Selective catalytic reduction (SCR) of nitrogen oxides (NOx) is a widely applied diesel engine exhaust gas after-treatment technology. For effective NOx removal in a transient operating automotive application, controlled dosing of urea can be used to meet the increasingly restrictive legislations on exhaust gas emissions. For advanced control, e.g. Model Predictive Control (MPC), of the SCR process, accurate state estimates are needed. We investigate the performance of the ordinary and the extended Kalman filters based on a simple first principle system model. The performance is tested through a series of simulation studies reflecting realistic challenges such as under-modelling and few gas composition sensors.
Proceedings of 8th Ifac Symposium on Advanced Control of Chemical Processes, 2012, p. 501-506
State Estimation; Kalman Filtering; Nonlinear processes; SCR process
Main Research Area:
8th IFAC Symposium on Advanced Control of Chemical Processes, 2012