A complex chemistry model of reduced size (65 species and 268 reaction steps) derived on the basis of n-heptane auto-ignition kinetics, low hydrocarbon oxidation chemistry, poly-aromatic hydrocarbon (PAH) and NOx formation kinetics together with a phenomenological soot model have been integrated with the KIVA code for multidimensional diesel simulations. A partially stirred reactor model is used to handle the turbulence-chemistry interaction. The results obtained from numerical simulations for a direct-injection (DI) diesel spray, which is injected into a constant-volume combustion vessel at engine-like conditions, show that the approach is able to reproduce the transient diesel auto-ignition and combustion processes as observed in many optical imaging studies.
The simulated results indicate that the auto-ignition of DI diesel spray occurs at a lean site close to the mean stoichiometric line for the cases tested. The ignited spot develops on the lean side, crosses over the mean stoichiometric line, and enters into the rich zone in a very short time. The prediction demonstrates that the post-ignition combustion process occurs in a lifted diffusion flame stabilized at a certain distance from nozzle exit. The spatial distributions of soot and NO in the predicted liftoff flame are similar to those proposed in Dec's conceptual diesel combustion model.
Further numerical investigation shows that, the lower the ambient gas pressure and temperature, the longer the auto-ignition delay times of the spray and vice versa. Increase in ambient gas pressure or temperature causes a reduction in the flame liftoff length. The results demonstrate that the flame liftoff length is more sensitive to the change in ambient temperature. The liftoff has a strong influence on the soot and NOx formation. The higher the flame is lift, the lower the emissions will be due to this effect.