2.1 Reacting Flows
- DNS for Turbulent combustion (B. Cuenot,
C. Jiménez, A. de Lataillade,
T. Poinsot)
- LES of turbulent combustion (J-P. Légier,
G. Lartigue, L. Selle, K. Truffin,
C. Prière, L. Gicquel, B. Cuenot,
C. Jiménez, F. Ducros, F. Nicoud,
T. Poinsot)
- Reduced kinetics for LES of combustion (K. Truffin,
G. Lartigue)
- Flame/acoustics interactions (A. Kaufmann,
L. Selle, F. Nicoud)
- Flame/wall interaction (F. Dabireau,
A. de Lataillade, B. Cuenot,
T. Poinsot)
- Atmospheric pollution (B. Cuenot,
R. Paoli, F. Laporte, F. Cousin)
2.1.1 DNS for Turbulent combustion (B. Cuenot,
C. Jiménez, A. de Lataillade,
T. Poinsot)
In 2000 two FP'5 European projects have started, and continued in 2001:
Glevel and STOPP.
Related to the topic of turbulent stratified combustion, the project Glevel
is now focusing on Direct Numerical Simulations of octane/air stratified
flames (see Fig. 2.1). Performing such DNS with complex chemistry
schemes is one of the unique capacities found in the CFD team
[Jiménez, 2000b].
The octane chemical kinetic scheme has been provided by RWTH Aachen, partner
of the project. Diagnostics are performed to analyse the effect of a
non-homogeneous reactant mixture on the primary flame front and on the
secondary or post-flame. The complexity of chemical kinetics makes it
difficult to identify simple structures in the secondary flame, and a
particular effort in modeling this reaction zone has to be done. In parallel
work is still going on with Prof. Dan Haworth (Penn State University) on
propane stratified combustion (2000 CTR Summer Program).
These two projects provide a unique data base of the processes involved in
stratified combustion. The next step is to develop models, which is the c
coming part of Glevel. Such modeling is already in progress in collaboration
with PSA and University of Southern California (Pr. Egolfopoulos) who studied
stratified combustion in Direct Injection engines
[Lauvergne, 2000].
The team is still involved in the american ITR/ACS project, first submitted
to NFS, later to AFOSR, coordinated by D.Haworth and involving Dr. S. Pope
(Cornell) and Dr. J. Chen (Sandia). The aim of this project is to implement
the ISAT technique in our DNS code, to compute complex chemistry at a reduced
cost.
The Research Training Network STOPP, coordinated by CERFACS, is also fully
operational. Seven young researchers have been hired up to now in the
Network, and at least two other persons should join in the coming months.
The scientific objective of the project is a better understanding of
pollution processes: pollutant emissions by industrial systems as well as
dispersion and reaction in the atmosphere are studied. The contribution of
CERFACS in this project is to study LES of turbulent combustion, and a Ph.D.
thesis (A. Kaufmann), has started in September 2000 to work on this topic.
The PRIDE project (FP'4) has been terminated in March 2000. Direct Numerical
Simulations of turbulent diffusion flames with NOx formation have been
achieved and used for assessment of turbulent combustion models. This project
has also given the opportunity to develop the chemistry reduction procedure
ICC, now coupled to the AVBP code for an operational use in complex
configurations computations.
Figure 2.1: DNS of stratified octane/air combustion: heat release.
Left: homogeneous mixture, right: stratified mixture
[Jiménez, 2000b].
2.1.2 LES of turbulent combustion (J-P. Légier,
G. Lartigue, L. Selle, K. Truffin,
C. Prière, L. Gicquel, B. Cuenot,
C. Jiménez, F. Ducros, F. Nicoud,
T. Poinsot)
The thickened flame model for LES of turbulent combustion developed in
collaboration with Ecole Centrale Paris
[Colin, 2000]
is now routinely used in AVBP calculations
[Selle, 2000;
Wintrebert, 2000].
However its assessment and validation still need some work, which will be
mainly based on DNS. An example is the work done on the problem of the
subgrid scale variance and dissipation of a scalar field
[Jiménez, 2000].
Another example is the study of diffusion flames stabilisation processes
in a turbulent environment
[Légier, 2000].
To incorporate more realistic chemistry into AVBP
[Lartigue, 2000a],
multi-reaction schemes are needed in order to reproduce not only the
pollutants formation but also the flame temperature (see Fig.
2.2). To this end, a new collaboration with IFP (A. Torres,
G. Gauthier) has started in early 2000. A prototype multispecies version
of AVBP, suitable for the implementation of multi-reaction kinetics, is
now available at CERFACS.
In the field of LES, flame transfer functions have been computed for
various geometries. This information is the key data needed in global
acoustic models for predicting the stability map of the combustion.
Transfer functions obtained from AVBP calculations are being compared to
measurements performed on both academic (ICLEAC european project and SNECMA
CIFRE thesis, experiment done at EM2C) and more industry oriented burners
(Siemens KWU, experiment done at Univ. of Karlsruhe).
CERFACS is also involved with Ecole Centrale de Paris (Laboratoire E.M2.C.)
and CORIA in a project supported by the COS (Comité d'Orientation
Supersonique) program, to study new combustion technologies for supersonic
flight. CERFACS and CORIA develop models and Ecole Centrale performs an
experimental investigation to validate LES models.
Figure 2.2: LES of combustion in a gas turbine (temperature field).
2.1.3 Reduced kinetics for LES of combustion (K. Truffin,
G. Lartigue)
Being able to include more complex kinetics into LES is a necessary condition
for the future. CERFACS has investigated various reduced schemes to reach
this goal, first for methane / air combustion, starting from two steps
schemes up to four steps. Results obtained with the classical Jones Linstedt
scheme have proved to be interesting since they provide correct flame
structures at a reasonable cost (Fig. 2.3). These schemes will
be implemented into AVBP in 2002.
Figure 2.3: Premixed flame structure for a methane / air flame computed with a
full chemical scheme (1) and with a four step scheme (4).
2.1.4 Flame/acoustics interactions (A. Kaufmann,
L. Selle, F. Nicoud)
In practice, it is possible to introduce acoustic perturbations in many
different ways in a stable calculation. Not all these techniques lead to
reliable results in terms of flame transfer function. Both analytical and
numerical work have been done in order to study the way a flame should be
excited in a LES
[Kaufmann, 2000].
2.1.5 Flame/wall interaction (F. Dabireau,
A. de Lataillade, B. Cuenot,
T. Poinsot)
CERFACS has developed a strong expertise in the field of flame/wall
interaction since 1995, both for piston engines and rocket engines
applications. These studies have addressed the case of premixed flames
interacting with cold walls. Since 2000, a new project has been started,
supported by CNES and SNECMA/DMF, to perform DNS's of H2/O2 flames
interacting with an isothermal wall. This mechanism is essential for the
design of nozzles in rocket engines.
2.1.6 Atmospheric pollution (B. Cuenot,
R. Paoli, F. Laporte, F. Cousin)
Two projects support the activity related to atmospheric pollution. One is
a regional project PREPA (coordinated by Météo-France, CNRM Toulouse),
started in january 2000. Work has been performed on the problem of the
subgrid scale segregation of reactive species
[Cousin, 2001].
This work is now continuing at CNRM and Laboratoire d'Aérologie
through a PhD thesis (supported by CNRS and ADEME).
The other project is supported in the framework of the Comité
d'Orientation Supersonique (COS). CERFACS is involved to perform
reactive simulations of airplane wake, as started in April 2001.
|
|
