Large Eddy Simulations

Large Eddy Simulations of Turbulent Combustion Basing on the Eulerian Stochastic Field Method Coupled to Flamelet Progress Variable Approach Using OpenFOAM Code

Default-Image-ProjectsDefault Image - ProjectHessen Agentur/Jürgen Kneifel


The present trend in energy consumption shows ever increasing energy demand to servemultitude of reasons such as: transportation, electricity, manufacturing, and heating. Eventhough there is a strong development towards the usage of renewable energy to phase outtraditional energy, most of the energy production will still be continued by fuel scombustion. The turbulent combustion process presents a challenging area for research andinvestigation, since the interaction between the turbulence and flame occurs in broad rangesof time and length scales that need to be considered while carrying out numerical modelingand simulation [1].


In this perspective, a novel, well-designed numerical combustion model is investigated andcompared in the current project to usual existing technique (β-PDF) in order to assess thegeneral prediction capability in reproducing main turbulent characteristics. This methodcombines a transported joint scalar probability density function (T-PDF) following theEulerian Stochastic Field methodology (ESF) on the one hand, and a flamelet progressvariable (FPV) turbulent combustion model under consideration of detailed chemicalreaction mechanism on the other hand [2,3,4,5]. It was implemented and tested inboth; Reynolds averaging-based numerical simulation (RANS) and large eddy simulation(LES) frameworks. The validation used case for this technique was the well-known air-piloted methane jet flame (Sandia Flame-D) in [6]. Additionally, applying the same novelapproach, an investigation on the predictions of combustion characteristics of a turbulentOxy-methane non-premixed flame operating under highly diluted conditions of CO2 and H2 in oxidizer and fuel streams respectively [7], is reported in the project. All numericalsimulations have been performed on the Lichtenberg cluster and high performancecomputing systems are needed as the detailed resolution of the Flame-D and the Oxy-fueldomains still require immense computational resources.


By applying both, the novel combustion method based on the transported probabilitydensity function and the FPV chemical technique (ESF/FPV) and the presumed-probabilitydensity function (β-PDF)-based FPV, on different Sandia Flame-D grids in RANS and LES frameworks, obtained results show a reliable agreement between simulated data and experimental ones. However, the novel hybrid approach removes the weaknesses of the P-PDF modeling where an over-estimation was noticed for some major species like the COand H2O and under-estimation of Temperature profile once turbulent Oxy-fuel non-premixed flames were investigated as application cases.


In this project, it was demonstrated so far that the combustion characteristics of a turbulentair-piloted jet flame (Sandia flame-D) in both RANS and LES frameworks can be predictedand analyzed using both, the novel hybrid ESF/FPV model and the presumed β-PDF. However, for more complex reacting turbulent cases presented in the oxy-methane non-premixed flame series operating under highly diluted conditions of CO2 and H2 ,employing the hybrid ESF/FPV model removed the weaknesses of the β-PDF anddemonstrated its high capability in capturing the main flame properties and flow fieldvariables applying RANS and LES turbulent models.


Nonetheless, the novel model needs to be further evaluated by using premixed and partiallypremixed FPV tables. Also chemical tables with a non-unity Lewis number and byconsidering advanced micro-mixing models can be further studied. These tasks are left forfuture work.

Last Update

  • Last Update: 2022-06-14 18:05