Application and development of dissipation based combustion models for conventional and unconventional combustion processes

Abstract

Concepts for burner operation have manifolded under the influence of increasing performance demands for combustion processes. This is to support the global effort to reduce pollution and greenhouse gas emissions radically. The trend to more efficient, environment-friendly process design has also led to increasingly complex burneroperating conditions and the growing importance of finite-rate chemistry phenomena such as flame thickening, local extinction etc. Computational Fluid Dynamics has the potential to effectively complement experimental research to achieve a higher level of understanding of the combustion process. However, the complex turbulence-chemistry interaction in modern combustion processes challenges the fundamental assumption of (infinitely) thin reaction zones (flamelets), which many turbulent combustion models build upon. The focus of this thesis is the application and development of dissipation-based combustion models that are capable of including finite-rate chemistry effects and relaxing the limiting thin flame assumption. One merit of dissipation-based models with finite-rate chemistry is their flexible applicability under a broad range of flame and flow conditions. The first part of the thesis deals with the application of the well-known Eddy dissipation concept (EDC) with finite-rate chemistry calculation to turbulent flames under varying conditions. Following some preliminary studies of the EDC in the RANS framework for conventional combustion, the main work was the modelling of a lab-scale MILD burner using the EDC with Large eddy simulation (LES). The aims of the corresponding Paper I are the evaluation of the EDC and a second dissipation-based combustion model, the Partially stirred reactor model (PaSR) in LES using measurements for the MILD burner, their direct comparison in terms of modelling performance and the discussion of the observed reacting flow to gain insights into this operational mode. An important conclusion from this study is that dissipation-based combustion models show competitive performance in predicting the reacting flow under MILD conditions when compared to other modelling approaches in the literature. Potential challenges arise, however, for the choice of proportionality constants in the context of LES, which is especially relevant for the EDC. An algebraic dissipation-based combustion model was developed and evaluated in a second comprehensive study within the thesis. The new combustion model addresses some of the challenges observed in the previous part. Motivations were to reduce computational expenses, improve compatibility with LES theory, relax presumptions on the flame structures, and avoid the necessity to adjust proportionality constants. The key output of this study is a new combustion model evaluated using measurements and numerical results from sophisticated, well-documented combustion models found in the literature for two different premixed flames. Papers II and III provide promising results concerning the applicability of the new model to complex premixed reacting flows showing competitive performance to capture moderate flame thickening and local extinction. Another advantage of the model is the strong coupling between the modelled heat release and reaction rate. Following the successful application of the new dissipation-based combustion model to two premixed air-methane flames, the results of a preliminary study investigating an oxy-methane flame are shown. Compared to the previous studies, the complexity increased considerably due to the non-unity Lewis number of the fuel in the investigated O2/CO2 oxidiser, the altered chemical activity, and the changed thermal behaviour of the oxidiser. Satisfactory results were achieved for predicting the reacting flow field. However, they also indicated that the characteristics of the oxy-fuel set-up need to be better taken into account by the modelling approach.