Prediction and Control of Long-Term System Degradation for a Light-Off SCR in an Ultra-Low NOx Aftertreatment System

<div class="section abstract"><div class="htmlview paragraph">The heavy-duty low NOx program funded by EMA at Southwest Research Institute (SwRI) evaluates a combination of engine and advanced aftertreatment systems to achieve a 0.035 g/bhp-hr tailpipe NOx standard. This work emphasizes improvements to the light-off SCR (LO SCR) model used for low NOx controls. Two key mechanisms drive these improvements: the first is a real-time feedback system that utilizes the LO SCR outlet NOx sensor for short-term corrections to the model state, and the second involves adjustments to the dosing mechanism based on long-term trends in dosing signals compared to predicted NH3 consumption, derived from LO SCR inlet and outlet NOx sensors, referred to as long-term trim. An algorithm is incorporated to differentiate the LO SCR outlet NOx sensor readings into NOx and NH3 components based on cross-correlation between inlet and out NO<i><sub>x</sub></i> sensors termed as speciation. The integration of this speciation algorithm with both short-term and long-term trim mechanisms significantly enhances the accuracy of the model estimated NH<sub>3</sub> storage state, as well as the prediction of outlet NOx, and NH3 levels under various transient conditions, including CFTP, HFTP, RMC, and LLC cycles. This improved accuracy in the LO SCR observer model enables more precise control of transient tailpipe NOx in the system.</div></div>