CORIA

Development of a Unified-LES methodology for high-speed turbulent flows with shocks

Lodato — 2018-02-20T10:06:06Z

Information :
Location : CORIA Laboratory, Rouen, France – www.coria-cfd.fr
Duration : 36 months
Contact : guido.lodato@coria.fr

Description of the scientific project

High speed flows in which turbulence and shock waves mutually interact remain a challenging task for numerical methods. This is especially true in the case of largeeddy simulation (LES), in which the spatial and temporal resolutions are selected such as to provide a direct representation of the energy-containing flow structures and where models for unresolved scales of different nature have to coexist.
In the case of high-speed flows, the development of efficient and accurate models for the unresolved sub-grid interactions can be extremely challenging due to the very different flow phenomena that can be involved. Typically, different models are adopted for turbulence and discontinuities, a segregated approach which leads to model interference and may produce unpredictable effects on the accuracy of the solution.

Numerical Platform

The research group has extensive experience in the context of high-order discontinuous finite elements (DFE) methods for unstructured meshes, with particular emphasis in the development of an MPI parallelized fortran 90 solver for compressible flows, based on the high-order spectral difference (SD) scheme for unstructured hexahedral elements. Recently efforts, in particular, have been focusing in developing robust and efficient shock-capturing and sub-grid scale modeling approaches which are specifically conceived to be used with DFE schemes. Both approaches make intensive use of modal detection techniques, which are particularly well suited for high-order DFE methods.

Objectives

The objective of the proposed PhD Thesis is to develop a unified-LES methodology to enable high-fidelity numerical simulations of high-speed turbulent flows. Extensions of such approach are also envisaged to deal with material interfaces for applications to detonations. The project will focus on the generalization of state-of-theart modeling techniques for DFE schemes to fully compressible flows and on the development of a unified modal detection approach for DFE methods which is able to detect and recognize under-resolved turbulence and shock waves, and to provide the most correct blend of dissipation to allow a physically consistent representation of the relevant mutual interactions.

Candidate profile

The candidate should hold a MSc degree in Engineering (Aeronautical, Mechanical), Physics, Applied Mathematics or other related fields showing evidence of deep knowledge of continuum transport models and numerical methods.
Knowledge in computer language and algorithm development is required. An experience in numerical simulation, high-performance computing, high-order schemes for computational fluid dynamics is a major plus.
Applicants must have good English communications and writing skills as demonstrated through sufficiently high TOEFL/TOEIC scores (or equivalent). Knowledge of French language is a plus, but not mandatory.

Post-Doctorat Turbulence et Mélange

Danaila, Varea — 2018-01-10T10:09:29Z

Date : 1^er février
Durée : 18 mois

Contact : voir ci-contre.

Contexte économique

La gestion énergétique durable des dispositifs industriels de nouvelle génération passe par l'utilisation de matériaux innovants pour assurer une régulation thermique économique et fiable. Les Matériaux à Changement de Phase (MCP) offrent la possibilité d'emmagasiner et de restituer une forte quantité de chaleur au moment de leur changement de phase : au-dessus d'une certaine température caractéristique ils se liquéfient en absorbant la chaleur extérieure et la restituent lorsque la température baisse. Les applications industrielles des MCP couvrent un large nombre de domaines d'activité, de l'électronique jusqu'au bâtiment.

Etat de l'art

Le changement de phase dans les systèmes physiques d'intérêt pratique (comme la fusion de l'acier ou la réfrigération des aliments) est généralement étudié en utilisant le modèle de Stefan, qui prend en compte un seul phénomène physique : la conduction thermique. Dans les matériaux à changement de phase, la convection dans la phase liquide joue un rôle crucial, ce qui nécessite une description théorique complexe, basée sur les équations de Navier-Stokes-Boussinesq. De nouveaux modèles ont été proposés récemment (voir l'ouvrage A. Faghri, Y. Zhang, Transport Phenomena in Multiphase Systems, Elsevier, 2006) et commencent à être étudiés théoriquement et numériquement. Les études mathématiques utilisant des modèles simplifiés (Cabrales et al., Analysis and optimal control of some solidification processes, Discrete and continuous dynamical systems. 2014) et les simulations numériques utilisant des méthodes de volumes finis (Wang et al., A comprehensive numerical model for melting with natural convection, Int. J. Heat Mass Transf., 2010) sont très récentes.

