APPROXIMATING COMPLEX SHAPES WITH INTELLIGENT STRUCTURES:EMBEDDED DESIGN INTELLIGENCE IN SYSTEMS FOR THE EARLY PHASES OF DESIGN

摘要

In contemporary architecture and structural engineering a trend toward the increased use of advanced geometryrnand computation can be observed. This poses new challenges for the structural engineer, the designed structuresrnand structural types, and the technology used to design, describe, model, calculate, engineer, communicate,rnproduce and assemble these structures.rnSystem concepts, like parametric associative design, can be used to capture the design intelligence in parameters,rnassociations, rules, objects (features) and programmatic components [1].rnAlthough these technologies have many features for the modelling the geometrical representation of a structure,rnbuildings consist of more than only geometry and generation. For free-form structures, surfaces should be meshed,rntessellated or populated to impose a grid-based structure consisting of these parametric elements. However, not allrnrestrictions of the structural design can be expressed in a single surface model of the intended design, since oftenrnthe structure imposes restrictions in the form of angles, lengths, stresses, etc. Therefore, surface population isrnoften not the appropriate technique for defining a structure.rnThis paper will present a computational method to model a geometrical and structural representation withrnembedded design intelligence in the form of simple rules, constraints and the generation process. With thernpresented method a designed shape can be automatically approximated in order to allow the architect and thernengineer to quickly apply and test a typical structure (for instance a space frame or a beam-column-structure) onrnthe designed shape in the early phases of the design, while taking into account restrictions from various aspects,rnsuch as the load bearing behaviour, production aspects, etc. Proper structural design and fine-tuning of the designrnstill has to take place by the designers, but this tool allows quick tests and indicative exploration during the firstrnconversations of the design.rnA prototype implementation has been created in a new parametric associative system, GenerativeComponents [1],rnby programming components in the C# programming language.