Arnaud Mura

Computational Fluid Dynamics Investigation of a Mach 12 Scramjet Engine

The description of combustion in high-speed turbulent flows, where turbulent mixing, compressibility effects, and chemical kinetics processes are competing, still remains a challenging issue for numerical simulations. The key features of such turbulent supersonic reactive flows are integrated into a turbulence–chemistry interaction closure, which relies on the partially stirred reactor framework. The corresponding closure, hereafter denoted as the unsteady partially stirred reactor model, is incorporated into the ONERA computational fluid dynamics code CEDRE. The computational model is used to investigate a Mach 12 rectangular-to-elliptical-shape-transition scramjet engine developed and operated at the University of Queensland. Results of Reynolds-averaged Navier–Stokes numerical simulations based on either the unsteady partially stirred reactor concept or the classical quasi-laminar closure are presented and compared. Similar results in terms of pressure distributions are obtained, thus confirming that t...

Influence of Residence and Scalar Mixing Time Scales in Non-Premixed Combustion in Supersonic Turbulent Flows

In high Mach number turbulent reactive flows, spontaneous ignition appears as a key ingredient for the stabilization of combustion. In our previous analyses devoted to such conditions, chemical kinetics as well as associated finite rate chemistry effects have already received considerable attention. However, the representation of flow time scales, such as residence and mixing time scales, still requires further work. The present modeling study is devoted to this peculiar point. Hence, a transport equation for the quantity of residence time is considered to evaluate the mean residence time associated with both oxidizer and fuel injection streams, while a modeled transport equation for the mean scalar dissipation rate (SDR) is considered to estimate the scalar mixing time scale. This allows us to improve the description of turbulent mixing, including the large-scale engulfment processes through the consideration of the residence time scale, as well as the small-scale molecular mixing processes, the intensity of which is set by the integral scalar (mixing) time scale. Some insights are gained from the analysis of the evolution of the different production terms that appear in the mean SDR transport equation. The model capabilities are evaluated through a comparison between numerical results and the data obtained from experimental studies devoted to supersonic coflowing jets of hydrogen and vitiated air. The first simulated test case corresponds to a detailed experimental database that includes Raman scaterring and laser-induced pre-dissociative fluorescence measurements. Albeit less documented, the second test case is retained to confirm the relevance of the proposed closure. Finally, the comparisons performed with the two distinct sets of experimental data establish that the main physical processes are well-described by the proposed approach.

Algebraic Models for Turbulent Transports in Premixed Flames

The thermal expansion induced by the chemical reactions taking place in a turbulent reactive flow of premixed reactants affects the velocity field so strongly that turbulent transports can be controlled by reaction rather than by turbulence. Moreover, thermal expansion is well-known to cause countergradient turbulent diffusion as well as flame-generated turbulence phenomena. In the present article, a splitting procedure of the velocity field is used that allows the identification of two different effects of the thermal expansion in the specific flamelets regime of turbulent premixed combustion: (i) the thermal expansion occurring through the local flames (direct effect) and (ii) the effect of thermal expansion on the velocity field associated to the growth of the flame surface (indirect effect). Algebraic closures for the turbulent transport terms of mass and momentum are proposed where the effect of the turbulent mixing (nonreactive effect) is modeled by classical closures, i.e., gradient law, while the contributions associated with thermal expansion are closed by taking advantage of flamelet relationships. Finally, this simple model is applied to the numerical simulation of a turbulent flame stabilized by the sudden expansion of a 2D channel. Corresponding results are satisfactorily compared with experimental data and confirm the ability of the model to represent the behavior of turbulent transports in premixed flames.

Modelling of Self-Ignition Processes in Supersonic Non Premixed Coflowing Jets Based on a PaSR Approach

The description of combustion in high-speed turbulent flows where turbulent mixing, compressibility effects and chemical kinetics processes are competing, still remains a challenging issue for numerical simulations. The key features of such turbulent supersonic reactive flows are highlighted and integrated into a new turbulence-chemistry interaction (TCI) model which relies on the partially stirred reactor (PaSR) framework. The corresponding model, hereafter denoted PaSR model, is integrated into the ONERA CFD code CEDRE. Preliminary results are reported and discussed. An extended closure denoted EPaSR (Extended PaSR) is also introduced. It features the ability to capture both unsteady and convection effects thus extending the PaSR closure. The corresponding effects are expected to play a crucial role in the ignition and stabilization processes associated with non-premixed supersonic turbulent flames. A Mach 2 hydrogen-air coflowing jet diffusion flame experiment is retained as a validation test case. The preliminary results of Reynolds Average Navier Stokes (RANS) numerical simulations based on the EPaSR concepts are presented.

