Automated design studies: Topology versus One-Step Evolutionary Structural Optimisation

Hofmeyer, H.; Davila Delgado, J. M.; Herm, Hofmeyer; Juan Manuel, Davila Delgado

Abstract

This paper presents the development of a virtual toolbox for the study of spatial-structural design processes. It will be able to transform a spatial design into a structural design. After structural optimisation, the structural design is interpreted as a spatial design. This spatial design is then modified to comply with the initial design requirements after which a new cycle starts. The transformation and optimisation processes within the toolbox can be altered, allowing automated design studies to be carried out. In this article, two processes for the structural optimisation are investigated to determine which is most suitable for specific conditions. These two processes are: (a) Topology Optimisation applied to complete structural systems for buildings; (b) Evolutionary Structural Optimisation for which only the first step is used. It can be concluded that: (a) although Topology Optimisation is formally more correct, One-Step Evolutionary Structural Optimisation will yield almost the same qualitative results, (b) quantitatively the methods cannot be compared exactly, however, it is likely that Topology Optimisation results in more efficient structures and (c) Topology Optimisation always leads to stable structures, whereas One-Step Evolutionary Structural Optimisation may yield a singular stiffness matrix, although this has no influence on the spatial design derived from the optimised structural design. It is intended to utilise the optimisation techniques in the virtual toolbox leading to design studies and to transcend additional transformation and optimisation processes in the virtual tool box via formal description. Fields of application are the academic study of spatial-structural design processes (i.e. design theory), the optimisation of all possible structural design types (i.e. design optimisation), and the generation of design instances (i.e. generative design). © 2013 Elsevier Ltd. All rights reserved.

• Journal Article Type: Article
• Acceptance Date: Jan 1, 2013
• Online Publication Date: Apr 28, 2013
• Publication Date: Oct 1, 2013
• Journal: Advanced Engineering Informatics
• Print ISSN: 1474-0346
• Publisher: Elsevier
• Peer Reviewed: Peer Reviewed
• Volume: 27
• Issue: 4
• Pages: 427-443
• DOI: https://doi.org/10.1016/j.aei.2013.03.003
• Public URL: https://uwe-repository.worktribe.com/output/927193
• Publisher URL: https://doi.org/10.1016/j.aei.2013.03.003