Formulation of Robustness in a CAE Design Model 2005-01-0813

As the computer efficiency and capability increase, so as the Computer Aided Engineering (CAE) technologies improve. Recently Robust Design or Reliability Based Design Optimization (RBDO) technologies have been utilized in all sorts of industries including automotive. The process generally involves identifying key input design variables and key performance output variables, determining a sampling plan for CAE simulations, building a response surface model (RSM), analyzing the results, and finding the optimized design that meets the reliability criteria. Yet little was addressed on the robustness of a CAE design model in the process. A systematic method of defining Robustness in a CAE design model was developed. How robust a CAE model is and how far away an optimized design is from the More Robust Region (MRR) are addressed in this paper. This method provides a clear measure of determining if a robust design within the performance variance is achievable in this CAE model and the size and location of the MRR in the design space. Numerical examples are used to illustrate the methodology.