Get Real towards Performance-Driven Computational Geometry: Fid-Bau Portal

Get Real towards Performance-Driven Computational Geometry

Oxman, Neri

In historic design conventions geometry has traditionally promoted descriptive manifestations of form. Beyond the realm of geometry, the concept of performance which may inform such manifestations also carries important potential for design generation. This work explores the relation between geometry and performance from a computational-geometry perspective. It does so by revisiting certain analytical tools offered in most of today's 3-D modelers which support the evaluation of any generated surface geometry specifically curvature and draft angle analysis. It is demonstrated that these tools can be reconstructed with added functionality assigning 3-D geometrical features informed by structural and environmental performance respectively. In the examples illustrated surface thickness (as a function of structural performance) is assigned to curvature values, and transparency (as a function of light performance) is assigned to light analysis values. In a broader scope this work promotes a methodology of performance-informed form generation by means of computational geometry. Vector and tensor math was exploited to reconstruct existing analytical tools adapted to function as design generators.