Robustness and reliability of composite structures: effects of different sources of uncertainty

Abstract

A study to evaluate the effects of different sources of uncertainty in the Reliability-based Robust Design (RBRDO) of composite laminate structures is performed. The goal is to understand how the set of Pareto-optimal solutions will change and the interaction between the design search and the reliability constraint. The RBRDO is executed by a newly proposed methodology exclusively based on Genetic Algorithms (GA), to guarantee higher levels of accuracy in the optimization procedure, avoiding local minima, common to gradient methods. Design optimization is considered as the bi-objective minimization problem of the weight (optimality) and the determinant of the variance–covariance matrix (robustness). Reliability assessment is made by a mathematical reformulation of the Performance Measure Approach, suitable for GA’s, as an inner-cycle of the design optimization. A numerical example of a fuselage-like composite laminate structure is presented. In the reliability assessment, the uncertainty of the system is considered only through the group of mechanical parameters. It is plausible that there exists an implicit functional relationship between feasibility robustness and the reliability constraint, on which the latter constrains the former, at least for the evaluated numerical example. Optimized weights vary between the same values. Tsai numbers and reliability indexes have similar distributions, for different sources of uncertainty. Only the thickness variables and the ply-angle seem to be affected by the structural feasibility robustness assessment. The distribution the Tsai numbers is affected by the reliability constraint, to respect the imposed reliability level.