Energy and Buildings 36 (2004) 353–366

Implementation and experimental survey of passive design specificationsused in new low-cost housing under tropical climates

François Gardea,∗, Laetitia Adelarda, H. Boyera, C. Ratba Laboratoire de Génie Industriel, Equipe Génie Civil Thermique de l’Habitat, IUT de Saint-Pierre, Université de La Réunion,

40 avenue de Soweto, 97410 Saint-Pierre, Ile de La Réunion, Franceb Agence Régionale de l’Energie à La Réunion, 40 avenue de Soweto, BP 226, 97456 Saint-Pierre Cedex, France

Abstract

This paper deals with a global approach in the thermal and airflow design of new dwellings and specially low-cost housing built in theFrench tropical overseas islands. The sizing of passive cooling techniques has been evaluated thanks to a simulation code. The solutionshave been edited in a reference document that serves as work-base for the design of dwellings developments in the French tropical territories.An experimental and sociological survey has been carried out over two years (from 1998 to 1999) in order to validate the impact of thesepassive cooling specifications on the comfort of tenants and to have an immediate feedback and correct the solutions if necessary. Thepaper presents the implementation, the experimental survey and the results for the two first housing projects built according to the passivedesign specifications. This step is important because the aim of the public utilities is to expand the ECODOM specifications on a muchbroader scale and to prepare to future housing thermal regulations. The results indicated that the solar protection of the roof remains oneof the main points in the thermal design of buildings in a tropical climate as well as the natural ventilation. The sociological study showedthat the tenants were globally satisfied with their flat as far as the thermal aspects are concerned. However, dissatisfaction resulted fromthe bad acoustic design of the flats.© 2004 Elsevier B.V. All rights reserved.

Keywords:Low-cost housing; Tropical climate; Experimental validation; Passive design; Sustainable development

1. Introduction

Managing the response to the demand for electricity inthe French Overseas Territories, known by their Frenchacronym ‘DOMs’, has long been a problem, and so thelocal government, with the encouragement of the Frenchstate electricity utility, has initiated a wide-ranging projectnamed ‘Experimental project ECODOM’, with the aim ofimproving the thermal design of new dwellings. Whereaslegislation pertaining to thermal standards—i.e. to combatwinter condition, is applied in mainland France, this leg-islation is not applied in the DOMs, which are in a humidtropical climate. There are currently no such rules govern-ing correct practice in a tropical climate. This lack of guid-ance has resulted in the construction of dwellings whichare completely unsuited to a tropical climate, and thereforeuncomfortable for the inhabitants. Apart from this thermaldesign problem, there is another issue common to every

∗ Corresponding author. Tel.:+262-262-96-28-96;fax: +262-262-96-28-99.E-mail address:garde@univ-reunion.fr (F. Garde).

island, where electricity generation is both limited and ex-pensive. Clearly, active measures are required to deal withthis problem of electricity supply and demand. In addition,the rising average standard of living of those living in theFrench DOM, the drop in price of air-conditioning units, therelatively rapidly growing population in these areas and theresultant need for new construction all point to the need fora wide-ranging initiative treating building thermal design.

One can think that the tenants will be less inclined to useactive air-conditioning systems if their dwellings are betteradapted to the tropical climate. Finally, the potential benefitsare not only energetic, but also environmental, economic andsocial.

2. A method of work based on a close co-operationwith people involved in the building process

The methodology of the ECODOM project relies on fourmain steps:

1. Setting-up of simple and pedagogical rules governing thedifferent passive components of the envelope of a typical

0378-7788/$ – see front matter © 2004 Elsevier B.V. All rights reserved.doi:10.1016/j.enbuild.2004.01.045

354 F. Garde et al. / Energy and Buildings 36 (2004) 353–366

dwelling from 1994 to 1997. Elaboration of what is callednow the ECODOM reference document[1].

2. Search for the first housing project to implement theECODOM passive solutions.

3. Experimental validation of the ECODOM specificationsthrough a technical and sociological monitoring on thefirst ECODOM projects to have an immediate feed backand modify the specifications if necessary (from 1996 to1998).

4. Setting up of building codes and thermal regulations forthe residential buildings in the French overseas territo-ries thanks to the ECODOM experience (currently inprogress).

The elaboration of passive standards has already beenpresented in a previous paper[2]. Section 3will make a shortreview of the ECODOM specifications but the paper itselfwill focus on steps 2 and 3—i.e. the implementation and theexperimental validation of the ECODOM specifications.

2.1. Elaboration of simple and instructive rules

This step lasted three years, from 1994 to 1997. A networkhas been set up since 1994 between the ECODOM partners(ministry of housing, architects, the French authority for en-vironment and energy conservation, energy consultants, theUniversity of Reunion Island, low cost housing companies,public utilities . . . ). This network of partners has been con-sulted at each step of the ECODOM project.

