Modeling and Simulation of Soot Formation
The reduction of pollutant emissions, and in particular soot emissions, is a key target for the development of next-generation combustion engines. Computational design and optimization have a high potential to significantly reduce the time scales and costs of the development processes of such engines, given sufficient predictive capabilities of integrated turbulence-combustion-soot models.
Soot evolution is a highly nonlinear multi-scale and multi-physics process, which makes the prediction of soot emissions a very challenging task. High-fidelity model components are required to describe the gas phase chemistry, heterogeneous chemistry on the soot particle surface, aerosol dynamics, and the turbulent flow and mixing field. While the development of each of these model components is a challenge in itself, they also need to be coupled with each other to provide an integrated model.
In order to exploit the potential of computational modeling for the reduction of particulate emissions in aircraft or internal combustion engines, models need to be tested and validated under engine-relevant conditions. However, in real engines, emission data are mostly available only at the engine outlet, which makes the analysis of model failures very challenging. As a result of this lack of detailed and reliable data under real engine conditions, model validation is often restricted to academic test cases such as laminar or turbulent jet flames. Extended efforts have therefore been undertaken by experimentalists to provide data bases which can bridge the gap between these more fundamental test cases and real world applications. At the same time, Large-Eddy Simulations (LES) have evolved as a powerful tool for accurate simulations of reacting flows in complex geometries.
In order to assess and improve the accuracy of the various model components of an integrated soot emission model, it is necessary to combine the analysis of configurations of varying complexity – ranging from one-dimensional laminar flames to system-scale setup – as well as simulations with varying level of fidelity.
Multi-Physics Simulation Methods
Aerosol dynamics modeling
Soot particles in a reacting flow constitute an aerosol. The evolution of the particle size distribution function is described by a population balance equation (PBE). At the ITV, advanced moment methods as well as Monte Carlo methods for the accurate and efficient solution of PBEs are developed.
Chemical kinetics of soot precursor formation
One of the least understood processes regarding the formation of soot emissions is the transition of large molecules that act as soot precursors into solid carbon particles. It is well established knowledge that the most relevant soot precursors are polycyclic aromatic hydrocarbons (PAH). However, the kinetic pathways and rates of their formation are associated with large uncertainties. Combined computational and experimental efforts are therefore undertaken to improve the accuracy of the soot precursor chemical kinetics.
Multivariate chemical soot modeling
As a result of aggregation processes, soot particles are characterized by a large morphological diversity. While incipient particles are typically spherical, larger soot aggregates are conglomerates of a large number of primary particles. In addition, the chemical composition of soot particles can be very different depending on the chemical processes contributing to their formation and growth. Therefore, we develop multivariate chemical soot models that account for this diversity and parameterize the particles by several parameters such as diameter, surface area, and chemical reactivity.
Finite-rate chemistry simulations of premixed and non-premixed laminar flames
Detailed chemical kinetic mechanisms consisting of more than 200 chemical species and several thousand elementary reactions are employed in one-dimensional and two-dimensional simulations of soot formation in laminar flames. Finite-rate chemistry is here combined with detailed chemical and statistical soot models and radiation models.
Large-Eddy Simulation of aircraft and internal combustion engines
Soot formation is highly sensitive to local fluctuations of the gas phase composition and temperature, which are partly resolved in LES. We use both structured and unstructured LES codes to compute the reacting flow in gas turbines and internal combustion engines using advanced flamelet-based combustion models coupled to detailed soot models.
Direkte Numerische Simulationen rußender Flammen
DNS of sooting non-premixed flames are perfomed and analyzed at the ITV. Large-scale DNS data are particularly useful to study small-scale effects and to develop of detailed models for turbulence-chemistry-soot interactions.