http://dx.doi.org/10.1016/j.rser.2011.11.037.

[4]

N. Khan, Y. Su, S.B. Riffat, A review on wind driven ventilation techniques, Energy Build. 40 (2008) 1586–1604. doi:10.1016/j.enbuild.2008.02.015.

[5]

M.N. Bahadori, A. Dehghani-sanij, A. Sayigh, A. Sayigh, Wind Towers: Architecture, Climate and Sustainability, Springer International Publishing, Switzerland; London, 2014. doi:10.1007/978-3-319-05876-4.

[6]

M. Khalili, S. Amindeldar, Traditional solutions in low energy buildings of hot-arid regions of Iran, Sustain. Cities Soc. 13 (2014) 171–181. doi:http://dx.doi.org/10.1016/j.scs.2014.05.008.

[7]

A.A. Dehghan, A.R. Dehghani, Experimental and theoretical investigation of thermal performance of underground cold-water reservoirs, Int. J. Therm. Sci. 50 (2011) 816–824. doi:10.1016/J.IJTHERMALSCI.2010.12.007.

[8]

A.A. Elmualim, Verification of design calculations of a wind catcher/tower natural ventilation system with performance testing in a real building, Int. J. Vent. 4 (2006) 393– 404. doi:10.1080/14733315.2005.11683717.

AC

CE

PT

ED

M

AN US

[1]

[9]

C. Karakatsanis, M.N. Bahadori, B.J. Vickery, Evaluation of pressure coefficients and estimation of air flow rates in buildings employing wind towers, Sol. Energy. 37 (1986) 363–374. doi:10.1016/0038-092X(86)90132-5.

[10] B.R. Hughes, H.N. Chaudhry, S.A. Ghani, A review of sustainable cooling technologies in buildings, Renew. Sustain. Energy Rev. 15 (2011) 3112–3120. doi:http://dx.doi.org/10.1016/j.rser.2011.03.032. 25

ACCEPTED MANUSCRIPT

[11] B.R. Hughes, C.M. Mak, A study of wind and buoyancy driven flows through commercial wind towers, Energy Build. 43 (2011) 1784–1791. doi:10.1016/j.enbuild.2011.03.022. [12] Y. Su, S.B. Riffat, Y.-L.L. Lin, N. Khan, Experimental and CFD study of ventilation flow rate of a MonodraughtTM windcatcher, Energy Build. 40 (2008) 1110–1116. doi:https://doi.org/10.1016/j.enbuild.2007.10.001.

CR IP T

[13] B.M. Jones, R. Kirby, Quantifying the performance of a top-down natural ventilation WindcatcherTM, Build. Environ. 44 (2009) 1925–1934. doi:10.1016/j.buildenv.2009.01.004. [14] A.A. Dehghan, M.K. Esfeh, M.D. Manshadi, Natural ventilation characteristics of onesided wind catchers: experimental and analytical evaluation, Energy Build. 61 (2013) 366–377. doi:http://dx.doi.org/10.1016/j.enbuild.2013.02.048.

AN US

[15] M.N. Bahadori, An improved design of wind towers for natural ventilation and passive cooling, Sol. Energy. 35 (1985) 119–129. doi:http://dx.doi.org/10.1016/0038092X(85)90002-7. [16] J.K. Calautit, B.R. Hughes, Measurement and prediction of the indoor airflow in a room ventilated with a commercial wind tower, Energy Build. 84 (2014) 367–377. doi:10.1016/j.enbuild.2014.08.015.

M

[17] J.G. Von Hardenberg, Considerations of houses adapted to local climate — a case study of Iranian houses in Yazd and Esfahan, Energy Build. 4 (1982) 155–160. doi:http://dx.doi.org/10.1016/0378-7788(82)90041-X.

ED

[18] J.K. Calautit, B.R. Hughes, Wind tunnel and CFD study of the natural ventilation performance of a commercial multi-directional wind tower, Build. Environ. 80 (2014) 71– 83. doi:http://dx.doi.org/10.1016/j.buildenv.2014.05.022.

PT

[19] H. Montazeri, F. Montazeri, R. Azizian, S. Mostafavi, Two-sided wind catcher performance evaluation using experimental, numerical and analytical modeling, Renew. Energy. 35 (2010) 1424–1435. doi:http://dx.doi.org/10.1016/j.renene.2009.12.003.