The final objective of this step was to have a tool thatallow to work with the architect from the first sketch rightthrough to completion of the project. This current problemhas been the subject of several recent publications, whichdiscuss the kind of design tool and types of solutions chosento help architects from their initial sketch[3–6]. Typically,the proposed methods are based on the use of simplifiedcomputing tools with interfaces which are as user-friendlyas possible. Holm[3] and Ellis and Mathews[6] state thatthe basic needs identified for architects are the user friend-liness of the design tool, the type of input format (buildingmaterials rather than thermal properties), short calculationtime and the option of ‘what if’ alternatives.

Aware of that, the working method used to launch theexperimental operation ECODOM is based on the useof a simple, instructive design tool, in the form of a pa-per document, giving passive solutions relating to all thebuilding components. Extensive computational simulationsusing detailed programs has been required to develop theexpert rules, as explained in[2] and [7], but once theserules have been established, it is no longer necessary touse a simulation program to design the building. The useof expert rules throughout the building process is justifiedbecause the buildings to be constructed (mostly low-costhousing) feature relatively simple architecture. This methodhas already been applied by Mathews and van Wyk[8] andMalama and Sharples[9] for the large-scale construction of

low-cost housing. Further, buildings in tropical climates are“passive” and designed to be used without air-conditioning.Conditions of comfort must be reached by modifying thethermal and airflow design of the envelope.

The design and size of each architectural component withregard to internal thermal comfort in buildings has been as-sessed with validated thermal and airflow building simu-lation software[10]. Simulations were run to estimate theimpact of each passive solutions on thermal comfort, i.e.the resultant temperature, on natural ventilation and on theenergy performances of housing buildings representative oflocal design habits. The reference document[1] containingall the passive cooling solutions and expert rules was editedin 1997 after much editing among the different partners totake into account the pedagogical aspects of the document.This document now serves as a work-base aid among all thepeople involved in the building process.

Finally,Fig. 1illustrates the elaboration of the expert rulesand the actual collaboration process between the architectand the building thermal engineer as a function of time.

The experience over the last five years has been pointingout that the working method proposed is efficient: more than800 dwellings have been built to the present time, respectingthe ECODOM specifications[7].

2.2. Implementation and validation (from 1997 to now)

The next step is to implement the ECODOM specifica-tions on new building projects and to carry out a technicaland sociological survey in the first dwellings built. The ex-perimental feedback gives informations on the thermal andairflow performance of the ECODOM dwellings and of thethermal comfort experienced by the tenants and allow tomodify the ECODOM specifications if necessary.

This experimental stage allows to answer several ques-tions which have only been treated numerically at the sim-ulation stage:

• Do the dwellings perform according to the resultant tem-perature objectives fixed at the simulation stage—i.e. doesthe resultant indoor temperature remain below the outdoortemperature?

• What effect have the various solar protection devices(roofing, walls and windows) had on the comfort and thetemperature?

• What kind of ventilation exists in the dwellings? What isthe order magnitude of the internal air velocity?

• What do the tenants think of the passive solutions? Arethere any particular problems caused by the open natureof the dwellings, and the airflow?

The technical and sociological validation period waslaunched in the first ECODOM operations and lasted twoyears thanks to the financial help of the French Ministryof Housing. The first two ECODOM projects were instru-mented during the ‘1998’ and ‘1999’ summer seasons. Thenames of these projects are ‘La Trinité’ and ‘La Découverte’.

F. Garde et al. / Energy and Buildings 36 (2004) 353–366 355

Fig. 1. Sequence of stages in the project: The expert rules generated using specialised programs are the only reliable tools used in practice by both thearchitect and the thermal engineer, with a close co-operation at each stage of the project[7].

Fig. 2. Location of the ECODOM projects ‘La Trinite’ (bottom left) and ‘La Decouverte’ (top right).

356 F. Garde et al. / Energy and Buildings 36 (2004) 353–366

Sections 3–5will describe the implementation of theECODOM specifications, the experimental setup of thedwellings and the main results. The entire results of thestudy are presented in four research reports[11–14].

Fig. 2.

3. Review of the specifications contained in thereference document [1]

The reference document contains five main sections, cor-responding to specific parts of the building. These sectionsconcern:

1. The location on site, i.e. vegetation around the building;2. The solar protection of the building, i.e. roof, walls and

windows;3. The natural ventilation, i.e. exploitation of trade winds,

optimised ratio of inside/outside air-permeability ofdwelling envelope, use of indoor louvers or mechanicalventilation such as ceiling fans;

4. The domestic hot water production, i.e. solar waterheaters, efficient gas water heaters or properly designedoff-peak hours servo-controlled high efficiency storageelectric water heaters;

5. As an option, high efficiency intermittent air conditioningsystems for bedrooms, i.e. air tightness of rooms, efficientcomponents, controls, etc.

For each section, a choice of technical solutions is pro-posed to guide the architect or the climate engineer in thedesign of the building without using a dynamic simulationtool.

