Design of extended Kalman filter observer coupled to ammonia cross-sensitivity model for an SCR system

Jinghua Zhao,Yutong Zhang,Haonan Wang,Yao Sun,Yunfeng Hu,Jianhua Yu,Fangxi Xie

DOI: https://doi.org/10.1177/14680874221140447

2022-12-19

International Journal of Engine Research

Abstract:International Journal of Engine Research, Ahead of Print. Commercial diesel vehicles that comply the China/EURO VI emission regulations are generally equipped with selective catalytic reduction (SCR) systems for reducing [math] emission and [math] sensors at its downstream for accurate closed-loop control. However, [math] sensors have significant ammonia cross-sensitivity pushing up their readings. Moreover, for the closed-loop control of SCR systems effectively estimating the critical states from the [math] sensor readings remains a further challenge. In this paper, an ammonia cross-sensitivity model with segmented sin and quadratic composite functions describing high frequency characteristics is established by analyzing the experimental data of an SCR system. A control-oriented three-states model of an SCR system is established, and based on it an extended Kalman filter (EKF) observer coupled to the ammonia cross-sensitivity model is proposed. The simulation results show that from the [math] sensor readings the EKF observer can effectively estimate the deterioration trend of the [math] and [math] emissions, and describe the dynamic change process of the ammonia coverage rate. The experimental results under the four non-excitation cycles show that the EKF observer can effectively estimate the deterioration trend, that the average estimated accuracy of [math] and [math] during their deterioration stage can respectively reach 86.16% and 54.87%, and that the observer has better robustness to disturbances from the operating conditions.