Bulletin of the American Physical Society
60th Annual Meeting of the Divison of Fluid Dynamics
Volume 52, Number 12
Sunday–Tuesday, November 18–20, 2007; Salt Lake City, Utah
Session JQ: Reacting Flows III |
Hide Abstracts |
Chair: V. Raman, University of Texas, Austin Room: Salt Palace Convention Center 251 E |
Monday, November 19, 2007 3:35PM - 3:48PM |
JQ.00001: Artificial neural networks for chemistry representation. Part 1: Generation of optimal ANNs using a pattern search algorithm Matthias Ihme, Christoph Schmitt, Heinz Pitsch Surrogate-based derivative-free optimization is applied to design an artificial neural network (ANN). Optimization is performed using a mixed variable extension to the generalized pattern search method. This method offers the advantage that categorical variables, such as the type of the neuron transfer function or the network connectivity, can be used as parameters in optimization. When used together with a surrogate, the resulting algorithm is highly efficient for expensive objective functions. Results from a chemistry example demonstrate the effectiveness of this method in optimizing an ANN for the number of neurons, the type of transfer function and the connectivity between layers. [Preview Abstract] |
Monday, November 19, 2007 3:48PM - 4:01PM |
JQ.00002: Artificial neural networks for chemistry representation. Part 2: Application in LES of turbulent reacting flows Christoph Schmitt, Matthias Ihme, Heinz Pitsch Numerical simulations of turbulent reactive flows often involve thousands of reactions among hundreds of species. Flamelet based combustion models or intrinsic lower-dimensional manifold (ILDM) methods employ a reduced set of scalars for the representation of all thermochemical quantities. In these methods, table look up techniques are often employed for the representation of all species. The major limitation of this method is the tremendous memory storage requirement when the number of independent scalars becomes larger than four. In order to overcome this limitation, a method for the generation of optimal artificial neural networks has been developed. The major advantage of this method is that an optimal architecture is automatically identified, which results in the lowest approximation error while guaranteeing considerable savings in memory storage. An optimal ANN is generated for the representation of a complex methane/air chemical mechanism, which is employed in LES of turbulent jet flame simulations. Results are compared with conventional tabulation techniques and experimental data. Sensitivity of the statistical flow field quantities are presented and accuracy requirements on the chemistry representation are highlighted. [Preview Abstract] |
Monday, November 19, 2007 4:01PM - 4:14PM |
JQ.00003: Large Eddy Simulation of a Sooting Jet Diffusion Flame Guillaume Blanquart, Heinz Pitsch The understanding of soot particle dynamics in combustion systems is a key issue in the development of low emission engines. Of particular importance are the processes shaping the soot particle size distribution function (PSDF). However, it is not always necessary to represent exactly the full distribution, and often information about its moments only is sufficient. The Direct Quadrature Method of Moments (DQMOM) allows for an efficient and accurate prediction of the moments of the soot PSDF. This method has been validated for laminar premixed and diffusion flames with detailed chemistry and is now implemented in a semi-implicit low Mach-number Navier-Stokes solver. A Large Eddy Simulation (LES) of a piloted sooting jet diffusion flame (Delft flame) is performed to study the dynamics of soot particles in a turbulent environment. The profiles of temperature and major species are compared with the experimental measurements. Soot volume fraction profiles are compared with the recent data of Qamar et al. (2007). Aggregate properties such as the diameter and the fractal shape are studied in the scope of DQMOM. [Preview Abstract] |
Monday, November 19, 2007 4:14PM - 4:27PM |
