| OCR Text |
Show INTRODUCTION Improved quantitative knowledge of the detailed chemistry of natural gas flames and ignition, including the processes responsible for formation of NOx and other pollutants, is essential to enable the gas industry to develop a deeper understanding of natural gas combustion. This knowledge is needed in order to use computer modeling methods confidently enough to describe combustion processes outside of the laboratory. With reliable chemistry, transport and fluid dynamics models, the industry could design equipment able to meet increasingly stringent demands for maintenance of environmental quality and to improve the efficiency of combustion. With this goal in mind the Gas Research Institute established a fom-Iaboratory collaboration to develop a comprehensively optimized chemical model of the high temperature combustion chemistry of natural gas. The approach we have employed begins with a comprehensive evaluation of the relevant thermochemistry, chemical mechanism, and temperatme and pressure dependent rate constants, and employs a systematic optimization procedure to produce a mechanism that provides good model predictability of experimental results from the literature data base of high temperature combustion research. This data base includes flame speeds, shock tube ignition delays, and species profiles in shock tube experiments and low pressure flames. The optimization procedure was as follows. We started with a data base of critically evaluated elementary reaction rate constant information from the literature together with a set of modeling targets selected so as to force the model to account for the phenomena observed under all combustion conditions where reliable quantitative measurements are possible. Detailed sensitivity studies were then undertaken for the purpose of identifying target data sets for automated optimization. Once these sets had been selected, response surfaces were generated describing in polynomial forms the relationships between rate parameters and simulated target responses. Each required the solution of a large set of ordinary or partial differential equations as appropriate for the target flame or ignition process. The objective function combining these response smfaces and the optimization target data was then explored in the areas around minima corresponding to optimal matches between model and experiment and appropriately subjected to constrained parameter optimization. Finally, extensive validation studies were performed to see how well the model behaved when tested against experimental results that had not been included in the optimization. In this report we describe the results of a comprehensive parameter optimization of the chemical reaction mechanism describing high temperature methane combustion, GRI- 2 |
