| OCR Text |
Show Models for emissions from practical burners do not currently exist, primarily because the flow fields are turbulent. The interaction of chemistry and turbulence is a current research topic in combustion science. Despite the fact that the chemical reaction mechanisms are often known, the task of including them in a turbulent flow simulation has not been accomplished. Conventional turbulence models that solve ~ time-averaged equations with statistical closure models are computationally very expensi ve -- even without considering a reacting flow. Furthermore, the closure models may be deficient in capturing the important physics of the entrainment and mixing. As a result, our general understanding of NOx and other emissions has relied upon observations of laboratory flames. Emissions data are only recently being generated, beginning with relatively simple, vertical jet-flames. Chemical kinetic models for emissions from these flames are needed to develop a better understanding of the experimental observations. Models are important in interpreting the data to distinguish the effects of physical phenomena. Chemical issues relating to local non-equilibrium can only be clarified by models that include the elementary kinetics. Demonstration of models not only confirms our understanding of the experiments, but provides a tool that can be extended to more practical uses, such as evaluating and designing practical burners. Since it is not currently possible to use both detailed kinetic mechanisms and a reliable turbulence model- such as large eddy simulation or direct numerical simulation -- one or the other must be simplified. Many investigators have used complex turbulence models, but reduced the chemistry model using simplifications such as partial equilibrium assumptions or reduced reaction mechanisms (see, for example, Whitelaw and Jones1, Chen, et al.2-3 or Pope4). While these models have found some success in predicting global behavior of flames, they cannot include the complex kinetics necessary for predicting pollutant formation. Similarly, most studies of detailed chemistry consider simplified fluid mechanical systems such as laminar flames or stirred reactors. In the present paper, we outline the classes of technical approaches that are often employed to address these complex modeling problems. Through these simplification strategies, it is possible to extract a very large amount of useful information that can lead to better combustion system design, but it is always important to fully understand the limitations inherent in the simplifying assumptions. The paper presents two examples: First, formaldehyde and carbon monoxide results from a perfectly stirred reactor model using refinery fuel gas. Secondly, predictions of NO, NO~ and CO emissions from a H2-CO turbulent jet flame using a model consisting of two stirred reactors. NUMERICAL MODELS Stirred reactor model A perfectly-stirred reactor is a highly simplified physical model, familiar to most chemical engineers, in which the fuel and oxidizer are assumed to mix very -2- |
