Abstract
The windshield de-icing is of vital importance to automobile manufacturers from a safety, as well as a comfort perspective. At Daimler, de-icing simulations have matured into a consistent performance evaluation method as described by Kurikesu et al. [1]. During the early stages of vehicle development, certain inputs supplied to the de-icing simulation may be approximate. Hence a sensitivity study of the predicted output is essential to be able to report the simulation results with increased confidence. This paper investigates the effect of variations in these inputs using a Robust Design Optimization (RDO) tool, optiSLang. The Design of Experiments (DOE) constructed within optiSLang has been evaluated using the CFD software, StarCCM+ to obtain a sensitivity map for the input parameters towards de-icing performance. This paper can be viewed as the next step in the evolution of de-icing simulation method applied to the vehicle development process. The know-how of the sensitivity of de-icing performance to various parameters can be applied towards robust decision making in the early design phases of subsequent vehicles and will also aid in effective optimization of the defrost systems.
KEYWORDS - De-icing simulation, Sensitivity Analysis, Robustness Design Optimization, StarCCM+