A. Sadiki

A digital filter based generation of inflow data for spatially developing direct numerical or large eddy simulations

In contrast to time-evolving turbulence, direct numerical or large eddy simulations of spatially inhomogeneous flows require turbulent inflow boundary conditions, that make the results strongly influenced by the velocity profiles to be prescribed. This paper aims to present a new approach for generating artificial velocity data which reproduces first and second order one point statistics as well as a locally given autocorrelation function. The method appears to be simple, flexible and more accurate than most of the existing methods. This is demonstrated in two cases. First, direct numerical simulations of planar turbulent jets in the Reynolds number range from 1000 to 6000 are performed. Because of the importance of the primary breakup mechanism of a liquid jet in which inflow influences are evident, the new procedure is secondly used, to study atomization in dependence of the flow inside the nozzle by means of a Volume of Fluid scheme.

Investigation of the influence of the Reynolds number on a plane jet using direct numerical simulation

Abstract Free jets represent a benchmark for research into the physics of turbulent fluid flow and are furthermore of great interest for many engineering applications. In the present work we investigate the influence of the Reynolds number on the evolution of a plane jet. The effect on the global jet characteristics, as well as on some spectral properties, is particularly addressed. A strong influence on the jet evolution is found for Re⩽6000, but also that the jet is close to a converged state for higher Reynolds numbers. Although it is believed that the jet reaches a universal self-similar state, there was early evidence that the inflow conditions can have a downstream effect on the development of the turbulent flow field. Therefore some results, concerning the influence of inflow boundary conditions on the simulations, are also reported.

Large-eddy simulation of a counterflow configuration with and without combustion

A large-eddy simulation in three dimensions was used to study the flow, mixing fields, and combustion in a counterflow burner. A non-reactive case (air/air jets) and a reactive case (methane/air jets) were investigated. Such a configuration is well suited to study and calibrate models for non-premixed flames because of its simplicity and versatility. In the numerical method, fluctuations of density in space and time were considered to depend only on chemistry, not on pressure. The effect of heat release was included by means of the mixture-fraction formulation. To represent the subgrid scale stresses and scalar flux, a Smagorinsky model was used, in which the Smagorinsky coefficient was determined by the dynamic Germano procedure. An equilibrium chemistry model was used to relate the mixture fraction to density, temperature, and species concentrations. The subgrid distribution of the mixture-fraction fluctuation was presumed to have the shape of a β-function. The computed results were found to be in overall agreement with experimental data for the non-reactive case. For the reactive case, in which a simple combustion model was used, a satisfactory agreement with measured data was achieved. A strong influence of combustion on turbulence mechanisms is apparent.

A Turbulence-Chemistry Interaction Model Based on a Multivariate Presumed Beta-PDF Method for Turbulent Flames

A turbulence-chemistry interaction model based on presumed probability density functions (PDF) is presented. It can be coupled with conventionally reduced mechanisms and is capable of capturing major and minor species distribution features in turbulent diffusion flames. Combined with a reduced mechanism using intrinsic low-dimensional manifolds (ILDM), the method in which the joint PDF is assumed to be a product of one-dimensional β-PDFs is successfully applied to model the turbulent mixing and scalar field of a turbulent piloted methane/air flame. Although the so-called flames E and F provide a superior test of a models ability to treat finite-rate chemistry, this work focusses on the flame D. A Reynold stress closure of second order is used for the turbulence description whilst gradient ansatz are postulated for scalar fluxes. Results of the simulations in form of means and variances of velocity and scalars and conditional means are compared to experimental data.

Prediction of finite chemistry effects using large eddy simulation

A large eddy simulation (LES) in three dimensions is applied to study flow, mixing, and combustion ina highly turbulent jet flame. Turbulence chemistry interactions, including finite rate chemistry effects, are investigated. The hydrogen fuel has been diluted with nitrogen to allow for both accurate numerical and accurate experimental investigation. In the numerical method, fluctuations of density in time and space are considered to depend only onthe chemical state, not on pressure. This low-Mach assumption greatly improves the efficiency of the code. Mixing and the effects of heat release are included by means of the mixture-fraction formulation. To model subgrid scale stresses and scalar fluxes, the Smagorinsky model is used since the dynamic Germano procedure did not show any particular advantage for this flame. To relate mixture fraction to density, temperature, and species concentrations, a steady flamelet model is used. To evaluate the performance of LES with steady flamelet chemistry, a comparison has been made to experimental data, as well as to the results of a probability density function simulation, with a five-step mechanism considering differential diffusion effects. This is done in terms of averaged quantities, scatter plots, and conditional averages. The LES results were found to be in good agreement with the existing data. For this stable flame, the influence of differential diffusion (inherent to hydrogen flames) seems to be negligible.

