Modelling and Numerical Simulation of Dual Fuel Lean Flames Using Local Burning Velocity and Critical Chemical Timescale

摘要

Addition of hydrogen to hydrocarbons in premixed turbulent combustion is of technological interest due to their increased reactivity, flame stability and extended lean extinction limits. However, such flames are a challenge to reaction modelling, especially as the strong preferential diffusion effects modify the physical processes, which are of importance even for highly turbulent high-pressure conditions. In the present work, Reynolds-averaged Navier-Stokes (RANS) modelling is carried out to investigate pressure and hydrogen content on methane/hydrogen/air flames. For this purpose, four different subclosures, used in conjunction with an algebraic reaction model, are compared with two independent sets of experimental data: (1) Orleans data consists of pressures up to 9 bar, with addition of hydrogen content by up to 20% in hydrogen/methane mixture, for moderate turbulence intensities. 2) The Paul Scherrer Institute data includes same fuels with higher volume proportion of hydrogen (40%), at much higher turbulent intensities at 5 bar. The first model Model I is based solely on the increased reactivity of the hydrogen/methane mixture under laminar conditions. It shows that the increase of unstretched laminar burning velocity (S_(LO)) is not sufficient to describe the increased reactivity in turbulent situations. This non-corroboration proves the importance of preferential diffusion effects in highly turbulent flames. Models II and III are formulated based on the localized increase in S_(LO), local burning velocity which is a strong function of local curvature and flow strain. Model II overpredicts the reactivity for higher pressures. Model III accurately predicts for nearly all studied flame conditions. Model IV is based on the leading point concept (LPC) that the leading part of the turbulent flame brush is more important than the rear part of premixed flame with the Lewis number less than unity. This model in its present formulation underpredicts the average reaction rate compared with experiments.