Multi-objective optimization of a composite stiffened panel for hybrid design of stiffener layout and laminate stacking sequence

Abstract

This paper presents a two-level approximation method for multi-objective optimization of a composite stiffened panel. The purpose is to seek the minimum structural mass and maximum fundamental frequency subject to given displacement constraints and manufacturing limitations. The design variables are the stiffener layout, and laminate stacking sequences for stiffeners and the panel skin. By introducing the concept of ground structure in both stiffener layout and laminate stacking sequence, the design problem is formulated with mixed discrete and continuous variables. Two types of discrete variables represent the existence of each stiffener and the existence of each ply in the laminate, respectively, with continuous ones for ply thicknesses. Considering the objectives are of different dimensions, a weighted min-max objective function is defined and minimized. The problem is firstly made explicit with branched multipoint approximate functions. Genetic algorithm (GA) is then adopted to optimize two types of discrete variables, determining which stiffeners/layers are deleted or retained. For fitness calculation in GA, a second-level approximation is built to optimize continuous ply thicknesses of the necessary layers that are retained. By giving different initial designs of stiffener layout and laminate stacking sequences, reasonable optimization results, which are tradeoffs between the considered two objectives, are obtained as design options. From the number of required structural analysis, it shows that the proposed method has a good efficiency in seeking rational solutions, which are tradeoffs between conflicting objectives and also feasible designs satisfying all considered constraints.