Measurement and prediction of the indoor airflow in a room ventilated with a commercial wind tower

Energy and Buildings 84 (2014) 367–377

Contents lists available at ScienceDirect

Energy and Buildings journal homepage: www.elsevier.com/locate/enbuild

Measurement and prediction of the indoor airflow in a room ventilated with a commercial wind tower John Kaiser Calautit ∗ , Ben Richard Hughes School of Civil Engineering, University of Leeds, Leeds LS2 9JT, United Kingdom

a r t i c l e

i n f o

Article history: Received 26 May 2014 Received in revised form 20 July 2014 Accepted 15 August 2014 Available online 23 August 2014 Keywords: Air change effectiveness Air change rate CFD Indoor air quality Mean age of air Wind tower/catcher Wind tunnel

a b s t r a c t This paper introduces a method for a full assessment of the internal air movement characteristics in a room ventilated with a commercial wind tower using scaled wind tunnel and computational fluid dynamics (CFD) analysis. An accurate geometrical representation of the experimental situation was recreated in the CFD simulation. The experimental investigation was carried out using a closed-loop low speed wind tunnel and the indoor air flow distribution was measured and compared with the CFD analysis. Good correlation between the numerical and experimental results was observed (10% average error). Flow visualisation was also conducted to further analyse the airflow structure within the room. The work assessed several ventilation parameters to describe the indoor airflow characteristics such as the mean age of air (MAA), air change rate, and air change effectiveness (ACE). The MAA was calculated using the steady-state method, by solving an additional partial differential equation describing the transport of the scalar “age of air”. The numerical analysis of the MAA allowed to detect less or insufficiently ventilated areas. The CFD code could be useful for the calculation of the MAA and ACE, allowing reduction in time and cost in the evaluation of indoor air quality in buildings ventilated with wind tower or similar device. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Buildings account for 40% of the world energy usage and are responsible for almost 40–50% of the global carbon emissions. Heating, ventilation and air-conditioning (HVAC) systems consumes more than 60% of total building energy utilisation [1,2]. The need to increase the energy efficiency of buildings, while providing comfortable and healthy indoor environment, has become a key challenge in the building sector. Great effort is being made to develop sustainable technologies that are capable of providing good indoor environmental quality (IAQ) and exploiting renewable energy resources [3,4]. A wind tower natural ventilation system is one of the green technologies that has the potential to tackle these challenges. In such a ventilation method, airflow through living spaces is achieved by using the natural driving forces of external wind and the buoyancy effect caused by air temperature differences between indoor and outdoor air. The wind tower technology has been around for many centuries in the Middle East and has been applied commercially in the UK over the last few years [5]. A wind tower can provide a

∗ Corresponding author. Tel.: +0044 75544158981; fax: +0044 75544158981. E-mail address: [email protected] (J.K. Calautit). http://dx.doi.org/10.1016/j.enbuild.2014.08.015 0378-7788/© 2014 Elsevier B.V. All rights reserved.

significantly higher airflow rate than in an equivalent area ventilated by an open window [6]. Wind towers can also provide the benefit of night-time cooling without posing a security risk and daytime ventilation without relying upon opening windows [7]. For a commercial multi-directional wind tower, the channel is typically divided into four quadrants with the cross-divider running the full height of the channel. As shown in Fig. 1, the outdoor fresh air flow is directed into the indoor environment via the windward openings, while the stale air is extracted through other external openings side due to negative wind pressure. As the direction of wind changes so do the function of each of the quadrant in the wind tower. This allows the wind tower to capture and supply fresh air irrespective of the wind direction [4]. Due to the increasing emphasis on the development and use of wind tower devices, there is constant scope for accurately analysing their performance. A number of studies have assessed the performance of modern or commercial wind towers using experimental, numerical and analytical analysis. Most of the research work have primarily focused on the geometry of wind towers (dimension, shape, control damper, diffuser and louvre) and achieved ventilation rates. Elmualim and Awbi [8] compared the natural ventilation performance of modern wind towers with square and circular cross-sections using experimental and numerical analysis. The experimental set-up consisted of a full scale wind tower

368

J.K. Calautit, B.R. Hughes / Energy and Buildings 84 (2014) 367–377

Nomenclature u X, Y, Z Re   Q g A P P Po Ps L W H t

velocity magnitude (m/s) Cartesian co-ordinates (m) Reynolds number air density (kg/m3 ) kinematic viscosity (m2 /s) volume flow rate (m3 /s) gravitational acceleration (m/s2 ) cross-sectional area (m2 ) total pressure loss (Pa) pressure (Pa) total pressure (Pa) static pressure (Pa) length (m) width (m) height (m) time (s)