JQ.00004: LES/filtered-density function approach for turbulent spray combustion Venkatramanan Raman, Heeseok Koo, Olivier Desjardins Spray systems provide a unique challenge in that both non-premixed and premixed regimes can coexist in the same flow. The nature of flame propagation is determined by, among other factors, the spray dispersion in the gas-phase and the rate of mixing prior to reaction. This leads to a rich variety of flame structures and compounds the modeling process. While much of the work on spray combustion has focused on understanding spray dispersion, a comprehensive model for the combustion process is yet to be formulated. In this work, we propose a novel large-eddy simulation/filtered-density function (LES/FDF) approach to modeling the spray combustion process. Here, unlike other methods, the chemical source appears closed. This unique property allows different combustion regimes to be represented using a single model. However, the sub-filter mixing term needs to be modeled. A Monte-Carlo based stochastic method is used along with novel algorithms to ensure temporal accuracy and theoretical consistency of the coupled LES/FDF approach. The liquid spray droplets are simulated using a Lagrangian particle tracking scheme. Canonical flow configurations are used to demonstrate the feasibility of this new approach. Further, experiments from the University of Sydney are used to understand the impact of sub-filter models on predictive accuracy of the method. [Preview Abstract] |
Monday, November 19, 2007 4:27PM - 4:40PM |
JQ.00005: Multi-Environment Conditional PDF model for extinction and re-ignition in non-premixed combustion Sean Smith, Rodney O. Fox The Multi-Environment Conditional PDF model was developed to address the deficiencies of other non-premixed combustion modeling techniques, which have difficulties with local extinction and re-ignition in highly turbulent combustion. The advantages of the MECPDF modeling approach and the formulation for multiple reaction variables with variable density have previously been presented. Recent model developments lead to model calculations that compare with results of direct-numerical simulations of temporally-evolving planar jet flames using detailed CO/H2 kinetics (E. R. Hawkes et al., Proc. of the Combust. Inst., 2007, 31, 1633-1640.) These recent developments with model validation will be presented. [Preview Abstract] |
Monday, November 19, 2007 4:40PM - 4:53PM |
JQ.00006: Verification of low-Mach number combustion codes using the method of manufactured solutions Lee Shunn, Frank Ham, Patrick Knupp, Parviz Moin Many computational combustion models rely on tabulated constitutive relations to close the system of equations. As these reactive state-equations are typically multi-dimensional and highly non-linear, their implications on the convergence and accuracy of simulation codes are not well understood. In this presentation, the effects of tabulated state-relationships on the computational performance of low-Mach number combustion codes are explored using the method of manufactured solutions (MMS). Several MMS examples are developed and applied, progressing from simple one-dimensional configurations to problems involving higher dimensionality and solution-complexity. The manufactured solutions are implemented in two multi-physics hydrodynamics codes: CDP developed at Stanford University and FUEGO developed at Sandia National Laboratories. In addition to verifying the order-of-accuracy of the codes, the MMS problems help highlight certain robustness issues in existing variable-density flow-solvers. Strategies to overcome these issues are briefly discussed. [Preview Abstract] |
Monday, November 19, 2007 4:53PM - 5:06PM |
JQ.00007: VSFMDF for LES of Sandia's turbulent piloted jet flame S. Mehdi B. Nik, M. Reza H. Sheikhi, Peyman Givi, Stephen B. Pope The ``joint velocity-scalar filtered mass density function'' (VSFMDF) [1] methodology is employed for large eddy simulation of Sandia Flame D [2]. This is a turbulent piloted nonpremixed methane jet flame. In VSFMDF, the effects of the subgrid scale chemical reaction and convection appear in closed forms. The unclosed terms in the VSFMDF transport equation are modeled in a fashion similar to PDF methods. The modeled VSFMDF is obtained by solving its transport equation by a hybrid finite-difference/Monte Carlo scheme. For this flame (which exhibits little local extinction), a simple flamelet model is employed to relate the instantaneous composition to mixture fraction. The results are compared with experimental data. It is shown that the method captures important features of this flame as observed experimentally. \newline \newline [1] Phys. Fluids (2007), in press. \newline [2] http://www.ca.sandia.gov/tnf. [Preview Abstract] |
Monday, November 19, 2007 5:06PM - 5:19PM |
