Model-Based Fault Detection and Fault-Tolerant Control of SCR Urea Injection Systems

This paper aims at developing integrated onboard diagnosis and fault-tolerant control methods with experimental validation for a urea selective catalyst reduction (SCR) aftertreatment system to reduce vehicle tailpipe emissions. Diagnostics are performed for an SCR urea injection system by estimating and monitoring the injected urea mass flow with no need for a costly physical flow sensor. The estimation is derived from a first-principle-based urea injection system model, and the model parameters are identified by using system identification. During vehicle transient maneuvers, a Kalman filter (KF) is formulated to further reduce the estimation noise and improve diagnostic robustness. Once an injection fault is detected, an adaptation control algorithm is applied to compensate the urea injection command, thus correcting certain types of urea under/overdosing faults and maintaining the SCR $\mbox{NO}_{x}$ conversion performance. These methods have been validated through vehicle tests by utilizing an onboard rapid prototyping control system.