Logowanie do System sprawozdań KDM
Turbulent premixed combustion modeling for large eddy simulations
Kierownik projektu: Bruno Savard
Wydział Mechaniczny, Energetyki i Lotnictwa
Instytut Techniki Cieplnej
The performance of combustion devices in terms of efficiency and pollutant emissions, and their adaptability to alternative fuels are limited by our understanding of the turbulent combustion processes involved. Turbulent combustion is a complex multi-physics (chemistry, fluid mechanics, thermodynamics) phenomenon that takes place over a large range of scales (both space and time). Despite all the research effort conducted in the past decades, our understanding of turbulent combustion is incomplete. As a consequence, numerical simulation tools either rely on models that are not verified for the regimes they are targeted for, or are too computationally expensive. Only a better understanding of turbulent combustion can lead to more appropriate models. These can in turn enable the development of fast and accurate predictive simulation tools needed for the effective design of more efficient and cleaner combustion devices.
Direct numerical simulations (DNS) with large detailed chemical mechanisms provide detailed resolution of the internal flame structure which is necessary to analyze the turbulence-chemistry interactions at high turbulence intensities and develop appropriate combustion models. We have previously performed such simulations of highly turbulent n-heptane and iso-octane flames over a wide range of unburnt temperatures and pressures relevant to transportation engines. These have provided well needed insight into the intricate interaction between small scale turbulence and chemistry. This interaction differs considerably from that found in low-turbulence methane/air flames, from which most combustion models have been developed. Unfortunately DNS are too computationally expensive to be used for practical simulations of combustion devices.
Detailed chemical mechanisms are generally not used in practical computational fluid dynamics (CFD) simulations for the following reasons: 1) they involve too many species (which need to be transported) and 2) they are associated with very challenging chemical source term closure for practical CFD, due to the complex coupling between the many species involved. These closures are necessary to model the unresolved quantities in practical CFD, typically in Reynolds-averaged Navier-Stokes simulations (RANS) or large-eddy simulations (LES). In contrast, chemistry tabulation relies on a low-dimensional manifold such that the chemistry is represented by only a few controlling variables (typically one to three). This allows a reduction in simulation cost by up to two orders of magnitude, and the source term closure for CFD is much simplified. However, the validity of chemistry tabulation relies on an important simplifying assumption: the flame structure can be represented by that of one-dimensional flames (flat or curved, stretched or unstretched).
Using the set of DNS we have performed previously, we have been able to identify important controlling parameters to represent highly turbulent flames with one-dimensional flames. A resulting chemistry tabulation approach for LES was proposed but not tested yet. This model accounts for the effect of high turbulence on the reaction zone, but this has yet to be verified for other fuels than single alkane species surrogates.
The objective of this work is to develop/validate this combustion model for large-eddy simulations of turbulent premixed combustion with transportation fuels at engine-relevant conditions. The approach taken is as follows: 1) to perform direct numerical simulations of highly turbulent flames with a single-component bio-fuel surrogate and multi-component gasoline surrogates to complement the set of available DNS, 2) to test a priori the proposed chemistry tabulation approach (for LES) by filtering DNS data and compare its performance to that of simpler combustion models, and 3) to test a posteriori the proposed chemistry tabulation approach by performing LES of the spherically expanding turbulent premixed flames of Prof. Chaumeix at CNRS Orleans (ICARE).
- Bruno Savard, Guillaume Blanquart, Effects of dissipation rate and diffusion rate of the progress variable on local fuel burning rate in premixed turbulent flames, Combustion and Flame -, (2017) -
- Bruno Savard, Andrzej Teodorczyk, Low-temperature chemistry in n-heptane/air premixed turbulent flames in the thin reaction zones regime, 2017 European Combustion Meeting -, (2017) -