60296 - Validating Soot Models in Les of Turbulent Flames: The Contribution of Soot Subgrid Scales Models to Prediction of Soot Production in an Aero-Engine Model Combustor 

Numerical simulations of soot production in an industrial system via an LES approach represent a great challenge. In addition to the fact that complex multi-scale coupled physical processes govern soot production in turbulent flames, the validation of newly developed models is difficult when considering sooting turbulent flames [1]. Indeed, the classical strategy used to validate gaseous phase models relying on a comparison of results from two simulations (one with the reference model, another with the newly developed model) would not be adequate for soot models.  

In this work, we highlight the difficulties specific to the validation of models for sooting turbulent flames by investigating the contribution of the soot intermittency subgrid model [2] to the prediction of soot production in the DLR aero-engine model combustor [3]. First, LES on grids with increasing refinements have been considered following a classical strategy for gaseous flames. It is found that negligible differences in the gaseous phase description due to different grid resolutions can lead to important discrepancies in soot volume fraction prediction. It is then not possible to discriminate the contribution of the soot subgrid scale model. A second classical strategy is then considered by performing two LES, one activating the soot subgrid scale model and the second without the model. It is proven that soot production occurs only for extremely rare local gaseous conditions and that this classical strategy is not efficient in quantifying the contribution of soot subgrid scale models since it is not possible to guarantee that the same rare local gaseous events are observed in the two LES due to numerical sensitivities. Obtaining a statistical representation of the solid phase would require the simulation of a much longer physical time compared to the gaseous phase [4] so that the classical strategy is extremely CPU-demanding.

Alternatively, a new strategy for the validation of soot models is proposed here. It is based on a unique LES simulation, where the set of equations describing the solid phase are doubled. One set accounts for the soot subgrid scale model, while the other set is treated with no soot subgrid scale model. This strategy enables the soot scalars from both sets to experience the same unique temporal and spatial gas-phase evolution, allowing to isolate the soot subgrid scale model effects from the uncertainties on gaseous models and numerical sensitivities. Although doubling the soot equations increases the CPU cost, only one LES is finally performed, leading to a smaller computational cost for the whole study compared to the original strategy. Thanks to this approach, the first indications of the soot intermittency model's contribution to soot prediction in the DLR burner are finally proposed. 

[1] L. Tardelli, P. Rodrigues, B. Franzelli, and N. Darabiha, 2019. “Impact of the reaction mechanism model on soot growth and oxidation in laminar and turbulent flames”. Proceedings of ASME TurboExpo 2019: Turbomachinery Technical Conference and Exposition.GT2019-90873.

[2] M. E. Mueller and H. Pitsch, 2011. “Large eddy simulation subfilter modeling of soot-turbulence interactions”. Physics of Fluids 23 (11).

[3] K. P.Geigle, R. Hadef, and W. Meier, 2013. "Soot formation and flame characterization of an aero-engine model combustor burning ethylene at elevated pressure". Journal of Engineering for Gas Turbines and Power 136 (2)

[4] Chong, S. T., M. Hassanaly, H. Koo, M. E. Mueller, V. Raman, and K. P. Geigle (2018). "Large eddy simulation of pressure and dilution-jet effects on soot formation in a model aircraft swirl combustor". Combustion and Flame 192, 452 – 472.