Reduction of kinetic models using dynamic sensitivities

The development of detailed chemical kinetic models is necessary for the design and optimization of complex chemical systems. However, it is also often desired to reduce the model size by excluding inconsequential chemical species and/or reactions for end-point applications, usually due to computational reasons. In this work, new model reduction methods based on dynamic sensitivities from the impulse parametric sensitivity analysis (iPSA) and the Green's function matrix (GFM) analysis have been developed. The iPSA and GFM were originally formulated to provide dynamical parameter-by-parameter and species-by-species information on how a system output behavior is achieved, respectively. The efficacies of the proposed reduction methods were compared with existing methods through applications to reduce detailed kinetic models of alkane pyrolysis and natural gas combustion (GRI Mech 3.0) and an ab initio kinetic model of industrial steam cracking of ethane.