Abstract

The present work shows the latest static structural non-linear non-parametric sizing, shape and bead optimization developments allowing industrial engineers to take advantage of innovative design opportunities permitted by applying non-linear realistic simulation for optimization. A commercially available non-parametric structural optimization program (Tosca) which automatically modifies the finite element input deck (Abaqus) in each optimization iteration, drives the optimization workflow. The present structural optimization disciplines are sizing of the sheet structures where the shell thicknesses are design variables as well as shape and bead optimization where the nodal positions of the finite elements are the design variables allowing non-parametric shape and bead optimization, respectively. To the best of our knowledge, this is the first work that shows results for non-linear sizing, shape and bead optimization using adjoint sensitivities of static CAE models including simultaneously the three modeling non-linearities as large deformations, plasticity as material non-linearity and contact. An application of sizing optimization is for a bumper model consisting of shell elements where each of the shell thicknesses represents one design variable. The contact between the bumper and a rigid component is included. The constitutive material modeling is elastoplastic as well as large deformation is also included. The present objective is to minimize the mass of the structure subject to static energy absorption requirements. Additionally, the same non-linear sizing optimization concept will be shown for a static side impact of a full car model and non-linear collapse of a jacket structure. Frequently, the design aim for non-parametric shape and bead optimization is to minimize or to constrain the plastic deformation of misuse load cases. Here automotive applications are shown where the geometrical non-linearities are also included for realistic model the bending forces and membrane force being principle for yielding correct optimized designs for the plastic deformation. - - - - - -