Research on cross sensitivity of NOx sensor and Adblue injection volume in accordance with the actual situation based on cubature Kalman filter

Abstract

Due to the nitrogen oxide (NOx) sensor's cross-sensitivity and inability to monitor ammonia coverage, it is critical to eliminate these flaws for a significant percentage of in-use diesel vehicles without NH3 sensors. In this study, a platform for multi-model co-simulation was established. The results showed that the downstream NOx concentration of the selective catalytic reduction (SCR) could be more accurately calculated using the filtering model based on cubature Kalman filter (CKF). Under European Steady-State Cycle (ESC) and European Transient Cycle (ETC), the prediction errors of the filtering model based on CKF were 4.6% and 7.7%, respectively. Under ESC, compared with the old control strategy, the prediction accuracy of temperature model of the new control strategy was increased by 2.6%, the average Adblue injection rate was increased by 9.1%, and the average NOx conversion efficiency was increased by 4.3%. Under ETC, the prediction accuracy of the temperature model of the new control strategy was increased by 2.5%, leading to a high average NOx conversion efficiency of 87.9%. Regarding the old control strategy, although the average NOx conversion efficiency reached 92.1%, the average ammonia slip concentration exceeded the threshold (10 ppm) by more than five times.