Motivation et sujet de l'étude

Les dispositifs industriels à base de MCP présentent un niveau de maturité faible et plusieurs questions pratiques sont de grande actualité : comment les MCP réagissent à un environnement confiné, quelle est la durée de vie, l'efficacité thermique de ces dispositifs ? Notre étude a comme objectif d'apporter des éléments de réponse à ces questions par la modélisation physique et par la simulation numérique. Elle sera organisée pour permettre l'application des connaissances acquises aux dispositifs industriels développés par nos collaborateurs.


img-c

Principales actions

Deux grandes actions sont prévues :

Modélisation physique de la dynamique temporelle du système diphasique. Les quantités modélisées seront, à chaque instant de temps : vitesse et taille des rouleaux de convection (l'écoulement n'est pas turbulent), température, épaisseur et position de la zone de mélange.

Réalisation expérimentale d'un système de MCP. La géométrie de la cellule devra permettre un écoulement 2D, ou quasi 2D, afin de faciliter la comparaison et la validation des résultats issus des simulations numériques. Le fluide considéré sera de la paraffine. La cellule de convection sera transparente pour permettre de visualiser l'avancement du front de solidification. Le haut et le bas de la cellule seront isolés thermiquement. Initialement, le fluide se trouvera à une température supérieure à la température de fusion (facteur 2). Un refroidissement -à une température inférieure à la température de fusion- sera appliqué d'un côté, soit à température constante, soit à flux constant. Cela conduira à la naissance du front solide qu'il s'agira de suivre au cours du temps. Différentes techniques de visualisation seront utilisées. En ce qui concerne la vitesse, la technique optique laser type PIV/PTV sera mise en place. Pour la partie thermométrie, on utilisera soit de la thermométrie sur phosphore qui offre une résolution excellente (proche de 0.1°C) liée à une technique ex-situ i.e. non intrusive. Des phosphores de type BAM BaMg2Al10O17:Eu semblent convenir pour les plages de températures considérées.

Références bibliographiques.

A. Faghri, Y. Zhang, Transport Phenomena in Multiphase Systems, Elsevier, 2006
R.Voller,M.Cross,N.C.Markatos,Anenthalpymethodforconvection/diffusionphasechange, Int. J. for Num. Meth. in Eng., 24 (1987) 271-284.
M. Okada, Analysis of heat transfer during melting from a vertical wall, Int. J. of Heat and Mass Transf., 27 (1984) 1986-2000.
S. Wang, A. Faghri, T. L. Bergman, A comprehensive numerical model for melting with natural convection, Int. J. of Heat and Mass Transf., 53 (2010) 2057-2066.
I. Danaila, R. Moglan, F. Hecht, S. Le Masson, A Newton method with adaptive ﬁnite elements for solving phase-change problems with natural convection, J. Comp. Physics, 274 (2014) 826-840.
R. Nourgaliev, H. Luo, B. Weston, A. Anderson, S. Schoﬁeld, T. Dunn, J.-P. Delplanque, Fullyimplicit orthogonal reconstructed Discontinuous Galerkin method for ﬂuid dynamics with phase change, J. Comp. Physics, 305 (2016) 964-996.
K. Luo, F. J. Yao, H. Yi, H. Tan, Lattice Boltzmann simulation of convection melting in complex heat storage systems ﬁlled with phase change materials, Applied Thermal Engineering, 86 (2015) 238-250.
L. Danaila, L. Voivenel, E. Varea, ”Self-similarity criteria in anisotropic ﬂows with viscosity stratiﬁcation,” Physics of Fluids, Vol.29, Issue 2, 020705, DOI : 10.1063/1.4974520.
Voivenel L., Varea E., Danaila L., Renou B. and Cazalens M., 2016, 'Self-similarity of variableviscosity ﬂows', Physica Scripta, 91 084007 (12pp).
Taguelmimt N., Danaila L. and Hadjadj A., 2016, 'Effects of viscosity gradients on mean velocity profile in temporal mixing layer', Journal of Turbulence, DOI http://www.tandfonline.com/doi/full/10.1080/14685248.2015.1131900.
Voivenel L., Varea E., Danaila L., Renou B. and Cazalens M., 2016, 'Variable-viscosity jets : entrainment and mixing process', 'Whither Turbulence and big data', Springer.
Taguelmimt N., Danaila L. and Hadjadj A., 2015, 'Effects of viscosity variations in temporal mixing layer', Flow, Turbulence and Combustion, DOI 10.1007/s10494-015-9649-6.
Krawczynski, J.F., Bouha A., Renou B. and Danaila L., 2015, 'Conserved scalar mixing in a confined opposed-jets flow', Journal of Turbulence, Vol. 16 (12), pp. 1139-1160.
Talbot B., Danaila L. and Renou, B. 2013, 'Variable-viscosity mixing in the very near field of a round jet', Physica Scripta, T155, p. 014006.

