posted on 2014-06-04, 00:00authored byAlessandro Stagni, Alberto Cuoci, Alessio Frassoldati, Tiziano Faravelli, Eliseo Ranzi
When
it comes to handling large hydrocarbon molecules and describing
the pyrolysis and combustion behavior of complex mixtures, the potential
and limitations of detailed chemistry require a careful investigation.
Indeed, as they involve a large number of species and reactions, detailed
kinetic mechanisms often make the model predictions computationally
expensive, thus strongly restricting their potential. Therefore, the
automatic generation of detailed mechanisms with several thousands
of molecular species and elementary reactions, very useful in many
circumstances and a prerequisite to derive reduced models, may become
useless from a more general application viewpoint. In order to overcome
these limitations, in this paper a proper strategy to obtain suitable
mechanisms for multidimensional computational fluid dynamics (CFD)
applications is presented and discussed. It couples the advantages
of two reduction techniques: chemical lumping of species and reactions
and a flexible and reliable reduction technique aimed at eliminating
unimportant species and reactions. It is shown that a central advantage
of semidetailed kinetic models, already reduced with a chemical lumping,
is their easier and more effective applicability to successive automatic
reductions. Kinetic schemes of n-heptane and n-dodecane oxidation, reduced to 100 and 120 species, respectively,
are obtained and compared with experimental data and with the complete
original model. These dimensions easily allow the successive use of
these reduced kinetic models particularly when the aim is to embody
them within computational fluid dynamic models. In particular, simulations
of steady-state, axisymmetric, laminar diffusion flames were performed
comparing the original and the reduced kinetic mechanisms. The results
demonstrated not only the reliability of the reduced mechanisms, but,
more importantly, highlighted the great computational advantages,
especially in terms of CPU times, of reduced kinetics for the simulation
of multidimensional systems. Similar or even more apparent benefits
are expected when reducing lumped kinetic schemes of combustion of
real transportation fuels, such as gasoline, jet, and diesel fuels.