A discrete model for the apparent viscosity of polydisperse suspensions including maximum packing fraction

Based on the notion of a construction process consisting of the stepwise addition of particles to the pure fluid, a discrete model for the apparent viscosity as well as for the maximum packing fraction of polydisperse suspensions of spherical, noncolloidal particles is derived. The model connects the approaches by Bruggeman and Farris and is valid for large size ratios of consecutive particle classes during the construction process, appearing to be the first model consistently describing polydisperse volume fractions and maximum packing fraction within a single approach. In that context, the consistent inclusion of the maximum packing fraction into effective medium models is discussed. Furthermore, new generalized forms of the well-known Quemada and Krieger–Dougherty equations allowing for the choice of a second-order Taylor coefficient for the volume fraction (ϕ2-coefficient), found by asymptotic matching, are proposed. The model for the maximum packing fraction as well as the complete viscosity model is...

Prediction of Swirling Confined Diffusion Flame With a Monte Carlo and a Presumed-PDF-Model

Abstract This work aims to compare numerical results obtained by using the Monte Carlo composition-PDF method and a presumed-β-PDF in order to reveal their effects on the prediction of flow and scalar fields in swirling confined methane diffusion flame. Using the intrinsic low dimensional manifolds method for modelling the chemistry and a second moment closure for the turbulence, it is shown that both PDF-methods provide a similar accuracy level of the prediction of mean quantities. While the presumed-β-PDF performs using reasonable computational efforts, the Monte Carlo-PDF allows to capture well the turbulence–chemistry interaction and strong finite-chemistry effects such as local extinction.

Towards a classification of models for the numerical simulation of premixed combustion based on a generalized regime diagram

The inner structure, and the physical behaviour of turbulent premixed flames are usually described, and classified by means of the regime diagram introduced by Borghi and Peters. Thereby properties related to both the flame and the (turbulent) flow are considered. In this work a diagram valid for all physical regimes, comprising suitable requirements for laminar simulations, direct numerical simulation (DNS), large-eddy simulation (LES), and Reynolds averaging based numerical simulation (RANS) is proposed. In particular the diagram describes essential situations within the validity limits of the “Borghi, Peters diagram” which physical phenomena are resolved by the simulation, and which have to be modelled. This information is used for systematic classification of various models by suggesting specific models that are appropriate depending on the regime and numerical resolution, and may provide guidance for numerical simulation methods and model development in turbulent premixed combustion. This might help users as a guideline in choosing appropriate models for a given device, and numerical effort available. The regime diagram suggested by Pitsch and Duchamp de Lageneste, which includes DNS and LES by explicitely accounting for the numerical related variable filterwidth, emerges here as one of the special two-dimensional cases possible. In contrast to the generalized regime diagram, their diagram does not include laminar simulations, and RANS based considerations, while transition between wrinkled and corrugated flamelets is not clearly established.

Study of interaction in spray between evaporating droplets and turbulence using second order turbulence RANS modelling and a Lagrangian approach

The objective of this work is to analyse the interaction between evaporating droplets and the turbulent flow of the carrier gas using a second order turbulence RANS modelling in an Euler-Lagrangian approach including two-way coupling. The analysis is performed in a vertical isopropanol polydispersed spray issuing into a co-flowing, heated air stream. The results mainly display the turbulence modulation effects besides the turbulent dispersion of evaporating droplets: 1) The influence of droplet diameter and interface transport on the gas phase turbulence as well as 2) The effect of the droplet evaporation on the mass and heat transfer processes are compared with experimental data. 3) The reciprocal influence of the modification of interface transport by turbulence, i.e., the influence of the turbulent fluctuations of different quantities (such as velocity, temperature, etc.) on the droplet evaporation (through the mass and heat evaporation rates, the droplets diameter distribution) are computed and discussed.

Large eddy simulation of combustion processes under gas turbine conditions

The method of large eddy simulation (LES) has a high potential to accurately predict complex turbulent flows. Gas turbine combustion systems feature a number of phenomena interacting with each other such as swirl with recirculation, complex turbulent mixing and combustion of premixed, non-premixed as well as partially premixed nature. We therefore tend to approach the simulation of real gas turbine combustors step by step. The aim of this paper is to document some of the progress made at EKT in assessing the capability of LES in flows separately exhibiting distinct features of gas turbine combustors. Some results from two configurations are presented and discussed: a non-confined isothermal swirl flow with precessing vortex core and a non-premixed bluff-body flame.