Embed Size (px)
Citation preview
GhulamMustafaMajal
The project aims to find methods for stimulating agglomeration of particles in internal combustion engine exhaustsystems by manipulation of the hydrodynamic and acoustic fields. An understanding of the evolution of particulatecharacteristics (e.g. particle mass, particle number, size distribution) as traveling along the exhaust system componentsand connections with relevance to modern engines is needed. The enhancement of the aforementioned understandingwill be developed with the usage of advanced computational tools and models.
Introduction and Motivation:Modern internal combustion engines show a tendency to form fineparticles that are prone to penetrate through the humanrespiratory system and cause cardiovascular as well as neurologicalproblems. This project aims at deriving predictive models for themanipulation of particles in the exhaust line through the use offorced pulsatile flow conditions and/or acoustic fields.
Setup:Preliminary simulations of particle agglomeration process have been firstperformed using simplified 1D models. The purpose was to carry outsensitivity studies for analyzing the effect of different flow andgeometrical parameters on particle agglomeration. The model considersan oscillating Stokesian flow in a corrugated pipe of particulardimensions. An additional simplification has been made by only takinginto account the drag force that acts on the particle.
The particle grouping process is characterised using two non-dimensionalparameters with this 1D model. The first (beta parameter) compares themean flow field with the amplitude of the flow oscillations. The secondnon-dimensional parameter (alpha parameter) encapsulates the inertialeffects due to the size of the particle.
In the second step, the simulations will be extended to 3D set-ups basedon CFD. Particular geometries and flow forcing conditions found topromote particle agglomeration will be investigated. Preliminaryverification and validation studies for the CFD solver have been carriedout using a benchmark case (flow over a backwards facing step).
Results:Starting from the model equation for the velocity of the hot gas, sinusoidal flow oscillations in time due to the engine pulses modulated by anexhaust pipe with periodically varying cross-section (period length𝜆) are considered. Grouping/non-grouping maps have been generated forvarious RPM and 𝜆. The left-most figures below show such maps at 5200 RPM and four different 𝜆. The right-most figure shows comparisonsbetween CFD data and experiments for the 3D backwards facing step flow study. The normalized mean streamwise velocity profiles at differentlocations downstream of the step are depicted, where H is the step size and 𝑈$ is the centerline velocity.
Summary and Conclusion:Tests have been carried out extensively to understand the impactof the aforementioned non-dimensional parameters on particlegrouping using a simple 1D model. Conditions have been identifiedwhere particle agglomeration can occur. Extensions of these verysimplified 1D studies will be made towards 3D flow – particlesimulations. Presently the 3D flow solver has been verified andvalidated using experimental data by Fessler et al. 1999. Along thisline, a much more realistic pipe geometry will be used to identifythe impact of flow oscillations (due to both engine pulses andgeometry constrictions) on particle agglomeration.
Acknowledgement:This project is supervised by Mihai Mihaescu (KTH-Mechanics),Mats Åbom (KTH-MWL) and Mikael Karlsson (KTH-MWL).The support received from the Swedish Energy Agency, Scania AB,Volvo GTT, BorgWarner, and Volvo Cars is highly appreciated.PartofsimulationswereperformedonresourcesprovidedbytheSwedishNationalInfrastructureforComputing(SNIC)atPDCCentreforHighPerformanceComputing(PDC-HPC).
Controlofparticleagglomerationwithrelevancetoafter-treatment
gasprocesses
