An experimental and simulation study on the cold start behaviour of particulate filters with wall integrated three way catalyst

• Temperature step experiments are used to compare the dynamic cold-start behaviour.
• A significantly delayed light-off is observed for the particulate filter.
• A numerical simulation model for the gasoline particulate filter is presented.
• The experimentally observed delayed light-off is well predicted by the model.
• Sensitivity analyses identify the physical effects contributing to this delay.

Upcoming legislation will most likely require the introduction of particulate filters for gasoline engines. One attractive technical solution combines the three way catalytic functionality and the filter in one device, the so called ‘catalysed Gasoline Particulate Filter’.The current study uses temperature step experiments and CO oxidation as a test reaction to compare the catalysed particulate filter and the conventional flow-through monolith with respect to their dynamic cold-start behaviour. Despite the fact that the two reactor configurations are tested with identical washcoat formulation, precious metal loading and thermal mass, experiments show a significantly delayed cold-start for the particulate filter. The resulting cumulated CO emissions of the catalysed filter exceed those of the open monolith by 190–300% in the temperature range between 250 °C and 325 °C.The experimental results are analysed by means of numerical simulation. In a first step a kinetic model of the CO oxidation is parameterised using only experimental data obtained for the conventional flow-through catalyst. The resulting kinetics are implemented in a model of the wall-flow filter. Without further modification of the kinetic parameters, this model quantitatively predicts the cold-start behaviour of the catalysed filter.Finally, the numerical model is used in a sensitivity analysis to identify and quantify the individual physical effects contributing to the experimentally observed difference in the light-off behaviour. It is shown that a part of the observed difference in the cold-start performance can be traced back to differences in cell density and the heat capacity of the plugs. Even at identical cell density and without the plug effect the filter shows significantly higher CO emissions. It is shown that this intrinsic difference between the filter and the conventional monolith can be quantitatively explained by differences in heat transfer, internal mass transfer and external mass transfer.

Figure optionsDownload as PowerPoint slide