Optimization of Engineering Systems

M.H. Rademacher, R.C. Averill, and R.S. Sidhu
Inverse Problems Symposium 2009, East Lansing, MI USA

Engineers are forced to design products that are higher performing and more complex due to increasing market pressure. This results in the need for more complex simulation models that take into account nonlinear phenomena. In these models, gradients are often unavailable analytically, so optimization must be performed using methods that do not require gradients. In this study, several problems are presented where the system was optimized using a new hybrid optimization strategy that does not need gradients. This strategy was applied to three problems. The first was the design of a chemical process, where a minimum input of energy was desired. The second application was in the design of an automotive front suspension system where toe and camber curves needed to be matched. The third and final problem presented was the design of a rubber automotive engine mount, where a nonlinear stiffness curve was matched.
 Download (PDF, 406K)