Air-breathing propulsion systems rely on a consistent and predicable flow of of air to maintain the combustion processes critical to producing thrust. Unstart in supersonic flow conditions is a phenomena in which the inlet mass flow for a duct exceeds the outlet mass flow. This results in the creation of a normal shock that propagates upstream until a mass flow equilibrium is achieved. Unstart in air-breathing inlets in an ongoing issue for maneuvering vehicles and is associated with large acoustic and thermal loads leading to significant vehicle risk.
In order to design effective control mechanism to prevent or reduce unstart phenomena, it is important to understand the dynamics of the fluid interactions. In practical situations, an effective reduction in duct area from boundary layer separation tends to be a major concern. For example, the boundary layer separation may be unintentionally induced by a shock-boundary layer interaction (SBLI) from a shock train intend to slow down and heat the air for combustion processes. It is important to note that even for proposed air-breathing hypersonic systems, the internal duct flows rarely exceed Mach 2-3.
Unstart is typically simulated with an area blockage in wind tunnels that chokes the duct flow. You can see in the figures below that for a cylindrical rod inserted in the the UTSI Mach 2 wind tunnel, the effective area blockage (e.g. more insertion) creates a normal shock that also exhibits SBLIs. We use our high-speed imaging capabilities and custom image processing tools to track how these features evolve in time and under different conditions (e.g. different blockage conditions).
We are also interested in the response of the normal shock to dynamic forcing, which we simulated here by actuating the rod during tunnel operation. You can see that the normal shock moves in a one-to-one manner with the change in blockage area.
We also use techniques likes pressure sensitive paint (PSP) to get global information about surface pressure fluctuations and distributions. These datasets are then compared with computational results to aid in proper modeling of the flow physics.