The solar protection devices sizing is made thanks to ta-bles that give geometric values for each component of thebuilding. For example,Table 1gives the correspondence be-tween the colour of the roof, the type of roof (terrace roof,etc.) and the minimum insulation thickness required to min-imize the solar gains.

It is recommended that all windows must be protected bysome kind of window shading system, such as horizontaloverhangs and other shading devices such as Venetian blindsor opaque, mobile louvers. If the architect decides to usehorizontal overhangs,Table 2gives geometric characteristicsof overhangs in relation to the orientation of the window.

Dwellings must have layouts allowing for effective crossventilation of all rooms and optimise trade winds. During

Table 1Solar protection of the roof. Minimum insulation thickness required func-tion of the roof colour

Roof colour Polystyrene typeinsulation (cm)l = 0.041 W m−1 K−1

Polyurethane typeinsulation (cm)l = 0.029 W m−1 K−1

Light (α = 0.4) 5 4Medium (α = 0.6) 8 6Dark (α = 0.8) 10 8

Table 2Solar protection of windows

Orientation

East South West North

Reunion 0.8 0.3 1 0.6Guadeloupe 0.8 0.6 1 0.3

Geometric values of the horizontal overhangd/(2a + h) (case 1) ord/h(case 2).

the hot season in a tropical island, trade winds are thermalwinds. They blow most of the time perpendicular to the sea.At each level or floor, there should be openings in each mainroom, on at least two opposite facades (main rooms beingbedrooms and the living room, the dining room and otherliving spaces). Interior layout should also be designed to al-low outside air to flow through main rooms from one facadeto the other through halls, corridors, doors and other internalopenings in room partitions. The first step is to calculate themean porosity. The internal and external porosities requiredby ECODOM standards are at least 25% of the mean of twoopposite facades of dwellings.

Dwellings are equipped with domestic hot water heatingsystems which are energy efficient, long-lasting and eco-nomic. Water heating can be solar, electric or gas.

The last section allows building owners to use air con-ditioning systems. In that case, specifications concern thecontrol of air renewal (a maximum of 1 ACH is tolerated),the performance of the system (minimum COP of 3 is re-quired for split-systems) and its set up (outdoor unit pro-tected against solar radiation and set up in a well ventilatedplace).

4. Implementation of passive design specifications ontwo building projects

Once the reference document was edited in 1997, welooked for some low-cost housing projects in which to ap-ply the ECODOM specifications. The first two projects to bestudied were the project of ‘La Trinité’ and ‘La Découverte’,in the area of the main town of the island called Saint-Denisand located 21◦ south and 55◦ east.

4.1. The location of the projects

The project ‘La Trinité’ is a low-cost housing project lo-cated in a very dense urban area and is not very well venti-lated by trade winds. The main facades face east and west.

F. Garde et al. / Energy and Buildings 36 (2004) 353–366 357

Fig. 3. ‘La Trinite’—western facade. Insufficient solar protection of the windows.

The project ‘La Découverte’ is located in the eastern partof Saint-Denis in the Sainte-Marie area. The site is well ven-tilated by trade winds thanks to the non-urban environment.The dwellings are low-cost dwellings and the orientationsof the main facades are north and south.

4.2. Implementation of the ECODOM specifications

‘La Trinité’ is the very first ECODOM project. Unfor-tunately, the design stage was already advanced when wehad our first contact with the architect in 1997. For thatreason, some of the ECODOM specifications had not beentaken into account, especially the solar protection of theFrench windows on the eastern and western sides which wasunder-sized compared to the ECODOM specifications (seeFig. 3). Moreover, this under-sizing is pointed out by themeasurements (seeSection 6, Fig. 11). The dwellings arecross-ventilated dwellings with walls and roof terraces madeof heavy concrete. The colour of external walls is orangeand red, and the colour of the roof is dark green.

The ECODOM specifications for ‘La Trinité’ are:For the solar protection of the dark-coloured terrace

roof—i.e. 0.2 of reflectance, it was recommended to put an10 cm insulation instead of the 6 cm one originally plannedas indicated in[1].

No insulation was planned for the solar protection ofmedium coloured walls. Even if the specifications requiredit, a dispensation was granted over this because of the ad-vanced stage of the project.

Concerning the solar protection of windows, horizontaloverhangs were planned for the main facade facing East(bedrooms) and West (living room). Nevertheless, the geo-metric values of the overhangs are undersized compared tothe ECODOM values. The ratio overhang/window is 0.4 in-stead of 0.8 and 1, respectively for East and West sides (asindicated inTable 2). A dispensation was granted over thispoint as well.

The natural ventilation of the dwellings complied withthe ECODOM values. The porosity of the openings wasmore than 25% of the mean of the facade surface, whichguarantees a good air exchange rate of each dwelling. Forsome dwellings, it was simply required to change the num-ber of sliding parts of the French windows (3 instead of 2)to increase the porosity (seeFig. 3). The internal porosity isrespected thanks to internal louvers located above the bed-room doors that increase the surface in indoor openings upto 2 m2 (seeFig. 4).

