This action might not be possible to undo. Are you sure you want to continue?
Daylight in atrium buildings
P J Littlefair, MA, PhD, CEng, MCIBSE and M E Aizlewood, BSc, MSc
Daylight is an essential component of a visually attractive and energy efﬁcient atrium building. However, the atrium itself needs enough light for circulation purposes and to sustain plant growth, good
control of electric lighting is vital for energy efﬁciency, and areas surrounding the atrium can suffer from poor penetration of natural light unless care is taken in design. Solar shading also needs to be considered. This
paper explains the issues involved and gives guidance. It will be of interest to architects, engineers and their clients.
The courtyard has been used successfully for thousands of years to bring air and light to the heart of a building. The development of strong, cheap panes of glass allowed the courtyard to be glazed over and transformed into the modern atrium. Perhaps the first of these was constructed in the London Reform Club in 1841. Since then, many atria have been built, but it was not till the 1970s that a boom in atrium construction took place, following the success of such buildings as the Hyatt Regency Hotel in Atlanta[1,2]. Atria are built for three main reasons. q They fulfil an architectural role in unifying a large building by providing a hub or focus. In a similar way an atrium between buildings can give them a shared sense of identity. q They provide a view from adjoining spaces. If the sky or patches of sunlight are visible then there is also a link with the outside world. This is a major force in some commercial developments where lettable values may be considerably higher for rooms with a view. q They are a passive solar feature which can save energy. To displace significant lighting energy, the daylighting aspects of the atrium need careful design. Within the atrium itself enough light
must be provided to give it an attractive daylit appearance and to sustain plant growth. However the key problem is often the lighting of work areas surrounding the atrium. Unless care is taken, light from the atrium may not penetrate very far into these areas. Solar shading also needs
Figure 1 At the Learning Resources Centre at Anglia Polytechnic University, Chelmsford, translucent fabric bafﬂes and light shelves shade areas off the atrium
8). For horizontal glazing 0. As equation 1 indicates. measured as shown in Figure 2. then DFav should be 5% or more.4 This equation is valid for an atrium of any shape but is less accurate for very deep ones. it does not give information about the detailed distribution of light.9. If the atrium sides are open. Typical average values of R are 0.8. However. A low well index means that the atrium is wide compared with its height.7 is typical. The advantage of the average daylight factor is that it can give a single number which summarises the overall daylit appearance of the atrium. not including the θ θ θ Atrium Atrium Atrium Figure 2 θ is the angle of visible sky (in degrees) viewed in a vertical section from the centre of the atrium roof aperture . and Bell and Burt). For a square or circular atrium the well index is the height divided by the width.3 double) will reduce the average. For many atria R is fairly low. Daylight distribution Daylight in the atrium The average daylight factor The British Standard Code of Practice for daylighting uses the average daylight factor DFav as a way to judge a daylit space (see IP15/88.15 single. atrium sides are open. their reflectance is taken as 0. For clean clear single glass. vertical 0. If an atrium has both roof and side glazing. the average daylight factor due to each is calculated separately and the two numbers added together. sloping glazing 0. they should be included in A by adding in the area of imaginary atrium walls R is the average reflectance of these surfaces. Average daylight factor can be calculated using the following formula (IP15/88): WTgTmTf θ DFav = A(1 . Provision must be made for cleaning atrium glazing Tf is a factor to allow for light blocked by the atrium roof structure (see next section) θ is the angle of visible sky in degrees. The daylight factor is the ratio of internal to external horizontal unobstructed illuminance. which is also important. Even if the atrium walls are white (reflectance 0.65 Tm is a maintenance factor to allow for dirt on the glazing. 0. the base of the atrium receives less light and little light penetrates the adjoining lower floors. both for the atrium itself and for the spaces surrounding it. This can be expressed in terms of its Well Index: Well Index (WI) = height × (width + length) 2 × length × width where W is the area of the atrium roof aperture (m2) Tg is the diffuse visible transmittance of the glazing. Changes in roof profile can reduce the amount of light falling either on the sides of the atrium or reaching its base (Figure 3). If an atrium is to look well daylit. A high well index means that the atrium is tall and narrow.3 to 0. The roof profile also has an impact on the light distribution. DFav is the average over the atrium base. the roof and side glazing (reflectance 0. If the The well index relates the light-admitting area of the atrium (length × width) to the surface area of the atrium walls (height × (width + length)).R2) …(1) The daylight distribution is affected by the geometry of the atrium. Figure 4 shows how the transmittance Tf of typical atrium roof structures (Figures 5 and 6).8 can be used. for double glazing 0. and adjoining spaces have a better chance of receiving light through it. For a completely unobstructed horizontal roof θ would be taken as 180° A is the total area of the atrium surfaces: roof. a value of 0.2 to be considered (Figure 1). floor. walls and windows(m2). the roof structure and glazing bars can have a big impact on the amount of light entering. and they can also affect the distribution of light in the atrium.
