Finite element deletion and topology optimisation for building structural optimisation

Buildings would not exist without a structural design. Although mechanical, electrical, and computer facilities now often put a much higher burden on the financial costs of a building than the structural design -related to both engineering and construction costs-, a building structure is unique in the sense that structural issues cannot be permitted, as they are lifethreatening. Thus a building structure should primarily be safe, but nevertheless also economical. To help structural engineers with these goals, computer-assisted methods exist to determine the stress distribution in structural designs (e.g. the finite element method) and to optimise design components (e.g. topology optimisation). However, research on the optimisation of (complete) building structures is still relatively rare. In this paper, two methods for structural design optimisation have been compared for the application to complete building structural designs. This via a so-called research engine, in which spatial designs are transformed in structural designs, and vice versa, to investigate preliminary design processes. The two methods are compared for their effectiveness of optimisation, which shows that the method topology optimisation is more effective than the method element deletion, and if structural optimisation is used for exploring a solution space and evaluating the design process outcomes, this is an important conclusion. Besides, during topology optimisation a structural design remains stable, whereas element deletion may render the design unstable. However, when structural optimisation is used to study the primarily design process (e.g. via the research engine), the qualitative effects of both methods can be compared, and element deletion is computationally more efficient. Because even an unstable design will be usable in the research engine, for this case the method of element deletion is preferred