PT Unknown
AU Gruhlke, P
TI Towards an improved prediction of CO and NOx emissions from gas turbines by LES
PD 01
PY 2021
DI 10.17185/duepublico/75285
LA en
AB The combustion of hydrocarbon fuels leads to the emission of toxic carbon monoxide (CO) and nitric oxide (NOx) pollutants. In order to reduce these emissions, increasingly restrictive emissions regulations and environmental protection measures have been established. For the development of low-emission gas turbine combustors, the fundamental mechanisms and sources affecting the formation of CO and NOx pollutants need to be understood. Due to the limited optical access in gas turbine engines for experimental investigations, numerical methods are a promising approach. In particular, large-eddy simulations (LES) fulfill the necessary requirements to investigate the formation of pollutants in technical combustion systems. The aim of this work is to improve the prediction of CO and NOx emissions from gas turbines by LES. For this purpose, the phenomena and effects that are relevant for the accurate prediction of CO and NOx formation are investigated. Different modeling approaches are evaluated and improved with respect to the associated requirements. Furthermore, the suitability of these numerical methods is assessed in terms of an accurate and numerically efficient prediction. In the first study of this work, the numerical LES framework is developed. A combustion model based on tabulated chemistry is extended for the prediction of CO and NOx emissions. The turbulence-chemistry interaction is modeled using the artificial flame thickening approach and a model for the unresolved flame wrinkling is used. The applied models and the simulation method are validated in a LES of an atmospheric turbulent premixed flame. Simulation results are analyzed and compared with the experimental data, with emphasis on CO and NOx pollutant concentrations. Subsequently, the modeling approach is evaluated by investigating a premixed high pressure jet flame. Heat losses in this configuration are taken into account requiring an additional transport equation for enthalpy and an extension of the chemistry tables. In this second study, a finite rate chemistry (FRC) combustion model is developed. A custom built reaction mechanism has been developed for the description of chemical kinetics suitable for the investigated conditions. Large-eddy simulations are performed with both combustion models. The simulation results are comprehensively compared with the experiment and special focus of the analysis is given to the pollutant formation and the stabilization of the flame. In the third study, a modeling approach is developed to efficiently predict finite rate chemistry effects in the LES of lifted high pressure jet flames. For this purpose, a new optimization criterion for reaction mechanism development is introduced and applied. To validate the developed reaction mechanism, 0D and 1D simulations are performed, followed by the application of LES. The simulation results are compared to those of a much more computationally intensive reaction mechanism. To gain a deeper understanding of the flame stabilization mechanism, further detailed analyzes follow. In the final study, the FRC combustion model is applied to the LES of a full-scale gas turbine prototype combustor operated at three different staged part load operating conditions at high pressure. The developed modeling approach is investigated in terms of its ability to accurately predict CO pollutants in a complex technical combustion system in comparison to the experimental data. Sources that eventually result in incomplete carbon monoxide oxidation at different operating points are identified. It is shown that the developed LES methodology is capable of accurately describing complex phenomena and effects, which are necessary for improved prediction of pollutants in gas turbines.
ER