As for the ‘La Découverte’ project, no modification wasmade regarding site location because gardens around thebuildings had already been planned. In addition, the projectalready had an optimal orientation as the main facades facenorth and south, with limited west and east wall exposureto morning and afternoon low incident sunrays. Thus theonly facade to be upgraded was the northern one by addingbalconies, as the southern one is seldom exposed: for thesouthern hemisphere latitude of Reunion Island, the middaysun only shines on southern facades in December.

Fig. 4. Internal louvers above bedroom doors.

358 F. Garde et al. / Energy and Buildings 36 (2004) 353–366

Fig. 5. The ‘La Decouverte’ project—northern facade.

The architect had already planned 5 cm insulation for redand blue roofs. According toTable 1, this is not enough.Therefore, he was asked to increase the insulation thick-ness up to 8 cm (the value for medium colours) and to usepolystyrene as insulation rather than mineral wool. Mineralwool is fairly cheap but not very well adapted to tropicalclimates: it loses its thermal properties when it absorbs am-bient humidity.

Turning to the solar protection of walls, the colour islight and materials used are hollow concrete blocks. In thereference document[1], it is specified to add 1 cm inter-nal polystyrene type insulation on East and West sides. Thesouthern facade was not insulated because overhangs pro-tect it. In addition, as the initial project did not providesolar protection for bedroom windows, it was specified toadd overhangs on exposed windows of the northern facade.However, the solar protection of the east windows of bed-rooms remained insufficient because of a poor roof overhang(seeFigs. 5 and 8). Moreover, there is no solar protection ofthe southern windows. A dispensation was made concerningthese points.

Concerning natural ventilation, ECODOM standards re-quire a 2 m2 porosity for bedrooms whereas the initial projectonly had 1.44 m2. This is why glazed doors have replacedall the original windows in order to increase the porosityto 25% of the mean of the surface area of the facades. Toreach the same porosity indoors, louvers were added abovethe bedroom doors. Balconies were also added all along thefacade for solar protection of windows and of walls (see

Table 3Synthesis of the experimental and sociological survey schedule

Location Experimental survey Sociological survey

La Trinite Non occupied (austral summer January to May 1998) 15 interviews (April 1999)Occupied (from March to April 1998)

La Decouverte Non occupied only austral summer (from January to April 1999) 15 interviews (March 2000)

Fig. 5). The position of the bedroom doors was modifiedcompared to the initial project in order to maximise the air-flow transfers inside the dwelling, so that the angle of thedoors and the walls was 45◦ (seeFig. 8—T4 dwelling).

5. The experimental and sociological survey

5.1. Choice of dwellings and main objectives

As mentioned before, two instrumentation campaignswere carried out for the first ECODOM projects imple-mented in Reunion. These instrumentation campaigns wereperformed in both empty and inhabited dwellings (only atthe ‘La Trinité’ site). The campaigns took place near theend of the construction;Table 3shows a synthesis of theschedule of the experimental and sociological survey.

The objective of the non-occupied campaign was to testthe performance of the buildings without the tenants. In thisway, it was possible to create various measurement scenar-ios and modify the configuration of the apartments withoutdisturbing the tenants. Four dwellings were instrumentedwith the agreement of the housing companies, in both cases‘La Trinité’ and ‘La Découverte’. The campaigns took placeduring two hot seasons—i.e. in 1998 for ‘La Trinité’ andin 1999 for ‘La Découverte’, in order to assess the thermalperformance of flats during extreme climatic spells. The hotand wet season in Reunion occurs from the middle of De-cember to the middle of April. Dwellings were selected inorder to allow for comparison and studies of various param-eters such as roof and wall thermal insulation, and naturalventilation. Thus a flat under a roof next to the gable and thesame flat at a lower level were selected to test the influenceof the roof and of the solar protection of the glazings.

Figs. 6 and 7, for example, show the two dwellings for thesite ‘La Découverte’, respectively named T4R1 and T4RdCfor the dwelling under the roof and the one just below. Thetwo other instrumented flats were identical in orientationand in level and floor. Also at La Découverte, the dwellingsconcerned are those labelled C1T3 and C2T3, shown inFigs. 6–8. This configuration allowed to test different sce-narios such as the effect of natural ventilation with one flatused as a control and the second in which we modified itsconfiguration.

As for the occupied period, 6 flats were instrumentedduring two months for ‘La Trinité’ only. The main objectiveswere to test the dwellings in a real situation and to correlate

F. Garde et al. / Energy and Buildings 36 (2004) 353–366 359

Fig. 6. Location of the instrumented dwellings on the overall building plan ‘La Decouverte’.

