Research on dynamic modeling and carbon load estimation of diesel particulate filter

Abstract

ABSTRACT Accurate estimation of carbon load in diesel particulate filters is an important basis for efficient and safe operation of active regeneration. Currently, model-based carbon load prediction has the advantages of high accuracy and low influence by working conditions. In this paper, a dynamic model of diesel particulate filters was developed based on the deep bed and cake layer trapping mechanism and regeneration mechanism. The trapping process was described using the spherical trapping mechanism, while the regeneration process was based on the non-catalytic oxidation process of the trapped particles. A pressure drop correction based on the extended Kalman filter was developed to correct the errors accumulated in the carbon load prediction during the integration. The maximum error of carbon load prediction was 0.3 g/L, and the average error was less than 0.17 g/L, the maximum error of pressure drop was 0.7 kPa, and the average error was less than 0.33 kPa. The carbon load prediction algorithm was proved to have good accuracy and can be used for start-stop judgment of DPF active regeneration, and the pressure drop and temperature calculated by the DPF model can be used for DPF status monitoring during regeneration to ensure safety and reliability.