Statistical methods applied to optimize perforated façade design for daylight availability

Abstract

Perforated solar screens (PSS) are an important consideration in building façade design due to its contribution to sustainability through daylighting. PSS design requires the consideration of many potential design alternatives that involves a large number of simulations. This paper presents a methodology in which the orthogonal and listing methods are integrated to predict a set of optimum PSS design variables to enhance the Daylight Availability in office buildings located in Seville, Spain. An orthogonal array is selected to perform a transverse comparison of the simulation factors mean effects and to find their statistical significance. Then, a standard level is fixed and used for further detailed analysis of a greater number of factor levels, measuring their daylighting contributions. The main advantage of the integrated method is the reduction of the number of simulations from 720 to 32, so it could save time considerably and would help designers to make early-design-stage decisions. With the optimization, the actual daylit area increased by 29-57% and the over lit area reduced by 36-57%, relative to reference models with no PSS.