Fig. 7. Location of the instrumented dwellings: the southern facades.

the thermal sensation of the tenants with the results of thesociological survey.

5.2. Data acquisition

A portable data-logger was set up close to the buildingsto record all the climatic data—i.e. external temperature andhumidity, direct and diffuse solar radiation, sky temperature,wind speed and wind direction. Data was collected every10 s and the average was calculated every hour.

During non-occupied periods, the sensors used werethermocouples for measurement of indoor air temperatureshielded against solar radiation by aluminium cylinders,globe temperature and surface temperatures on walls androofs; thermo-hygrometers for the relative humidity, andaccurate anemometers for indoor air speed. Data was col-lected every minute and the average was calculated everyhour. Sensors were located in the middle of each room. Themethodology used for the data acquisition and location ofthe sensors is the same than the one used for validationpurposes[14–16].

During occupied periods, the sensors used resemblesmall “white boxes” with internal memory and were wall-mounted to avoid inconvenience. They give temperatureand hygrometry measurements every 30 min over severalmonths.

The recorded data allow not only to check the perfor-mance of the instrumented dwellings but also serve as adatabase for the experimental validation of our buildingthermal-ventilation software. It is always useful to haveaccess to measurements taken at a real scale site. Thanksto these data, the modelling of diffuse radiation in the

360 F. Garde et al. / Energy and Buildings 36 (2004) 353–366

Fig. 8. Plan view of the instrumented dwellings (left side: C1T3 and C2T3, right side T4RdC and T4R1).

building simulation code[10] has been improved by tak-ing into account the shading devices in these calculations[17].

5.3. The sociological survey

The measurement results and analyses have been supple-mented for both projects by a sociological survey in inhab-ited dwellings in order to understand how thermal comfort,acoustic comfort (in tropical climates, it might be difficultto obtain thermal comfort and acoustic comfort at the sametime), visual comfort and dwelling environment are experi-enced by the tenants.

The interviews were conducted one year after the apart-ments were first occupied (the sociological survey was per-formed for the project ‘La Trinité’ in April 1999 and for ‘LaDécouverte’ in March 2000). The tenants had therefore hadthe time to get used to their new dwellings and to weighup the pros and cons of their homes. A professional surveywas carried out on a sample of 30 people, in conjunctionwith the University of Reunion Island, aiming to cover thefollowing points:

• reasons for choosing the apartment;• feeling or otherwise of well-being;• knowledge of the ECODOM experiment and its contents;• a systematic assessment of the apartment, with particular

regard to thermal comfort; and• points of dissatisfaction.

The results of the questionnaires, together with their anal-ysis and the experimental results are presented in the nextsection.

6. Results and discussion

The following sub-sections summarise the main conclu-sions of Refs.[11–13].

6.1. Solar protection of the roof

To demonstrate the influence of the roof, two identicalapartments were instrumented. They are strictly the sameas the oneFig. 8 for ‘La Découverte’ but with one underthe roof and the other one at a lower level. In this case,when the two apartments are closed, the gap between theindoor resultant temperatures is only due to the energeticcontribution of the roof.

Figs. 9 and 10show respectively the evolution of theindoor resultant temperatures in two apartments of ‘LaDécouverte’ and of ‘La Trinité’. A consistent 1–1.5◦Coffset is observed between the two apartments during thewhole period of measurement. The difference in the evolu-tion of the discrepancy between the two figures comes fromthe inertia of the roof. In the case of ‘La Découverte’, theroof is made of corrugated iron with insulation. The inertiais light. In the case of ‘La Trinité’, the roof is a terraceroof with an external insulation. Thus, the inertia inside thedwelling is high. That is why a constant gap is observed inFig. 10while the gap fluctuates between a maximum valueof more than 1◦C when the solar radiation is maximum andzero during the night.

The conclusion of the study is that a flat under the roofwith good insulation corresponding to ECODOM alwayshas a resultant temperature above 1–1.5◦C higher than theflats which are not under the roof. Other measurements that

F. Garde et al. / Energy and Buildings 36 (2004) 353–366 361

Fig. 9. La Decouverte—comparison of the resultant temperature in two dwellings (one under roof and the other one at intermediate level). Note that thesolar radiation is important between the 1/5 and the 4/5.

are not showed here pointed out that differences of morethan 3◦C have been observed between a dwelling with awell insulated roof and another one with no insulation[14].This type of measurements confirms the major importanceof good insulation of the roof in tropical climate and con-clude that no changes can be made to reduce the minimuminsulation required by ECODOM.

6.2. Solar protection of windows

Concerning ‘La Découverte’, the poor solar protection ofthe eastern and southern windows was highlighted by themeasurements and particularly byFig. 11. This figure showsthe evolution of the globe temperature discrepancies betweenthe front bedroom and the back bedroom of the T4R1 andT4 RdC dwellings during the month of April 1999. The sunremains on the north side all day long. The analysis showsthat the overheating comes from the front bedroom whichis facing north with more than 2.5◦C at peak time.

