Flow distribution effects in the loading and catalytic regeneration of wall-flow diesel particulate filters

Abstract

The study of catalytic regeneration characteristics of porous ceramic diesel particulate lters (DPFs) is of growing interest to industry as diesel soot emissions are limited by legislation to levels below 0.01 g/km (for passenger cars). More speci cally, pressure drop computations and correlations are important factors employed in the design and control of diesel lter systems. However, in numerous cases, computations and models fail to match the experimentally observed evolution of soot combustion in the lter. In this paper, the role of ow maldistribution in this issue is investigated, by means of full-scale tests of the loading and regeneration behaviour of a particulate lter installed on a modern diesel engine run on catalyst-doped fuel. Loading tests were performed at three characteristic engine operation points with markedly di erent levels of engine exhaust gas mass owrate. In these tests, it becomes apparent that complex ow maldistribution phenomena exist during the loading phase, which are not directly re ected by the behaviour of the pressure drop versus time curve. However, these phenomena are shown to a ect the distribution of collected soot mass in the di erent channels of the lter and, consequently, the regeneration behaviour. The evolution of ow maldistribution was also studied in a number of regeneration experiments. It was con rmed that the variation of the volatile organic fraction in the lter and the associated partial catalytic regenerations at low temperatures interact with ow and soot maldistribution in a complex way. The conclusions from this study set the scene for future, more detailed investigations that are expected to improve understanding and modelling of diesel lter pressure drop and regeneration characteristics. Keywords: diesel exhaust emissions, diesel particulate lters, catalytic soot incineration, volatile organic fraction, ow maldistribution m dynamic viscosity (Pa s) NOTATION r density (kg/m3) (rk) p soot layer density times permeability A area (m2) product (kg/m) E substrate thickness (m) k permeability (m2) m mass (kg)

Experimental investigation of catalytic soot oxidation and pressure drop characteristics in wall-flow diesel particulate filters

“…As is shown in Table 3.9 a constant, typical value of (ρk) p = 2.25E-13 was proven successful in most simulations. The discrepancy between the experimental and calculated values could be partly attributed to soot and flow maldistribution effects [106] that cannot be taken into account in the 1D model. Concerning the wall permeability, a constant value of the order of 2E-13 was found capable to match the pressure drop curves of the whole range of validation experiments.…”

“…However a certain degree of inaccuracy in predicting the initial heating phase of the regeneration is observed. It is believed that this is a 3D phenomenon related to the change of average mass flowrate, due to the reduction in flow resistance in the central channels [106].…”

“…This fact hints to the existence of flow maldistribution at filter inlet during the loading phase (center -vsperiphery). The existence of flow maldistribution is investigated in [106] by means of flow velocity measurements at filter exit, during the loading phase and the regeneration phase. …”

“…As can been seen a difference of the order of 90 o C is appeared between the temperatures measured at the periphery where the higher peak temperatures recorded and those measured at the center. This is attributed to the flow maldistribution phenomena that enforce the flow to pass mainly through the central channels of the filter [106]. The higher flowrate through the central part means higher convection that is responsible for the lower wall temperatures of this region.…”

“…These points result to a variation in exhaust gas mass flow rate in the range from 28 to 100g/s and represent the real conditions met under City and Extra-Urban driving conditions. The appearance of secondary effects related with the flow maldistribution phenomena affected by the flow rate during regeneration is expected to play significant role on the simulation results [106]. Each regeneration process was initiated after loading at the respective engine operation point, until the same level of filter loading is achieved (6g/l filter) by a combined step variation in engine load and engine intake airflow obtained with adjustment of turbocharger and EGR valves.…”