Inferential NOx emission prediction in feedforward reductant control for SCR

To adequately control the reductant flow for the selective catalytic reduct ion of NOx in diesel exhaust gas a tool is required that is capable of accu rately and quickly predicting NOx emissions from the engine's operating var iables. Two algorithms for non-linear modelling are evaluated: neural netwo rks (Solla et al., Adv. in Neural Information Processing Systems 12 (MIT Pr ess, Five Cambridge Center, Cambridge, MA, 2000)) and the split & fit algor ithm (Bakker et al., submitted for publication to NIPS). Measurements were carried out on a transient automotive diesel engine and a semi-stationary d iesel engine. Both algorithms gave excellent predictions with a short compu tation time (0.03-0.13 ms). This makes them very promising tools in automot ive catalytic NOx emission control.