Traditionally the NOx prediction models use to take into account exclusively the NOx formation process. Nowadays diesel engines use to operate with high EGR rates and high fuel/air equivalence ratios, in opposition with what it was standard in the past years. In such conditions a considerable amount of NOx generated in the previous cycle (coming from the recirculated exhaust gases) or in the previous combustion is re-entrained by the flame, where a highly reducing region exists (due to a lack of oxygen and the presence of hydrocarbons and a high temperature level). Face to this fact a question arises: what happens with these re-entrained NOx?
A first experiment in a single-cylinder engine where the fuel/air equivalence ratio was progressively increased (and, consequently, the re-entrainment of exhaust gases containing NOx was more and more important) showed a surprising trend in total NOx emissions that a model based only on the NOx formation process and validated with engine operating conditions with low NOx re-entrainment was unable to reproduce. This was a first trace of the possible existence of a NOx reduction mechanism inside the diffusion flame. A second and more specific experiment in the same engine, where a large amount of synthetic NO was added to the intake air, demonstrated the existence of NOx reduction, as a diminution of the exhaust NOx mass fraction compared to the intake NOx mass fraction was observed.
If a simple model taking into account this NOx reduction process is integrated in the original NOx prediction model based exclusively in the NOx formation process, the predictions are now able to reproduce more accurately the experimental behaviors in the operating conditions where a high NOx re-entrainment exists.