1 INTRODUCTION

PDEC originates from the vernacular architecture of the middle-east [1] and is still a common cooling device in hot-arid regions worldwide. It is a low cost and passive/low energy consump-tion cooling system that can improve thermal comfort and reduce CO2 emissions [2].

The device consists of single or multiple towers equipped with a water/vapour supply placed on the top. During the constant injection of water, droplets descend through the tower and con-ditions close to saturation persist throughout its length. Cool air descends the tower and exits at its base where it is delivered to the adjacent spaces..

The concept is based on the relatively large amount of energy required to convert water from its liquid form to its gaseous form within a local thermal imbalance with subsequence differ-ences in air density. This leads to the movement of air from a zone of high pressure, where air is hot and less dense (top of the tower) to a zone of lower pressure, where air is colder and denser (bottom of the tower).

When the system uses only passive means, and the local thermal imbalance is caused by “free convection”, it is considered passive downdraught. The movement of air inside an evaporative cool tower may occur as a result of a “forced convection”, produced when air is deflected by a solid object or driven by mechanical means such as an electric fan. The system is no longer passive.

Since PDEC is a natural and ventilated system is important to ensure that air flow rates are vertically and horizontally balanced to ensure uniform suppplies of fresh and thermally treated air. This is specially important to verify humidity stratification and local air movement to reduce risk of Legionella, condensation and microbiological growth in stagnant zones that can lead to fabric deterioration and air contamination. It is also important to prevent from high rates of exit air speed that occur as the result of draughts produced when water downdraught is activated.

The best way to solve the Navier-Strokes (N-S) problem of conservation of energy, move-ment and mass of the fluid is using a computacional program, such as a CFD. Today there are few programs that allow a PDEC simulation because there is the need to recognize top-down airflows and driven forces. PDEC is often described as a “reverse chimney” because the column of cool air falls. STAR-CCM+ is a CFD program that analyses particle routine of the water and air inside the PDEC tower and adjacent room in a tridimensional way. A bidimensional CFD program can also be used altough it does not reproduce the interior environment in such an ac-

Dynamical and Thermal Modelling of PDEC: using traditional chimney in Moura and new dwelling in Castelo Branco as case studies

Ana Cláudia Martins de MeloInstituto Superior Técnico, Universidade Técnica de Lisboa, Lisboa, Portugal

ABSTRACT: The present paper is a study on the importance of thermal and dynamic simula-tion of Passive Downdraught Evaporative Cooling (PDEC) towers once applied to both new dwellings and refurbishments in hot-dry areas. The first step was an experiment using PDEC in a hot-dry area in Alentejo where a PDEC shower system was introduced and tested on an exist -ing brick chimney traditionally used for summer ventilation through “chimney” effect. The next step was to validate the collected data with mathematical, thermal and Computer Fluid Dynam-ics (CFD) models. Finally the results helped dimensioning a PDEC tower for a private house in Castelo-Branco where this evaporative cooling system was applied in a complex volumetric room. In this case a tridimensional CFD was used to ensure a good particle routine inside the whole room thus revealling efficiency of PDEC maintaining thermal comfort conditions in-doors.

0,9

2 m0 0,5 1

0,33

0 2 m10,5

1–exterior warm air replacement 2-inertial shower 3-lower exit air temperature from downdraught cooling 4-infiltration/leakage due to opening gaps

curate way.In order to evaluate thermal and dynamical modelling of PDEC two case studies were used to

nalyse cooling and air routine efficiency of PDEC in two hot-arid regions of Portugal: in Moura (38º13N:07º13W), a PDEC shower system was introduced on an existing traditional chimney adjacent to a regular room with a dense occupancy. In Castelo-Branco (39º43’N:07º15’W), a PDEC system was proposed as a cooling device in a private house with complex geometry and architectural form but low internal gains.

On the first case measurements and collected data were first validated through Givoni’s math-ematical model [3] and tested conditions were reproduced on a thermal and a CFD model. On the second case, a thermal model helped defining thermal performance of the building thus help-ing dimensioning PDEC with Givoni´s and Perlmutter et al [4] mathematical models and definig envelope conditions. Such conditions were inputed in a CFD model to evaluate cool air move-ment within the room. In both cases internal conditions were reproduced according to an occu-pancy pattern (number of users of the space/building, heat gains from people, machinery, etc) and physical properties of the construction (envelope, heat losses from envelope, etc). More in-vestigation on case study 2 will include a CFD simulation of downdraught evaporative cooling effect inside the tower.

