Soft computing model for prediction of EGR effects on particle sizing at CR Diesel engine exhaust

Use of EGR (Exhaust Gas Recirculation) and aftertreatment devices allows diesel engines to comply with actual emission regulations. In order to satisfy future emission standards a careful analysis of peculiarities and limits of the current systems for pollution control and of their possible influence on production of other harmful substances will be necessary. Engine control maps determine optimal EGR considering a trade-off between NOx and smoke emissions. However, actual control strategies do not consider, in the definition of optimal EGR, its effect on particle number density, which has a great importance for the optimal functioning of after-treatment systems. In this paper a soft computing model that gives real-time information on the characteristic of exhaust particles, is proposed. The model, by using a neural network approach, is able to provide information on the effect of EGR on particulate mass concentration and particle size distribution. The proposed model can be employed for advanced real-time engine controls which, acting on the amount of recirculated exhaust gas, can lead to a reduction of exhaust emission optimizing at the same the particulate size distribution. The experiments have been carried out at the exhaust of a Common Rail 1.9 l, 16 v diesel engine for different engine operating conditions. Particle number size distributions in the range 7 nm-10 \gmm have been measured.