Detailed chemical kinetic reaction mechanisms are developed for all nine chemical isomers of heptane (C--H16), following techniques and models developed previously for other smaller alkane hydrocarbon species. These reaction mechanisms are tested at high temperatures by computing shock tube ignition delay times and at lower temperatures by simulating ignition in a rapid compression machine. Although the corresponding experiments have not been reported in the literature for most of these isomers of heptane, intercomparisons between the computed results for these isomers and comparisons with available experimental results for other alkane fuels are used to validate the reaction mechanisms as much as possible. Differences in the overall reaction rates of these fuels are discussed in terms of differences in their molecular structure and the resulting variations in rates of important elementary reactions. Reaction mechanisms in this study are works in progress and the results reported here are subject to change, based on model improvements and corrections of errors not yet discovered.

Detailed chemical kinetic reaction mechanisms are developed for combustion of all nine isomers of heptane (C--H16), and these mechanisms are tested by simulating autoignition of each isomer under rapid compression machine conditions. The reaction mechanisms focus on the manner in which the molecular structure of each isomer determines the rates and product distributions of possible classes of reactions. The reaction pathways emphasize the importance of alkylperoxy radical isomerizations and addition reactions of molecular oxygen to alkyl and hydroperoxyalkyl radicals. A new reaction group has been added to past models, in which hydroperoxyalkyl radicals that originated with abstraction of an H atom from a tertiary site in the parent heptane molecule are assigned new reaction sequences involving additional internal H atom abstractions not previously allowed. This process accelerates autoignition in fuels with tertiary C-H bonds in the parent fuel. In addition, the rates of hydroperoxyalkylperoxy radical isomerization reactions have all been reduced so that they are now equal to rates of analogous alkylperoxy radical isomerizations, significantly improving agreement between computed and experimental ignition delay times in the rapid compression machine. Computed ignition delay times agree well with experimental results in the few cases where experiments have been carried out for specific heptane isomers, and predictive model calculations are reported for the remaining isomers. The computed results fall into three general groups; the first consists of the most reactive isomers, including n-heptane, 2-methyl hexane and 3-methyl hexane. The second group consists of the least reactive isomers, including 2,2-dimethyl pentane, 3,3-dimethyl pentane, 2,3-dimethyl pentane, 2,4-dimethyl pentane and 2,2,3-trimethyl butane. The remaining isomer, 3-ethyl pentane, was observed computationally to have an intermediate level of reactivity. These observations are generally consistent with knocking tendencies of these isomers, as measured by octane ratings, although the correlations are only approximate.

Using CARM (Computer Aided Reduction Method), a computer program that automates the mechanism reduction process, a variety of different reduced chemical kinetic mechanisms for ethylene and n-heptane have been generated. The reduced mechanisms have been compared to detailed chemistry calculations in simple homogeneous reactors and experiments. Reduced mechanisms for combustion of ethylene having as few as 10 species were found to give reasonable agreement with detailed chemistry over a range of stoichiometries and showed significant improvement over currently used global mechanisms. The performance of reduced mechanisms derived from a large detailed mechanism for n-heptane was compared to results from a reduced mechanism derived from a smaller semi-empirical mechanism. The semi-empirical mechanism was advantageous as a starting point for reduction for ignition delay but not for PSR calculations. Reduced mechanisms with as few as 12 species gave excellent results for n-heptane/air PSR calculations but 16-25 or more species are needed to simulate n-heptane ignition delay.

In general, combustion is a spatially three-dimensional, highly complex physi co-chemical process oftransient nature. Models are therefore needed that sim to such a degree that it becomes amenable plify a given combustion problem to theoretical or numerical analysis but that are not so restrictive as to distort the underlying physics or chemistry. In particular, in view of worldwide efforts to conserve energy and to control pollutant formation, models of combustion chemistry are needed that are sufficiently accurate to allow confident predic tions of flame structures. Reduced kinetic mechanisms, which are the topic of the present book, represent such combustion-chemistry models. Historically combustion chemistry was first described as a global one-step reaction in which fuel and oxidizer react to form a single product. Even when detailed mechanisms ofelementary reactions became available, empirical one step kinetic approximations were needed in order to make problems amenable to theoretical analysis. This situation began to change inthe early 1970s when computing facilities became more powerful and more widely available, thereby facilitating numerical analysis of relatively simple combustion problems, typi cally steady one-dimensional flames, with moderately detailed mechanisms of elementary reactions. However, even on the fastest and most powerful com puters available today, numerical simulations of, say, laminar, steady, three dimensional reacting flows with reasonably detailed and hence realistic ki netic mechanisms of elementary reactions are not possible.

A detailed chemical kinetic reaction mechanism is used to study the oxidation of n-heptane under several classes of conditions. Experimental results from ignition behind reflected shock waves and in a rapid compression machine were used to develop and validate the reaction mechanism at relatively high temperatures, while data from a continuously stirred tank reactor (cstr) were used to refine the low temperature portions of the reaction mechanism. In addition to the detailed kinetic modeling, a global or lumped kinetic mechanism was used to study the same experimental results. The lumped model was able to identify key reactions and reaction paths that were most sensitive in each experimental regime and provide important guidance for the detailed modeling effort. In each set of experiments, a region of negative temperature coefficient (NTC) was observed. Variation in pressure from 5 to 40 bars were found to change the temperature range over which the NTC region occurred. Both the lumped and detailed kinetic models reproduced the measured results in each type of experiments, including the features of the NTC region, and the specific elementary reactions and reaction paths responsible for this behavior were identified and rate expressions for these reactions were determined.

Several reduced chemical kinetic mechanisms for combustion of ethylene and n-heptane have been generated using CARM (Computer Aided Reduction Method), a computer program that automates the mechanism-reduction process. The method uses a set of input test problems to rank species by the error introduced by assuming they are in quasi-steady state. The reduced mechanisms have been compared to detailed chemistry calculations in simple homogeneous reactors and experiments. Reduced mechanisms for combustion of ethylene having as few as 10 species were found to give reasonable agreement with detailed chemistry over a range of stoichiometries. Much better agreement with detailed chemistry was found for ethylene ignition delay when the reduced mechanism was tuned through selection of input test problems. The performance of reduced mechanisms derived from a large detailed mechanism for n-heptane was compared to results from reduced mechanisms derived from a smaller semi-empirical mechanism. The semi-empirical mechanism was clearly advantageous as a starting point for reduction for ignition delay, but the differences were not as notable for perfectly-stirred-reactor (PSR) calculations. Reduced mechanisms with as few as 12 species gave excellent results for n-heptane/air PSR calculations but 16-25 or more species are needed to simulate n-heptane ignition delay.