Model predictive control of SCR aftertreatment system

The urea-water-solution based selective catalyst reduction(SCR) system is one of the effective device to reduce the NO_x emission of diesel engines. In order to realize the SCR technology in automotive application, a model-based approach is used for controller design in this paper. NO_x emission must be regulated while NH₃ slip is maintained at acceptable level even during the transient driving cycle. The existence of SCR model, pointwise-in-time constraints and multiple control objectives motivate the application of model-based optimization-control approaches such as model predictive control(MPC). The closed-loop simulation result shows that the thermal dynamics in the system must be captured in order to achieve optimal conversion efficiency and minimum NH₃ slip. The closed-loop simulation based on engine-out data in a FTP test cycle shows an average NO_x reduction of 93%, while the ammonia slip is kept at an average value of 10 ppm.