Optimization in a realistic structural engineering context: Redesign of the Market Hall in Ghent. (1st February 2021)

• Record Type: Journal Article
• Authors: Dillen, Wouter; Lombaert, Geert; Mertens, Ruben; Van Beurden, Hanne; Jaspaert, Dirk; Schevenels, Mattias
• Is Part Of: Engineering structures. Volume 228(2021)
• Publication Date: 2021-02-01
• Subjects: Structural optimization -- Size optimization -- Genetic algorithm -- Metaheuristic -- Eurocode design -- Steel design -- Material minimization; Structural engineering -- Periodicals; Structural analysis (Engineering) -- Periodicals; Construction, Technique de la -- Périodiques; Génie parasismique -- Périodiques; Pression du vent -- Périodiques; Earthquake engineering; Structural engineering; Wind-pressure; Periodicals; 624.105
• DOI: 10.1016/j.engstruct.2020.111473
• Languages: English
• ISSNs: 0141-0296

Abstract

Highlights: A real structure is modelled and optimized with feedback from the designers. The optimized design uses 15% less steel. Material efficiency reduces costs, environmental impacts and engineering efforts. Abstract: Structural optimization has been extensively studied for decades, motivated by expected savings in material costs and engineering efforts in the design of loadbearing structures. The potential of optimization methods in the construction sector is high, but practical applications remain scarce, which may be explained by the fact that most test problems in academic literature are not representative of reality. The aim of this article is threefold: (1) assessing possible weight savings in a realistic structural design context, (2) examining the feasibility of structural optimization accounting for all practical constraints, and (3) providing a realistic test case to developers of structural optimization algorithms. In order to simulate a real design context, the steel structure of the Market Hall in Ghent (Belgium) is optimized, taking into account all relevant Eurocode provisions and practical requirements. By maintaining the structural member groups of the original design, the dimensionality of the optimization problem is small enough to be solved by a Genetic Algorithm. Using this approach, a design is obtained that consumes 15% less material than the original design with no increase in structural complexity, implying equivalent savings in material costs and embodied environmental impact.