Integrated reliability and sizing optimization of a large composite structure

In this paper, we present the application of probabilistic design modeling and reliability-based design optimization (RBDO) methodology to the sizing optimization of a composite advanced submarine sail structure under parametric uncertainty. With the help of probabilistic sensitivity analysis, the influence of individual random variables on each structural failure mode is examined, and the critical modes are treated as probabilistic design constraints under consistent lower bounds on the corresponding reliability indices. Whereas the failure modes are evaluated for structural components in the solution of the RBDO problem, the overall system reliability is also evaluated as a post-optimization step. The results indicate that in comparison to a deterministic-optimum design, the structural mass of the probabilistic optimum design is slightly higher when consistent probabilistic constraints are imposed, and the overall structural stiffness is found to be more critical than individual component laminate ply thicknesses in meeting the specified design constraints. Moreover, the post-optimality analysis shows that the overall system failure probability of the probabilistic optimum design is more than 50% lower than that of the deterministic optimal design with less than 5% penalty in structural mass. (C) 2008 Elsevier Ltd. All rights reserved.