The daylight factor approach, based the assumption of uniform sky luminance distribution, has served the purpose of daylighting design in buildings for a long time. With the availability of cheaper and abundant computing powers, it is now possible to make use of more sophisticated algorithms, which can make use of the graded distribution of luminance in the sky vault. The variation of luminance distribution in the sky vault is of particular interest in ascertaining the glare problem in the application of daylight in the tropics. The daylight data of horizontal and vertical illuminances and irradiances in Singapore were analyzed by Ullah [1,2]. The illuminance data were applied to predict the potential of energy conservation with the integration of daylight with the indoor artificial lights[3]. The study of daylight in Singapore continued with more detailed observations using the sky scanner, which divides typically the sky into 145 circular patches. The luminance value of every single patch is captured by a computer-controlled scanner fitted with a luminance meter; simultaneous solar irradiance is also recorded with a solarimeter. There is a built-in safety device in the scanner that closes the meter aperture, whenever the scanner directly points to the sun. Thus, the scanner data are meant for evaluating the diffuse luminance distribution in the sky vault. The scanner is being used to collect the data continuously from 1996 till now at National University of Singapore, except for some lapses for repairs and maintenance. In 1998, Lam et al4 made use of part of these data to evaluate six sky luminance prediction models as an effort to adopt a sky luminance models for Singapore. Kittler, Darula and Perez5 proposed a set of 15 sky models for the world-wide use. Sky patch luminance and horizontal diffuse illuminance models were formulated as functions of solar and sky patch geometry for the sky luminance and of solar geometry for the horizontal diffuse illuminance, both normalized as the ratios to the zenith luminance. While the basic morphology of the function was kept, various sky types were represented by assigning differing values to the constants used in the functions. These models were also evaluated by Tregenza6 with the data for Fukuoka (Japan), Garston (UK) and Sheffield (UK) and again with a limited amount of data for Singapore. This paper attempts to make a thorough evaluation of the 15 sky models proposed by Kittler etal5 by calculating the sky patch luminance and horizontal diffuse illuminane using their equations and comparing them against the sky scanned measured data in Singapore and to obtain the equations which fit the best of these data.
Read full abstract