Abstract
This study involves simultaneous section shape and gauge optimization of vehicle bodyin-white (BIW) or similar structural sub-systems for multiple performance requirements such as global torsional stiffness, local free-free stiffness (FFK), natural frequency and others. The optimization formulation is setup using Optistruct as a mass minimization problem with performance constraints. Achieving global mass minimum is a challenge using a single optimization run that combines both section shape and gauge variables due to the limitations of gradient-based algorithms converging to a local-minima. Such an optimization problem is inherently non-convex and global optima cannot be guaranteed. Furthermore, the default optimization setting in Optistruct tend to converge to sub-optimal and non-intuitive design solutions. In order to improve convergence characteristics of simultaneous section shape and gauge optimization, the following techniques are explored in Optistruct version 14.0: (i) choice of optimization algorithm, (ii) parameter set for optimizer, (iii) numerical preconditioning of design variables. The best combination of optimizer settings improved mass reduction from 4% to 8% for BIW structure, and from 1% to 24% for IP sub-system structure. The results show that simultaneous section shape and gauge optimization in Optistruct is viable and has potential to identify lower mass configuration.