Coupled multiobjective optimization of crashworthiness and modal frequencies in structures using genetic algorithms.

摘要

This dissertation presents an approach to concurrent optimization of structures by using an island injection Genetic Algorithm for: (1) crash energy management, (2) modal vibration frequency response, (3) peak crush force, (4) weight reduction, and (5) manufacturability, while considering stochastic variability of design variables and loading conditions. Existing automated design technology has never simultaneously addressed all of these competing mechanisms. In addition, this dissertation presents the development of a geometrically nonlinear first-order zig-zag sublaminate theory and finite element model which accounts for moderately large displacements and moderate rotations using a Total Lagrangian formulation. The accuracy of the model is demonstrated by comparing its structural response predictions with results from previous experimental investigations and with numerical tests. Finally, the optimal design of sandwich panels was approached with a simple Genetic Algorithm and the developed geometrically nonlinear first-order zig-zag sublaminate finite element model.