Defining the minimum density of a sky luminance grid based on scale model measurements without the sun

The aim of this paper is to present a discussion on the complexity behind the implementation of sky luminance distribution models to correctly predict the illuminance values in indoor premises. In 2003, the standard CIE S 011/2003 was published, describing 15 CIE sky types and providing necessary equations to model the luminance distribution over a sky vault. As a basis, a sky hemisphere is divided into at least 145 quasi-equal, solid angle patches. Today, because practically unlimited computational power is available, the hemisphere could be divided into tens of thousands of quasi-equal, solid angle patches. However, the question remains whether this is really necessary. In this paper, the results of indoor illuminance simulations using different numbers of elements are shown to describe the sky luminance distribution and the errors between simulation results and scale model measurements. The results are somewhat surprising because the density of the sky grid has very small influence on indoor illuminance calculations if the sun is not taken into account. •Complexity of implementing sky luminance distribution models discussed.•Indoor illuminance simulations performed with different numbers of elements.•Errors between simulation results and scale model measurements shown.•Density of sky grid has small influence on indoor illuminance without the sun.