connected to a model test room. Due to the size of the wind towers (1.5 m × 0.5 m × 0.5 m and 1.5 m × 0.55 m diameter), the experimental investigation was carried out in an open test section (2 m × 2 m) wind tunnel. The work concluded that the sharp edges of the square wind tower created a larger region of flow separation and high pressure difference across the device openings, making it more effective than the circular shape wind tower. Furthermore, the CFD code predicted a reasonable air flow rate compared with the wind tunnel result, despite the limitations of the test setup. Later, Elmualim [8] studied the effect of volume control dampers and diffusers on the ventilation rates of a commercial wind tower using the same method in the previous work [6]. The achieved results showed that the damper and diffuser reduced the air flow by approximately 20% at 3 m/s external wind velocity and 50% at 1 m/s. Hughes and Ghani [9] analysed the effect of varying the wind tower louvre angle on the ventilation pefromance. Eight CFD models were generated with the louvre angle increased by 5◦ increments, for a range of 10–45◦ . Performing the same analysis using physical models would have required more time and effort for the design adjustments. Hence, CFD was used as the primary tool for the investigation. Results showed that a louvre angle of 35◦ provided the optimum ventilation performance. Liu et al. [10] performed a similar study but evaluated the effect of varying the number and length of louvres on the ventilation rates instead. The results indicated that the flow rate of air induced into the wind tower increased with the number of louvre layers and

the highest ventilation rate was achieved when the louvre length equated with the reference length. Like other earlier CFD-based studies on commercial wind towers, the numerical results were also validated against the experimental data of [6]. A good correlation between both methods was observed, although the numerical calculation domain in their CFD analysis did not accurately represent the experimental situation. An outdoor far field wind was considered instead of an open section wind tunnel. Calautit et al. [11] used CFD to compare the performance of a row house model integrated with a traditional and commercial wind towers. The study used grid adaptation to verify the programming and computational operation of the computational model. Su et al. [5] measured the net flow rate of a commercial wind tower using an experimental set-up which included an inlet cone flow meter, balancing chamber and variable-speed blower fan. The measured data was compared with CFD results, and a good agreement between the two methods was achieved. Furthermore, CFD modelling of the wind tower was carried out to simulate the conditions similar to the situation of an outdoor far field wind. The calculated extract flow rate of the wind tower in a far field wind was approximately double that for the situation using a blower fan. A few studies [6,12] have also assessed the performance of a commercial wind tower using analytical modelling. Jones and Kirby [12] proposed a semi-empirical approach in which a comprehensive analytical model was coupled with the data from the controlled experiments of [6] to quantify the flow rates. The semi-empirical model performed well against the CFD model of [9]. The latest research [13–16] on wind towers were mainly focused on traditional systems. 1.1. Indoor air distribution characteristics assessment Fig. 2 shows the full assessment methodology flow chart for the evaluation of the indoor air distribution characteristics in buildings integrated with a commercial wind tower or similar ventilation device. The method uses indices of air change rate, mean age of air, and air change effectiveness. The wind tower geometry and the indoor domain were modelled using a commercially available CAD modelling software and then imported to ANSYS DesignModeller to generate a computational model for the CFD analysis. The same CAD model was used for the rapid prototyping of the wind tunnel model. The work combined the approach of rapid prototyping with wind tunnel testing in order to perform efficient wind tower experimentation in terms of cost-effectiveness and time management. Additionally, the creation of an accurate scaled wind tunnel prototype was essential for the experimental study. The wind tower geometry featured a variety of unconventional and complex parts such as the external louvers and cross-dividers. A model of this

Fig. 1. Schematic of a commercial multi-directional wind tower system.

J.K. Calautit, B.R. Hughes / Energy and Buildings 84 (2014) 367–377

369

Fig. 2. CFD-wind tunnel full assessment methodology.

type made using conventional technology would have taken more time to construct. The wind tower was scaled down by a factor of 10 so it can be tested inside the 0.5 × 0.5 m close test section and ensure that the blockage ratio was small enough, below 5%, that there won’t be significant increase in the flow velocities around the model. In order to have a valid comparison between CFD and experiment, an accurate geometrical representation of the wind tunnel test set-up was recreated in the computational modelling. Care was taken to create a high-quality grid, set consistent boundary conditions and compare the CFD results with the experimental measurements. Following the successful validation of the CFD method, the air change rate, mean age of air, and air change effectiveness values were calculated to evaluate the indoor air distribution characteristic. Air change rate/hour : ACR =

Q × 3600 V

(1)

where Q is the volumetric flow rate of air into the space and V is the interior volume of the room. User-defined scalar was used to implement the steady-state calculation of the mean age of air distribution in the code [17]. To calculate the transport of an arbitrary scale ˚i , one additional convection-diffusion equation was solved, taking the following general form:





∇ × u˚i − ˚i ∇ ˚i = S˚i

(2)

where ˚i is the diffusion coefficient of the scalar ˚i , and S˚i is the source term of the scalar ˚i . The air change effectiveness is defined as the age of air that would occur throughout the room if the air was perfectly mixed, divided by the average age of air where occupants breathe.



