Jet combustion applications
Kerstein, Alan
All,
This is a discussion of the ODT jet combustion model development path that might interest some of you.
Alan
-----Original Message-----
From: Kerstein, Alan
Sent: Friday, December 12, 2008 8:59 AM
To: 'Michael Oevermann'
Cc: Christina Schrödinger; Heiko Schmidt
Subject: RE: ODT question
Dear Michael,
This has not been done for premixed but it is certainly doable. Of course, one of the features of ODT is that it is regime-independent, which is a popular term these days among state-space (vs. physical space) combustion modelers who recognize this as a worthy, but from their perspective distant, goal.
I will mention some substantial improvements since the 2001 C&F paper. The first step was development of a planar spatially developing formulation (Ashurst & Kerstein Phys. Fluids 2005), which required introduction of an adaptive mesh (second step) for numerically accurate application to combustion in planar free shear flows. A paper describing this (with application to a planar jet diffusion flame), submitted to Combust. Sci. Tech., is attached. The adaptive mesh enabled a cylindrical spatially developing formulation (third step) that Niveditha Krishnamoorthy developed and applied to turbulent round jet diffusion flames. Her Ph.D. thesis describing this is downloadable via the ODT google group site, and I have her code. David Lignell's almost-ready c++ code (previous codes were fortran) has better mesh adaption than anything previous and a general CANTERA interface (fourth and fifth steps). It is a planar time developing code, but using Niveditha's code as a model, can be converted straightforwardly to spatial development and/or cylindrical geometry. This effort would be relatively modest because it would be a synthesis of demonstrated capabilities, but it would have disproportionate payoff because David's code is designed for extensibility to a variety of interesting flow configurations and small scale processes (multiphase, etc.). Admittedly there would be a long learning curve, but that applies to ODT in general, and many students and postdocs (not to mention Heiko!) have handled this well. If Christina pursues this, a wide range of premixed, non-premixed, and intermediate regimes would be within reach, initially for jets but subsequently for other applications.
But why go beyond the 2001 C&F methodology? Two (of many) benefits of spatial development are (1) residence time is inversely proportional to vertical velocity, so slow chemistry such as for NOX in the far field has the appropriate residence time and therefore avoids underestimation of NOX production as can occur using a time developing formulation, (2) As illustrated in the attached, continuity-equation effects such as near-field flame necking in plumes (buoyancy dominated) are captured. Cylindrical vs. planar is important because it affects similarity laws governing the streamwise development of mean velocity, centerline mixture fraction, jet width, etc. Coincidentally, temporal planar jets have the same similarity laws as spatial round jets, but spatial planar jets are different.
I've written here mainly in the context of diffusion flames because that is our main experience to date, but these considerations also apply to premixed flames.
I think this is of general interest, so I will copy it, without the attachment, to the google group, also to Suresh Menon, who has done premixed applications of LEM.
Best wishes,
Alan
-----Original Message-----
From: Michael Oevermann [mailto:michael....@tu-berlin.de]
Sent: Friday, December 12, 2008 3:03 AM
To: Kerstein, Alan
Cc: Christina Schrödinger
Subject: ODT question
Dear Alan,
I just had a look at your ODT Combustion & Flame Paper from 2001 with Tarek Echekki, Thomas Dreeben, and Jyh-Yuan Chen. I was wondering if something similar (I mean construction of a 2D solution by advecting the 1D ODT domain with the flow) has been done for premixed combustion system? If not and if you think it is possible to do that might actually be a very interesting problem to look at for Christina.
Best regards
Michael
--
==========================================
Dr. Michael Oevermann
TU Berlin, Fak. III, Prozesswissenschaften Institut für Energietechnik Fasanenstr. 89, 10623 Berlin
Tel.: ++49 (0) 30 3142 2452
Fax : ++49 (0) 30 3142 2157
mailto: michael....@tu-berlin.de
==========================================