Cavalcanti Miranda, Flavia (2019)
Large Eddy Simulation of Turbulent Reacting Flows With Radiative Heat Transfer.
Technische Universität Darmstadt
Dissertation, Erstveröffentlichung

Kurzbeschreibung (Abstract)

Combustion is the most common method of energy conversion constituting about 85 % of the primary energy consumption. However, combustion is responsible for emissions of CO, NOx , CO2 , soot and others pollutants. Hence, having the expected increase of the global energy demand in mind as well as the challenge of a global reduction in the greenhouse gases emissions and the current difficulties in developing renewable energy sources for a foreseeable future, combustion science will continue being a very important topic and a very active field of technology. Moreover, since the parameters involved in combustion systems are affected by heat transfer, the understanding and development of mathematical models for analyses of heat transfer is crucial. Studying such flows is not a straightforward task because of the high nonlinear interaction among the involving processes, which include chemical kinetics, turbulence and thermal radiation. In this context, Large Eddy Simulation (LES) is an outstanding approach and has become a common model to deal with such complex flows. In this approach, the instantaneous form of the governing equations are filtered. As a consequence of the filter procedure, unclosed terms appear that correspond to effects of Turbulence-Chemistry Interactions (TCI) and Turbulence Radiation Interactions (TRI). The importance of the TCI has long been recognized and it is a very active research topic in the combustion community. Furthermore, the relevance of TRI has been gaining recognition but just a few LES studies considered their effects of TRI. This thesis deals with the simulation of turbulent flames by taking into account radiative heat transfer. The focus of this work is on the development and application of a radiation solver for computing turbulent reacting flows. In this study, the role of the TRI in the context of LES is analyzed for two important and widely investigated configurations: Sandia flame D and bluff-body stabilized non-premixed flame. The radiation solver was implemented by considering the complete radiative transfer equation, including the emission, absorption and scattering terms. The finite volume method, which is a variation of the Discrete Ordinates Method (DOM), was applied to discretize this equation. To account for the spectral behavior of the combustion gases involved, the Weighted Sum of Gray Gases (WSGG) method is used. To include thermal radiation in the LES framework, the filtered radiative source term should be computed. The contribution of the resolved scales can be explicitly calculated, whereas the terms involving the subgrid-scale contributions are unclosed and require approximations. The Optically Thin Fluctuation Assumption (OTFA) is applied for approximating the filtered absorption term and the Eulerian Stochastic Field (ESF) method is employed for representing the emission TRI. Following, the importance of considering the subgrid-scale contributions is analyzed. For this aim, simulations are performed by considering and neglecting these contributions. The Sandia flame D, bluff-body flame and their corresponding four times scaled flames are the configurations studied here for analyzing the TRI. The scaled flames are additionally investigated in order to have a more pronounced radiation effect. For all cases, the difference between the radiative source term computed by accounting and neglecting the subgrid-scale contribution is not significant, which indicates that considering these terms is not important in the context of LES. Additionally, the radiative source term is computed with the mean fields of temperature and species concentrations to show the difference to Reynolds Averaged Navier-Stokes equations (RANS), since this procedure corresponds to compute the radiative source term in a RANS framework. As expected, the results for this case presented a significant difference to the remaining procedures, which demonstrates that considering the subgrid-scale contributions are relevant for RANS simulations.

Frage zum Eintrag

Redaktionelle Details anzeigen