Results led the author to argue about the importance of PDEC as a cooling and ventilation system in buildings that should be protected from such an extreme natural evironment.

2 CASE STUDY 1 – MOURA´S TRADITIONAL CHIMNEY

The chimney object of the case study is located on the first floor of ATENEU MOURENSE, a small cultural centre in Moura.It has a cylinder shape measuring 4,8 m in height (measured from the floor of the room) and 0,9 m in diameter. It is located on the first floor adjacent to a room with approximately 19 m2

floor area and 63 m3 of volume. The room has tree old wooden doors (two interior and one ex-terior), and one north-facing window with single glazing and alumin-um frame. Internal and external walls are 60 cm thick, made from adobe and finished with lime. The chimney walls are in adobe 0,30 m thick. On the upper limit there are a few openings. There is no lime

rendering on it. The chimney is adjacent to a fireplace that has its own smoke conduct. This con-firms its use as a ventilation tower producing coolness by stack effect.

2.1 PDEC inertial shower

FIELD REPORT

A single shower was placed 4,5 m high above floor plan, with coarse water. The water was provided from an existing tap on the exterior that could be easily controllable. The water

flow rate was 10,5 l/min. Water did not circulate continuously in order to avoid flood problems and the shower operated for only a few minutes. Testing conditions were not ideal, especially due to the use of coarse instead of fine spray.

Figure 1 – Plant and Section of the chimney and adjacent room

0,9

2 m0 0,5 1

0,33

0 2 m10,5

1–exterior warm air replacement 2-inertial shower 3-lower exit air temperature from downdraught cooling 4-infiltration/leakage due to opening gaps

Measured Parameters / time (hour) 11:30 12:30 13:30 14:30 15:301 Exterior conditions

1.1 Air Flow Velocity (m/s) 0,15 0,3 0,2 0,15 0,151.2 Dry Bulb Temperature (ºC) 40 37,1 42,3 41,8 40,71.3 Relative Humidity (%) 19,7 23,4 14,2 15,5 17,2

2 Room conditions before evaporative cooling shower2.1 Air Flow Velocity (m/s) 0,2 0,15 0,15 0,2 0,22.2 Chimney Temperature (ºC) 31,2 28 31,5 32,3 33,62.3 Chimney Relative Humidity (%) 46,7 53,6 43,7 39,8 34,82.4 Room Temperature (ºC) 28,8 27,9 30,4 30,8 32,62.5 Room Relative Humidity (%) 48,5 50,3 42,2 39,6 31,9

3 Room conditions after evaporative cooling shower3.1 Air Flow Velocity (m/s) 0,5 0,6 0,6 0,6 0,53.2 Chimney Temperature (ºC) 26,5 25,5 22,0 26,6 23,43.3 Chimney Relative Humidity (%) 62 79,7 95,5 74 94,23.4 Room Temperature (ºC) 26,6 25,3 22,2 25,2 23,73.5 Room Relative Humidity (%) 56,6 75 91,7 78,6 92

2.2 Measurements

The measurements took place in the 6th of August, every hour, between 11:30 and 15:30. The room´s exterior and interior atmospheric conditions are registered on row 1 and 2 of Table

1. The drop of temperature and relative hu-midity registered at 12:30, is due to a breeze. Results show that after the shower system was activated, both exit air temperature and RH were not stable on the measured

periods . Temperature ranged from 22 oC to

26.6.oC and RH from 62 % to 95.5 %. To validate the experiments measured values, the outdoors climatic data and chim-ney physical properties were first tested

Table 1-Collected data on Givoni´s inertial shower mathematical

model [3]. Values obtained are very similar with the ones measured, except for those registered

at 13:30. The differential is more than 4oC, when the remaining is approximately 0.5oC .

2.3 Modelling

2.3.1Thermal Model

TAS 8.4 was used to create a thermal model with an envelope similar to Moura´s chimney and adjacent room. It was tested according to the weather profile of the experimental day with

input data from 4 people. The program does not include Moura´s weather

data, so a place with similar climatic conditions was used: Phoenix, Arizona, USA because of its hot dry cli-mate and similar latitude(33º26N:112º01E). Also, the test day was not the 6th of August but 29th of July, more similar to the climate parameters measured on the ex-periment day. Output data was similar to those verified on the field report.

