Analysis and modelling of particulate matter filtration in internal combustion engine exhaust aftertreatment systems

The filtration of Particulate Matter (PM) in the Exhaust Systems of the Internal Combustion Engines (ICE) is one of the main aftertreatment solutions that allowed to comply with progressively more restrictive regulations on engine emissions. While very efficient to this aim, Particulate Filters may be the cause of different drawbacks such as additional fuel penalties, increased costs and impact on the system durability. Therefore, a proper balance among these different issues is key for the design of such emission control components and systems. Although the experimental analysis is still required to guarantee the real performances of aftertreatment systems, numerical simulation seems to be the most convenient strategy to cope with their development, optimization and control of one in this case. In fact, after a specific validation by comparing results with a limited set of experimental data, a virtually unlimited number of combination of the design and operating parameters may be numerically tested with a remarkable reduction of time and costs. However this is advantage may be only apparent, as the technical challenges presented by advanced numerical modeling are more than a few: the Particulate Filter behavior is intrinsically characterized by processes acting on a very wide set of characteristic space and time scales. The latter in fact ranges from the description of the exhaust gases flow occurring at a scale in the order of the length of the device, down to the nanoparticle deposition process occurring at the micro-pores scale of the porous media. Furthermore, the design of any Exhaust Aftertreatment component or system requires both the understanding of the effects of parameter design and control on long-term operation as well as a detailed analysis of microscopic fluid dynamic processes (e.g. Brownian motion effects on nanoparticle filtration). All the mentioned effects must be properly represented in the model. The main objective of this PhD dissertation consists thus of the development of two alternative numerical models which allow to cope with this multiscale problem: a simplified macroscale and a detailed multiscale one. While each of the model is able to describe problems on the reference time and space scale, their coupling may be effectively used for a complete design procedure of such components within Diesel or gasoline ICE exhaust systems. The simplified macroscale model will be first introduced to study a simplified long-term operation of a Particulate Filter, or to implement real-time control strategies; then a detailed multiscale one, capable of analyzing basic filtration and fluid dynamic processes in the component, will be described. The effects of key design parameters such as channel shape, porous media specifications and fuel type will be commented for both the approaches, along with the description of the processes and of the main parameter trade-offs in terms of efficiency, fuel consumption, costs and durability.