Navigation and service

Investigations into Turbulent Flames with Detailed Chemistry and Detailed Transport

Principal Investigator: Andreas Kempf, Institut für Verbrennung und Gasdynamik, Lehrstuhl Fluiddynamik, Universität Duisburg-Essen (Germany)
HPC Platform: JUQUEEN of JSC
JSC Project ID: hdu18

World energy demand is mainly covered by burning fossil fuels. A detailed understanding of the complex physical and chemical processes involved is required to develop clean and efficient combustion devices. Numerical analysis increasingly helps to understand combustion, exploiting the rapid increase of computational power. As combustion in technical devices occurs under turbulent conditions, the numerical investigation of the interplay between turbulence and chemical reactions is of high interest.

The fundamental problem of numerical turbulent combustion research is that fully resolving all relevant scales of turbulence and chemistry in a technical combustion device requires computational resources unavailable in the foreseeable future. One promising approach is the large eddy simulation (LES), where most of the relevant flow scales are resolved and only the smallest scales are modelled. This is justified by the smallest scales generally showing a universal behaviour. However, combustion mostly occurs on the unresolved scales where the chemistry interacts with the smallest eddies, making modelling a difficult task.

Within this project, which has been performed on JUQUEEN, scientists of the Universität Duisburg-Essen investigated the interplay between turbulence and chemistry by two different methods in the context of LES for well measured laboratory experiments. The first method (M1) is based on an approach where the small scale mixing and reaction is described by means of a particle based Monte Carlo method. The second method (M2) relies on a simplified description of the chemistry, which however allows to fully resolve the interaction between simplified chemistry and turbulence.

The simulations of partially premixed and premixed experimental configurations with M1 showed a good agreement with the experiments. A detailed comparison of the turbulence-chemistry interaction between simulation and experiment showed that the interaction is well reproduced by the simulation for both the diffusion and premixed dominated combustion modes.
The second method M2 provided highly resolved velocity and scalar fields, which were used in two ways. First, the full information was used to investigate the interaction of turbulence and combustion chemistry on the smallest scales. The importance of the wrinkling of the flame and the alignment between flame normal direction and velocity gradient for the flame propagation speed was shown for a lab-scale flame. In a second step, the fields were filtered so that detailed statistics of the sub-filter flame wrinkling could be generated. These statistics enabled the validation and further development of sub-filter models, which can be used for precise LES predictions of technical combustors.

Investigations into Turbulent Flames with Detailed Chemistry and Detailed Transport

Figure 2: Monte Carlo particles representing the scalar fields (M1) in plane cutting though the flame in the Darmstadt turbulent opposed jet burner. From left to right: temperature, type of chemical model, methane mass fraction, water vapor mass fraction. Approximately 32x103 particles are shown per sub-plot. © Universität Duisburg-Essen

Scientific Contact:

Prof. Dr.-Ing. Andreas Kempf
Institut für Verbrennung und Gasdynamik, Lehrstuhl Fluiddynamik
Universität Duisburg-Essen
Carl-Benz-Str. 199, D-47057 Duisburg (Germany)
e-mail: andreas.kempf @

JSC Project ID: hdu18

July 2016