Figure 2 – Section (not on scale) and output data from TAS model

70% 80% 90% 100%

2.3.2 CFD Model

AMBIENS is a CFD two dimensional Cartesian grid system, in which a transversal “slice” 1m tick including exit air from chimney, was modelled. The grid represents the computational cells that will be the basis of the simulation. Building elements were not placed reproducing reality thus maximizing output information to better understand indoors thermal comfort. There-fore window and outlets (corresponding to doors gaps) are placed in the same plan, although this does not happen in any 1m slice.

Features represented are:a) one window b) one cooling source (air inlet) c) two door gaps (air inlet and air outlet) Please note that inlet air volume should equal outlet air volume.

Definitions of modelling included:

a) each pixel is equivalent to 0.1m height or width and 0.1m2 area b)4 pixels for each

opening, equivalent to 0.41m2 assuming 2.5m2 in reality Input data for modelling included:

a) 63m3 of room volume b) 2 ACH of infiltration, equivalent to 0.035m3/s of Airflow Rate

in the whole room, therefore 5.83 x 10 -3m3/s in the “slice” c) air speed of 1 pixel equivalent to 0.0075m/s at 0.6m/s or 8 pixels at 0.3m/s i.e., “slice” volume divided by area of each pixel d) 0.6m/s definition of exit air speed, either 4 pixelsTemperature surface (obtained from TAS model) included:

roof - 34o; floor - 31oC; walls - 32oC ; window - 40oC

Figure 3–Relative Hu-midity Figure

4–Air Temperature Figure 5 – Air Speed

2.3.3 Results

According to AMBIENS model, there is an excessive production of humidity inside the chimney (between 90% - 100%) than can lead to mold growth, Legionella and fabric corrosion inside chimney. Junctions of chimney walls and ground are especially affected and water particles that did not fully evaporate before reaching the ground can warmer situation on the bottom PDEC tower. Air beside a window has a higher RH than the rest of the room, excluding the chimney since its surface has a higher temperature. (Figure 4). Here risk of condensation is higher. However, air is distributed in a very satisfactory way within the room, and door gaps are important to help extract “old” air.

Exit air openings can play an important role defining water flow rate. Results also show that even though thermal comfort benefits from reducing on exit air speed, increase of temperature penalizes it. By reducing the water flow rate by 2.5l/min, exit air temperature will increase 1oC. Regarding ventilation supply, the system is efficient since 25 people require approximately 13 ACH (assuming 8l.person.second), and the system can supply 38 ACH.

3 CASE STUDY 2 – CASTELO BRANCO PRIVATE HOUSE

This building is located in a rural area of Caste-lo-Branco. It includes a private house and ateliers for sculputure and peinture. PDEC is part of an holistic environmental strategy for hot and dry periods that in-cludes proper shading, a well insulated envelope and high thermal mass on the walls. Ground works as a temperature stabilizer therefore there is no insolation or slab on it.

Figure 6 – Aerial view of the building

The object of this study is the living room located on the southwest part of the building. It has a sitting romm 4.4 m heigh, a dinning room and kitchen 2.4 m heigh and a corredor 4.4 m heigh that ends on stairs and a door and links to the rest of the dwelling. It has a zenital window on it.

Floor Area 90 m2 Volume – 311 m3

Figure 7 – Scheme of ground floor and upper floor of the living room

3.1 Thermal Model

First thermal performance of the build-ing was tested on RCCTE whose results led to an estimation of energy con-sumption of 11.46 Kwh/m2.year. Ecotect was used to create a thermal model to obtain thermal performance of the room on an extreme hot-dry condi-tion. Since there was no valid metere-ological data for Castelo Branco it was used Cáceres (39º5N:06º5W) with similar latitude. Test day was 21st of August, an extreme hot-dry day.It was assumed 4 occupants.

Figure 8 – Ecotect thermal modelInput data includes:

a) u - value external walls 0.43 W/m2ºC b) u-value floor 0.79 W/m2ºC c) u-value roof 0.27 W/m2ºC d) u-value windows 2.7 W/m2º e) 5 and 7 w/m2 sens-ible and latent heat gains

3.2 Dimensioning PDEC tower

Table 2 – Input parameters for dimensioning PDEC

Further developments of PDEC in Castelo Branco should detail design of the tower. It should include a recycling water system from domestic sources since in PDEC all types of water can be used .In a first stage and taking in account energy consumption from RCCTE [5] and Ecotect it was assumed an aluminium frame cilindrical inertial shower PDEC tower according with parameters from Table 2. A pond needs to be placed at the bottom to collect non evaporated water particles.

