Flow dynamics in corrugated pipes: Effect on particle agglomeration
Time: Fri 2020-10-23 10.00
Subject area: Engineering Mechanics
Doctoral student: Ghulam Mustafa Majal , Teoretisk och tillämpad mekanik, KTH CCGEx
Opponent: Professor José Galindo, Universitat Politècnica de València, Spanien
Supervisor: Docent Mihai Mihaescu, Teoretisk och tillämpad mekanik; Professor Mats Åbom, Marcus Wallenberg Laboratoriet MWL; Docent Lisa Prahl Wittberg, Teoretisk och tillämpad mekanik; Dr. Mikael Karlsson, Marcus Wallenberg Laboratoriet MWL
Abstract
In recent times, the deleterious impact of particulate emissions from road vehicles on the well-being of humans has garnered a significant amount ofattention. Regulatory bodies have enacted legislation in order to counter act particulate emissions. In order to meet the proposed legislative requirementsthe engineers have to come up with a suitable after treatment treatment technology. It has been understood that typically the emitted particles are distributed such that the particles largest in terms of number are the smallest in size. Recent legislations have increasingly emphasized controlling boththe particle mass and number of the emissions. The present thesis employs a numerical approach in order to study the transport of particles in an exhaust flow. A particle agglomeration concept is considered as a means to shift the particle size distribution. In order to increase the likelihood of interactionbetween the particles the host gas is manipulated in such a way that it successively accelerates and decelerates.As a starting point a 1D model is considered. Extensive parameter studies are performed in order to determine the appropriate flow characteristics,which promote particle grouping. The term grouping refers to particles moving together in a group or a cluster, thereby increasing the likelihood foragglomerating with one another. It is revealed that higher pulse frequency and geometric wavelength promotes particle grouping. A comparison between different pulse shapes highlighted that smoother pulse shapes are marginally better for particle grouping. Finally, it is observed that an idealized sinusoidalpulse form underestimated the extent of particle grouping when compared to an actual engine pulse.In the next step, a 3D computational fluid dynamics (CFD) study is performed on an idealized axi-symmetric sinusoidal pipe-like geometry. The geometrical parameters of the geometry are based on the earlier completed 1D study. Firstly, for a continuous inlet flow scenario the observed flow structures are highlighted. Proper orthogonal decomposition (POD) revealed that for a geometry with a sufficiently large maximum cross sectional diameteran asymmetric oscillatory mode is present. This mode is caused by the dynamics of the recirculation bubble and the shear layer. In a geometry witha much smaller maximum cross sectional diameter only axial modes driven by the shear layer are observed. In the case of pulsatile inflow conditions an additional axial mode, driven by the pulsation frequency, is observed on top of the earlier observed asymmetric oscillatory mode. Particles are later added into this geometry and it is observed that most of the particles have a small residence time within the geometry under both continuous and pulsatile inflow conditions. This is attributed to the narrow recirculation regions within the idealized pipe-like geometry. A second geometry that is utilized in this study is the experimental corrugated pipe-like geometry. Particles are injected again under continuous andpulsatile inflow conditions for this geometry. Earlier experimental work on this experimental pipe prototype revealed that there is no significant difference in the particle distribution obtained using a straight pipe and an experimental pipe prototype. The numerical study conducted revealed that despite the presence of large recirculation zones, within the cavities of the pipe, most of the particles tended to pass through contracted regions of the pipe without entering the recirculation zones.