A Lagrange parameterization of lamination parameters which are used for optimization of variable stiffness laminates is presented. The advantages of the approach are: a) the design variables become independent of the ...
详细信息
A Lagrange parameterization of lamination parameters which are used for optimization of variable stiffness laminates is presented. The advantages of the approach are: a) the design variables become independent of the finite element mesh and, b) the smoothness of the solution is inherently guaranteed. Due to independency of design variables of the finite element mesh, the reduction in the number of design variables is drastic, once variable stiffness laminates usually demand a fine mesh. The lamination parameters formulation allows a more precise and concise approach to laminate design, removing difficulties related to fiber direction and stacking sequence, increasing the chances to find a global optimal solution. In this paper, the Lagrange parameterization is used for the maximization of the bucklingload of a variable stiffness composite plate.
In this paper a revised algorithm, derived by the Stud GA, to optimize the stacking sequence of layered composite structures is presented. The constrained combinatorial problem is formulated and hence solved by means ...
详细信息
In this paper a revised algorithm, derived by the Stud GA, to optimize the stacking sequence of layered composite structures is presented. The constrained combinatorial problem is formulated and hence solved by means of the algorithm presented herein. Unlike the Stud GA, the proposed algorithm uses a fitness-proportionate selection to find potential mates to the Stud and a self-adaptive extinction to avoid premature convergence. Elitism is also allowed. Standard benchmark problems have been used to fully assess the validity of the proposed algorithm. Results in terms of reliability and efficiency are compared against three GAs and a Permutation Search (PS) algorithm. It has been demonstrated, by means of a parametric study, that a low probability of extinction is sufficient to ensure fast convergence as well as reliable results. Finally the optimization of a more complex aeronautic structure is addressed aiming to prove the validity of the proposed algorithm in a two level optimization strategy. Results in terms of efficiency are compared against those of an in-house standard GA.
暂无评论