Optimization of Computational Laminate Composites for the Prediction of Performance Metrics Using Finite Element Simulation

摘要

In order to address the increasing demands of strength, weight, reliability, and thermal management in aero-structure design, composite materials have been developed and implemented to replace conventional materials in space-access vehicles. These new advanced materials are complex, non-homogeneous, anisotropic, and highly sensitive to small perturbations of their parameters. Because of this, current metrics are computationally difficult and expensive to obtain. Due to the excessive computational cost associated with optimization, metamodels are often utilized as a resource to alleviate the computational cost associated with repeated evaluations of full-scale simulataions through lower-order representations of the design space. Using Response Surface Methodology, these metamodels can be constructed and optimized efficiciently with various software packages. It is shown in this work that by applying Response Surface Methodology to various model geometries, laminated composites can be optimized for various parameters with very low computational cost. The validation and accuracy of this approach is demonstrated on two small scale metamodeling problems and a much larger scale wing-box optimization model.