The Knock Propensity of Carbon Dioxide-Containing Natural Gases: Effect of Higher Hydrocarbons on Knock-Mitigating Influence of Carbon Dioxide

To assess the effect of the presence of carbon dioxide (CO₂) in natural gases on the knock resistance of fuel, the knock behavior of a lean-burn, high-speed medium Brake Mean Effective Pressure (BMEP) Combined Heat and Power (CHP) engine fueled with CH₄ + 8 mole% C₃H₈ mixtures. The engine experiments are supplemented with ignition measurements and simulations of ignition and cylinder processes for various fuel compositions. The engine results show that increasing the fraction of CO₂ results in an increase in knock resistance. The analysis of simulations of cylinder processes shows that for binary mixtures (CH₄/CO₂) and ternary mixtures (CH₄/C₃H₈/CO₂) the increase in knock resistance with increasing CO₂ fraction is caused by the reduction in peak pressure/temperature, which consequently increases the autoignition delay time of the mixture. For the ternary mixtures, the presence of 8 mole% propane in the fuel mixture partially compensates the reduction in peak pressure/temperature by decreasing the autoignition delay observed in the Rapid Compression Machine (RCM) measurements and ignition calculations. In contrast, for binary mixtures, the autoignition delay time increases with increasing CO₂ content. Consequently, the effect of increasing fractions of CO₂ in methane results in a substantially larger increase in the knock resistance as compared to the CH₄/C₃H₈ mixtures studied.

The characterization of the changes in knock resistance using the Propane Knock Index Methane Number (PKI MN) shows an agreement with the Knock-Limited Spark Timing (KLST) measurements within the experimental uncertainty. Traditional methods for computing an MN, such as AVL and MWM, base the computation for CO₂-containing gases on extrapolation of measured data for binary CH₄/CO₂ mixtures. As a result, these methods strongly overpredict the change in MN when CO₂ is added to the propane-containing fuel, underestimating the potential risk for an engine. This result shows that caution is advised when extrapolating the binary CH₄/CO₂ data to gas mixtures containing higher hydrocarbons.