CE

[20] Abbas Ali Elmualim, H.B. Awbi, Wind Tunnel and CFD Investigation of the Performance of ―Windcatcher‖ Ventilation Systems, Int. J. Vent. 1 (2002) 53–64. doi:10.5555/ijov.2002.1.1.53.

AC

[21] H. Montazeri, Experimental and numerical study on natural ventilation performance of various multi-opening wind catchers, Build. Environ. 46 (2011) 370–378. doi:10.1016/j.buildenv.2010.07.031. [22] P. Nejat, J.K. Calautit, M.Z.A. Majid, B.R. Hughes, I. Zeynali, F. Jomehzadeh, Evaluation of a two-sided windcatcher integrated with wing wall (as a new design) and comparison with a conventional windcatcher, Energy Build. 126 (2016) 287–300. doi:10.1016/j.enbuild.2016.05.025. [23] S.H. Hosseini, E. Shokry, A.J. Ahmadian Hosseini, G. Ahmadi, J.K. Calautit, Evaluation of airflow and thermal comfort in buildings ventilated with wind catchers: Simulation of 26

ACCEPTED MANUSCRIPT

conditions in Yazd City, Iran, Energy doi:http://dx.doi.org/10.1016/j.esd.2016.09.005.

Sustain.

Dev.

35

(2016)

7–24.

[24] J.K. Calautit, D. O’Connor, B.R. Hughes, Determining the optimum spacing and arrangement for commercial wind towers for ventilation performance, Build. Environ. 82 (2014) 274–287. doi:10.1016/j.buildenv.2014.08.024.

CR IP T

[25] S.A. Ameer, H.N. Chaudhry, A. Agha, Influence of roof topology on the air distribution and ventilation effectiveness of wind towers, Energy Build. 130 (2016) 733–746. doi:https://doi.org/10.1016/j.enbuild.2016.09.005. [26] M. V. Cruz-Salas, J.A. Castillo, G. Huelsz, Experimental study on natural ventilation of a room with a windward window and different windexchangers, Energy Build. 84 (2014) 458–465. doi:10.1016/j.enbuild.2014.08.033.

AN US

[27] H. Montazeri, F. Montazeri, CFD simulation of cross-ventilation in buildings using rooftop wind-catchers: Impact of outlet openings, Renew. Energy. 118 (2018) 502–520. doi:10.1016/j.renene.2017.11.032. [28] H. Montazeri, R. Azizian, Experimental study on natural ventilation performance of onesided wind catcher, Build. Environ. 43 (2008) 2193–2202. doi:http://dx.doi.org/10.1016/j.buildenv.2008.01.005.

M

[29] B.R. Hughes, J.K. Calautit, S.A. Ghani, The development of commercial wind towers for natural ventilation: A review, Appl. Energy. 92 (2012) 606–627. doi:http://dx.doi.org/10.1016/j.apenergy.2011.11.066.

ED

[30] I.N.C. Fluent, FLUENT 6.3 user’s guide, Fluent Doc. (2006).

PT

[31] J.I. Perén, T. van Hooff, B.C.C. Leite, B. Blocken, CFD analysis of cross-ventilation of a generic isolated building with asymmetric opening positions: Impact of roof angle and opening location, Build. Environ. 85 (2015) 263–276. doi:10.1016/j.buildenv.2014.12.007.

CE

[32] R. Ramponi, B. Blocken, CFD simulation of cross-ventilation for a generic isolated building: Impact of computational parameters, Build. Environ. 53 (2012) 34–48. doi:10.1016/j.buildenv.2012.01.004.

AC

[33] T. van Hooff, B. Blocken, Y. Tominaga, On the accuracy of CFD simulations of crossventilation flows for a generic isolated building: Comparison of RANS, LES and experiments, Build. Environ. 114 (2017) 148–165. doi:10.1016/j.buildenv.2016.12.019. [34] Task Committee on Manual of Practice for Wind Tunnel Testing of Buildings and Structures of the Committee on Aerodynamics of the Aerospace Division of the American Society of Civil Engineers, Anon, Wind Tunnel Model Studies of Buildings and Structures., in: Manuals Reports Eng. Pract. Am. Soc. Civ. Eng., American Society of Civil Engineers, 1987: pp. 5–8. [35] S. Gowing, Pressure and shear stress measurement uncertainty for DARPA SUBOFF experiment, DAVID TAYLOR RESEARCH CENTER BETHESDA MD SHIP 27

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN US

CR IP T

HYDROMECHANICS DEPT, 1990.

