The purpose of this thesis is to study how an uncertain knowledge in the definition of the input parameters of a computational model propagates through the outputs. To do so, we will develop efficient methods based on the polynomial expansion of the output from a deterministic model, also known as Polynomial Chaos Expansion. We will compare this class of methods with other approaches for uncertainty propagation and use this new-found knowledge first on a low-fidelity hypersonic model and then to a high-fidelity model from computational fluid dynamics in the high Mach number regime. For this high-fidelity system, a computational analysis was performed using US3D and STABL2D software to predict the mean flow and the boundary layer transition location for a two-dimensional cone. This system was influenced by design conditions for the new Notre Dame Mach 6 Quiet Tunnel (ANDLM6QT) to provide groundwork for future simulations.