A reactor network model for predicting NOx emissions in gas turbines

The numerical prediction of NOx emissions from gas turbines is addressed in this paper. Generated from Computational Fluid Dynamics (CFD), a Reactor Network (RN) is defined to model the NOx formation with a detailed chemistry. An optimized procedure is proposed to split the reactive flow field into homogeneous zones considered as Perfectly Stirred Reactors (PSR). Once connected together, they result in a Chemical Reactor Network (CRN) that yields a detailed composition regarding species and temperature in the combustion chamber. Sensitivity studies are then performed to estimate the influence of air humidity and gas turbine load on NOx predictions. The NOx emissions predicted are in good agreement with the measured data in terms of levels and trends for the case studied (a gas turbine flame tube fed with natural gas and functioning at a pressure of 15 bar). Finally, the RN methodology has shown to be efficient estimating accurately NOx emissions with a short response time (few minutes) and small CPU requirements.