An optimization framework for composite structure design with bounded uncertainties

This paper proposes a new optimization approach for stacking sequence design of laminated composite structures with uncertain-but-bounded parameters. The problem is first formulated based on the definition of an initial stacking sequence design, involving both discrete 0/1 variables and continuous variables. Discrete 0/1 variables are used to represent the existence of each ply in the initial stacking sequence, and continuous variables are associated with layer thicknesses. A two-level approximate method is developed to efficiently address this problem with mixed variables, where genetic algorithm is employed to handle discrete variables, and the dual method is adopted to optimize continuous variables. When considering bounded uncertainties, the uncertainty analysis is realized based on the convex model. By integrating the convex modeling technique with the two-level approximate method, an optimization framework is developed for laminated composite structure design considering bounded uncertainties. Numerical examples are presented to demonstrate the validity and practicality of the newly established framework. The optimization results show that the developed optimization framework is efficient in dealing with optimal design of composite structures with bounded uncertainties.