Non-convex layup optimization of the composite laminates with fast and robust FEA model and modified SPEA-II

Abstract

The biggest design challenge regarding composite laminates is the selection of a stacking sequence such that the cost and/or weigth of the laminate to be minimized subject to constraints like failure criteria. The layup optimization of the composite laminates involves a very complex multidimensional solution space which is usually non-exhaustively explored using heuristic computational methods like genetic/evolutionary algorithms. To ensure the convergence of the prefered method it is necessary to evaluate a lot of layup configurations during the optimization process which involves that the evaluation of a single configuration should be fast enough to keep the overall optimization time to an acceptable level. On the other hand the mechanical behaviour of the composite laminates is very complex and it is analyzed with expensive computational tools such as finite element analysis. In this paper we propose a computational efficient framework which combines a very fast and robust Matlab-based FEA model with a non-convex evolutionary algorithm (strength pareto evolutionary algorithm II). We speed up the convergence by modifying the standard algorithm to generate new potentially pareto-optimal individuals. Also, a detalied numerical example is presented in order to highligth the steep computational time improvement between our method and a convex genetic algorithm.