News

2017-12-19T13:05:13Z

New results from EM2C

December, 19^th 2017


SoufflageCanon

Combustion instability coupled to a 1A1L standing azimuthal mode of the MICCA-Spray annular combustor leading to partial blow off of the spray flames along the nodal line. Instabilities appear in the form of bursts. (a) At low instability amplitude, all flames are visible at t=0 ms. (b) and (c) Flames diametrically opposed along the nodal line are blown off as the instability level increases during about 40 ms. (d) All flames reattach at t=60 ms when the pressure level drops at the end of the instability burst. Operating conditions, global equivalence ratio φ = 0.85, thermal power P = 111 kW. For more information see : DOI : 10.1115/1.4037824

CORIA

2017-09-28T07:59:34Z

Françoise Baillot, Jean-Bernard Blaisot, Eric Domingues, Marcos Caceres, Gilles Godard

The response of a linear array of injectors to axial and transverse forcing will be examined at CORIA on TACC-Spray bench equipped with one to three injectors. Effects of neighbouring flame interactions will be examined by modifying the distance between injectors (the spacing ratio). The response of the two-phase flow and flame to transverse forcing will be examined from a pressure to a velocity antinode via an intensity antinode where the acoustic flux gradient is maximum. The gas flow dynamics will be examined to highlight the possible excitation of convective modes and global modes of the swirling jet and flames, and reveal the main hydrodynamic/acoustic processes taking place. Particular attention will focus on the liquid flow dynamics by following liquid elements. The relevance of FTF and FDF will be studied for neighbouring flame interactions with axial and transverse forcing in the two-phase flow configuration. For that purpose, measurements of velocity by hot wire in the gas jet injection system, of acoustic pressure in the chamber and in the plenum, of excited radicals as CH* and OH* and visualizations by fast camera and by pulsed source will be possibly completed by advanced laser diagnostics (PIV for the gaseous jet, PDA for the liquid phase, LIF-OH). These data will be used by CERFACS for the corresponding simulations of the forced TACC-Spray chamber. The mechanisms identified and FTF/FDF measured in the TACC-Spray chamber will be compared to the results in the single EM2C burner and to the observations and measurements made in the MICCA-Spray chamber to interpret the self-excited instabilities of that chamber.


a


b


c


d

Figure 1 : (a) Acoustic transverse wave and atomization interaction visualization. (b-d) sub-harmonic flame response to a transverse forcing f0 : jet and flame (b) at a pressure antinode (null pressure gradient) and (d) at an intensity antinode (maximum acoustic flux gradient) ; (c) PSD of CH* emission (linked to the heat release rate).

EM2C

2017-09-27T15:20:11Z

Thierry Schuller, Clément Mirat, Jérôme Beaumier, Daniel Durox, Sébastien Candel

Self-sustained instabilities of spray flames coupled by azimuthal modes will be first analysed in the annular MICCASpray chamber at EM2C as a function of the flow operating conditions by recording the chemiluminescence and acoustic pressure signals in the plenum and chamber associated to the different injectors. These data will be completed by images of the flame dynamics with a high-speed camera at the different injector locations. The flame response to longitudinal flow disturbances will be determined in the single burner setup with a single injector of the annular chamber. Boundary and flow characterizations will be made to provide the data needed by CERFACS to simulate these acoustic responses. The resulting FDF (Fig. 1d) determined by only considering longitudinal forcing will be used to examine the linear and nonlinear stability of the annular combustor and compare predictions from low-order models originally developed for gaseous fuels with measurements gathered at limit cycles in the annular chamber operated with liquid fuels. These comparisons will allow determining the validity of this approach and identifying conditions where multiple flame interactions and transverse response of flames determined in the TACC-Spray chamber need to be considered. Interactions with CORIA in analysing the combined response of these flames to longitudinal and transverse forcing will then provide elements to guide new models including effects of injector separation and transverse forcing.

a	b
c	d

Figure 1 : MICCA setup : (a) with swirling injectors, (b) operating with gaseous fuel and a metallic inner cylindrical wall. (c) Setup used for FDF determination. (d) Example of FDF from a liquid fuel injector.

