Turbulence models based on the RANS (Reynold-averaged Navier-Stokes) equations are still widely spread in industry to simulate turbulent flows. However, since accurate results with these models cannot be expected in many cases, more exact but costly simulations such as Large-eddy simulations (LES) are gradually implemented. This is made possible by the increasing computing power of modern high-performance computers. Figure 1 shows a comparison between a LES simulation and an experiment of a swirl-stabilized spray burner. Special attention was paid to the coupling of a multi-component vaporization model with a detailed approach for the reaction kinetics. This is particularly important for the combustion of complex, liquid fuels (in this case kerosene). The main processes take place in the vicinity of the burner. In the numerical simulation, the composition of the gas phase and the heat release were determined with the detailed chemistry solver in THETA. It solves a transport equation for each individual chemical species and treats the reactions directly in Arrhenius form. The trajectories and the evaporation of the individual components of the liquid phase were calculated with SPRAYSIM. Both solvers are coupled and exchange data in both directions (THETA <-> SPRAYSIM). As can be seen in the figure, the qualitative comparison of the flame images is very good. Nevertheless, many uncertainties remain in the LES of reactive multi-phase flow.
Significant further developments for the internal CFD code platform THETA-SPRAYSIM deal with LES. One of the objectives in the department is to support these developments with new SGS (sub-grid scale) models for reactive multi-phase flows. In combination with high-resolution measurements in well-defined flows with drops, DNS (Direct Numerical Simulation) is the only way to develop realistic SGS models for reactive multi-phase flows.
As part of the DLR internal project R2F, DNS calculations of drops were carried out using the DLR THETA code. The challenge is that on the one hand a boundary condition for the momentum exchange on the surface of the drop and on the other hand a model for the heat and mass transfer during vaporization with chemical reactions has to be implemented before an exact SGS model can be developed. Results of simulations of the drop’s surrounding flow under laminar conditions can be seen in Figure 2. Figure 2 a) shows that the flow is symmetrical even for the highest calculated Reynolds number and that only a small separation bubble forms in the trail of the droplet. Figure 2 b) demonstrates the very good agreement of the results of the drag coefficient with literature values.
1 G. Eckel, J. Grohmann, L. Cantu, N.A. Slavinskaya, T. Kathrotia, M. Rachner, P. Le Clercq, W. Meier, and M. Aigner, LES of a swirl-stabilized kerosene spray flame with a multi-component vaporization model and detailed chemistry. Combustion and Flame, 2019
2 Edelmann, Christopher A. and Le Clercq, Patrick C. and Noll, Berthold (2017) Direct Numerical Simulation of Evaporating Drops at Laminar and Turbulent Conditions. 3rd International Conference on Numerical Methods in Multiphase Flows, ICNMMF-III, 26.-29.06.2017, Tokio, Japan.
Numerical simulations have the potential to significantly reduce the duration and costs of the approval of alternative fuels. In addition, virtualization and numerical methods can substantially support the design of new fuel-flexible and low-emission combustion chambers. However, this raises the question of how reliable the information obtained by numerical simulations is and how far it can serve as a basis for decisions with severe consequences for the safety of people and the environment. In this context, the MAT department is carrying out research on new methods for the validation of numerical models. The focus is on a risk-informed decision making, based on data obtained from simulations. On the one hand, it is fundamental to quantitatively and systematically determine the accuracy of the models in their entire range of application. On the other hand, it is essential to be able to assess the reliability of the model prediction. This includes the effect of uncertainties in the input variables on relevant target variables of the investigation.
As an example, the figure compares the numerical prediction of the diameter of an evaporating drop with experimental data. Areas shaded in gray represent the uncertainty of the numerical prediction.
B. Rauch, Systematic accuracy assessment for alternative aviation fuel evaporation models, VT-Forschungsbericht 2018-01
B. Enderle, B. Rauch, F. Grimm, G. Eckel, and M. Aigner, Non-Intrusive Uncertainty Quantification in the simulation of turbulent spray combustion using Polynomial Chaos Expansion: A case study. Combustion and Flame, 2020