On the same way, the bad design of the solar protectionof windows of ‘La Trinité’ induced a measured overheatingof more than 3◦C [11].

It is therefore essential that the recommendations relatingto the openings and windows in the reference document befollowed, for all orientations.

Fig. 10. La Trinite—comparison of the resultant temperature in two dwellings (one under roof and the other one at intermediate level).

6.3. Solar protection of walls

The impact of solar protection of walls is of less impor-tance than the solar protection of windows. Therefore, if aproject is too expensive, the preferred initial source of econ-omy is the wall protection.

Nevertheless, the experimental feed-back in[11] for ‘LaTrinité’ highlighted that an overheating of 2◦C was observedin rooms with coloured concrete walls exposed to the solarradiation compared to rooms with no walls exposed. Nooverheating is observed for ‘La Découverte’ thanks to thegood overall thermal protection of walls (insulation of theeastern and western gables and balconies on the northernfacades).

The most important surfaces to treat therefore are thosemost exposed to the sun, that is to the north and west, and toa lesser extent, the eastern and southern surfaces. The othersurfaces could be granted dispensations, or else given lessonerous requirements.

6.4. Natural ventilation

Cross ventilation is of prime importance in humid trop-ical climates as well as the solar protection of the roof.As for these two projects, natural ventilation of dwellings

362 F. Garde et al. / Energy and Buildings 36 (2004) 353–366

Fig. 11. Evolution of the discrepancies between the globe temperatures of the front and back bedrooms (T4RdC and T4R1).

is efficient during the hot and humid season. The naturalventilation is better in the country site of ‘La Découverte’rather than the urban site of ‘La Trinité’., most of the indoortemperature and humidity readings pairs of ‘La Découverte’remain inside the comfort zone for a velocity of 1 m s−1

although they are on average 10% outside the comfort zonefor ‘La Trinité’, as shown byFig. 12.

The air velocity study inside the two dwellings of ‘LaDécouverte’—i.e. C1T3 and C2T3 both open at their max-imum openings, point out that C2T3 is better ventilated inthan C1T3 (seeFig. 12). This can be explained by the threeupwind buildings which stop the effects of trade winds onthe C1T3 building. Meanwhile, the indoor average velocityis about 1 m s−1 for C2T3 and 0.4 m s−1 for C1T3. Thanksto these values, the comfort conditions are reached 100% of

Fig. 12. Difference of comfort in the dwellings of ‘La Decouverte’ and ‘La Trinite’ during the hot and humid season. The temperature/humidity readingpairs remains inside the comfort zone for ‘La Decouverte’ whereas 10% of them are outside the comfort zone for ‘La Trinite’.

the time. These results were confirmed by the inhabitantsduring the sociological survey.

Fig. 13shows that for both dwellings, the indoor air tem-perature remains 1.5◦C below the outdoor temperature dur-ing the day and 1◦C above during the night. This is quite agood result because the main objective of the building de-sign in tropical climates is to avoid the overheating of theindoor temperature by keeping it at least below the outdoortemperature.

The effect of cross ventilation was demonstrated by clos-ing one of the opposite openings of the C1T3 and C2T3 liv-ing room respectively on the 27th and on the 29th of January.The comparison of the indoor air temperature shows a gapof more than 1.5◦C between the cross-ventilated dwellingand the other one.

F. Garde et al. / Energy and Buildings 36 (2004) 353–366 363

Fig. 13. Evolution of the indoor air velocity and the simultaneous difference of temperature between air inside and air outside—C1T3 and C2T3 dwellings.

6.5. Thermal comfort during occupied period (project ‘LaTrinité’)

The results presented here are obtained in an occupiedflat of the operation ‘La Trinité’. The period of instrumenta-tion lasted two months—i.e. in March and April 1998. Theinstrumented flat is on a split level. The first level includestwo bedrooms—i.e. bedroom 1 and bedroom 2, the kitchenand the living room. The upper level under the roof includesonly a third bedroom—i.e. bedroom 3. This flat failed theECODOM specifications in two respects: the first concernsthe insulation of the roof which was only 5 cm thick whereasECODOM requires 10 cm (the roof is grey in colour). Thesecond concerns the ventilation of the bedroom under theroof: the porosity is twice lower than the one required byECODOM (1 m2 instead of 2 m2).

Fig. 14 shows the set of temperature/humidity readingsover the period of instrumentation for the rooms on the firstlevel and for the room under the roof—i.e. bedroom 3.

It is observed that those rooms which are not under theroof and also naturally well ventilated have the same globalthermal and airflow behaviour. The comfort conditions—i.e.

the percentage of temperature and humidity readings pairsinside the comfort zone for an air velocity of 1 m s−1 arereached 90% of the time. Bedroom 3 is very hot. The comfortconditions are reached only 80% of the time. The distributionof the points on the psychometric chart is spread over the axisof the resultant temperatures. Although this room is naturallyventilated and is often used by the tenant, the indoor airtemperature rises above 33◦C although it remains below30.5◦C in the lower rooms. It is not in fact the life stylethat influences the thermal behaviour of the room, but rathera problem of insufficient heat insulation coupled with aninsufficient cross ventilation.