In all of the above approaches, the key element relies on a FTF or FDF for stability and limit-cycle predictions. To model FTF or FDF, state-of-the-art today points to Large Eddy Simulations (LES) of acoustically isolated burners forced with acoustic waves at the frequencies of interest, the combustion response to these solicitations being measured directly from the unsteady simulations. This method has proved successful in capturing fundamental properties of the FTF/FDF for gaseous and swirled flames. Extension and systematic validation to swirling flames fed by liquid fuel are however missing underlining the importance of the present work proposal.

CERFACS

2017-09-27T15:08:11Z

Laurent Gicquel, Gabriel Staffelbach

CERFACS proposes the following methodology : (1) LES of multiphase flow single sectors for FTF/FDF evaluations : issues pertaining to the fundamental liquid phase modelling and effects of poly-dispersity are indeed to be investigated on the basis of existing turbulent combustion models that can cope with partially-premixed flames ; (2) FTF/FDF and flame interactions : in multi-injection systems, flame elements can locally be supposed isolated while other parts are the results of neighbouring flame interactions. Constructing FTF/FDF based on a single injector domain postulating it is representative of all possible engine configurations may be questionable ; (3) Acoustic evaluation of the full annular burner stability : having assessed the determination of FTF/FDF on basic flame configurations, acoustic solvers developed at CERFACS will be applied to gauge the overall numerical strategy by determining its capacity in predicting the MICCASpray annular chamber stability as well as the establishment of the observed limit-cycle.


Tools developed at CERFACS: for combustion instabilities :
low-order models (ATACAMAC),
Helmholtz simulations (AVSP) and
LES (AVP).

The potential to bring new results in that domain is great since no lab-scale experiments of two-phase flow combustion with multiple injectors disturbed by longitudinal and transverse waves can be found in the literature, whereas it is an important situation found in practical applications. In this respect, simulations devoted to such configurations remain rare. There are yet no comparisons between simulations and experiments on FTF/FDF of flames fed by liquid fuels, no systematic analysis of effects of flame interactions between neighbouring injectors and no attempt was made as yet to compare simulations of azimuthally coupled modes with experiments in an annular systems. These ultimate comparisons are essential to validate numerical strategies for combustion instability predictions in aeronautical engines. Thus, this project will constitute a major advance for fundamental knowledge as well as for methodologies for industrial applications.

LEGRAND

Legrand — 2018-03-07T14:48:08Z

Spécialité Énergétique

Numerical and Modeling Methods for Multi-level Large Eddy Simulations of Turbulent Flows in Complex Geometries

Large-Eddy Simulation (LES) has become a major tool for the analysis of highly turbulent flows in complex geometries. However, due to the steadily increase of computational resources, the amount of data generated by well-resolved numerical simulations is such that it has become very challenging to manage them with traditional data processing tools. In Computational Fluid Dynamics (CFD), this emerging problematic leads to the same "Big Data" challenges as in the computer science field. Some techniques have already been developed such as data partitioning and ordering or parallel processing but still remain insufficient for modern numerical simulations. Hence, the objective of this work is to propose new processing formalisms to circumvent the data volume issue for the future 2020 exa-scale computing objectives. To this aim, a massively parallel co-processing method, suited for complex geometries, was developed in order to extract large-scale features in turbulent flows. The principle of the method is to introduce a series of coarser nested grids to reduce the amount of data while keeping the large scales of interest. Data is transferred from one grid level to another using high-order filters and accurate interpolation techniques. This method enabled to apply modal decomposition techniques to a billion-cell LES of a 3D turbulent turbine blade, thus demonstrating its effectiveness. The capability of performing calculations on several embedded grid levels was then used to devise the multi-resolution LES (MR-LES). The aim of the method is to evaluate the modeling and numerical errors during an LES by conducting the same simulation on two different mesh resolutions, simultaneously. This error estimation is highly valuable as it allows to generate optimal grids through the building of an objective grid quality measure. MR-LES intents to limit the computational cost of the simulation while minimizing the sub-grid scale modeling errors. This novel framework was applied successfully to the simulation of a turbulent flow around a 3D cylinder.