Air change effectiveness :

ACE =

Q/V Qage



(3)

The suggested approach will advance the numericalexperimental modelling of commercial wind towers and provide solution to the current limitations found in the review of previous

works. Numerous works have already investigated the effects of different geometry configurations on the performance of a commercial wind tower but only a few have investigated the internal conditions of rooms installed with the device. This study will address this by providing data for the indoor air speed and flow patterns, static pressure distribution, ventilation rates, air change rate, mean age of air (MAA) and air change effectiveness (ACE). 2. Methodology 2.1. Computational fluid dynamics (CFD) setup The CFD simulations were performed with the commercial software ANSYS Fluent 12. The Navier–Stokes governing equations were discretised by a Finite Volume Method (FVM) and the flow fields were estimated using the 3D Reynolds Averaged Navier–Stokes (RANS) model in combination with the standard k–ε turbulent model. The overall accuracy of prediction by the standard k–ε turbulent model was proved acceptable in the simulation of wind tower ventilation [18,19]. The second-order upwind scheme was adopted for the convection term, and a Semi-Implicit Method for Pressure-Linked Equation (SIMPLE) algorithm was used for steady-state analyses. Convergence was monitored and the iterations was ended when all residuals showed no further reductions with increasing number of iterations. The governing equations are detailed below: Mass conservation :

∂ + ∇ × (u) = 0 ∂t

(4)

where  is density, t is time and u refers to fluid velocity vector. Momentum conservation :

∂(u) + ∇ × (uu) = −∇ p + g + ∇ × (∇ u) − ∇ × t ∂t

(5)

where p is the pressure, g is vector of gravitational acceleration,  is molecular dynamic viscosity and  t is the divergence of the turbulence stresses which accounts for auxiliary stresses due to

370

J.K. Calautit, B.R. Hughes / Energy and Buildings 84 (2014) 367–377

velocity fluctuations. Turbulence kinetic energy (k) :

∂(k) + ∇ × (ku) = ∇ × [˛k eff ∇ k] + Gk + Gb − ε ∂t

(6)

Energy dissipation rate (ε) :

∂(ε) ε + ∇ × (εu) = ∇ × [˛ε eff ∇ ε] + C1ε (Gk − C3ε Gb ) − C2 k ∂t (7) where Gk stands for source of turbulent kinetic energy due to average velocity gradient, Gb is source of turbulent kinetic energy due to buoyancy force, ˛k and ˛ε are turbulent Prandtls numbers, C1ε , C2ε and C3ε are empirical model constants. 2.1.1. Computational geometry and grid The commercial wind tower geometry and the micro-macro climate domain were modelled in Solid Edge ST3 software and then imported to ANSYS DesignModeller to generate a computational model. A flow domain representation of the physical geometry of the wind tower under investigation is shown in Fig. 3a. The wind tower with an internal cross-sectional area of 1 m2 was integrated to a micro climate with the length, width and height of 5, 5, and 3 m3 representing a small classroom of 15 people [20]. The internal volume of the device was divided into four equal quadrants by a diagonal cross divider which allowed for air flow supply, regardless of the direction of the wind. The wind tower was modelled with seven louvres angled at 45◦ . The wind tower was assumed to be supplying at 100% or fully open, therefore the volume control dampers was not added to the model [10]. Furthermore, the macro-climate with the length, width and height of 10, 5, and 5 m3 was created to simulate the external wind velocity. A non-uniform mesh was applied to volumes of each of the computational models. The mesh arrangement consisted of around 4 million mesh elements. The generated computational mesh of the wind tower and test room model are shown in Fig. 3b. The grid was modified and refined around critical areas of interests in the simulation. 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 of the velocity fields. Inflation parameters were set for the complex geometry face elements to generate a finely resolved mesh normal to the wall and coarse parallel to it. The two-dimensional faces elements at the selected wall or boundaries were inflated into 3d prism elements which resolve boundary layer properly at relatively less computational cost [21].