it is possible to assess the visual impact of the system as well as measure the light admitted. The reflectances of the atrium surfaces are also important. Usually the roof structure will cut off more light to the top of the atrium sides than it does to the base of the atrium. With the eggcrate roof (left) most of the light goes straight down and little reaches the atrium sides. It also reduces solar glare on the top floors and increases daylight on the poorly lit lower floors. A small area of glazing on the top floors of the atrium increasing to a large area of Model studies In scale-model studies (IP23/93). Three techniques can be used: q Model studies q Computer simulation q Analytical formulae Figure 4 Graph of transmittance Tf of the roof structure against viewing angle in two atria: Birmingham School of Jewellery and Centre Court glazing. an overcast sky study in an artificial sky is easiest. and assessment made visually. the Figure 5 Birmingham School of Jewellery: roof detail Figure 6 Centre Court shopping centre: roof detail . Predicting the distribution of daylight levels in an atrium and its adjoining spaces can be difficult. The internally reflected component of the daylight has increasing impact in deep atria. For models used in daylight measurement. With sloping apertures (right) more light is directed to the sides and less to the atrium ﬂoor glazing on the bottom floors will help to reflect light to the base of a deep atrium. coupled with internal luminance measurements if required. To assess solar shading or glare protection.3 Figure 3 Changing the roof proﬁle affects daylight distribution. To check whether enough daylight reaches the interior and that there are no gloomy areas. varies with angle of view. the real sun or a heliodon can be used with a sundial.
8 3.8 2. which gives values for three types of atrium side: open. All surfaces modelled are divided into a number of elements. Tropical plants are used to 12 hours of light each day throughout 35q q s s v v v v v v v v v v v v v v v v Daylight factor (%) 30 25 20 15 10 5 0 0. correction for roof structure Tf = 0. For simple spaces with no significant areas of specular (mirror-like) reflection this is reasonably accurate and quick to compute. both in (scale) size and reflectance 6 Model window details such as atrium roof structure if possible 7 Choose the scale of the model with care.0 1. dark. where the atrium walls are extensively glazed or have a surface like brickwork. maintenance factor Tm = 0. specular or of any semi-specular nature. This technique requires far more computer power than a radiosity calculation. For the sky component at the base of the atrium. multiply the daylight factor by the fraction of the roof area glazed.6 1. where there is no wall between the atrium and adjoining spaces. the daylight factor at the mid-line of the vertical walls is given in Figure 7. 1:30 is a good compromise 8 Plan measurement positions in advance (not too many) Computer simulation Some computer programs can calculate the daylight factor at the base of an atrium. Comparison of these analytical predictions together with measured data from our model shows a good match for the sky component.) Figure 7 gives values for two atrium depths corresponding to well indexes of 1 (a cube-shaped atrium) and 3. almost as a ‘fitting out’ issue.6 2. For the reflected component the analytical prediction underestimates for lighter colours in a similar way to computer simulation.2 2.0 2. or taken at the very end of the design process. and light. Most use the ‘point to point’ or ‘radiosity’ algorithm. One such program is RADIANCE.4 2. For a square atrium. The decision to have plants is sometimes automatic. To model complex spaces with significant specular reflection ‘ray tracing’ programs have been developed. However.2 q q s s q q s s Well Index 3 s Open walls q Dark walls v Light walls Well Index 1 s Open walls q Dark walls v Light walls v v v q q s s q q s s q q s s q q s s q q s s s s q q q s v q s v q s v v q s v q s v v q s v q s v v q s v q s v v q s v q s v v v q s q s q s q s q s q s q s q s q s 0. which for a square atrium is the height below the top of the atrium divided by the atrium width. Plants in an atrium Plants soften hard edges and give visual pleasure. Vertical daylight factors are given in terms of well index