Rapporteur
Y. DUBIEF, Professeur associé à l'Université du Vermont USA
F. NICOUD, Professeur à l'Université de Montpellier, IMAG
Examinateur
L. Danaila, Professeur à l'université de Rouen
G. Balarac, Maître de conférence, LEGI
A. Larat, Chargé de recherche CNRS, EM2C
V. Brunet Chef du département CFD, Safran
Directeur de thèse
A. BERLEMONT, Directeur de recherche CNRS, CORIA
Co-encadrant de thèse
V. MOUREAU, Chargé de recherche CNRS, CORIA

KANJANASAKUL

Kanjanasakul — 2018-05-30T09:48:39Z

Physique et Energétique

Analyse des propriétés optiques de gouttes de différents fluides dans un environnement gazeux haute pression par réfractométrie d'arc-en-ciel

La caractérisation des propriétés optiques de gouttes de combustible aux alentours de la pression critique est un défi. L'objectif de la thèse est de mesurer la taille et l'indice de réfraction de fluides dans des conditions de pression élevée proche du point critique. A cet effet, une installation expérimentale d'injection de gouttes à haute pression a été conçue. La réfractométrie d'arc-en-ciel a été développée et adaptée pour travailler d'abord sur des gouttelettes d'eau et d'éthanol dans le domaine de pression 1 - 40 bar. Des évolutions de leur indice de réfraction avec la pression ont été obtenues. Pour les mesures se rapprochant d'un fluide proche du point critique, l'éthane a été sélectionné car son point critique est à 48,7 bar et 32,2°C. Des mesures de réfraction d'index sur des gouttes liquides d'éthane ont ensuite été réalisées dans le domaine 40 - 46 bar et 18 - 25°C. Une mesure de l'indice de réfraction de gouttes d'éthane donne alors une valeur de 1,255 ± 0.002.

Analysis of the optical properties of droplets of different fluids in high-pressure environments by rainbow optical diagnostic

The characterization of the optical properties of fuel drops around the critical pressure is a challenge. The aim of the thesis is to measure the size and the refractive index of fluids under high pressure conditions close to the critical point. For this purpose, an experimental installation of injection of high-pressure drops has been designed. Rainbow refractometry has been developed and adapted to work first on water and ethanol droplets in the pressure range 1 - 40 bar. Evolutions of their index of refraction with the pressure were obtained. For measurements approaching a fluid near the critical point, ethane was selected because its critical point is 48.7 bar and 32.2 ° C. Index refractive measurements on liquid drops of ethane were then carried out in the range 40 - 46 bar and 18 - 25 ° C. A measurement of the refractive index of ethane drops then gives a value of 1.255 ± 0.002

Rapporteur
Mme Grazia LAMANNA, Directeur de Recherche, Institut für Thermodynamik der Luft- and Raumfahrt, Germany
M. Jeroen VAN BEECK, Professeur, Von Karman Institute for Fluid dynamics, Belgium

Examinateur
Mme Christine MOUNAIM ROUSSELLE, Professeur, PRISME/Polytech'Orléans
M. Guillaume RIBERT Maître de Conférences, INSA Rouen Normandie - CORIA

Directeur de thèse
M. Frédéric GRISCH, Professeur, INSA Rouen Normandie - CORIA

Codirecteur de thèse
M. Gérard GRÉHAN, Directeur de Recherche CNRS, UMR 6614 - CORIA