2.1.2. Grid verification The accuracy of the results achieved from the numerical modelling is highly dependent on the mesh quality, which equally have implications on the convergence of the model. Grid verification was used to validate the programming and computational operation of the computational model [18]. The numerical grid was refined and locally enriched using the h-method grid adaptation technique [22,23]. This procedure of evaluation requires the use of different mesh sizes (mesh sizes ranging from 1.5 to 7 million elements) and a posterior error indicator (average velocity). The grid was refined until the posterior estimate error becomes insignificant between the number of elements, computational iterations and the posterior error indicator. The maximum error for average velocity was recorded 4.38%. The discretisation error was found to be the lowest at over 7 million cells for the indicated variable. 2.1.3. Boundary conditions In order to have a valid comparison between CFD and experiment, the set boundary conditions for the numerical model were identical to the conditions in the wind tunnel experiment [24]. Fig. 4 displays the computational domain containing the indoor and outdoor fluid volumes. A wall boundary condition was used to create a boundary between each region. The macro-climate (outdoor) fluid volume, used to simulate the external velocity flow field, generates a velocity into the wind tower. One horizontal plane was set as a velocity inlet with the opposite boundary wall set as pressure outlet, to simulate a velocity flow field. The velocity inlet was varied between 1 and 5 m/s to investigate the effect of different outdoor wind speeds. To simulate different wind angles, the macro-climate was rotated around the wind tower model with, the micro-climate remained stationary. 2.2. Wind tunnel experimental set-up and measurement procedure The experimental experiment was carried out in a subsonic wind tunnel in the building physics laboratory of the School of Civil Engineering of the University of Leeds. Full details of the verification and characterisation of the wind tunnel are available in [24]. The uniform flow wind tunnel has a test section of the height, width, and length of 0.5, 0.5, and 1 m3 . Based on the blockage ratio formula in [25], the 1:10 model generated a test section blockage of 4.8% at 0◦ wind, and no corrections were made to the indoor airflow measurements obtained with these tests. The wind tower model was mounted on top of a 0.5 × 0.5 × 0.3 m3 test room representing the indoor space (Fig.3a). In order to measure the indoor velocity at the same points as in the CFD model, several holes were drilled into the walls of the test room. The room model walls were made of

Fig. 3. (a) Commercial wind tower scale model inside the wind tunnel (b) generated computational mesh.

J.K. Calautit, B.R. Hughes / Energy and Buildings 84 (2014) 367–377

371

Fig. 4. CFD flow domain of the macro-micro climate environment and set boundary conditions.

perspex sheet to allow smoke visualisation testing as well as to be able to accurately position the hot-wire anemometer sensors along the measurement points. In aerodynamic studies, the air flow patterns around a structure or a building and thus wind loads on it are a function of the Reynolds number. Therefore, wind tunnel testing on scaled models should ideally be performed at the same Reynolds number as would be experience by the full scale model, thus satisfying Reynolds number similarity. Strict scaling of wind and turbulence Reynolds number for the simulated flow is generally not possible for wind tunnel model testing of building and structures, even in the largest, high speed and most expensive wind tunnels. However, the equality of model and full-scale Reynolds number, based on the mean wind speed and a characteristic dimensions of the structure, is not necessary for sharp edged structure, provided that the model Reynolds number is not less than 10,000 [25]. The flow separation points are fixed at these sharp corner location regardless of Reynolds number, so that wind responses tends to be less sensitive to Reynolds number.

Furthermore, the geometric scale of the model of a structure should be selected to maintain, as close as possible, equality of model and prototype ratios of overall building dimensions to the important meteorological lengths of the simulated wind [25]. This was easily achieved, all the relevant dimensions of the prototype wind tower model and test room were equally scaled down by the appropriate factor. 2.2.1. Measurement procedures In this study, the indoor airflow velocity was measured using a traversing hot-wire anemometer. Nine measurement points in an equally spaced 3 by 3 grid were located within the test room at a height of 1.5 m. Additionally, two measurement points were positioned below the supply and exhaust quadrants of the wind towers and at the bottom of the room (below point 5). Fig. 5 shows the measurement setup and summarises the coordinates of the measurement points. The values of the velocity were obtained from the three components of the vector (X, Y, and Z). The tests were carried out at an inlet wind speed of 3 m/s. The velocity was

Fig. 5. Test room setup and measurement points inside the test room model.

372

J.K. Calautit, B.R. Hughes / Energy and Buildings 84 (2014) 367–377

Fig. 6. Wind tunnel smoke visualisation set-up.

confirmed during the setup and configuration of the wind tunnel during commissioning [24]. The airflow inside the wind tunnel was allowed to normalise before measurements were taken. The effect of the urban boundary layer on the ventilation performance was not investigated in the study. The sample for each point was taken and averaged over a 2 min period with the results and start/finish times recorded. The uncertainties associated with the velocity readings (Testo 425) were estimated to be ±1.0% of reading at speeds lower than 8 m/s and ±0.5% of reading at higher speeds (8–20 m/s). Flow visualisation test was carried using an AFA-10 wind tunnel smoke generator to further investigate the flow pattern inside the test room model and identify the supply and extract quadrants prior to airflow measurements (Fig. 6). The wind tower model was exposed to a free stream air velocity of 3 m/s to obtain smoke of a sufficiently high concentration. A high speed camera was used to capture the movement of smoke-visualised airflow paths.

3. CFD model validation 3.1. Indoor airflow distribution Fig. 7 compares the measured and CFD predicted results of the airflow velocity inside the test room. As expected, highest airflow speed was measured below the wind tower supply opening (point 11) and at the middle of the room (point 5). The speed reduced as it approached the bottom (point 10) and spreads from all directions across the room. The trend (points 1–12) showed that the CFD model was capable of predicting the indoor airflow distribution. Average error across the points was 9% with highest error at 12% (point 2) and lowest at 1.2% (point 11). The model did not predict the velocity located in the recirculation zones as accurate as the other points, this was one of the limitations of the k-epsilon turbulence model. Using a similar justification

Fig. 7. Comparison between predicted and measured indoor air velocity (points 1–12) with outdoor wind speed set at 3 m/s.

