A NEW SUBGRID MODEL FOR LARGE-EDDY SIMULATIONS OF MIXING AND CHEMICAL REACTION IN TURBULENT FLOWS

CONVENTIONAL, TURBULENT-MIXING MODELS BASED ON GRADIENT-DIFFUSION ASSUMPTIONS ARE NOT CAPABLE OF ACCURATELY PREDICTING MIXING AND REACTION RATES IN MOST PRACTICAL COMBUSTION DEVICES. FURTHERMORE, AT THE SMALL SCALES, MOST CONVENTIONAL MODELS MAKE NO DISTINCTION BETWEENTURBULENT CONVERSION AND MOLECULAR DIFFUSION. THIS DISTINCTION IS CRITICAL FOR THE ACCURATE DESCRIPTION OF THE MIXING PROCESS. IN ADDITION, IT IS KNOWN THAT TURBULENTMIXING AND ENTRAINMENT PROCESSES IN SHEAR FLOWS ARE DOMINATED BY UNSTEADY, LARGE-SCALE, VORTICAL MOTIONS. THE SPATIAL AND TEMPORAL EVOLUTION OF THESE LARGE-SCALE STRUCTURES CANNOT BE MODELED AND MUST BE EXPLICITLY COMPUTED FOR ACCURATE PREDICTIONS. PHASE I WILL EXPLORE SUBGRID MODELING TECHNIQUES FOR USE IN LARGE-EDDY SIMULATION (LES) OF REACTING FLOWS. IN PARTICULAR, A MODEL FOR MIXING AND CHEMICAL REACTIONS AT THE SUBGRID LEVEL IN BOTH LOW- AND HIGH-SPEED FLOWS WILL BE DEVELOPED BASED ON KERSTEIN'S LINEAR-EDDY APPROACH. LES OF INCOMPRESSIBLE, TWO-DIMENSIONAL MIXING LAYERS WILL BE PERFORMED, AND THE RESULTS WILL BE COMPARED WITH HIGH-RESOLUTION, DIRECT NUMERICAL SIMULATIONS AND AVAILABLE EXPERIMENTAL DATA TO ASSESS THE PROPOSED SUBGRID MODEL. THIS MODEL WOULD BE EXTENDED TO STUDY THREE-DIMENSIONAL, COMPRESSIBLE REACTING FLOWS WITH HEAT RELEASE IN PHASE II. CONVENTIONAL, TURBULENT-MIXING MODELS BASED ON GRADIENT-DIFFUSION ASSUMPTIONS ARE NOT CAPABLE OF ACCURATELY PREDICTING MIXING AND REACTION RATES IN MOST PRACTICAL COMBUSTION DEVICES. FURTHERMORE, AT THE SMALL SCALES, MOST CONVENTIONAL MODELS MAKE NO DISTINCTION BETWEENTURBULENT CONVERSION AND MOLECULAR DIFFUSION. THIS DISTINCTION IS CRITICAL FOR THE ACCURATE DESCRIPTION OF THE MIXING PROCESS. IN ADDITION, IT IS KNOWN THAT TURBULENTMIXING AND ENTRAINMENT PROCESSES IN SHEAR FLOWS ARE DOMINATED BY UNSTEADY, LARGE-SCALE, VORTICAL MOTIONS. THE SPATIAL AND TEMPORAL EVOLUTION OF THESE LARGE-SCALE STRUCTURES CANNOT BE MODELED AND MUST BE EXPLICITLY COMPUTED FOR ACCURATE PREDICTIONS. PHASE I WILL EXPLORE SUBGRID MODELING TECHNIQUES FOR USE IN LARGE-EDDY SIMULATION (LES) OF REACTING FLOWS. IN PARTICULAR, A MODEL FOR MIXING AND CHEMICAL REACTIONS AT THE SUBGRID LEVEL IN BOTH LOW- AND HIGH-SPEED FLOWS WILL BE DEVELOPED BASED ON KERSTEIN'S LINEAR-EDDY APPROACH. LES OF INCOMPRESSIBLE, TWO-DIMENSIONAL MIXING LAYERS WILL BE PERFORMED, AND THE RESULTS WILL BE COMPARED WITH HIGH-RESOLUTION, DIRECT NUMERICAL SIMULATIONS AND AVAILABLE EXPERIMENTAL DATA TO ASSESS THE PROPOSED SUBGRID MODEL. THIS MODEL WOULD BE EXTENDED TO STUDY THREE-DIMENSIONAL, COMPRESSIBLE REACTING FLOWS WITH HEAT RELEASE IN PHASE II.