VU

Vu — 2017-12-20T13:44:35Z

Spécialité Physique Energétique

Abstract
Liquid fuel atomization is crucial for the performance of internal combustion engines. Through an injector, the liquid is delivered into the combustion chamber and breaks down into droplets. The finer the drops, the quicker their evaporation and the more proper their mixing with air. A proficient combustion could hence be expected, with low pollutant emissions.
Atomization quality is primarily affected by the injector design and the operating conditions which shape the internal flow structure, the turbulence level, the velocity profile at the nozzle outlet, the cavitation and so forth. All these features are determinants of the breakup of the external liquid flow.
Another key parameter to optimize the atomization process is the fuel physical properties. One can think of, among others, the dynamic surface tension controlled by the diffusion of the surfactants on the liquid-gas interface or the extensional viscosity which makes a liquid to become more resistant to the stretching, thereby affecting the breakup.
Effects of the injector design, the operating conditions and the liquid properties on the atomization are inter-dependent. Analyses of experimental data help us to understand the involved mechanisms and their interactions. On the one hand, this is useful for the numerical developments which should be carried out depending upon the configuration. On the other hand, quantitative criterion could be established to validate the simulation results.
Following the above research methodology, we aim to study the disintegration of planar turbulent liquid sheets produced by a triple-disk injector. Experimental measurements provide the sheet images, used as input for a multi-scale analysis. We investigate, thanks to the latter, the behaviours of the liquid sheet, the ligaments appearing on its edges and the resulting droplets. Moreover, two immersed boundary methods are developed, aiming to simultaneously solve the nozzle flow and the breakup process. We carry out two applications, the first one on a liquid jet ejected by a cylindrical nozzle and the other a planar sheet issuing from a triple-disk injector.

Résumé
Le processus d'atomisation est important pour la performance des moteurs à combustion interne. Grâce à un injecteur, le carburant liquide est admis dans la chambre de combustion et se divise en gouttelettes. Plus petites les gouttes, plus rapide leurs évaporation et meilleur le mélange air-carburant. Une meilleure combustion pourrait être obtenue, avec faibles émissions polluantes.
La qualité de l'atomisation est influencée principalement par la géométrie de l'injecteur et les conditions opératoires qui forment la structure de l'écoulement interne, la turbulence, le profil de vitesse à la sortie de l'injecteur, la cavitation, etc. Tous ces aspects sont déterminants pour la rupture de l'écoulement externe.
Un autre paramètre clé pour optimiser le processus d'atomisation est les propriétés physiques des carburants. On pense, parmi autres, à la tension de surface dynamique contrôlée par la diffusion des surfactants sur l'interface liquide-gaz ou à la viscosité extensionnelle qui rend un liquide plus résistant à l'étirement, influençant donc la rupture.
Les effets de la géométrie de l'injecteur, les conditions opératoires et les propriétés physiques des liquides sur l'atomisation sont inter-dépendents. Les analyses expérimentales nous aident à comprendre les mécanismes impliquant et leurs interactions. D'une part, elles sont utiles pour les simulations numériques qui devraient être réalisées suivant la configuration considérée. D'autre part, les critères quantitatifs pourraient être établis afin de valider les résultats numériques.
En suivant cette méthodologie de recherche, nous souhaiterons étudier l'atomisation d'une nappe liquide turbulente produite par un injecteur triple-disque. Les mesures expérimentales fournissent les images de la nappe, utilisées comme l'entrée d'une approche multi-échelle. Nous étudions, grâce à la dernière, les comportements de la nappe, des ligaments qui apparaissent à ses bords et des gouttelettes. De plus, deux méthodes de frontières immergées sont développées pour résoudre en même temps l'écoulement interne et le processus d'atomisation. Nous réalisons deux applications, la première pour un jet liquide éjecté par un injecteur cylindrique et la deuxième pour une nappe plane produite par un injecteur triple-disque.

Rapporteur
M. Stéphane VINCENT, Professeur, MSME, Université Paris-Est Marne-la-Vallée
M. Rudy BAZILE, Maître de Conférences, HDR, IMFT, Toulouse

Examinateurs
Mme. Françoise BAILLOT Professeure, CORIA, Université de Rouen Normandie, Rouen
Mme. Marie-Charlotte RENOULT, Maître de Conférences, CORIA, INSA de Rouen
M. Jérôme HELIE,Expert Senior, Responsable Equipe Nozzle & Sprays, CONTINENTAL, Toulouse

Directeur de thèse
M. Christophe DUMOUCHEL, Directeur de Recherche, CORIA, Université de Rouen

Co-directeur
M. Thibaut MENARD Maître de Conférences, CORIA, Université de Rouen