depth. However. plants place big constraints on the lighting of an atrium and the decision about which plants to include should be made early on in the design process.7. this is a matter of some concern 40 v v because there is no clear explanation for it. Often models are too light coloured which can be a big source of error 4 Make the model light-tight 5 Model external obstructions accurately. Table 1 shows the recommended illuminances for various species of plants. It assumes that the roof is fully glazed with clear double glazing (transmittance Tg = 0. Although the relative error was not large. Analytical formulae Another way to estimate daylight in an atrium is to use an analytical formula. assumed to be perfectly diffuse.65.8 1.2 1.6 0. there was complete agreement between them. (For a partially glazed roof.4 following guidelines need to be followed (in order of importance): 1 Include all the building surfaces 2 Provide access inside the model for measurement and visualisation 3 Make reflectances correct.7).4 0. a very big or very small model is hard to test.0 Well Index depth Figure 7 Daylight factor on the sides of a square atrium . RADIANCE underestimated the reflected light in deep. high reflectance atria. where there is a substantial area of white or nearly white panelling on the walls. We carried out measurements in an atrium model and compared them with RADIANCE predictions. but in principle should be more accurate since surfaces may be diffuse.4 1.
lowering the average reflectance of atrium surfaces. for example. Figure 8 shows the number of hours for which some artificial lighting will be required in an atrium to provide a minimum of 1000 lux for a minimum of 12 hours each day. Plants also absorb much of the light falling on them. for double glazing 0.65. UK .8 can be used. for a European Fanpalm to grow the atrium must be designed with very high daylight levels in mind.5 the year. For less than 1000 lux or a seasonal variation where less than 1000 lux is acceptable in winter.0 metres tall Araucaria excelsa (Norfolk Island Pine) Eriobotyra japonica (Chinese Loquator. The data are for Kew. In practice. the atrium will require increasingly large amounts of supplementary lighting to sustain plant growth. some artificial lighting may be required. Table 1 Recommended illuminances for acclimatised plants Examples Illuminances (lux) Trees 1.0 As is the total area (m2) of the room surfaces: ceiling.R 2) s(1 s …(2) where AW is the net area of the glazing between the space and the atrium(m2) Ts is the diffuse visible transmittance of this glazing. Daylight in a space adjoining an atrium can be estimated by a twostage process. q If the daylight factor falls below 15%. Plants from other parts of the world experience seasonal variations: less than 12 hours at some times of year. First. including those to the atrium Rs is the average reflectance of the room. and they need care in periodical cleaning of the leaves and watering.5 for a light-coloured space Often an adjoining room would also be lit from the outside of the building. UK (Hunt). Even then.5–3. many will survive on less light and less variation but.6–1. while daylight factors above 25% will save relatively small amounts of artificial lighting. some artificial light will always be required if we try to provide 1000 lux for 12 hours/day. The average daylight factors from each set of glazing can simply be added together. for no glazing 1. There are two curves. more than 12 hours at others.8 metres tall Brassaia actinophylla (Schefﬂera) Chamaedorea erumpens (Bamboo Palm) Chamaerops humilis (European Fan Palm) Dieffenbachia amoena (Giant Dumb Cane) Ficus elastica ‘Decora’ (Rubber Plant) Phoenix roebelinii (Pygmy Date Palm) Pittosporum tobira (Mock Orange) Yucca elephantipes (Palm Lily) Table or desk plants Aechmea fasciata (Bromeliad) Aglaonema ‘Pseudobacteatum’ (Golden Aglaonema) Asparagus sprengeri (Asparagus Fern) Cissus rhombifolia (Grape Ivy) Citrus mitis (Calamondin) Dieffenbachia ‘Exotica’ (Dumb Cane) Hoya carnosa (Wax Plant) Philodendron oxycardium (Common Philodendron) 250–750 250–750 750–2000 750–2000 Above 2000 750–2000 750–2000 750–2000 750–2000 250–750 Above 2000 750–2000 750–2000 750–2000 Above 2000 Above 2000 Above 2000 750–2000 750–2000 750-–2000 Ficus benjamina ‘Exotica’ (Weeping Java Fig) 750–2000 Above 2000 Dracaena