The transition to carbon-neutral energy has renewed interest in hydrogen as a gas turbine fuel in the form of fuel blends with hydrocarbons. However, the distinct fluid properties and chemical kinetics of hydrogen and hydrocarbon blends necessitate redeveloped combustor designs. While conventional combustor design and emissions estimation through computational fluid dynamics (CFD) is preferred, it is computationally intensive and impractical for system-level simulations. To alleviate this, a thermofluid network model was developed to predict the performance of a MGT combustor operating on pure and fuel blends of propane and hydrogen. It incorporates sub-component pressure losses and heat transfer, and presents the first implementation of well-stirred and plug-flow reactors into Flownex SE. A 3D CFD study of the combustor revealed that hydrogen addition improved combustion efficiency and reduced wall temperatures. However, despite producing less CO2, it leads to 70 % more CO and 80 % more NO than for propane-only operation. Validated against the 3D CFD data, the network model predicted the combustor's outlet total temperature and pressure within 0.5 % and 0.26 %, respectively. The change in total pressure across subcomponents ( 6%) and the mass flow distribution showed similarly strong agreement. Major species mass fractions, CO2 and H2O, were predicted accurately. However, by assuming that the temperature and composition are uniform within combustion zones, zone and wall temperatures and pollutant predictions deviated considerably. NO was overpredicted by a factor of 8.2–10.7, and CO was overpredicted for propane-only, but underpredicted for blended cases. The network model achieved this performance 420 times faster than CFD, making it suitable for rapid design exploration.
Fenner et al. (Mon,) studied this question.