A radiation model based on the discrete ordinates method was developed and integrated into a multi-dimensional engine simulation code, KIVA-ERC. The predicted total heat loss in the diesel engine used in this work increased by about 30% when radiation effects were considered in addition to convection. Application of radiation has the effect of lowering bulk gas temperatures resulting in a decrease in soot oxidation rates and a substantial decrease in predicted NOx production rates.
The radiation model was validated by comparing the predicted radiative heat flux at a point on the head of a diesel engine to experimentally measured values. The predicted radiative heat flux compared well with the experimental data both in phasing and peak intensity. However, the experimentally measured heat flux tapered off more rapidly than the predicted levels, suggesting that the soot oxidation model was under-predictive.
Using a thermophoretic soot deposition model, it was found that radiation from the soot in the near-wall region had only a very small effect on engine heat loss. However, thermophoretic soot deposition was found to account for 1?3% of the total exhausted soot.
A finite element code, HCC, was used to predict the temperatures of the metal engine components, which were then used as a refined boundary condition in KIVA. This increased the accuracy of both the radiation and soot deposition models. The predicted peak temperatures of the engine components could differ by more than 60 K when radiation was considered in addition to convection.
The effect of radiation at very high load conditions was investigated with a parametric study. Equivalence ratios of 0.5 and 0.7 were tested over a range of start-of-injection timings using the maximum attainable air flow rate available in the engine test cell. Peak piston temperatures were found to approach the melting point of the piston material indicating that the engine might be damaged if it were run at that particular operating condition. Consideration of radiation effects with a non-uniform surface temperature distribution was found to increase predicted engine-out soot levels by 13% and decrease NOx levels by 46%.