The present research investigates the details of mixture preparation and combustion in a two-stroke, small-scale research engine with a numerical methodology based on large eddy simulation (LES) technique. A major motivation to study such small-scale engines is their potential use in applications requiring portable power sources with high power density. The investigated research engine has a rectangular planform with a thickness very close to quenching limits of typical hydrocarbon fuels. As such, the combustor has a high aspect ratio (defined as the ratio of surface area to volume) that makes it different than the conventional engines which typically have small aspect ratios to avoid intense heat losses from the combustor in the bulk flame propagation period. In most other aspects, this engine involves all the main characteristics of traditional reciprocating engines. A previous experimental work has identified some major design problems and demonstrated the feasibility of cyclic combustion in the high aspect ratio combustor. Because of the difficulty of carrying out experimental studies in such small devices, resolving all flow structures and completely characterizing the flame propagation have been an enormously challenging task. The numerical methodology developed in this work attempts to complement these previous studies by providing a complete evolution of flow variables. Results of the present study demonstrated strengths of the proposed methodology in revealing physical processes occurring in a typical operation of the high aspect ratio combustor. For example, in the scavenging phase, the dominant flow structure is a tumble vortex that forms due to the high velocity reactant jet (premixed) interacting with the walls of the combustor. LES gives the complete evolution of this flow structure, from its beginning to its eventual decay after the scavenging period is over. In addition, LES is able to predict the interaction between the bulk flow at top dead center (TDC) and the turbulent flame propagation. The success of this depends on the ability of the model in predicting turbulent flow structure including its length and velocity scales.
| Identifer | oai:union.ndltd.org:GATECH/oai:smartech.gatech.edu:1853/14134 |
| Date | 20 November 2006 |
| Creators | Kirtas̀§, Mehmet |
| Publisher | Georgia Institute of Technology |
| Source Sets | Georgia Tech Electronic Thesis and Dissertation Archive |
| Language | en_US |
| Detected Language | English |
| Type | Dissertation |
| Format | 27074244 bytes, application/pdf |
Page generated in 0.0631 seconds