Improving direct solar radiation in complex building envelopes with a computational genetic algorithm

This thesis addresses the issue of relating the shape of building envelopes to solar positioning and direct solar radiation. Existing strategies for maximizing solar gains assume that most building surfaces are flat. By examining contemporary architecture, and the computer modeling tools utilized to produce it, a trend towards complex building forms beyond the simply planar emerges. The presented research outlines a method of applying computer genetic algorithms for improving direct solar radiation intake of complex non-planar surfaces. Experiments performing specific searches for optimal forms with improved direct solar radiation performance establish a relationship between surface shape and solar position as a function of the intensity of direct radiation throughout the year. A modeling tool was developed that allows for the parametric manipulation of surfaces for better solar performance.; Using known techniques for solar positioning and for calculating clear sky direct radiation a method was developed to measure solar performance of Non-Uniform Rational BSpline Surfaces (NURBS). Using the developed surface rating method experiments were performed to determine the relationship between surface morphology, surface area, and the direct solar radiation received according to the movement of the sun. To this end a computational Genetic Algorithm (GA) is utilized as a search method to find better performing surfaces. Conclusions are drawn as to the geometric relationship between the developed surfaces and the positioning and direct radiation intensity of the sun at various hours, days, months, and seasons.; The result is a prototype plug-in tool with which designers can select parameters to improve the solar gain performance of surface designs. The availability of a plug-in tool that is embedded inside commonly utilized modeling software is of significance for designers who seek to develop early-stage shapes that are energy efficient. This method is geared at designers modeling shapes that help in the reduction of energy consumption. Conclusions are drawn as to the new ability the tool provides and new directions are suggested for continuing research and development.