3.3 CFD Model

3.3.1 Validation

STAR-CCM+ was first subject to a verification exercise whereby a heated cavity flow was sim-ulated. Experimental measurements were taken from Cheeswright R, King KJ, Zini S [6] who took a cavity with a 5:1 aspect ratio and a temperature difference between the longer vertical sides of 45.8 º. The top and the bottom of the cavity were assumed to no-slip adiabatic walls. The other sides are set to a simetry condition.

3.3.2 STAR-CCM + Model

Radiant temperature from building envelope was necessary to build STAR-CCM+ model and this values where obtained from ECOTEC. It was assumed always the same temperature for each building element since variations ac-cording to solar orientation were negli-gible. Therefore glazing - 30ºC; walls- 27.5 ºC , roof 29ºC . It was assumed floor was adiabatic since it uses earth as a tem-perature stabilizer.It was used 42ºC as a standard outdoor temperature and it was assumed that all the top windows would be 30% open, ex-cept UW1 and ZW1 window, that would be tottaly open. UW4 was disregarded.

Air Flow [7] on each window (Figure 7) UW1 – 0.08 m3/s UW2 – 0.05 m3/s UW3

– 0.25 m3/s Figure 9 – STAR CCM+ model

Minimum diameter of chimney 0,8 m Height of shower 6 mMinimum area of chimney 0,5 m2 Water flow rate 6,66667 l/minVelocity of air flow 0,5 m/s Exit opening 1 m2Outdoor temp 42 degC Air flow rate 1800 m3/hrDesired exit temp 25,36 degC Volume of room 311,5 m3outdoor wetbulb temp 23 deg C ACH 5,77849Indoor max RH% 70 %Rh Water consumption

Abs Hum ambient air temp t 10 g/Kg Energy consumption per year 1035 kwhAbs hum max at that temp 18 g/Kg Money savings per year €

Expected wet bulb depression 13,3 degC Water Consumption l/minWater Consumption (13 h) m3/day

Delivered power in ambient conditions 3,765216 kW Water Consumption (7 h) m3/day

Figure 10 (clockwise) – Temperature, Relative Humidity, Air Speed, Air Speed

The model included a mesh with 123280 cells and 655088 interior faces. The working fluid was air, taken as an ideal gas. The Reynolds-averaged form of the Navier-Stokes equations was solved with the Realizable K-Epsilon turbulence model implemented in STAR-CCM+. Buoyancy was considered via the inclusion of an additional acceleration term in the momentum equations, something which is already an option in STAR-CCM+. Relative humidity was simulated as a passive scalar, to insure conservation and efficiency reasons. A second-order accurate discretization scheme was employed on all convective terms.

3.3.3. Results

Regarding STAR CCM+ results, a clear stratification of the indoor air happens on account of the window openings. Due to the PDEC a good mixture of the lower layer is also achieved, as can be seen in the streamline plots. Please note that the streamlines in the air coming from the windows is colored red to better show the layering. Altough this air does not mix with fresh air from PDEC it act as a “buffer” (with lower RH and higher T) helping maintaining stable and comfortable conditions (26ºC T and 70% RH) on the thermal comfort zone. This “buffer” also protects roof from excessive production of RH thus fabric detioration due to condensation, cor-rosion and microbiological growth. Zenital window is important to extract “old” air. Here air velocity is high (0.8 m/s) but this phenomena occurs above thermal comfort zone. This is due to

its placement on a “corner” of diferent walls with different heights that act as obstacles. Also air in contact with higher temperature increses its velocity. Results sugest zenital window should be placed on the end of the stairs corredor to minimize air turbulence. This kind of turbulence occurs in all the room but with air velocities much slower (0.08 m/s – 0.2 m/s) than the maxim-um recommend (0.5 m/s) [8]. Results also show that exit air from PDEC has a high speed. Therefore pavement and walls sor-rounding tower and pond need to be impermeable and protected from water droplets that may not evaporate before reaching the ground.

