Design of a Museum Facade Bracing System for Changing Performance Requirements Using Multiobjective Optimization

Abstract

Facade bracing systems are applied all over the world in structural design to limit deflections and guarantee stability. Efficient distribution of bracing over a structure is an important concern for structural design professionals and is often based on intuition and previous experience. Meanwhile the limited amount of academic research on this topic often focuses on one aspect of the design, neglecting the practical design process itself. This research presents a topology optimization procedure for cable bracing of the hanging steel facade of a new museum in the United States. In this procedure the use of a multiobjective Genetic Algorithm allows for flexibility during design modifications and accounts for uncertainty of deflection constraint values. Since the allowed position of the cables is limited, and only certain standard cable sections can be used, the design variables are discrete, and the cost function is easily defined. The presented method achieves practical solutions to a series of cost minimizing problems, giving the designer a range of optimal bracing configurations which can be selected in response to the continuously changing structural and architectural requirements throughout the design process. The application of the method to the design of the museum facade, demonstrates the strength of the proposed approach. This research aims at stimulating discussion on optimization methods which are capable of taking the design process into account, and the possibility of using multiobjective optimization to deal with these practical design uncertainties. INTORDUCTION X-bracing systems are some of the most widely applied structural systems, particularly in steel construction. Schodek [1] and Taranath [2] describe various bracing systems for frame structures including X-braced, V-braced, K-braced and Chevron-braced (inverted V) systems, of which X-bracing is the most common. One of the main difficulties when designing braced systems is the choice of the location of the bracing components. This selection is often based on intuition and previous experience. While an approach based on experience is invaluable, it does not necessarily produce the best results. The cost associated with these bracing systems lies primarily in the connections. Reducing unnecessarily large numbers of bracings is thus of great interest to make a design cost