Abstract
A numerical investigation of pollutant emissions of a novel dry low-emissions burner for heavy-duty gas turbine applications is presented. The objective of this work is to develop and assess a robust and cost-efficient numerical setup for the prediction of NOx and CO emissions in industrial gas turbines and to investigate the pollutant formation mechanisms, thus supporting the design process of a novel low-emission burner. To this end, a comparison against experimental data, from a recent experimental campaign performed by BHGE in cooperation with University of Florence, has been exploited. In the first part of this work, a Reynolds-averaged Navier–Stokes (RANS) approach on both a simplified geometry and the complete domain is adopted to characterize the global flame behavior and validate the numerical setup. Then, unsteady simulations exploiting the scale adaptive simulation (SAS) approach have been performed to assess the prediction improvements that can be obtained with the unsteady modeling of the flame. For all simulations, the flamelet generated manifold (FGM) model has been used, allowing the reliable and cost-efficient application of detailed chemistry mechanisms in computational fluid dynamics (CFD) simulation. However, FGM typically faces issues predicting flame emissions, such as NOx and CO, due to the wide range of time scales involved, from turbulent mixing to pollutant species oxidation. Specific models are typically used to predict NOx emissions, starting from the converged flow-field and introducing additional transport equations. Also CO prediction, especially at part-load operating conditions could be an issue for flamelet-based model: in fact, as the load decreases and the extinction limit approaches, a superequilibrium CO concentration, which cannot be accurately predicted by FGM, appears in the exhaust gases. To overcome this issue, a specific CO-burn-out model, following the original idea proposed by Klarmann, has been implemented in ANSYS fluent. The model allows to decouple the effective CO oxidation term from the one computed by FGM, defining a postflame zone where the source term of CO is treated following the Arrhenius formulation. In order to support the design process, an indepth CFD investigation has been carried out, evaluating the impact of an alternative burner geometrical configuration on stability and emissions and providing detailed information about the main regions and mechanisms of pollutants production. The outcomes support the analysis of experimental results, allowing an indepth investigation of the complex flow-field and the flame-related quantities, which have not been measured during the tests.