Mixing of Fuel Jet in Supersonic Crossflow: Estimation of Subgrid-Scale Scalar Fluctuations

Non-reactive large-eddy simulations (LES) of a hydrogen jet injected into a transverse supersonic flow of vitiated air at Mach 2 are conducted. The present investigation is focused on the intertwin...

Relevance of Two Basic Turbulent Premixed Combustion Models for the Numerical Simulations of V-Shaped Flames

ABSTRACT The present study is devoted to the analysis of basic turbulent premixed combustion closures applied to the numerical simulations of V-shaped flames. It is well known that an important parameter for the numerical simulation of such premixed turbulent flames is the description of the departure from the bimodal limit (thin flame limit), which is associated to the maximum value of the progress variable segregation rate, i.e., S = 1. The evolution of this segregation rate is often deduced from a modeled transport equation written for the progress variable variance. However, the closure of such a transport equation does involve many additional sub-models, which are related to the mean and variance progress variable fluxes and mean scalar dissipation rate of the progress variable variance. In the present work two original closures for the mean chemical rate are considered. Special emphasis is also placed on algebraic closures for S that circumvent the difficulty associated to the modeling of the second moment of the progress variable. In the first step of the analysis these closures are analyzed in the case of the propagation of a one-dimensional turbulent premixed flame brush. They are subsequently applied to the numerical simulation of premixed V-shaped flames that have been studied experimentally by Galizzi (2003) and by Degardin et al. (2006). It is found that these closures provide a satisfactory representation of the turbulent premixed flames.

Density Variations Effects in Turbulent Diffusion Flames: Modeling Of Unresolved Fluxes

Unresolved fluxes in turbulent diffusion flames are investigated by introducing the specific volume to analyze the effects of density variations. Unresolved fluxes are found to be related to scalar correlations involving this specific quantity. The algebraic models proposed for the turbulent scalar and momentum fluxes allow to recover and generalize well-known previously established relations and highlight the possible occurrence of non-gradient diffusion. These correlations are evaluated from the consideration of strained laminar diffusion flames and chemical equilibrium conditions. Finally, the calculations performed confirm that such non-premixed flames may exhibit a strong production of turbulence near stoichiometric conditions.

A Layered Description of a Premixed Flame Stabilized in Stagnating Turbulence

ABSTRACT In the thin-flame regime of turbulent premixed combustion, instantaneous progress variable gradients are essentially fixed by propagating flamelets. The laminar flamelet internal structure thus imposes, at least to some extent, the progress variable one-point one-time statistics, i.e., the progress variable PDF . In the present study a generalized presumed probability density function (PDF) shape is considered to account for possible departures from the thin-flame limit. The parameters that characterize this PDF shape depend on the ratio of the Kolmogorov length scale to laminar flame width. The corresponding framework may be associated to a layered description of the turbulent flame brush (TFB) with the thicknesses of the different sub-layers determined from the knowledge of both the Reynolds and Karlovitz numbers, ReT and Ka. It incorporates boundary layers on both sides of the TFB, which are associated to the finite thickness of the local flamelets, i.e., small but finite values of the Karlovitz number. This description leads to a modeling proposal for premixed turbulent combustion. A tabulation is constructed based on canonical one-dimensional planar unstrained laminar premixed flame computations and the different quantities are tabulated as functions of ReT and Ka. The final closure is applied to the numerical simulations of premixed flames stabilized in stagnating turbulence.

Importance of Physical Modeling for Simulations of Turbulent Reactive Flows

The physical models implemented in practical computational tools are not systematically required for the numerical simulations of turbulent flows. When the grid is sufficiently refined, satisfactory numerical results can be obtained even if the smallest characteristic scales are not solved. However, when reactive flows are considered, the physical mechanisms occurring at the smallest scales may control the main characteristics of the flow such as the flame velocity propagation. Therefore, the development of new physical models is still needed for practical numerical simulations of turbulent reactive flows. A recent work that describes the inner structure of turbulent flames as composed of different layers is presented. This study also evidences the necessity to understand in details the transition between a slow chemistry layers to a fast chemistry layer. The behavior of the scalar variance and turbulent scalar flux between these two limit cases is presented.