Surrogate modeling in design optimization of stiffened composite shells

The development of an optimization design procedure for stiffened composite shells has emerged as an essential component in overcoming the challenge of deriving design guidelines with a reasonable amount of tested structures. The optimization procedure is based on the building of surrogate models employing the experimental design and response surface methodology. Surrogate models are built using continuous and discrete quantitative parameters and approximated with global and local nonparametric approximations. The numerical data obtained from finite element simulations of composite stiffened shells subjected to buckling and post-buckling is used for the building of surrogate models. Additionally a data set of natural experiments has been implemented into surrogate models showing high efficiency in the estimation of the post-buckling behavior of the stiffened composite shells (panels). The resulting design procedure provides an effective optimal design tool for the preliminary study of composite stiffened shells. (c) 2005 Elsevier Ltd. All rights reserved.