Uncertainty Propagation and Quantification in Direct Simulation Monte Carlo Calculations of Hypersonic Reactive Cylinder Flows

摘要

Uncertainty quantification of inflow and model parameters introduced in the Direct Simulation Monte Carlo (DSMC) method remains an important issue, in particular, in hypersonic flows with chemical reactions. This paper is intended to investigate the uncertainty propagation and quantification for DSMC calculations, using the Probabilistic Collocation Method in which the sparse gird technique was introduced to reduce the collocation points to 109 for six coupled input uncertainties. Several features of the uncertainty propagation were obtained on the basis of molecular kinetic dynamics. For instance, the molecular energy transfer associated with different internal modes is responsible for the increase in translational and rotational temperature uncertainties before the shock wave. Compared results indicate that the chemical reactions have no distinct effects on the propagation of uncertainties in three temperatures, but induce a distinct decrease in uncertainty introduced on the wall heat flux because the generation of massive N and O atoms results in less molecules participating in the propagation of uncertainty in the vibrational accommodation fraction. The global sensitivity analysis demonstrates that the surface fluxes are the most and secondarily sensitive to two inflow uncertainties. Meanwhile, the chemical reactions significantly enhanced the sensitivity of the stagnation point heating to three model uncertainties in DSMC method.