6.6. Results of the sociological survey

The main results of the sociological survey are hereinpresented. The results often confirm some of the conclusionsof the technical survey.

The inhabitants of ‘La Trinité’ suffer from the heat in thebedrooms under the roof. Concerning the solar protectionof windows, tenants of ‘La Trinité’ complain of the lack ofwindow solar protection. Tenants of ‘La Découverte’ find

364 F. Garde et al. / Energy and Buildings 36 (2004) 353–366

Fig. 14. Temperature/humidity reading pairs of the different rooms of an occupied flat—‘La Trinite’. The rooms located at the same level have the samethermal behaviour. The bedroom 3 located under the roof with insufficient insulation and natural ventilation is more uncomfortable.

F. Garde et al. / Energy and Buildings 36 (2004) 353–366 365

that the indoor air temperature is high in the bedrooms lo-cated on the eastern side.

The tenants appreciate the cross ventilation despite thedissatisfaction resulting from the bad acoustic design of theflats (only for the tenants of ‘La Trinité’ because of the highurban density). Indeed, one of the major problems of designin tropical climates is the conflict of thermal and acousticdesign. On the one hand, the dwelling must be kept open forthermal comfort, but on the other hand, acoustic comfort isachieved by closing the windows.

For the ‘La Découverte’ site which is in a windy spot anda semi-urban environment, the tenants are more than sat-isfied with the natural ventilation of the dwellings and donot complain about acoustic problems. People often missthe possibility of adjusting the airflow outside the dwellingwith mobile louvers (bedrooms were equipped with “classic”windows). People from ‘La Découverte’ appreciate the useof mobile louvers above the doors of the bedroom to im-prove the internal ventilation. By contrasts, people of ‘LaTrinité’ did not use the louvers (perception of bad acousticperformance and propagation of odours) mostly because noinformation was given about the way to use them.

When it comes to thermal comfort, even if, in both cases,it was hardly a determining factor when choosing an apart-ment, it is nonetheless important for tenant satisfaction.Overall, the level of satisfaction with thermal comfort to ac-ceptable for more than 90% of the interviewed people.

It seems however that the ECODOM guidelines weremore obvious to the tenants of ‘La Découverté’, and weretherefore better used, resulting in a higher level of appre-ciation. One can see that even where a successful thermalproject has been put it place, it is still essential to properlyinform the inhabitants. It is necessary to arrange a demon-stration of the advantages of ECODOM, and to give prac-tical advice (on opening the vertical louvers and the otheropenings to control the ventilation, etc.).

Finally, the experimental operation ECODOM is an ex-ercise greatly appreciated by the tenants. Provided the rec-ommendations are respected and given the investment thatthis requires, the unanimous approval of the tenants can bereported, despite some dissatisfaction resulted from the badacoustic design. This can be achieved by nothing more thanmaking them aware of these guidelines and giving themsome training.

7. Conclusion

A method of work has been set-up to allow the large scaleconstruction of ‘passive’ low-cost housing under tropicalclimates. The ECODOM reference document serves as auseful tool among the different partners involved in the build-ing process. 800 dwellings have been built to the presenttime.

Monitoring has been set up for the first ECODOMdwellings constructed in order to experimentally validate

the impact of passive cooling specifications on the comfortof tenants. This experimental work has allowed real scalefeedback.

The solar protection of the roof remains one of the mainpoints in the thermal design of buildings in a tropical cli-mate. Even with the insulation required, a consistent 1.5◦Coffset is observed between one flat under the roof and an-other one at an intermediate level. It is also necessary to up-hold the specifications concerning the minimum thicknessof insulation needed.

The solar protection of the opaque surfaces is withoutdoubt the most controversial point, in as much as its effecton the thermal comfort is less pronounced than the windowsor the roofing. In any case, a dispensation is often made overthis protection when the other forms of protection becometoo expensive.

The results concerning the solar protection of windowsshowed an overheating of 1–2◦C in the bedrooms of ‘LaDécouverte’ with bad solar protection. It is therefore essen-tial that the recommendations relating to the openings andwindows in the reference document be followed, for all ori-entations.

Natural ventilation proved itself to be effective in thetested flats except at night in the case of very light nightbreezes when an overheating effect of 1◦C was observedcompared to the outside temperature. Also, the ECODOMpercentage of openings of 25% is a required minimum. Theindoor temperature of the cross-ventilated dwellings remainsbelow the external temperature by 1◦C although it remains1◦C higher during the night. A cross-ventilated dwelling hasalways is internal temperature below 1.5◦C compared to anon-cross-ventilated dwelling with a mean indoor velocity of0.4 m s−1. The comfort conditions are reached 100% of thetime, what is the main objective of the ECODOM project.

