ACCURATE PREDICTIONS OF CHEMICAL PHENOMENA IN CHEMICALLY REACTING FLOWS

Abstract. In the past decade, the computational fluid dynamics (CFD) community has put increasing attention on incorporating complex chemistry and physics into CFD simulations. However, for many practicing engineers today there is still much lacking in the effectiveness of reacting-flow simulation. Particularly in the area of combustion, the results can often be either too time consuming, or still too inaccurate to provide practical solutions for applications where chemistry details are critical. There are important alternatives to forcing such detailed phenomena directly into CFD. Such alternatives leverage CFD advances but also provide more effective and more accurate chemistry capabilities. These approaches build on reduced-geometry models that provide higher fidelity chemistry predictions, while establishing indirect coupling to the transport environment that is well predicted by CFD. The solvers employed in reduced-order models easily handle large chemistry reaction mechanisms and accurately resolve both trace and major species over disparate time scales. Links to CFD can be established through determination of zonal mapping to equivalent reactor networks or through table look-up methods