Development of the chemical kinetic mechanism and modeling study on the ignition delay of liquefied natural gas (LNG) at intermediate to high temperatures and high pressures

•Develop a chemical kinetic mechanism for liquefied natural gas combustion at low temperature conditions.•Determine the correlation between ignition delay times and methane number.•Increasing the methane amount in the mixtures reduces the reactivity of the mixtures that leads to a higher ignition delay time.•The activation energy of different liquefied natural gas mixtures are nearly identical, while each reference mixtures has a different activation energy from other reference mixtures and the liquefied natural gas mixtures. In the past decade, liquefied natural gas (LNG) has attracted attention as a sustainable energy resource. The composition of LNG is one of the primary factors that has an influence on the engine knock and determines the fuel reactivity and the heat release in the combustion processes due to the presence of C1-C5 alkanes. The knocking behavior of LNG in gas engines is derived by the methane number (MN) which is directly related to the composition of the mixtures and the operating conditions. The determination of the methane number therefore demands a comprehensive characterization of a fuel mixture towards its auto-ignition properties such as ignition delay time (IDT) based on chemical kinetics of the fuel oxidation. In this study, a detailed chemical kinetic mechanism was developed to discover a correlation between the MNs and IDTs. The comparison of the modeling and experiments results of the ignition of different LNG mixtures containing C1-C5 alkanes and N2, and reference mixtures containing only CH4 and H2 with known MN were carried out in the temperature range 850–1550 K, at pressures of 10, 20 and 40 bar, and equivalence ratios (ϕ) of 0.4 and 1.2. The newly developed model demonstrates very good performance in simulating the IDT of various LNG mixtures, while the improvement in reduction of the maximal discrepancies between model and experiments is from factor 7–10 to 1–2. Moreover, the modeling studies provided a deeper understanding of the oxidation chemistry of the investigated mixtures.