SIMULATION BASED DESIGN OPTIMIZATION OF A CFRP FUSULAGE PANEL ACCORDING TO DRAPING PROCESS OF CARBON FIBER TEXTILES USING EVOLUTIONARY ALGORITHMS AND RESPONSE SURFACE METHODS

Design optimization plays a vital role in Aerospace Industry due to its challenging requirements that are not only based on mechanical aspects but also on manufacturing processes. Besides, latter cause imperfections and variations on the structure which could be estimated by advanced simulations that result in expensive computations. On the other hand, it has been an issue to obtain global optima of the objective function of structures such as CFRP stiffened fuselage panels whose mechanical system responses are generally evaluated by Finite Element calculations.

This paper proposes a global optimization process on an innovative CFRP fuselage panel which simultaneously covers stability criterion and draping simulation of fiber textiles that is considered as a manufacturing effect. Genetic algorithms were chosen as an optimization technique to reach global optima. During the optimization process, a modified Response Surface Method that is based on artificialneural networks (RBF-ANN)[7] is carried out in order to reduce the computational effort and to couple the simulation inputs and outputs in the optimization frame.

A parametric FE model generator is developed under periodic boundary conditions in order to obtain buckling modes and corresponding failure criterion during the parameter variation. The shear angles of the fibers arising due to draping process are mapped into FE model which is represented by RBF-ANN surrogate model in the optimization loop. This method presents an optimization  frame that covers the production deviations of draping process and its influence on structural design.