To accurately predict emissions as well as combustion phasing in
a homogeneous charge compression ignition (HCCI) engine, detailed
chemistry needs to be used in Computational Fluid Dynamics (CFD)
modeling. In this work, CFD simulations of an Oak Ridge National
Laboratory (ORNL) gasoline HCCI engine have been performed with
full coupling to detailed chemistry. Engine experiments using an
E30 gasoline surrogate blend were performed at ORNL, which included
measurements of several trace species in the exhaust gas. CFD
modeling using a detailed mechanism for the same fuel composition
used in the experiments was also performed. Comparisons between
data and model are made over a range of intake temperatures. The
(experiment & model) surrogate blend consists of 33 wt %
ethanol, 8.7 % n-heptane and 58.3 % iso-octane. The data and
simulations involve timing sweeps using intake temperature to
control combustion phasing at a constant fuel rate. The modeling
uses a detailed chemical kinetic mechanism consisting of 428
species and 2378 reactions. This mechanism was obtained by a
targeted mechanism reduction of a well validated master kinetics
mechanism for multiple gasoline surrogate-fuel components, which
consists of 3553 species and 14904 reactions. A 15-degree sector
mesh consisting of 53,800 cells at IVC has been used for the closed
valve simulations.
The CFD simulation employs the newly developed FORTÉ simulation
package, which was designed to take advantage of advanced chemistry
solver methodologies as well as advanced spray models. In this
study, there is no spray model used, since the fuel is atomized and
quickly vaporized during port injection. However, parallel
computing, dynamic adaptive chemistry and dynamic cell clustering
methods have been used to minimize the chemistry related
computational time while maintaining accuracy in the kinetics
predictions. These methods allow inclusion of the relatively large
(428 species) detailed kinetics mechanism directly in the
simulation, while keeping the overall simulation time reasonable
for production work.
Comparisons with the engine data include the trends of
combustion phasing as a function of intake temperatures. Emissions
of several species are also compared with engine data. The ORNL
engine experiments included detailed exhaust measurements of
NOx, CO, formaldehyde, acetaldehyde, methane, ethylene,
propene, iso-butylene and the overall unburned hydrocarbons. All of
these exhaust measurements have been compared with the modeling
results, as a function of intake temperatures. The results agree
well with the engine data, and the agreement provides confidence in
the predictive capability of the model for studying chemistry and
fuel effects.