We systematically reevaluated the rates and corresponding uncertainties of H + O2 = O + OH (R1) based on our latest H2-O2 kinetic mechanism [named as Yang’s mech, Xueliang Yang et al., Journal of Physical Chemistry A, 2021, 125, 10223-10234] using reported shock tube experimental data from Ryu et al. (1995), Yuan et al. (1991), Masten et al. (1991) and Hong et al. (2011). The response of matching was altered to be the parameters like the maximum slope of the radical profile which was more robust with interference of the impurities regardless of the initial process. The chain-termination reaction of H + O2 ( + M ) =HO2 ( + M ) (R2) was reevaluated as well using Yang's mech for further constraint of the uncertainty of data interpretation. The weighted least-square fitting on all the experimental data with the updates on specific rates and uncertainties was conducted, which came up with new rate expressions of R1 and R2 and updated Yang's mech.
Simulations with above updated kinetic mechanism were performed to obtain the fundamental combustion characteristics, e.g. laminar flame speed (Mass burning rate), ignition delay time, profiles of species, etc. The results were compared with the experiments and the predictions of other mechanism, like HP Mech, Aramco Mech 3.0 and FFCM-1. The predictions of the updated mechanism at the same conditions are mostly within the error range of experimental data of mass burning rate.
The experiments at cryogenic temperatures were performed by the Fellow’s students at his home institution. The tests not only extend the observation of hydrogen explosion at extreme conditions, but also give broader targets for numerical model validation. Strong flame acceleration (FA), fast flame and the subsequent galloping detonation of hydrogen-oxygen mixtures in tube are investigated with different initial pressures at cryogenic temperature of 77 K.
Both 2D and 3D numerical model of flame propagating in closed tubes were built on the open-source platform OpenFOAM. The combustion solver XiFoam developed based on the flame wrinkling combustion model was selected. The large eddy simulation (LES) model can be computationally inexpensive while reproducing well the size and velocity of the annular vortex created before the propagating flame front and the evolution of the flame shape and structure. For large eddy simulation (LES), XiFoam solves the Navier-Stokes conservation equations of mass, momentum, and energy for unsteady compressible and reactive flows. The time integral was the first-order Euler format. For the gradient term, diffusion term, and Laplace, the Gaussian second-order linear scheme was used. The interpolation Schemes sub-dictionary contains terms that are interpolations of values typically from the grid body center to the grid front face center. The setting of Interpolation Schemes was linear as default in OpenFOAM. The PIMPLE algorithm was used to solve the discretization equation. The pressure equation and the velocity correction equation were iterated 20 times, and the pressure residual (tolerance) was 1×10-6.
