Explorative structural design

Abstract

The thesis proposes a new way of thinking about structural design software. The current state of computational structural design in practice is assessed, and a review of relevant literature and existing software identifies both the strengths of existing approaches and areas in which contributions can be made. A new approach is proposed which combines the strengths of architectural modeling software with structural analysis software, and an original object-oriented framework for the development of next-generation structural design tools is presented. The thesis shows that the field of structural optimization, long maligned by engineers for its impracticalities for engineering practice, can be made relevant and beneficial in providing techniques to explore the design space in an optimally-directed way, leading to the discovery of unexpected and novel structural designs which are easier to build, use less material, and cost less than structures designed by conventional software. The software framework is extended to include these optimization components and to facilitate the future inclusion of new algorithms by users. A fully functional design environment is developed and presented as an implementation of the work of the thesis.