J.K. Calautit, B.R. Hughes / Energy and Buildings 84 (2014) 367–377

373

Fig. 8. (a) experimental smoke visualisation inside the test room with the wind tower (b) combined CFD predicted streamlines and smoke visualisation with outdoor wind speed set at 3 m/s.

recommended by [26], it can be claimed that the validation of the CFD modelling study was acceptable. 3.2. Wind tunnel smoke visualisation Fig. 8a and b compares the visualised and CFD airflow pattern inside the test room. The air flow passed around the wind tower with some of the air entering the wind tower through the louver openings. A more visible amount of smoke was seen being displaced at the point of entry (below the wind tower) which indicated higher air speed at this part of the room, this was consistent with the results in Figs. 7 and 8b. The airflow was directed towards the bottom surface of the room and spread outwards in all directions. As the airflow hits the bottom surface the air slows down and flows through the side walls. Although a similar pattern was observed, clearly, the smoke test did not capture the full flow structure (especially the large recirculation occurring at the bottom left area) which showed the advantage of using CFD streamlines when analysing internal flows in buildings. The smoke was less visible on the left side of the ventilated space, which was due to the exhaust quadrant pulling the air out of the room. The smoke visualisation tests also helped to detect small air short-circuiting (air entering through the supply quadrant and immediately leaving through the exhaust without flowing inside the room). Since the smoke was injected near the wind tower inlet (0.2 m distance) and the high speed camera was configured to clearly visualise the internal flows which were at a much lower speed than the external flows, the recirculation zone and flow pattern at the leeward side of the wind tower were less visible. 4. Results and discussion 4.1. Overall airflow distribution Fig. 9 depicts the airflow path predicted by the numerical model. From the contour plot, the airflow enters the inlet boundary wall on the right and the airflow splits with some entering the wind tower and some exiting to the pressure outlet. As expected, large recirculations were present on the leeward side of the wind tower. Clearly, locating another wind tower device in this region won’t be advantageous [20]. The airflow entering the wind tower was accelerated as it hits the cross-dividers, reaching a maximum speed

of 2.5 m/s and forced the flow down into the space. Since there were no volume control dampers present in the simulated model, a larger throw (maximum distance the air can reach effectively at a given supply velocity) and medium spread (width of the air stream) air stream shape was observed. A larger throw of a supply outlet was important because the air must reach into the room as far as possible for good mixing. However, in a typical occupied room, high draft speed at the center of the room ranging from 0.3 to 2.5 m/s (1–5 m/s external wind) can be considered a comfort issue. Therefore, it is important to locate the wind tower and direct the air stream properly depending on the occupancy to avoid undesired sensation of a draft. The airflow reduced speed (1 m/s) as it approached the test room floor and spreads outwards in all directions, causing air recirculation inside the room and further reducing the air velocity to 0.20 m/s. At an external wind speed of 3 m/s, the average velocity of the airflow leaving the wind tower quadrant was 1.62 m/s while the average velocity in the test room was 0.55 m/s. 4.2. Overall static pressure distribution Fig. 10 shows the static pressure contour of the mid-plane inside the test room with the wind tower. The air striking the wind ward surface exerted a positive pressure on the inlet face, which forces air through wind tower openings and into the interior of the room. At the same time the windward surface was receiving positive pressure, the side and rear surfaces were receiving negative pressure; therefore, air within the room was drawn out at openings in these other surfaces. The highest pressure was obtained upstream of the louvers with a maximum value of 6.5 Pa. Negative pressure was observed at the back and upper side of the wind tower with a minimum value of −6.0 Pa. The room under negative pressure indicated that less air was supplied to the room than exhausted which was the case for a multi-directional wind tower with the external wind at 0◦ angle; there were three exhaust quadrants and only one supply quadrant. 4.3. Effect of wind conditions on ventilation rate Fig. 11 shows the numerical results of the supply and exhaust airflow through the wind tower quadrants at various wind angles (0–90◦ ). Fig.11a displays the horizontal contour plot drawn below the wind tower channel. It was observed that at 0◦ angle, a large

374

J.K. Calautit, B.R. Hughes / Energy and Buildings 84 (2014) 367–377

Fig. 9. CFD velocity contour plot of a cross sectional plane in the computational domain.

volume of the wind tower was used for extract purposes. While the wind tower oriented at 45◦ into the prevailing wind had a larger area available to capture the wind. In this case, two windward quadrants were available for the air flowing into the tower and two leeward quadrants for the air flowing out of the tower. Developing regions of vortices were observed inside the exhaust channel of the wind tower angled at 30◦ and 60◦ which reduces the performance of the wind tower. In Fig. 11b, the supply and the

extract segments are recognised by positive and negative values of airflow rate. A volumetric airflow rate of 0.32 m3 /s was achieved through the supply quadrant 1 at 0◦ for an external wind speed of 3 m/s. As the wind angle increased, the supply airflow through quadrant 1 decreased. Exceeding the wind angle over the transition angle (>70◦ ), caused a change in airflow direction into quadrant 1. At 45◦ wind angle, a net volumetric flow rate of 0.47 m3 /s was achieved through combined supply quadrants 1 and

Fig. 10. Static pressure contour plot of a cross sectional plane in the computational domain.

