Over the past three decades, extensive research on atrium buildings has been undertaken and resulted in peer-reviewed publications, conference proceedings, research reports and handbooks. Most notably, Richard Saxon’s books “Atrium Buildings: Development and Design” (1983) and (1986), and “The Atrium Comes of Age” (1994) present a historical development of the modern atrium and include notable case studies. Furthermore, Michael Bednar in his noteworthy book “The New Atrium” (1986) illustrates the role of atria in key building types: hotels, shopping and leisure developments, office buildings, public buildings
(IEA), the American Society of Heating, Refrigerating and Air-conditioning Engineers, Inc. (ASHRAE) and the Building Research Establishment (BRE) that have resulted in guidelines on the design of atrium buildings.
The importance of daylight in an atrium’s environmental performance has led to several investigations of daylighting in atria and their adjoining spaces. Case studies, scale models, simple formulas and computer programmes have been used by various authors to provide design aids and simple guidance quantifying the effects of varying daylight linked atrium parameters.
In an atrium well, Daylight Factor (DF) comprises of the sky component (SC) and the internally/atrium reflected component (IRC/ARC) from the atrium’s enclosing surfaces (walls and floor). Therefore, wall reflectance has a direct and significant impact on the inter- reflectance occurring inside the light well and determines the distribution of light in the space, and the amount of light that reaches the adjoining spaces.
1.6.1
Atrium Surface Reflectances
Letherman and Wright (1998) highlighted the increasing impact that the internally reflected component of daylight has in deep atria. Mabb (2008) confirmed that light levels in the adjoining spaces are affected by the geometry, reflectivity and glazing of the atrium and its adjoining space. For design calculation purposes, the range of reflectances in an atrium is usually represented in terms of a single, area-weighted mean reflectance for estimating the average daylight factor (ADF) (Littlefair and Aizlewood, 1998), or the ARC of the DF (BRE Digest 310, 1986). Although this approach simplifies the calculation procedure, it does not help in identifying how the different distributions of reflectances around an atrium that are evident in real buildings actually produce different values of daylight factor (DF) or atrium reflected component (ARC). Most atria will consist of bands of different reflectances, both in value and in surface properties.
Letherman and Wright (1998) suggest that the influence of surface reflectances on daylight levels in atria and their adjoining spaces is complicated to model mathematically and most standard daylight calculation techniques do not transfer easily to atrium buildings. Consequently, many studies on this subject have been carried out using physical scale models or computer simuations. Oretskin (1982); Willbold-Lohr (1989); Navvab and Selkowitz (1984); Cartwright, (1986); Aschehoug, (1986); Liu et al., (1991); Baker et al. (1993); Iyer (1994); Boubekri (1995); Aizlewood et al. (1996); Aizlewood et al. (1997); Clarke et al. (1999); Matusiak et al. (1999); Fontoynont (1999a); Calcagni and Paroncini (2004); Mabb (2008); Lau and Duan (2008); and Du and Sharples (2009b) demonstrated that higher atrium wall reflectances improve daylight levels in an atrium building.
Aizlewood et al. (1996) carried out parametric physical model studies of the atrium surface characteristics, the atrium geometry and the geometry of the adjoining spaces. An approximate analytical expression for the atrium reflected component, ARC, was also developed. Comparison between predicted and measured ARC values suggested that the analytical expression had the correct general form, but that it underestimated ARC values for high reflectance surfaces. In a second paper Aizlewood et al. (1997) performed a similar study but now also compared their data with predictions from the computer program, RADIANCE. Sharples and Lash (2007) in their review paper observed that “Despite the simplicity of their models Aizlewood et al., (1997) failed to correlate measured ARC values with calculated values, demonstrating the complex and as yet poorly understood behaviour of reflected flux, particularly when highly reflective surfaces are used”.
Fontoynont (1999a) confirmed that indoor finishes and glazing materials contribute to illuminance levels in buildings and in particular they can be major contributors to the daylight availability in areas that are further away from the apertures. For atria of several different well indices, Calcagni and Paroncini (2004) evidence that whilst increase in the wall
the above studies is that the effect on the atrium reflected component (ARC) of varying the distribution of the reflectances around the atrium well was not investigated. Reflectance patterns are altered by introducing bands of openings, this inevitably produces changes in the area-weighted reflectances of the atrium surfaces. Most atria will consist of bands of different reflectances both in value and in surface properties. This is due to the fact that atrium facades comprise of a sequence of horizontal bands of glazed openings and opaque surfaces that correspond to the several floors of the adjoining spaces they envelope.
This study investigates how the different reflectance distribution patterns, for the same overall area-weighted reflectance value, affects DF and ARC in a simple four sided, top lit atrium model using physical scale models. The experiments are repeated using the computer simulation program, RADIANCE, to justify its use for the subsequent experimental work undertaken in this thesis.
