| description abstract | Alkanes such as methane, ethane, and propane make up a large portion of most natural gas fuels. Natural gas is the primary fuel used in industrial gas turbines for power generation. Because of this, a fundamental understanding of the physical characteristics such as the laminar flame speed is necessary. Most importantly, this information is needed at elevated pressures to have the most relevance to the gas turbine industry for engine design. This study includes experiments performed at elevated pressures, up to 10 atm initial pressure, and investigates the fuels in a pure form as well as in binary blends. Flame speed modeling was done using an improved version of the kinetics model that the authors have been developing over the past few years. Modeling was performed for a wide range of conditions, including elevated pressures. Experimental conditions include pure methane, pure ethane, 80/20 mixtures of methane/ethane, and 60/40 mixtures of methane/ethane at initial pressures of 1 atm, 5 atm, and 10 atm. Also included in this study are pure propane and 80/20 methane/propane mixtures at 1 atm and 5 atm. The laminar flame speed and Markstein length measurements were obtained from a high-pressure flame speed facility using a constant-volume vessel. The facility includes optical access, a high-speed camera, a schlieren optical setup, a mixing manifold, and an isolated control room. The experiments were performed at room temperature, and the resulting images were analyzed using linear regression. The experimental and modeling results are presented and compared with previously published data. The data herein agree well with the published data. In addition, a hybrid correlation was created to perform a rigorous uncertainty analysis. This correlation gives the total uncertainty of the experiment with respect to the true value rather than reporting the standard deviation of a repeated experiment. Included in the data set are high-pressure results at conditions where in many cases for the single-component fuels few data existed and for the binary blends no data existed prior to this study. Overall, the agreement between the model and data is excellent. | |
