Numerical Simulation of In-Flight Icing Under Uncertain Conditions

The simulation of in-flight ice accretion requires the implementation of a computational model resulting from a set of various modeling hypotheses. The epistemic and the aleatoric uncertainty inherent in the reality we aim at describing make the simulation process at least questionable, if not just poorly credible. For this reason, the numerical modeling of in-flight ice accretion cannot escape the utmost need for quantifying the uncertainty associated with computer predictions. The uncertainty analysis is fundamental for delivering a proper verification & validation of simulation software, to confirm the robustness and the predictiveness of the computerized icing model. The quantification process should account for both the epistemic uncertainty inherent in the physical model and the aleatoric uncertainty affecting the specification of the simulation parameters. This chapter reports a preliminary effort toward developing the necessary tools for validating simulation software for investigating in-flight ice accretion. In this work, we gather several reference test cases concerning the simulation of actual experiments targeting the investigation of in-flight icing. Experiments are reproduced numerically, and observations are compared to simulations to assess the accuracy of computational tools. Numerical predictions are complemented with uncertainty bars related to the randomness of the experimental operating conditions. The uncertainty analysis uses a non-linear approach to improve the approximation of the complete icing model behavior in domain regions scarcely affected by ice build-up. Several test cases are considered. Results generally confirm the robustness and the predictiveness of the considered computational icing model.