Methods for the Uncertainty Quantification of Aircraft Simulation Models

Abstract

The paper deals with the propagation of uncertainty in input parameters through the aircraft model in clean cruise configuration triggered by the elevator pulse. Assuming aerodynamic coefficients as random variables and processes, the evolution of uncertainties in the aircraft state is estimated with the help of efficient nonintrusive procedures—stochastic collocation and the nonintrusive Galerkin approaches, here contrasted to the slow convergent Monte Carlo integration. These numerical methods are implemented by using the flight simulator in a black-box manner. In this way, the set of samples of aircraft states is simply obtained by solving the corresponding systems of deterministic ordinary differential equations. Additionally, the paper provides the variance-based sensitivity analysis of a flight model carried out with the help of the polynomial-chaos approach.