• Dr. James Driscoll
  • University of Michigan, Ann Arbor
  • 109 Surge Building
  • 1:00 p.m.
  • Faculty Host: Dr. Lin Ma

Three turbulent combustion experiments are discussed:  (a) a lean premixed-prevaporized (LPP) gas turbine device that uses a GE TAPS fuel injector, (b) a scramjet experiment, and (c) our fundamental high Reynolds number HiPilot flame.  Movies obtained using kilohertz laser imaging diagnostics allow us to assess some theoretical concepts that form the fundamentals of turbulent combustion theory.

In the gas turbine model combustor, the unsteady liftoff of the flame base leads to undesirable engine “growl”; measurements are explained by a simple Helmholtz model.  In the scramjet and fundamental flames, we investigate the boundaries that define the regimes of thickened flamelets, broken flamelets and distributed reactions. These boundaries determine which subgrid modeling approach is most appropriate.   The high speed imaging data also indicate how the theory of flame stretch and the theory of hydrodynamic instabilities apply to explain turbulent wrinkling of a flame surface and local flame extinction.  The role of auto ignition chemistry becomes especially important in the highly-preheated air flows within the scramjet and gas turbine experiments.   An “auto-ignition assisted flame” is observed, as evidenced by images of formaldehyde within distributed reaction zones that exist far upstream of the main heat release reactions.  A jet-in-cross flow flame was studied in which a fuel jet is injected perpendicular to a heated high-speed air flow. Both distributed reactions and broken flamelets were observed.  Some comparisons are presented of the Michigan measurements to DNS computational results obtained at LBL and Sandia Labs.