Soot Loading Estimation Accuracy Improvement by Filtration Layer Forming on DPF and New Algorithm of Pressure Loss Measurement

A diesel particulate filter (DPF) is a key component for reduction of engine soot emission. The soot collected in the DPF is periodically burned off, so-called DPF regeneration, and a behavior of the pressure drop increased by the soot loading is generally utilized to estimate the amount, which must be a trigger of the regeneration. However, it is said that the estimation of the soot loading amount has considerable dispersion caused by two main reasons. One is hysteresis of the transient pressure drop resulted from the combination of so-called deep-bed and cake filtration modes. The other is a fluctuation of exhaust gas temperature and flow rate as well as a pulsation from the engine. In this study, the accurate estimation method of the soot amount accumulated in the DPF was proposed in combination with filtration layers (FLs) technology and a new algorithm based on fast Fourier transform (FFT) technology. The FLs enable to prevent the deep-bed filtration mode and bring an advantage of repeatability of the transient pressure drop. The new algorithm allows us to transform the measured pressure drop into the exact values by eliminating the aforementioned fluctuation and pulsation. According to verification tests of these technologies through an engine bench, it was found that the former and the latter improved the hysteresis and the linearity of the pressure drop respectively. As a result, the dispersion of the estimation of the soot amount was considerably improved and became less than ± 1 g/L.