A two-step optimization scheme for global optimization design of composite frame

Fiber-reinforced composite frame structure is considered to be an ideal material and structure for aerospace structure to realize ultra-lightweight and large-span demanding. Filament winding parameters (filament winding angle, winding thickness and winding layer order) have a significant effect on performance of the composite frame structure. However, when using the filament winding angle as design variables, the objective function feasible domains are non-convex and the traditional gradient-based optimization algorithm is not effective to get the global optimal solution. The present paper proposes a two-step optimization scheme for fiber-reinforced composites frame. When considering structural stiffness parameters (tension stiffness, bending stiffness, torsional stiffness) as the design variables, the optimization feasible domains are convex. Based on the derivation of equivalent stiffness of the composite beam with circular sections, the new method realizes the conversion of the design space from a non-convex domain to a convex domain. The proposed scheme can effectively solve the non-convex problem when the filament winding angles are considered as the design variables, which provides an effective method to obtain a global optimal solution in the optimization of composite frames. Numerical examples examine the computational efficiency of the proposed approach.