In this paper different strategies to search for a robust design are presented and investigated with respect to their efficiency and applicability to time consuming numerical models. After starting with deterministic optimization we introduce different measures to define the robustness of a design. An iterative Robust Design Optimization (RDO) is proposed where deterministic optimization is combined with variance-based robustness analysis and final reliability proof. The iterative procedure is compared to coupled RDO approaches, where the robustness or reliability measures are calculated for each optimization design. For such a procedure often global approximation models are used in order to enable the application for more complex problems.
For full paper, view: Robust Design Optimization in Industrial Virtual Product Development – White Paper (PDF)