fragrans ‘Massangeana’ (Corn Plant) 250–750 Daylight in surrounding areas The contribution of the atrium Average daylight factor can be used to assess the amount of daylight in a room. walls and windows. Second. Japan Plum) Ficus Iyrata (Fiddleleaf Fig) Ficus retusa nitida (Indian Laurel) Ligustrum lucidum (Waxleaf) Floor plants 0. assess the vertical daylight factor DFv on the atrium wall (Figure 7). one is for just diffuse daylight. The figure shows that: q In the UK. As an example for plant growth in atria. calculate the contribution from the atrium to the average daylight factor DFsav in the adjoining space: 2A TsDFv DFsav = A W . The average daylight factor measures the amount of daylight in the adjoining Figure 8 Hours of artiﬁcial lighting required to sustain plants at 1000 lux for 12 hours/day at Kew. For clean clear single glass. floor. a value of 0. then some lighting energy can be saved. A typical value is 0. the other includes sunlight as well.
However the distribution of light is also very important. On the lower ﬂoors. The penetration of daylight into adjoining spaces can be improved by: q Increasing the head height of the apertures between the adjoining spaces and the atrium q Higher reflectances in both the atrium and adjoining spaces to increase the amount of reflected light q Innovative glazing systems (Littlefair. They are particularly important in atria because without proper control all the electric lighting could be on when there were very high levels of daylight (Figure 11). The obvious solution is to introduce blinds or baffles between the atrium and the adjoining space. both in the atrium itself and in adjoining spaces. photoelectric on/off control switches off the lighting when a particular daylight illuminance is reached at the control point. externally mounted shading devices are better at controlling incoming solar heat gain. The depth to which daylight penetrates into an adjoining space can be roughly estimated using the no sky line (Bell and Burt. particularly with splayed atrium walls where the top floors are set back. appropriate lighting controls are essential (Digest 272. In its simplest form. Provision must be made for cleaning and maintaining shading devices. 1991). The sun adds light warmth. Littlefair. particularly those on the top floor of the atrium building. and reduces daylight on dull days. and switches it on again when the daylight illuminance drops below this control value. Light from an atrium may not go very far into an adjoining space. from those which cannot receive any. Areas beyond the no sky line will generally look gloomy. Alternatively. and supplementary electric lighting will be required. sunlight entering an atrium can reach adjoining spaces. however. Normally. If sunlight is not required in the atrium there are various options: q Motorised blinds (effective but expensive) q Fixed louvres or baffles q Diffusing glazing or fabric (Figure 1). Because of the high daylight levels commonly achieved. IP6/96). internally mounted shading may stay clean longer. The no sky line divides those areas of the working plane which can receive direct skylight. though this Lighting controls For daylight to make a real contribution to energy efficiency it is not enough that it should just be admitted into the building. 1996) like light scoops and shelves (Figure 1) can distribute light from the atrium into adjoining spaces. Improving the penetration of daylight will give an asymmetrical distribution of light q Prismatic glazing (Littlefair. Often the occupants will need some form of shading to control it. This automatic switch-on may not be ideal No sky line No sky line No sky line Figure 9 The no sky line in an atrium building. and managers are often too busy to make minor adjustments to it. However. Figure 9 shows typical no sky lines in an atrium building. some form of daylight-linked lighting control involving either photoelectric switching or photoelectric dimming is often appropriate. These can be used in combination with occupancy sensing where the atrium is less frequently used. though this may itself be a source of secondary glare q A northlight-type sawtooth roof. it may be possible to modify the geometry of the atrium glazing to reduce sunlight penetration (see Figure 10). IP6/96 classifies atria as ‘managed’ or ‘unowned’ spaces: occupants