Method of Kinetic Model Reduction for Computational Fluid Dynamics Applications

A new method for kinetic model reduction is proposed. This method consists of two steps: skeletal model reduction using the directed relation graph method, followed by tuning of the remaining rate constants using the solution mapping method. The method can generate reduced models without the need for nonelementary reactions or non-Arrhenius rate constants. The method is demonstrated on a hydrogen oxygen model and a methane-air model. The results for models at different levels of reduction are compared with the original full models. Results show that the reduced models that underwent optimization are capable of replicating the temperature profile produced by the full model for constant-pressure stream tube reactions, over a limited range of initial conditions, and can now replicate laminar flame speeds at the condition specified by the current optimization. However, perfectly stirred reactor temperatures cannot be accurately predicted. Recommendations are made for further refinement of the technique.