Hybrid Solution of Unsteady Flow Problems Using a Merged Euler/Kinetic Theory Approach

Wednesday, May 24, 2000 - 2:00pm - 2:25pm
Keller 3-180
David Goldstein (The University of Texas at Austin)
An adaptive computational technique that couples Nadiga's Adaptive Discrete Velocity method for solving the Euler equations of fluid flow with Bird's direct simulation Monte Carlo molecular method has been used to analyze the unsteady jet evolution from a slit subject to a pressure differential. This work is motivated by the need to study thruster plume impingement on spacecraft. The method adaptively decomposes the domain according to the degree of local translational non-equilibrium that is quantified by appropriate breakdown parameters. Disconnected patches that employ the direct simulation Monte Carlo solution deform adaptively to track non-equilibrium regions of the flow. The approach allows one to resolve complicated transient flow structures through the concentration of a large number of simulated particles in non-equilibrium regions. The process of interaction between the continuum and particle domains relies on the continuous exchange of properties at a common boundary. The novel application of ghost cells decreases the statistical noise of the Direct Simulation Monte Carlo property signal across the interface. Simulations of the jet show that a weak shock wave initially propagates into the flow followed by the high density jet core. The flow accelerates through an expansion region bounded by two shear layers that form a characteristic barrel structure. The behavior of the flow is investigated as the jet strikes a target plate. One-dimensional solutions of a shock tube problem may also be presented. The hybrid, adaptive shock tube simulation demonstrates the stability of the interface and the capability of the adaptive algorithm to capture physical phenomena such as shock waves, expansion waves and contact discontinuities.

Joint work with Roberto Roveda and Philip Varghese.