La Simulation aux Grandes Echelles (SGE) est appliquée à des brûleurs industriels pour prédire de nombreux phénomènes physiques complexes, tel que l’allumage ou la formation de polluants. La prise en compte de réactions chimiques détaillées est alors indispensable pour obtenir des résultats précis. L’amélioration des moyens de calculs permet de réaliser des simulations de brûleurs avec une chimie de plus en plus détaillée. La principale problématique est le couplage entre les réactions chimiques et l’écoulement turbulent. Bien que la dynamique de flamme soit souvent bien reproduite avec les modèles actuels, la prédiction de phénomènes complexes comme la formation de polluants reste une tâche difficile. En particulier, des études ont montré que l’influence du plissement de sous-maille sur la structure chimique des flammes n’était pas prise en compte de manière précise. Deux modèles basés sur le filtrage explicite des fronts de flammes sont étudiés dans cette thèse afin d’améliorer la prédiction de polluants en combustion turbulente prémélangée : (i) le premier modèle met en jeu une méthode de déconvolution des variables filtrées ; (ii) le second modèle implique l’optimisation de la chimie pour obtenir des flammes turbulentes filtrées. L’objectif de la thèse est d’obtenir une prédiction précise des polluants à coût de calcul réduit. / Large Eddy Simulation (LES) is applied to industrial burners to predict a wide range of complex physical phenomena, such as flame ignition and pollutants formation. The prediction accuracy is tightly linked to the ability to describe in detail the chemical reactions and thus the flame chemical structure. With the improvement of computational clusters, the simulation of industrial burners with detailed chemistry becomes possible. A major issue is then to couple detailed chemical mechanisms to turbulent flows. While the flame dynamics is often correctly simulated with stateof- the-art models, the prediction of complex phenomena such as pollutants formation remains a difficult task. Several investigations show that, in many models, the impact of flame subgrid scale wrinkling on the chemical flame structure is not accurately taken into account. Two models based on explicit flame front filtering are explored in this thesis to improve pollutants formation in turbulent premixed combustion: (i) a model based on deconvolution of filtered scalars; (ii) a model involving the optimization of chemistry to reproduce filtered turbulent flames. The objective of the work is to achieve high accuracy in pollutants formation prediction at low computational costs.

This work is motivated by the question: how much complexity is appropriate for a cloud parameterization used in general circulation models (GCM). To approach this question, cloud parameterizations across the complexity range are explored using general circulation models and theoretical Monte-Carlo simulations. Their results are compared with high-resolution satellite observations and simulations that resolve the GCM subgrid-scale variability explicitly. A process-orientated evaluation is facilitated by GCM forecast simulations which reproduce the synoptic state. For this purpose novel methods were develop to a) conceptually relate the underlying saturation deficit probability density function (PDF) with its saturated cloudy part, b) analytically compute the vertical integrated liquid water path (LWP) variability, c) diagnose the relevant PDF-moments from cloud parameterizations, d) derive high-resolution LWP from satellite observations and e) deduce the LWP statistics by aggregating the LWP onto boxes equivalent to the GCM grid size. On this basis, this work shows that it is possible to evaluate the sub-grid scale variability of cloud parameterizations in terms of cloud variables. Differences among the PDF types increase with complexity, in particular the more advanced cloud parameterizations can make use of their double Gaussian PDF in conditions, where cumulus convection forms a separate mode with respect to the remainder of the grid-box. Therefore, it is concluded that the difference between unimodal and bimodal PDFs is more important, than the shape within each mode. However, the simulations and their evaluation reveals that the advanced parameterizations do not take full advantage of their abilities and their statistical relationships are broadly similar to less complex PDF shapes, while the results from observations and cloud resolving simulations indicate even more complex distributions. Therefore, this work suggests that the use of less complex PDF shapes might yield a better trade-off. With increasing model resolution initial weaknesses of simpler, e.g. unimodal PDFs, will be diminished. While cloud schemes for coarse-resolved models need to parameterize multiple cloud regimes per grid-box, higher spatial resolution of future GCMs will separate them better, so that the unimodal approximation improves.