Skeletal Kinetic Mechanism Generation and Uncertainty Analysis for Combustion of Iso-octane at High Temperatures

A detailed mechanism for combustion of iso-octane with 116 species and 754 reactions has been reduced using a directed relation graph with error propagation (DRGEP) and DRGEP with sensitivity analysis (DRGEPSA) methods under high-temperature conditions. Two skeletal mechanisms, i.e., a 63-species mechanism with a maximum error of 7.2% and a 51-species mechanism with a maximum error of 28.5% on autoignition delay times have been generated. These two skeletal mechanisms are shown to reproduce ignition delays, laminar flame speeds, species and temperature profiles in good agreement with those of the detailed mechanism. Uncertainty in the ignition predictions by detailed and two skeletal mechanisms induced by the uncertainties in reaction rate coefficients has been studied. Probability distribution of autoignition predictions demonstrated that the 63-species mechanism can still keep the uncertainty characteristics, while the 51-species mechanism has significant discrepancy compared with the detailed one. Further analysis of autoignition shows that the structure and integrality of the reaction system in the 51-species mechanism has changed. Global sensitivities of 63-species and detailed mechanisms on ignition have been investigated using the high-dimensional model representation (HDMR) method. The highly important reactions for ignition in the detailed mechanism are the same as those in the 63-species mechanism, and sensitivity coefficients of the listed reactions agree well with each other. The most important reactions in the first-order sensitivity on autoignition in the detailed mechanism are the same as those in the 63-species mechanism, especially for the five most important reactions. The most important 10 reactions contribute almost 75% to the overall variance in ignition delay under the present conditions, while the second-order effects are quite small and almost negligible. The top ranked reactions show that small-molecule chemistry (C-0-C-4) contributes significantly to uncertainties in the ignition predictions at high temperatures.