Finally, the sociological study showed that the tenantswere globally satisfied with their flat as far as the thermalaspects are concerned (even if thermal comfort was not animportant issue for the tenants). However, dissatisfactionresulted from the bad acoustic design of the flats.

The sociological survey has been linked to the results ofthe technical survey while both studies will allow the build-ing physicists, the engineers and the architects to improvethe thermal and airflow design of buildings by proposingnew superior products but also taking into account people’sway of life.

The work in progress is aimed at softening the passiverules by proposing performance levels to be reached. Thiswill allow a measure of flexibility in a project, in case ofbudget constraints. Another aim is to adapt the method toteaching establishments (nursery schools, secondary schoolsand university buildings) because classrooms have similarrequirements to dwellings: both need good thermal design ofthe envelope and optimised natural ventilation. A first draftbased on the ECODOM feed back will be presented by theyear 2003 for the introduction of thermal regulations in theFrench overseas territories.

366 F. Garde et al. / Energy and Buildings 36 (2004) 353–366

References

[1] R. Célaire, O. Jourdan, Opération expérimentale ECODOM, Pre-scriptions techniques, Antilles et les bas de La Réunion, Documentde référence ECODOM, 1997, 32 pp.

[2] F. Garde, H. Boyer, J.C. Gatina, Elaboration of global quality stan-dards for natural and low energy cooling in French tropical islandbuildings, Building and Environment 34 (1999) 71–83.

[3] D. Holm, Building thermal analyses: what the industry needs: thearchitect’s perspective, Building and Environment 28 (4) (1993) 405–407.

[4] L. Lutton, Hyperex—a generic expert system to assist architectsin the design of routine building types, Building and Environment30 (2) (1995) 165–180.

[5] E. Shaviv, A. Yezioro, I.G. Capeluto, U.J. Peleg, Y.E. Kalay, Sim-ulation and knowledge-based computer-aided architectural design(CAAD) systems for passive and low energy architecture, Energyand Buildings 23 (3) (1996) 257–269.

[6] M.W. Ellis, E.H. Mathews, A new simplified thermal design tool forarchitects, Building and Environment 36 (2001) 1009–1021.

[7] F. Garde-Bentaleb, P. Depecker, F. Miranville, H. Boyer, Bringingscientific knowledge from research to the professional fields: thecase of the thermal and airflow design of buildings under tropicalclimates, Energy and Buildings 34 (2002) 511–521.

[8] E.H. Mathews, S.L. van Wyk, Energy efficiency of formal low-costhousing in South Africa’s Gauteng region, Energy and Buildings 24(1996) 117–123.

[9] A. Malama, S. Sharples, Thermal and economic implicationsof passive cooling strategies in low-cost housing in tropical

upland climates, Architectural Science Review 39 (1996) 95–105.

[10] H. Boyer, F. Garde, J.C. Gatina, J. Brau, A multi model approachof thermal building simulation for design and research approach,Energy and Building 28 (1998) 71–78.

[11] F. Garde, F. Pignolet-Tardan, T. Mara, H. Boyer, Evaluation tech-nique de l’opération expérimentale ECODOM, Rapport intermédi-aire: opération “La Trinité” Convention d’étude Secteur Pilote Inno-vation Outre Mer, DDE Réunion, 1998, 82 pp.

[12] F. Garde, L. Adelard, T. Mara, P. Poudroux, H. Boyer, Evaluationtechnique de l’opération expérimentale ECODOM, Rapport intermé-diaire: opération “La Découverte” Convention d’étude Secteur PiloteInnovation Outre Mer, DDE Réunion, 2000, 60 pp.

[13] F. Garde, L. Adelard, T. Mara, Evaluation technique et sociologiquede l’opération expérimentale ECODOM à l’Ile de La Réunion, Notede synthèse LGI/IPSOS O.I. Convention d’étude Secteur Pilote In-novation Outre Mer, DDE Réunion, 2000, 20 pp.

[14] L. Seauve, Evaluation technique et sociologique de l’opération ex-périmentale ECODOM en Guadeloupe, Synthèse, convention d’étude98000094, SPIOM DDE, Janvier 2000, 105 pp.

[15] K.J. Lomas, IEA task VIII Empirical validation: a critical appaisal,BRE, Reserch Report 3, 1991.

[16] T. Mara, F. Garde, H. Boyer, M. Mamode, Empirical validationof a solar test cell, Energy and Buildings 33 (6) (2001) 589–599.

[17] A.J.P. Lauret, T.A. Mara, H. Boyer, L. Adelard, F. Garde, A valida-tion methodology aid for improving a thermal building model: howto account for diffuse radiation in a tropical climate, Energy andBuildings 33 (7) (2001) 711–718.