Reduction of Detailed Chemical Mechanisms Using Reaction Class-Based Global Sensitivity and Path Sensitivity Analyses

Reduced chemical
mechanisms with a small size and good performance
are very important for the simulation of advanced combustion engines.
In the present study, a new reduction method of detailed chemical
mechanisms was proposed using reaction class-based global sensitivity
and path analyses. During the reduction process, the influence of
the species and reactions was determined according to the contribution
of their corresponding reaction classes to the prediction uncertainties
by calculating the nominal sensitivity index and the path sensitivity
coefficient of each reaction class from the detailed mechanism. Furthermore,
the dependence of the prediction target on the operating temperature,
pressure, and equivalence ratio was studied. After establishing the
initial reduced mechanism, the refinement of the rate coefficients
in the fuel-specific submechanism was conducted to improve the nominal
predicted value of the reduced mechanism covering broad temperature
conditions. Based on the proposed method, a reduced n-heptane mechanism with 89 species and 276 reactions is obtained
from a detailed one comprising 645 species and 2827 reactions. By
comparing the calculated value of the reduction targets from the reduced
mechanism and the detailed mechanism over broad operating conditions,
the reliability of the reduced mechanism was examined. Good agreements
for the predicted data between the reduced and detailed mechanisms
indicate the advantages of the present reduction method. Compared
to the other methods, the reduced mechanism built using the present
method was capable of better reproducing the prediction performance
of the detailed mechanism with a more compact size.