J.K. Calautit, B.R. Hughes / Energy and Buildings 84 (2014) 367–377

45

30

0/90

(a)

Quadrant 1

Quadrant 2

375

60

Quadrant 3

Quadrant 4

(b) 0.4 Flow rate (m3/s)

0.3 0.2 0.1 0 -0.1 -0.2 -0.3 -0.4 0

10

20

30

40

50

60

70

80

90

Wind direction (deg.) Fig. 11. (a) Horizontal velocity contour plot below the wind tower model (b) volumetric airflow through the wind tower quadrants at different wind directions.

3, with the exhaust flow rate from the opposite quadrants at its maximum. The simulation models were run for varying wind speeds (0.5–5 m/s). The supply rates, indoor velocity and static pressure readings were taken from the weighted-average of the horizontal diffuser surface (Fig. 11a) and indoor points (Fig. 5). The results of the simulations are summarised in Table 1. Approved Document F1A suggests that a minimum air supply rate per occupant of 10 L/s per person [27] is required for a small classroom of 15 people [20, 28]. Based on the review of [29] on the associations of indoor ventilation rates and carbon dioxide concentrations, most of the studies found that ventilation rates below 10 L/s per person in all building types were associated with statistically significant worsening in one or more health or perceived air quality outcomes. Furthermore, increase in ventilation rates above 10 L/s per person, were associated with further significant decrease in the prevalence of sick building syndrome (SBS). Although the 1 × 1 m2 wind tower device analysed in this study did not meet this recommendation

for an external wind velocity of 1 m/s and below; the system surpassed the recommendation exponentially as the external velocity increased (2 m/s and above) as shown in Table 1. 4.4. Mean age of air The application of the theoretical models is most suited to mechanical ventilation with constant volume flow rates (magnitude and direction). In this study, the natural ventilation flows was treated as steady in the mean and in that sense there was no difference from mechanical system. Accepting that the main aim of analysis was to achieve a certain flow pattern, the age of air distributions and the air change effectiveness can, in principle, be calculated for a given pattern [30]. Mean Age of Air (MAA) is defined as the average lifetime of air at a particular location in the room relative to the time when it first entered the room. It is used to assess the quality of ventilation and identify poorly ventilated areas [31]. This study used a validated method described by [17,32] to

Fig. 12. Mean age of air contours inside the test room with external wind set at 3 m/s.

376

J.K. Calautit, B.R. Hughes / Energy and Buildings 84 (2014) 367–377

Table 1 Summary of the simulation results. Air change rate [1/h]

Supply rate [L/s]

Supply rate [L/s/person] 15 occupants

Supply rate [L/s/m2 ] Area = 25 m2

Average indoor velocity [m/s]

Average indoor pressure [Pa]

0.50 1.00 2.00 3.00 4.00 5.00

3.00 6.48 13.20 19.44 27.60 34.68

62.50 135.00 275.00 405.00 575.00 722.50

4.17 9.00 18.33 27.00 38.33 48.17

2.50 5.40 11.00 16.20 23.00 28.90

0.09 0.19 0.40 0.55 0.81 0.99

−0.05 −0.12 −0.61 −1.28 −2.41 −3.64

Mean Age of Air (seconds)

Inlet speed [m/s]

900 800 700 600 500 400 300 200 100 0

Indoor Average

0.5

1

1.5

2 2.5 3 3.5 External wind speed (m/s)

Maximum

4

4.5

5

Fig. 13. Effect of varying external wind speed on the mean age of air (MAA).

calculate the MAA by means of an numerical procedure using the CFD code Fluent (detailed in Section 2). Fig. 12 shows the cross-sectional contour plot of the MAA in the ventilated space. The MAA distribution followed the airflow pattern in the room increasing gradually from the inlet along the supply jet and reaching the highest value in the recirculation zones. The MAA was lowest in the central area (below the wind tower supply) indicating a fresher air in that location. Furthermore, areas with a lower ventilation air distribution or stagnant regions have a mean age of air that was higher than the space average of

178 s. The space ventilated by a wind tower exhibits MAA values between 100 s (center) and 200 s (near walls) at the height of 1 m, which was representative of the breathing height of a sitting person. Fig. 13 shows the calculated values of mean age of air (MAA) at different external wind speeds, with lower values being more favourable. The maximum MAA was 23–30% higher than the indoor average MAA. This difference appears to get smaller as the external wind speed increases which indicated that more mixing was present at higher speeds.

