To meet the stringent emission regulations on diesel engines,
engine-out emissions have been lowered by adapting new combustion
concepts such as low-temperature combustion and after-treatment
systems have also been used to reduce tailpipe emissions. To
optimize the control of both in-cylinder combustion and the
efficiency of an after treatment system to reduce NOx, the amount
of real-time NOx emissions should be determined.
Therefore, in previous studies, the authors developed a
real-time NO estimation model based on the in-cylinder pressure and
the data available from the ECU during engine operation. The model
was evaluated by comparing its results with a CFD model, which
agreed well. Then, the model was implemented on an embedded system
which allows real-time applications, and was verified on a
2.2-liter diesel engine. The model showed good agreement with the
experimental results at various steady-state conditions and simple
transient conditions.
In this paper, to verify the performance and to investigate the
characteristics of the real-time estimation of the model, the
engine-out NO emissions measured by a fast NOx analyzer and the
estimated NO emissions were compared during ECE-15 and EUDC cycles.
Furthermore, to extend the NO model to a complete NOx prediction
model, an empirical NO₂ prediction model is proposed based on the
experiments under steady-state conditions. The in-house EGR
prediction model was also applied in the NOx prediction model for
accurate cycle-by-cycle prediction and used as an input during
transient engine operations.