Correlation of combustor lean blowout performance to supercritical pyrolysis products

摘要

Studies to assess correlations between fuel pyrolysis products and fuel-to-air equivalence ratio at combustor lean blowout (LBO) were conducted. Eighteen fuels, including conventional and alternative fuel formulations and blends, single compounds and binary surrogates of hydrocarbon compounds, were studied. The pyrolysis tests consisted of thermally cracking the fuels in a flow reactor at supercritical conditions of 4.14 MPa and up to 625 degrees C, at varying residence times. The liquid and gaseous pyrolysis products were analyzed via gas chromatography to correlate yields to the fuel LBO data acquired in a previous study on a single-nozzle swirl-stabilized combustor. Focus was given to correlate LBO with primary decomposition products, since it is hypothesized that fuel combustion characteristics, particularly ignition and flame extinction, are related to the rate and type of products formed upon initial thermal decomposition of the parent fuel. Therefore, only the range of conversions for which low molecular weight products had a near-constant formation rate was considered. Results show that the highly branched fuels primarily decomposed to branched C-4 and methane, while fuels with high concentration of linear alkanes show greater yields of C-2-C-4 alkanes and 1-alkenes. Regression analyses of the pyrolysis gas and LBO data show very strong correlations (R-2 > 0.90) for several pyrolysis products for the conventional and alternative fuels, with ethane, ethylene and iso-butylene yields providing the best correlations to LBO. The calculated LBOs based on regression parameters and concentrations of the pyrolysis gases were within 2% of the measured LBO for most fuels tested. These initial results suggest that despite the vast differences in fuel reaction timescales in the flow reactor (similar to 0.2-11 s) and combustor environments (< 0.5 ms), fuel pyrolysis species in a flow reactor environment are relevant and may be used to predict fuel combustion behavior. If further validated, this method provides a capability to identify fuel components that can improve combustion performance via their primary pyrolytic product slates, potentially providing a more cost-effective alternative to combustion testing for determination of fuel combustion performance metrics. Results of the pyrolysis tests of all fuels, and future efforts are discussed.