Luminance pattern simulation: Design of an artificial BRE Average Sky

The BRE Average Sky has recently been proposed as a means of evaluating the energy effects of new forms of daylighting design. However, the complex geometry of some innovative design solutions means that predictive calculations can be extremely difficult and direct measurement under an artificial sky may be preferred. This paper describes the outcome of a feasibility study to design an artificial sky to simulate the BRE Average Sky. The problem is essentially one of producing a particular luminance distribution within a space, and is therefore relevant to other lighting design problems where a controlled pattern of luminance is required. First the difficulties of setting up a luminance pattern are briefly discussed. These include the problems of internal reflection, and parallax errors which occur when the luminance pattern is viewed from different points within the space. Two alternative designs for the artificial sky are outlined. The first uses a specular paraboloid which can reduce parallax error considerably and causes negligible inter-reflection problems. However, it is best suited to unilateral fenestration. The second uses the more traditional matt hemisphere, but with a time-saving computer technique to optimise the simulation of the required luminance pattern. This technique could be used in a wide variety of lighting design problems, whenever a highly controlled luminance pattern is required.