Eulerian-Lagrangian modeling of particle motion in exhaust gas aftertreatment systems

Licentiate thesis, 2009

In the work to design and optimize new aftertreatment system devices for NOx reduction and particulate matter removal in lean engine exhaust (such as diesel exhaust), it is of great help to be able to perform quick, robust and reliable computer simulations of system performance. A suitable platform for such simulations is Computational Fluid Dynamics (CFD) software. The work within this thesis deals with such simulations. A promising technique for the reduction of NOx in lean engine exhaust is the urea-SCR system. The influence of the most common plausible modeling choices in CFD-simulations of urea-SCR systems has been assessed. It has been shown that modeling choices may affect the predicted extent of wall hit, which types of droplets that are predicted to hit the wall, and most possibly also the outcome of wall-wetting if a wall film model is to be employed. The influence of the lift force, the thermophoretic force and the history force was shown to be negligible. It was also shown how difficult it is to account for turbulent dispersion effects, as the random walk models commonly used together with two-equation turbulence models are very sensitive to deviations from isotropic turbulence. Removal of particulate matter from the exhaust can be accomplished with a wall-flow or a flow-through filter. A mathematical model for an arbitrary flow-through filter has been assembled. The model has been used to determine the particle trapping efficiency of a metallic substrate with protrusions. It was found that – due to the interaction of two dominating particle trapping mechanisms – there exists a minimum in particle trapping efficiency for intermediate sized-particles (150 nm) in this device. These interacting trapping mechanisms are Brownian diffusion (mainly affecting the smallest particles) and inertial impaction (mainly affecting the largest particles). The fact that different sized particles are captured by different mechanisms also reflects in that different sized particles are trapped in different regions of the filter. Whereas the trapping of small particles is increased with a lowering of the gas flow velocity (due to an increased retention time), the trend is reversed for large particles (due to a decreased system response time). Medium-sized particles are least influenced by changes in gas flow velocity. The proposed model can be used for design optimization of flow-through devices for trapping of particulate matter.