Fig. 14. (a) Air change effectiveness (ACE) contours at 1 m height (breathing height) with external wind speed set at 3 m/s (b) effect of varying external wind speed on the ACE.

J.K. Calautit, B.R. Hughes / Energy and Buildings 84 (2014) 367–377

4.5. Air change effectiveness Air change effectiveness (ACE) is defined as the ratio of a nominal time constant to a mean age of air [33]. The nominal time constant was calculated as a ratio of the room volume (5 × 5 × 3 m3 ) to the supply air volume flow rate (m3 /s) to the room. In other words, the relative air-change effectiveness is a measure of how effectively supply air was used to ventilate the zone. If the air within a space is perfectly mixed, then the values of local air change effectiveness at all locations within the space and the air change effectiveness for the space will equal to 1.0. An ACE value of less than 1.0 indicates that the air distribution within the space is less than perfect mixing and above 1.0 indicates better ventilation performance [34]. Fig. 14a displays the calculated values of ACE at breathing height (1 m) ranging from 0.86 (recirculation areas) to 2.1 (supply jet). Values of the ACE in the occupied space were generally close to 1.0, which was consistent with good mixing of the ventilation air within the ventilated space. Fig. 14b shows the calculated values of indoor average ACE at different external wind speeds. 5. Conclusions In this study, CFD analysis and scaled wind tunnel testing were conducted to perform detailed analysis of the indoor air flow in a test room with a commercial multi-directional wind tower. The CFD code Fluent was used to evaluate the airflow in the test room which represents a small classroom. An accurate geometrical representation of the wind tunnel test set-up was recreated in the numerical modelling. Care was taken to create a high-quality grid, set consistent boundary conditions and compare the CFD results with detailed wind tunnel measurements. The CFD simulations were generally in good agreement (0–12%) with the wind tunnel measurements. The study provided extensive data for the air change rate, supply rates, indoor velocity and indoor static pressure distribution. The 1 × 1 m2 wind tower analysed in this study did not meet the recommendation of 10 L/s per occupant when the external wind velocity was at 1 m/s and below, however, the system surpassed the recommendation exponentially as the external velocity increased. The mean age of air (MAA) was calculated using the steady-state method. At 3 m/s external wind, the space exhibits MAA values between 100 s (centre) and 200 s (near walls) at the height of 1 m. The calculated values of MAA at different outdoor wind speeds showed that the maximum MAA value was 23–30% higher than the indoor average. This difference appears to get smaller as the external wind speed increases which indicated that more mixing was present at higher speeds. Air change effectiveness (ACE) was determined based on the calculated mean age of air. Values of the ACE in the occupied space were generally close to 1.0, which was consistent with good mixing of the ventilation air within the ventilated space. The CFD model could be very useful for the calculation of the MAA and ACE, allowing reduction in time and cost in the evaluation of indoor air quality in buildings ventilated with wind tower or similar device. Acknowledgement The support by the School of Civil Engineering, University of Leeds is gratefully acknowledged. The statements made herein are solely the responsibility of the authors. References [1] L. Pérez-Lombard, J. Ortiz, C. Pout, A review on buildings energy consumption information, Energy and Buildings 40 (2008) 394–398.

