OCR Text |
Show instantaneous semiquantitative snapshot of the spatial distribution throughout the flame. Most popular for such purposes are the OH and CH radicals. Such measurements have provided a great deal of insight into turbulent flows, a subject of considerable interest in practical combustion systems. However, a turbulent flame is far too complex to gain a quantitative understanding of the combustion chemistry itself. Instead, chemical details should be investigated in a much simpler system, a laminar flame. Once established in this manner, and incorporated into a flame chemistry model, the chemical description can be used as needed (often condensed) in a separate model of the turbulent flows. Laminar flame investigations employing LIF are often undertaken at low pressure, to achieve a high degree of spatial resolution of the experimental results, and thereby a more stringent test of the combustion chemistry models. Finally, LIF can be used in a semiquantitative way.4 The mere identification of some compound in a combustion system can reveal a good deal of information about the chemical mechanism. Examples are the NS radical in flames containing surrogates of both fuel nitrogen and sulfur, linking NOx and SOx chemistry, and the discovery of formaldehyde in an internal combustion engine, confirming ideas concerning the chemistry of early ignition. We adopt a two-part experimental philosophy for understanding the chemistry of NO in practical combustion systems, as described below. First are detailed measurements in fully premixed low pressure flames, against which combustion chemistry models may be tested. Second are less detailed but rigorous studies of flame structure in a annospheric pressure, partially premixed flame. TIlls will provide data for the testing of two dimensional models incorporating realistic ~ows and chemistry. Both systems are studied by LIF. Combustion Chemistry Models The future will see the addition of detailed chemical kinetics to flame models of many dimensions, ultimately including turbulent flows. At the present time, only laminar flow codes can include large chemical mechanisms, which are responsible for a large ~ction of the time needed for computation. Numerical integration packages which include full chemistry are routinely available for one dimensional flows and progress is well underway for the two dimensional case. Sensitivity analysis allows the modeller to examine some prediction (e.g., NO concentration at some point in the flame) as a function of input variables such as rate coefficients. This enables meaningful physical interpretation of the output, places confidence limits on extrapolation to conditions pertinent to other applications, and pinpoints the causes of uncertainty in the models. 3 |