An OBD strategy to estimate SCR ageing and detect urea injection faults

Abstract

The most recent regulation for diesel engines with regards to NOx emissions and SCR systems considers stringent OBD requirements. In this paper, an observer of the ageing state is developed considering NO, NO2 and NH3 concentrations, the surface coverage ratio θNH3, and the ammonia storage capacity Ω as states. The observer is based on an extended Kalman filter and the SCR downstream NOx and NH3 measurements. Accuracy and convergence time are presented for the observer validation in simulation in a dynamic WHTC at warm starting conditions. As the ageing state estimation is based on the residual between a model and measurements, an additional indicator avoids associating an error inherent of urea injection faults to ageing. The performance of this indicator is independent from the ageing state of the SCR. The highly detailed, physics-based 1D SIMCAT model is used as the SCR plant for the ageing estimation and as the model to track the urea injected quality. Then, a control-oriented model is also used in simulation to allow the state-space representation required for the observer. Accuracy of both models is argued, showing the limitations and advantages of the control-oriented model.