RESEARCH ARTICLES CURRENT SCIENCE, VOL. 119, NO. 3, 10 AUGUST 2020 473 *For correspondence. (e-mail: nabeel.k91@gmail.com) Identifying the design skies for Indian tropical climatic conditions Nabeel Ahmed Khan*, Paras Malik and Bishwajit Bhattacharjee Indian Institute of Technology Delhi, Delhi 110 016, India The current Indian design sky model is inadequate for daylighting design and calculation of annual daylight energy savings in buildings due to which designers are compelled to resort to other sky luminance prediction models like Perez and International Commission on Illumination (CIE) sky models. Perez and CIE sky models are derived from the sky scan data of mostly sub-tropical and temperate regions where the climatic conditions are different from the tropical climatic conditions. The present study attempts to check and compare the adequacy of Perez and CIE sky models for prediction of sky luminance in Indian tropical climatic conditions by comparing model-predicted values with the measured values from the past studies. Further, a framework has been proposed for the iden- tification of CIE design skies for passive window design in the locations where measured sky luminance data is not available. CIE design skies have been rec- ommended for the major climatic zones of the tropical climate of India using the proposed framework. Keywords: Design sky, passive window design, day- lighting, Indian tropical region. INDIAN Standard Code IS 2440 (ref. 1) recommends a design sky model for passive window design in India. Accordingly, the sky luminance at an altitude angle of θ is given as L v = L z cosecθ, (1) where L v is the luminance of the sky element at the altitude angle of θ and L z is the sky luminance at zenith. The formula is applicable for sky elements lying between 15° and 90° sky altitude with uniform brightness below 15° sky altitude. Uniform luminance is assumed along the azimuthal direction for a given sky altitude angle. The design sky given in the code has been derived based on the measurements taken at CBRI (Central Building Re- search Institute, Roorkee, India) at 15° solar altitude in 1975 (refs 1, 2). Basic hypothesis for above design sky is: 15° altitude represents the time corresponding to that at one hour after sunrise (7 a.m. or so) or one hour before sunset (5 p.m. or so), covering complete working hours. The daylight availability would be minimum during working hours at that time and represents the lower limit of sky luminance for design purpose. Passive window de- sign with a low solar altitude design sky ensures daylight availability through fenestration throughout the working hours and is, therefore, a reasonable assumption. It was further assumed that fenestration design for daylight would prefer window on north wall to exclude direct sun light so as to avoid glare and heat admission. Thus dif- fused sky on north-east or north-west quadrant of the sky hemisphere opposite to the sun’s position was considered in the morning and afternoon respectively. Hence the assumption of azimuthal uniformity was adopted. But assuming azimuthal uniformity may lead to an inefficient design, which is explained as follows. According to the measured sky luminance data of other tropical countries and established sky luminance prediction models 3,4 , the sky luminance varies as the distance from the sun changes (both in azimuth and alti- tude). A good passive window design requires precise prediction of sky luminance in every direction of the sky dome. Apart from this, Indian design sky does not take into consideration the change in sky conditions with the change in climatic zones. It is also inadequate for calcula- tion of annual energy savings due to daylight as it gives a single sky luminance distribution for a low solar altitude angle and leads to underestimation of the daylight avail- ability. The shortcomings of the current Indian design sky model as described above are azimuthal uniformity, climate invariant and inadequacy for calculation of annual energy savings due to daylight. These shortcom- ings compel the designers to employ other sky luminance prediction models for window design purposes and annual building energy load calculations. Perez 5 and International Commission on Illumination (CIE) sky models 6 are widely accepted models for the prediction of sky luminance distribution. Daylight simulation softwares used by architects like Gendaylit Sky Generator in Radi- ance 7 , Daysim 8 and Daylight Visualizer 9 use Perez and CIE sky models for prediction of sky luminance distri- bution. The coefficients used for the prediction of sky lumin- ance in Perez and CIE sky models are derived from the regression analysis of the measured data from countries mostly falling in subtropical and temperate regions. The climatic conditions of tropical regions differ considerably