Well Index (WI) is an indicator of the geometrical proportions of an atrium space, where a higher well index means the atrium space is deep and narrow. Conversely, a low well index indicates that the atrium space is shallow and wide. Letherman and Wright (1998) state that as the WI decreases, the ARC potentially increases due to the increase in the relative size of the atrium walls with respect to the atrium floor. However, the view factor between the atrium’s walls and sky vault is small resulting in a lower wall luminance. As the WI becomes very low however, the ARC would be expected to decrease, due simply to the fact that the opportunity for inter-reflectance is reduced significantly. However, it is vital to understand how the atrium geometry influences the ARC and therefore DFs in an atrium building and to establish the range of well indices over which reflectance distributions can affect the daylight levels. Willbold-Lohr (1989); Baker et al. (1993); Boubekri (1995); Boubekri and Anninos (1996); Aizlewood et al. (1996); Aizlewood et al. (1997); CIBSE (1999); Calcagni and Paroncini (2004); Mabb (2008) and Du and Sharples (2009b) have examined the influence of both atrium geometry and atrium enclosing surface reflectances, on the daylighting conditions in atrium buildings. While Oretskin, 1982; Willbold-Lohr, 1989; and Baker et al. 1993 show that the wells of square plans receive better illumination than rectangular/linear
plans at a given level, Liu et al. (1991), Matusiak et al. (1999), Calcagni and Paroncini (2004), Lau and Duan (2008) and Du and Sharples (2009a) demonstrate that whilst keeping height the same, increasing the length of the atrium increases the light-admitting area (or reduces WI) and consequently the DFs. Therefore, there is a lack of agreement in the findings of the previous studies in relation to the atrium well indices and geometries in which the DFs are affected due to the atrium wall surfaces. Additionally, these studies do not examine the effects of varying reflectance distributions on DFs in atria of different well indices.
1.6.2
Fenestration
Several authors (Willbold-Lohr, 1989; Cole, 1990; Aschehoug, 1992; Szerman, 1992, Iyer, 1994; Boubekri 1995; Matusiak et al., 1999) suggest that the proportion of window area feeding light into the adjoining spaces should vary between the floors of the atrium. Since most daylight is available at the top of the atrium, adjoining spaces need the smallest windows to achieve desired daylight levels. A progressive increase in the amount of openings from upper to the lower floors can lead to higher DFs available at the bottom of the atrium. Willbold-Lohr (1989) studied different facade apertures in square shaped atria and demonstrated that at the base of the atrium, in comparison with white facades, facade aperture with 50% window openings reduced the ARC by half, and with 100% glazing the ARC reduced to third and was mostly dependant on the skylight. While Cole (1990) concluded that variable openings in atrium facade with 100% opening on the first floor, 80% on the second, 60% on the third, 40% on the fourth and 20% on the top floor was most effective in terms of bringing daylight on the lower floors adjoining a square atrium in comparison with a 100% glazed and a 50% glazed atrium facade. Aschehoug’s (1992) recommended optimum glazing ratios for a glazed street of infinite length of 100% on the first floor, 70% on the second floor, 60% on the third floor and 50% glazing on the fourth floor
Undertaking physical model studies for a linear atrium, Matusiak et al. (1999) evidence that varying glazing area or glazing type results in a small but important increase in daylight on the atrium floor and improves the balance of lighting in the adjoining spaces. Equations were established to estimate the DFs in the adjoining spaces. Calcagni and Paroncini (2004) provided a relationship between the main architectural components of an atrium (geometry, material properties, fenestration system, atrium roof) and the daylight conditions in the adjoining space and on the atrium floor.
Whilst there is general consensus in terms of the positive influence of progressive increases in openings from the top to the atrium floor on the daylighting conditions in the adjoining spaces, an area of continued uncertainty is whether a particular incremental approach to fenestration from the roof to the floor of an atrium’s facade might be advantageous in terms of improved daylighting in an atrium building.
Therefore, in summary, several studies have identified atrium surface reflectances as one of the key factors impacting on the daylight performance of atrium buildings; it is this parameter that the thesis concentrates on. It also seeks to gain a better insight on the effects of different fenestration distributions on the daylight conditions in an atrium and its adjoining spaces for an open, four sided, top lit, square atrium building under overcast sky conditions through a series of related parametric studies. The reflectances and well indices used in this study are representative of the built atria as identified by Liu et al.’s (1991) survey. The four- sided square atrium is chosen as it provides the least opportunity in terms of admitting daylight in comparison with a two-sided or a three–sided atrium, and therefore the study examines the worst case scenario and the possible improvements that can be achieved in it.