The scope of this University of Michigan computational effort addresses the development of a fully validated large-eddy simulation (LES)-modeling capability to predict unstable combustion of high hydrogen content (HHC) fuels. To incorporate effects of preferential diffusion, pressure variations, and variations in mixture composition, an unsteady flamelet-based LES combustion model will be extended. The integrated LES-validation effort includes (1) an a priori analysis of critical modeling assumptions using a Direct Numerical Simulation (DNS) database of jet-in-cross-flow configurations, and (2) a posteriori model validation in LES application of a swirl-stabilized gas turbine combustor. The LES-combustion model will be used to develop detailed simulations to characterize facility-induced nonidealities in flow-reactor experiments. Effects arising from high-Reynolds number turbulence transition, mixture stratification, and other mechanisms associated with turbulence/ chemistry interaction on the autoignition behavior will be quantified through parametric calculations. The information gained from these efforts will be used to develop a low-order model that can be utilized for chemical-kinetics investigations and for guiding and improving future flow reactor designs in order to reduce facility effects.

The experimental effort includes high-pressure measurements of HHC fuel combustion in a dual-swirl gas turbine combustor, development of a comprehensive experimental database for LES model validation by considering stable and unstable gas turbine operating conditions, and obtaining improved understanding about fundamental combustion-physical mechanisms that control flame-holding, liftoff, and flashback for HHC fuels. A range of pressures, HHC fuel compositions, and equivalence ratios will be investigated experimentally.