Structural robust design based on hybrid probabilistic and non-probabilistic models

Yuan-sheng CHENG,Yu-xiang ZHONG,Guang-wu ZENG

DOI: https://doi.org/10.3969/j.issn.1007-4708.2005.04.022

2005-01-01

Abstract:The paper described the strategies for feasibility robustness and objective function robustness respectively subjected to both probabilistic and non-probabilistic uncertainties in detail and proposed hybrid probabilistic and non-probabilistic models for structural robust design. In the study, non-probabilistic variables were first taken as design parameters at their nominal values while the conventional statistical methods were employed to compute the means and standard deviations of constraint functions and the objective function. The effects on the means and standard deviations of constraint functions and the objective function due to variation of non-probabilistic variables were then taken into account by either an optimization procedure or a worst-case analysis with respect to the non-probabilistic uncertain variables. The mathematical models of conventional feasibility robustness and objective function robustness were consequently modified to achieve the robustness of constraint functions and the objective function under both probabilistic and non-probabilistic uncertainties. The proposed method was applied to the minimum weight design and the displacement robust design of a 10-bar truss structure subjected to bar cross-sectional areas with probabilistic properties and uncertain applied loads with non-probabilistic properties. The insensitive design schemes with respect to uncertain variables are obtained based on proposed method.