Natural gas is a promising alternative fuel which can be used in internal combustion engines to achieve low carbon emission and high thermal efficiency. However, at high compression ratio, super-knock due to detonation development might occur. In this study, the autoignitive reaction front propagation and detonation development from a hot spot were investigated numerically and the main component of natural gas, methane, was considered. The objective is to assess the performance of different kinetic models in terms of predicting hot spot–induced detonation development in methane/air mixtures. First, simulations for the constant-volume homogeneous ignition in a stoichiometric methane/air mixture were conducted. The ignition delay time, excitation time, critical temperature gradient, thermal sensitivity and reduced activation energy predicted by different kinetic models were obtained and compared. It was found that there are notable discrepancies among the predictions by different kinetic models. Then, hundreds of one-dimensional simulations were conducted for detonation development from a hot spot in a stoichiometric CH4/air mixture. Different autoignition modes were identified and the detonation regimes were derived based on the peak pressure and reaction front propagation speed. It was found that even at the same conditions, different propagation modes can be predicted by different kinetic models. The broadest detonation development regime was predicted by the reduced GRI mechanism, while a relatively narrow regime was predicted by the recent kinetic models such as FFCM-1 and Aramco 3.0. The present results indicate that super-knock prediction strongly depends on the kinetic model used in simulations. Therefore, significant efforts should be devoted to the development and validation of kinetic models for natural gas at engine conditions.

No abstract

In explosion accidents, inert layer(s) can be used to dampen or suppress detonation propagation. In detonation engines, the detonation may propagate in an inhomogeneous mixture with inert layer(s). Here, the detonation propagation in hydrogen/oxygen/nitrogen mixtures with a single inert layer normal to the detonation propagation direction was investigated. Six hydrogen/oxygen/nitrogen mixtures with different amounts of nitrogen dilution and at different initial pressures were considered. The emphasis was placed on assessing the effects of nitrogen dilution and pressure on detonation across an inert layer. It was found that successful detonation reinitiation occurs only when the inert layer thickness is below some critical value. The detonation reinitiation process was analyzed. The interactions of transverse waves, the reactive–inert layer interface, and instabilities jointly induced local autoignition/explosions and detonation reinitiation. Counterintuitively, it was found that a thicker inert layer is required to quench a weaker detonation (with more nitrogen dilution or with lower-energy density at lower pressure). With the increase of nitrogen dilution or the decrease of initial pressure, the induction length and cell size of the detonation became larger, which unexpectedly resulted in the larger critical inert layer thickness.

Laminar flame speed (LFS) is used as an important input in certain turbulent premixed combustion modelling for spark ignition engines. At engine-relevant temperatures and pressures, the LFS is difficult to be measured and it is usually obtained from calculations using detailed or reduced chemical mechanisms. However, at very high temperature in the range of 700~1000 K, it is difficult to get the converged solutions during LFS calculation. In this study, the LFS at enginerelevant conditions are obtained from simulating propagating spherical CH 4 /air flames in a closed spherical vessel. The results from different chemical mechanisms are obtained and compared. It is found that there is significant difference between the LFSs predicted from different chemical mechanisms, especially at high temperatures and pressures. It is also shown that the reduced mechanism has very good performance in terms of predicting the LFS at engine-relevant conditions. Besides, the performance of difference chemical mechanisms in terms of predicting the ignition delay time and the excitation time is assessed and significant difference is also observed.