A Computational Problem Solving Environment for Creating and Testing Reduced Chemical Kinetic Mechanisms,
- C. J. Montgomery, D. A. Swensen, T. V. Harding, M. A. Cremer, M. J. Bockelie
- Advances in Engineering Software,
Reduced chemical kinetic mechanisms developed and tested using the CARMPSE software have been implemented into a variety of CFD codes. When used in the commercial code Fluent, a CARM-PSE created reduced mechanism performs well for predicting the major species and temperature of the Sandia D flame. Predictions for radical and pollutant species are not as good. Much of this error may be due the the inadquacies of the EDC turbulence-chemistry model used in these simulations. Reduced mechanisms representing combution of Diesel fuel and gasoline perform well in simple reactors, but are presently hindered numerical innaccuracies within the engine combustion CFD simulation. Numerical optimization techniques have been implemented into CARM-PSE to automatically select the best species to include in reduced mechanisms for a given set of conditions. The genetic algorithm has been demonstrated to be robust and successful at creating reduced mechanisms that are improvements over previous species selection methods. A new algorithm, based on forming in-vs.-out pseudo-gradients has been developed and implemented. When procedures are implemented to deal with reduced mechanisms that crash the test codes, this algorithm is both efficient and robust.