Abstract
Suspended structures such as cable roofs and bridges are tensile spatial systems. The objective of this paper is to describe an automated robust design methodology that can be used to evaluate suspended structures. Numerical simulations combine dynamic relaxation for the nonlinear structural
analysis with a non-dominated sorting genetic algorithm (NSGA-II) for multicriteria optimization. The formulation used is general and adaptable to allow for handling of multiple objectives and constraints concurrently. Robust designs are obtained by including random uncertainties in the methodology.
Uncertainties are assigned to model inputs which yields outputs with associated uncertainties. Polynomial chaos expansion (PCE) is utilized to create reduced-order stochastic structural analysis models. These models allow statistical robust measures to be obtained with reasonable computational
time. A polyester-rope suspended footbridge case study is analyzed to show how the methodology handles both static and dynamic parameters. Test cases in which Young's Modulus and prestress are taken as random variables are examined. Two objectives (maximization of the lowest in-plane natural
frequency and minimization of rope volume) and two static constraints (maximum stress and maximum slope) are considered simultaneously. Best compromise solution sets, also named Pareto fronts, for the deterministic and robust designs are compared and found to be similar for all test cases
examined. Thus, for this case study, the deterministic solution is the most robust solution. The design methodology described in this paper can be used to evaluate other suspended systems subject to different constraints, objectives, uncertainties, etc. Consequently, this methodology has the
potential to be a powerful computational tool for designing robust suspended structures.