A DNS Study on CO Emissions in Stationary Gas Turbine Combustors with Cooled Walls

Principal Investigator:
Prof. Dr.-Ing. Heinz Pitsch

Institute for Combustion Technology, RWTH Aachen University, Germany

Local Project ID:

HPC Platform used:
SuperMUC-NG of LRZ

Date published:


A direct numerical simulation (DNS) with finite rate chemistry has been performed to investigate the influence of flame-wall interaction (FWI) on carbon monoxide (CO) emissions in very lean turbulent premixed methane flames. CO emissions are affected by the mean strain rate of the turbulent flow, the FWI, and the interactions of the flame with the recirculation zones of the flow. The CO production and consumption in the turbulent flame differ strongly from the reaction rates in a freely propagating flame. A flamelet-progress variable (FPV) model has been extended by an enthalpy defect term for wall heat loss, a transport equation for the CO mass fraction and the mass fraction of the OH radical as a marker for turbulent strain.

The reduction of pollutant emissions is one of the main challenges in the development of gas turbine combustors. Due to increasingly stringent emission limits, improvements of combustors are required with respect to emissions. In this context, numerical simulations, in particular large-eddy simulations (LES), can make an important contribution to reducing development cost, accelerating development processes, and to improving the technology. Therefore, the aim of this project is the development of improved emission models for LES of both aircraft engines and stationary gas turbines.

The reduction of emissions is a challenge within the development process of gas turbine combustors. Due to increasingly stringent emissions regulations, optimization of combustion chambers regarding emissions is required. Numerical simulations, in particular Large-Eddy Simulations (LES), will play an important role in the development process while reducing development costs. This project aims to develop predictive LES emission models for aircraft engines as well as stationary gas turbines.

In this work, a systematic model development approach is deployed, which combines results of Direct Numerical Simulations (DNS) and experiments for model analysis and validation. Systematic tools such as the optimal estimator analysis are used to analyze the results of DNS and to develop the modelling approach.

For stationary gas turbines, the focus is on the impact of convective heat losses at combustor walls on CO emissions. These heat losses can perturb and slow down the oxidation of CO. A series of DNS investigations, focused on turbulent flame-wall interaction (FWI), has been performed. The data has been systematically analyzed and modeling approaches derived. As in the soot part of the project, a flamelet-progress variable (FPV) model combined with tabulated chemistry is proposed. The chemistry table is extended with additional dimensions to consider the pollutants in more detail. Transport equations are solved for the most relevant parameters. Finally, a-priori investigations into the model performance have shown a significant improvement in CO prediction.

The simulation domain consisted of 130 million grid points for the adiabatic cases and 3.1 billion grid points for the cases with isothermal walls. The production runs were performed on up to 768 nodes with 36,864 cores, using a total of about 40 million core hours.

References and Links

Kai Niemietz, Lukas Berger, Michael Huth, Antonio Attili, Heinz Pitsch, Direct numerical simulation of flame-wall interaction at gas turbine relevant conditions, Proceedings of the Combustion Institute, 2022,