Kinetic model reduction

The applicability of detailed kinetic models in the large-scale CFD simulations of combustion devices is often limited by their excessive size, in terms of number of species and reactions. In order to address this, flux-based methodologies are traditionally coupled to sensitivity analysis techniques, in order to retain accuracy for specific target properties (e.g. ignition delay time). Nevertheless, this does not ensure an accurate prediction of the remaining combustion features, like flame propagation or extinction. The objective of this thesis consists in extending the DoctorSMOKE++ reduction framework, developed at CRECK modeling group, in order to include further target properties and reduction methods, thus obtaining compact and reliable kinetic models to be adopted for the downstream CFD simulations