377

[2] Y. Wang, F. Zhao, J. Kuckelkorn, D. Liu, J. Liu, J. Zhang, Classroom energy efficiency and air environment with displacement natural ventilation in a passive public school building, Energy and Buildings 70 (2014) 258–270. [3] M. Santamouris, A. Synnefa, M. Asssimakopoulos, I. Livada, K. Pavlou, M. Papaglastra, N. Gaitani, D. Kolokotsa, V. Assimakopoulos, Experimental investigation of the air flow and indoor carbon dioxide concentration in classrooms with intermittent natural ventilation, Energy and Buildings 40 (2008) 1833–1843. [4] J.K. Calautit, B.R. Hughes, H.N. Chaudhry, S.A. Ghani, CFD analysis of a heat transfer device integrated wind tower system for hot and dry climate, Applied Energy 112 (2013) 576–591. [5] Y. Su, S. Riffat, Y. Lin, N. Khan, Experimental and CFD study of ventilation flow rate of a MonodraughtTM windcatcher, Energy and Buildings 40 (2008) 1110–1116. [6] A. Elmualim, Effect of damper and heat source on wind catcher natural ventilation performance, Energy and Buildings 38 (2006) 939–948. [7] N. Khan, Y. Su, S. Riffat, A review on wind driven ventilation techniques, Energy and Buildings 40 (2008) 1586–1604. [8] A. Elmualim, G.B. Awbi, Wind tunnel and CFD investigation of the performance of Wind catcher ventilation systems, International Journal of Ventilation 1 (2002) 53–64. [9] B.R. Hughes, S.A. Ghani, A numerical investigation into the effect of Windvent louver external angle, Building and Environment 45 (2010) 1025–1036. [10] S. Liu, C.M. Mak, J. Niu, Numerical evaluation of louver configuration and ventilation strategies for the windcatcher system, Building and Environment 46 (2011) 1600–1616. [11] J.K. Calautit, B.R. Hughes, S.A. Ghani, Numerical investigation into the feasibility of integrating green building technologies into row houses in the Middle East, Architectural Science Review 56 (2013) 279–296. [12] B.M. Jones, R. Kirby, Quantifying the performance of a top–down natural ventilation WindcatcherTM , Building and Environment, 2009 (44) (2009) 1925–1934. [13] V.A. Reyes, S.L. Moya, J.M. Morales, F.Z. Sierra-Espinosa, A study of air flow and heat transfer in building-wind tower passive cooling systems applied to arid and semi-arid regions of Mexico, Energy and Buildings 66 (2013) 211–221. [14] A.A. Dehghan, M.K. Esfeh, M.D. Manshadi, Natural ventilation characteristics of one-sided wind catchers: experimental and analytical evaluation, Energy and Buildings 61 (2013) 366–377. [15] S.M. Hosseinnia, H. Saffari, M.A. Abdous, Effects of different internal designs of traditional wind towers on their thermal behavior, Energy and Buildings 62 (2013) 51–58. [16] L.C. Haw, O. Saadatian, M.Y. Sulaiman, S. Mat, K. Sopian, Empirical study of a wind-induced natural ventilation tower under hot and humid climatic conditions, Energy and Buildings 52 (2012) 28–38. [17] G. Gan, Effective depth of fresh air distribution in rooms with single-sided natural ventilation, Energy and Buildings 31 (2000) 65–73. [18] B.R. Hughes, J.K. Calautit, S.A. Ghani, The development of commercial wind towers for natural ventilation: a review, Applied Energy 92 (2012) 606–627. [19] J.K. Calautit, B.R. Hughes, Wind tunnel and CFD study of the natural ventilation performance of a commercial multi-directional wind tower, Building and Environment 80 (2014) 71–83. [20] J.K. Calautit, D. O’Connor, B.R. Hughes, Determining the optimum spacing and arrangement for commercial wind towers for ventilation performance, Building and Environment (2014), http://dx.doi.org/10.1016/ j.buildenv.2014.08.024. [21] J.K. Calautit, B.R. Hughes, H.N. Chaudhry, S.A. Ghani, Comparison between evaporative cooling and heat pipe assisted thermal loop for a commercial wind tower in hot and dry climatic conditions, Applied Energy 101 (2013) 740–755. [22] J.K. Calautit, B.R. Hughes, S.A. Ghani, Numerical investigation of the integration of heat transfer devices into wind towers, Chemical Engineering Transactions 34 (2013) 43–48. [23] B.R. Hughes, H.N. Chaudhry, J.K. Calautit, Passive energy recovery from natural ventilation air streams, Applied Energy 113 (2013) 127–140. [24] J.K. Calautit, B.R. Hughes, H.N. Chaudhry, L.F. Sim, A validated design methodology for a closed-loop subsonic wind tunnel, Journal of Wind Engineering and Industrial Aerodynamics 125 (2014) 180–194. [25] N. Isyumov, Wind Tunnel Studies of Buildings and Structures, ASCE Publications, Virginia, 1998. [26] Z. Zhang, W. Zhang, Z. Zhai, Q. Chen, Evaluation of various turbulence models in predicting airflow and turbulence in enclosed environments by CFD: Part 2—Comparison with experimental data from literature, HVAC and R Research 13 (2007) 871–886. [27] Building Regulations, Building Regulations 2000: Approved Document F1A: Means of Ventilation, Building Regulations, London, 2010. [28] D. O’Connor, J.K. Calautit, B.R. Hughes, A study of passive ventilation integrated with heat recovery, Energy and Buildings (2014), http://dx.doi.org/10.1016/j.enbuild.2014.05.050. [29] O.A. Seppänen, W.J. Fisk, M.J. Mendell, Association of ventilation rates and CO2 concentrations with health and other responses in commercial and institutional buildings, Indoor Air 9 (1999) 226–252. [30] D. Etheridge, Natural ventilation of buildings: theory, in: Measurement and Design, Wiley, London, 2011. [31] H. Awbi, Ventilation of Buildings, Routledge, West Sussex, 2003. [32] N. Kobayashi, Q. Chen, Floor-supply displacement ventilation in a small office, Indoor and Built Environment 12 (2003) 281–292. [33] AHSRAE, ANSI/ASHRAE Standard 129-1997, Measuring Air-Change Effectiveness, AHSRAE, Atlanta, GA, 1997. [34] AHSRAE, ASHRAE Handbook—Fundamentals (SI), AHSRAE, Atlanta, GA, 2005.