A comprehensive study on laminar burning velocity and flame stability of oxy-producer gas mixtures. Part-2: Laminar burning velocity and Markstein length analysis

Abstract

Part-1 reported the extent of variation in each species in producer gas (PG) mixture(s) after a thorough literature study. Forty-five PG variants were identified to analyse the effect of each species on the combustion characteristics of PG-O2 mixtures. Spherical flame method was used for measuring laminar burning velocities and burned gas Markstein lengths of PG at 1 bar, 300 K and ϕ = 0.8 in a 4.5 L spherical chamber. Extrapolation (linear and non-linear) and non-extrapolation based schemes were used to estimate unstretched flame speed. LBV extracted from all three schemes were in good agreement within ±4.92%. Only H2/CO-rich-PG-O2 mixtures having mole fraction of combustible components (CC) less than 20% displayed negative Lb values and their effective Lewis numbers were less than the respective critical values and were corroborated with experimental results. CH4-rich-PG-O2 mixtures were unconditionally stable. Their LBV was the highest at most conditions due to high flame temperatures, with exception to LBV of H2-N2-rich-PG-O2 mixtures at 45%CC owing to large thermal diffusivities. LBV varied non-linearly as CO2 in non-combustible components (NCC) increased beyond 50% and its suppressing effect was the highest in CH4-rich-PG. Simulations were performed using 1D expanding spherical flame model in COSILAB with GRI-Mech3.0. Simulated LBV showed a deviation of ±13.2% with experimental values at 45%CC. Unstretched LBV obtained from non-extrapolation-mapping method which involved simulation results showed good agreement with experiments. Combustion of PG with higher H2, CO and N2 content and relatively lesser amounts of CH4 and CO2 were stable to preferential diffusion effects and predictable to fluctuations in composition.