do not in general expect to have control over the lighting. Excess heat can be vented out if the atrium roof has opening lights.6 space. variety and sparkle to the atrium space. IP5/87. They can also decrease the contrast between the very bright areas next to the atrium and the darker areas further in q Changing the roof profile (Figure 3) Solar shading A key decision in atrium design is whether sunlight is welcome. sunlight will be welcome in the atrium itself. Sometimes atria contain reception or other working areas. only a tiny area may have a direct view of the sky . though diffuse daylight penetration will also be reduced. A canopy enclosing these areas will control high daylight levels and make it easier to see computer screens if there are any (IP11/95). However. 1996) Note that tinted window film will reduce solar heat gain but is relatively poor at controlling solar glare. In general.
Sometimes a single controller dims large areas of a building. Carefully placed opaque elements can limit it (bottom) at the cost of reduced daylight penetration for energy saving. Separate control of different areas of the atrium is best. Photoelectric switching can be distracting to occupants when the daylight illuminance fluctuates around the control value. mounted either outdoors. and are less obtrusive. like on/off control. when this is switched on the main lighting may dim down. This type of sensor should be screened from direct light from the windows. and are difficult or impossible to use with some types of lamp. Open-loop control involves a sensor receiving daylight alone. Top-up controls save more energy than on/off controls. In closed-loop control the sensor is placed in the interior and needs to allow for the output of the lighting system which it controls. An indoor sensor is easier to maintain and relates more closely to the daylight actually received in the space. the extreme approach being a ‘solar reset’ control that only switches off photoelectrically at a specific time of day (11. For both types of photoelectric system. Even when daylight levels are very high. and the extra dimming facility may not be used much. here solar reset or a daylight-linked time delay can help to avoid switching off when the lamps will be required again soon after. are more difficult to install. Frequent switching may cause special problems with discharge lamps which have a finite restrike time. This is not a good idea since the dimming will be adjusted to provide enough light at the worst-lit point even if other areas are severely overlit. Commercial on/off controls incorporate a time delay or illuminance differential to try to lessen this problem. Out of hours control also needs planning. either using separate sensors or switching at different signal levels from the same open-loop sensor. but they cost more. Where the atrium is very well daylit. lighting may be left on along circulation routes because routes in the atrium are on the same circuit as those elsewhere. This will depend on the type of control algorithm used. Circulation lighting is another important area. more sophisticated systems only switch back on if an occupancy sensor detects movement. or for ‘safety’ reasons. a key decision is the location of the sensor. Closed-loop control is inappropriate where there is a lot of extra decorative lighting. or indoors facing out of the atrium roof glazing. Often decorative ‘sparkle’ lighting is left without automatic control. but it can look silly and superfluous when sunlight floods an atrium. But when the daylight illuminance is below this value this control will also increase or dim the artificial lighting so that the controlled illuminance is kept constant. photoelectric switching may keep the lighting off nearly all the time anyway. Dimming or top-up control will. sunlight dominated illuminances. It is wasteful to leave all the atrium .00 h for example). This should be unnecessary for properly designed photoelectric controls which default on if the sensors fail. switch the lighting off when the daylight illuminance is above a particular control value. One solution is to leave it on a photoelectric circuit set to switch off under high.7 Opaque elements Figure 11 Unnecessary lighting in a daylit atrium Figure 10 Uncontrolled sunlight can sometimes enter the top ﬂoor of an atrium building (top).