4 DISCUSSION

In both studies simulation led to argue about PDEC advantages and disadvanatges helping making some decisions about tower design. Simulation proved that PDEC provides appropriate rates of coolled and ventilated air at the bottom of the rooms where it can be mixed with ambi-ent air within the thermal confort zone. Efficiency of PDEC depends on the replacement air cycle since it deals with high rates of humidity and indoor environment can easily get saturated, i.e., air can not hold more moisture. Therefore an exaust (like an uper window or an electric fan) should always be placed opposite to PDEC tower to extract “old” air thus renewing the air and creating a replacement cycle. Also placement of an microbial bio film on all tower surfaces in contact with water is essential , since these films can consume certain inhibitors and prevent ac-cess of inhibitors;

Neverthless in such an extreme environment (hot-dry) PDEC seems an excellent way to nat-ural ventilate a room. As an example, cross ventilation is not recommended since opening a window would induce warm exit air and increase interior air temperature. No ventilation at all should be avoided since internal environment becomes stagnant and no renewall of air causes thermal discomfort.

Also in PDEC natural replacement air is introduced at a high heigh where air is cleaner than in at the pedestrian level thus minimizing introduction of dust and other particles. Please note that the tower can be used as well as for night cooling through chimney or Venturi effect without PDEC activation.

However PDEC is a specific technique can help reducing the impact of global warming thus providing thermal comfort in building and occupants. In 2004 the annual temperature of Por-tugal increased 0.6 ºC[ 9] and recent studies predict that the dry condition will affect 1/3 of the Portuguese territory [ 10] .

5 ACKNOWLEDGEMENT

The author is gratefull to Nelson Marques, MechEng, PhD, from blueCAPE, for his support concerning STAR CCM+ modelling.

6 REFERENCES

1. Fathy, Hassan “Natural Energy and Vernacular Architecture: Principles and Examples with Reference to Hot Arid Climates , UNU, 1986

2. “Passive Downdraught Evaporative Coolig (PDEC) applied to the central atrium space within the New Stock Exchange in Malta”, WSP Environmental Ltd

3. Givoni, Baruch “Passive and Low Cooling of Buildings” Van Nostrand Reinhold, New York, 19944. Pearlmutter et all Pearlmutter D, Erell, E Etzion, Y Meier, I A, Di H “Refining the Use of

Evaporation in an Experimental Down-draft Cool Tower”,Energy and Buildings 23, 19965. “Regulamento Geral das Edificações Urbanas”, Decreto-Lei nº 40/90 de 6 de Fevereiro6. Cheeswright R, King KJ, Zini S “Experimental Data for the Validation of Computer Codes for the

Prediction of Two-Dimensional Buoyant Cavity Flllows” in ASME Winter Annual Meeting Ana-heim, HTD-60, December 1986.

7. Q = A Cd (2 ΔT H g) / (273 + T)

where Q= air flow rate (m3/s); A (m2)= area of the window; Cd = 0.61; ΔT (ºC)= 12; H (m) = height of window mesured from its center ; g (m/s)= 9.81; T(º)= 42

8. “Environmental Design - CIBSE GUIDE A” The Chattered Institute of Building Services Engineers, London, 1999ASHRAE STANDARD - ASHRAE Guideline12-2000, Minimizing the Risk of Le-gionellosis Associated with Building Water Sytems, Atlanta 2000: American Society of Heating, Refrigerating, and Airconditioning Engineers, Inc.

9. http://www.INM.pt10. Dagmar Schröt , Wolfgang Cramer , Rik Leemans , Colin I. Prentice, Miguel B. Araújo ,

Nigel W. Arnell , Alberte Bondeau , Harald Bugmann , Timothy R. Carter , Carlos A. Gracia , Anne C. de la Vega-Leinert , Markus Erhard , Frank Ewert , Margaret Glendining , Joanna I. House , Susanna Kankaanpää , Richard J. T. Klein , Sandra Lavorel , Marcus Lindner , Marc J. Metzger , Jeannette Meyer , Timothy D. Mitchell , Isabelle Reginster , Mark Rounsevell , Santi Sabaté , Stephen Sitch , Ben Smith , Jo Smith , Pete Smith , Martin T. Sykes , Kirsten Thonicke , Wilfried Thuiller , Gill Tuck , Sönke Zaehle , Bärbel Zierl ,

“Ecosystem Service Supply and Vulnerability to Global Change in Europe” , in http://www.sciencexpress.org, 27/10/2005