Uncertainty Analysis of the Chemical Reactor Network Approach for Predicting the Pollutant Emissions in a Double-Swirl Combustor

The purpose of this study is to predict the pollutant emissions generated within an aero-engine combustor model using the computational fluid dynamics-chemical reactor network (CFD-CRN) approach by modeling combustion in highly swirled flows. The selected test case is a laboratory double swirled combustor that came with an extensive experimental database from previous works for CH4/air diffusion flames at atmospheric pressure. The CFD-CRN modeling approach is initiated by solving Reynolds-averaged Navier–Stokes (RANS) equations for a 3D computational domain. The numerically achieved time-averaged values of the velocity components are in good agreement with the experimental data for two different thermal power. The CRN is obtained by dividing the flow field into ideal chemical reactors using various filters on the CFD results. The temperature, axial velocity, CH₄, and O₂ mass fractions distributions are selected as the splitting criteria for constructing the CRN. An uncertainty analysis is carried out to investigate the effects of different splitting approaches for the temperature criteria since it significantly affected the pollutant emissions in the gas turbine combustor. The simulations of the pollutant emissions are performed via the detailed gas-phase chemical kinetic mechanism of GRI-Mech 3.0. The nonlinear distribution of the temperature intervals result in lower uncertainty and provide reliable results even with a small number of ideal reactors. Also, it is observed that the CRN can be used in different operating conditions and provide suitable results if it is constructed with exceptional consideration. Moreover, a parametric study is performed by varying the equivalence ratio and air inlet temperature to investigate the trends of the NO and CO emissions.