28

ACCEPTED MANUSCRIPT

Table 1. Dimensions of various parts of the geometry. Dimensions 51×5×12 (cm) 35.2×4×2.2(cm) 2.2×15.8(cm) 4×15.8(cm) 8.6×6.5×12.1(cm) 19.4×15.8×21.7(cm) 27×39×52(cm)

Table 2. Summary of the CFD simulation boundary conditions.

Velocity inlet (m/s) Pressure outlet Incident angle (º)

AC

CE

PT

Walls

ED

Interior Gravity (m/s2) Time

M

All walls of the model (no-slip)

Condition V=25.5 (constant) Turbulence intensity= 0.28 % Atmospheric 0-180º Roughness Constant= 0.5 Roughness height= 1e-5 Interior -9.81 Steady state All walls (no-slip) Roughness Constant= 0.5 Roughness height= 1e-5

AN US

Boundary

CR IP T

Parts Wind tower Channels Single opening Double openings Wind tower room Parlor Courtyard

29

ACCEPTED MANUSCRIPT

Table 3. Overall ventilation performance of the wind tower connected to parlor and courtyard at different wind incident angles. Number of supply channels

Number of extract channels

1 1 1 3 3 2 (one channel =0) 2 2 3 3 5 5 5

5 5 5 3 3 3 4 3 (one channel =0) 3 3 1 1 1

CR IP T

Wind tower dominant performance suction suction suction equal equal equal equal equal suction suction suction suction suction

AN US

The ratio of exhaust to supply (α) 1.81 1.66 2.42 0.94 1.00 1.09 1.07 1.11 1.20 2.00 4.06 3.92 3.90

AC

CE

PT

ED

M

Angle of attack 0 15 30 45 60 75 90 105 120 135 150 165 180

30

AN US

CR IP T

ACCEPTED MANUSCRIPT

M

Fig. 1. The compact urban texture of Yazd city and the typical style of houses, which benefit

AC

CE

PT

ED

from a wind tower connected to parlor and courtyard.

31

AN US

CR IP T

ACCEPTED MANUSCRIPT

Fig. 2. Full CFD assessment methodology flowchart considered for the investigation of the wind

CE

PT

ED

M

tower performance.

AC

Fig. 3. (a) Computational domain for the simulation of external wind; (b) The wind tower and the building model; (c) computational grid.

32

ED

M

AN US

CR IP T

ACCEPTED MANUSCRIPT

Fig. 4. Wind tunnel setup and the location of pitots and statics at the upper and lower parts of the

AC

CE

PT

wind tower’s channels.

33

CR IP T

ACCEPTED MANUSCRIPT

Fig. 5. CFD validation, Comparison of semi-empirical and CFD predicted flow rates of all wind

AC

CE

PT

ED

M

AN US

tower channels at different wind incident angles.

34

M

AN US

CR IP T

ACCEPTED MANUSCRIPT

ED

Fig. 6. (a) Wind incident angles, the wind tower components, including edges, faces, channels, and openings; (b) Extract and supply flow rate of the wind tower channels at 13 wind incident

AC

CE

PT

angles.

35

ACCEPTED MANUSCRIPT

Wind angle

Velocity magnitude / flow streamlines

Contour of Z velocity

CR IP T

0˚

45˚

AN US

90˚

CE

PT

180˚

ED

M

135˚

AC

Location of contour plots

Fig. 7. (Left column) Wind incident angles; (Middle column) velocity contour plot overlaid with flow streamlines around the wind tower; (Right column) Z velocity magnitude contour plot of the wind tower’s channels. 36

CR IP T

ACCEPTED MANUSCRIPT

Fig. 8. The contour plot of velocity magnitude overlaid with vectors at 0º wind angle; (left side)

AC

CE

PT

ED

M

AN US

the magnified area of parlor; (right side) the area of separation point at inlet channel.

Fig. 9. Z velocity profile extracted from all wind tower channels at 0º and 180º wind angles, showing the impact of positive pressure imposed to the mouth of the parlor.

37

CR IP T

ACCEPTED MANUSCRIPT

Fig. 10. The contour plot of velocity magnitude overlaid with vectors at 180º wind angle; (left

AC

CE

PT

ED

M

AN US

side) the magnified area of parlor; (right side) the area of separation point at inlet channel.

Fig. 11. Comparison of total extract and total supply flow rate of wind tower at different wind incident angles.

38