London. Isaac K A and Littlefair P J. Daylight in atria: a comparison of measurements. Atrium buildings development and design. particularly on the lower floors. pp 270–276. Improving reflectances. Technical enquiries to: BRE Advisory Service Garston. Kelvin Isaac also carried out the work on roof structure transmission. ASHRAE Conference. Low-level safety lighting of key circulation routes coupled with occupancy sensing for other areas can be appropriate here. London. EC1R 4QX. The new atrium. British Standard BS 8206:Part 2:1992. sent free to subscribers.  Maddock E.uk Tel: 0171 505 6622 Fax: 0171 505 6606 Full details of all recent issues of BRE publications are given in BRE News.co. 1992. Amsterdam. Perth.  British Standards Institution. San Diego. Proceedings CIBSE National Lighting Conference 1992. Code of practice for daylighting.8 lighting on when only one or two people are working somewhere in the building. Thanks go to Kelvin Isaac and Julian Butt who carried out measurements and analysis of daylight in the scale-model atrium. Watford.  Bednar M J. Proceedings Lux Europa. References  Saxon R.emap. Designing with innovative daylighting Acknowledgements This work was funded by the UK Department of the Environment. but solar shading may be needed to control sunlight coming into these areas. McGraw-Hill. 1986. A scale model study of daylighting in atrium buildings. Transport and the Regions as part of its Internal Environmental Issues (InEI) programme. WD2 7QG . Butt J D. The daylighting of atria: a critical review. 1986. For current prices please contact: Construction Research Communications Ltd. Daylighting in architecture: a European reference book. Availability of daylight Littlefair P J (1991). pp 459–472. June 1995. Herts. The Radiance lighting simulation and rendering system. E-mail: crc@construct. by choosing: q An appropriate atrium shape (not too deep or narrow) q High reflectances q Appropriate glazing areas (using the average daylight factor technique) q Plants that tolerate low light levels q Effective electric lighting controls Getting light to spaces surrounding the atrium. PO Box 202. BSl. Architectural Press. The effect of control algorithm and photosensor response on the performance of daylight-following lighting systems. Watford. Lighting controls and daylight use Information Papers IP5/87 Lighting controls: an essential element of energy efﬁciency IP15/88 Average daylight factor: a simple basis for daylight design IP23/93 Measuring daylight IP10/95 Daylighting design for display-screen equipment IP6/96 People and lighting controls Reports Bell J and Burt W (1995). Australia. 1983. Building Research Establishment Digest 272 Conclusions It is possible to have a well daylit atrium that does not use excessive lighting energy. © Copyright BRE 1998 ISBN 1 86081 194 9 Published by Construction Research Communications Ltd by permission of Building Research Establishment Ltd Applications to copy all or any part of this publication should be made to Construction Research Communications Ltd. The atrium landscape environment for plants. Fanchiotti A and Steemers K. Computer Graphics: Proceedings Annual Conference Series. Isaac K A and Littlefair P J. changing the roof profile and raising window head heights may help. McLean. James & James. Proceedings 2nd International Daylighting Conference.  Ward G J.  Aizlewood M E.  Aizlewood M E. 151 Rosebery Avenue London. The no sky line (Figure 9) can be used as a check. theory and simulation. Designing buildings for daylight Hunt D R G (1979). WD2 7JR Tel: 01923 664664 Fax: 01923 664098 Digests Good Building Guides Good Repair Guides Information Papers are available on subscription. Site layout planning for daylight and sunlight Littlefair P J (1996). November 1996. Cable Assoc. VA. is more difficult. 1993. Proceedings 41st Annual Convention of the IESANZ. 1997. London. Long Beach.  Baker N. 1994.  Aizlewood M E. New York. pp 30–45. Larson G W and Verderber R R.  Rubinstein F M.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.