JQ.00008: Exact results and field-theoretic bounds for randomly advected propagating fronts, and implications for turbulent combustion Jackson R. Mayo, Alan R. Kerstein One of the authors previously conjectured that the wrinkling of propagating fronts by weak random advection increases the bulk propagation rate (turbulent burning velocity) in proportion to the 4/3 power of the advection strength. An exact derivation of this scaling is reported. The analysis shows that the coefficient of this scaling is equal to the energy density of a lower-dimensional Burgers fluid with a white-in-time forcing whose spatial structure is expressed in terms of the spatial autocorrelation of the flow that advects the front. The replica method of field theory has been used to derive an upper bound on the coefficient as a function of the spatial autocorrelation. High precision numerics show that the bound is usefully sharp. Implications for strongly advected fronts (e.g., turbulent flames) are noted. [Preview Abstract] |
Monday, November 19, 2007 5:19PM - 5:32PM |
JQ.00009: DNS of $H_2$/Air Combustion using Complex Chemistry Jeff Doom, Krishnan Mahesh Direct numerical simulation (DNS) is used to study reacting, laminar, vortex rings and turbulent diffusion flames. A novel, all--Mach number algorithm developed by {Doom et al} ({\it J. Comput. Phys.} 2007) is used. The chemical mechanism is a nine species, nineteen reaction mechanism for $H_2$ and Air from Mueller at el ({\it Int. J. Chem. Kinet.} 1999) and the extended Zel'dovich mechanism was used to account for the formation of $NO$. Simulations were performed for three dimensional vortex rings where diluted $H_2$ at ambient temperature (300 K) is injected into hot air (1200 K). The effect of Damkohler number and stroke length will be discussed. Simulations of a three dimensional turbulent diffusion flames were performed. Isotropic turbulence is superimposed on an unstrained diffusion flame where diluted $H_2$ at ambient temperature interacts with hot air. Results of the simulation will be discussed. [Preview Abstract] |
Monday, November 19, 2007 5:32PM - 5:45PM |
JQ.00010: Subgrid-scale scalar diffusion in turbulent partially premixed flames Jian Cai, Chenning Tong, Robert Barlow, Adonios Karpetis The conditionally filtered scalar diffusion used in large-eddy simulation of turbulent combustion is studied experimentally in the Sandia flames. One dimensional filter is implemented along line images to obtain filtered variables. For nearly fully burning samples the conditionally filtered diffusion for the mixture fraction and reactive scalars such as temperature and methane is consistent with the well mixed and highly nonpremixed SGS mixture fraction fields for small and large SGS scalar variances, respectively. Comparisons with IEM model predictions show that in addition to the inconsistency with the nonlinear functional form of the experimental results the model generally over-predicts the diffusion for small SGS variance but under-predicts it for large SGS variance, a result of the SGS mixing time scale used. For reactive scalars the use of their own SGS mixing time scales rather than the SGS mixture fraction time scale generally improve the magnitude of the prediction, suggesting that the current LES results can be improved by using reactive scalar time scales. [Preview Abstract] |
Monday, November 19, 2007 5:45PM - 5:58PM |
JQ.00011: Direct Numerical simulation of premixed flame fronts Francesco Salvadore, Carlo Massimo Casciola Turbulent flames are the core of the major part of combustion devices for applications due to the fast energy release they realize. For premixed flames, the energy release is directly related to the global speed of the flame front propagating towards the fresh mixture. However, it is the local behavior which controls the overall dynamics. We employed the Direct Numerical Simulation of reacting Navier-Stokes equations to investigate in detail the physics of the turbulent premixed fronts. We specifically addressed the statistically planar configuration, due to the rich dynamics involved and to its numerical suitability. Spectral and compact discretization schemes have been adopted to accurately reproduce the whole spectrum of turbulent scales. On the other hand, the chemical modeling is a simple single-specie/single-reactant model. The stabilization of the mean position of the planar front has been studied in detail in order to obtain long-lasting simulations. We simulated two flames subjected to high turbulent stretching and having different Lewis numbers. The local analysis showed a high resistance of the thin front structure, compared to classical predictions. The correlations between the turbulent burning speed and the flame surface area show a complex dynamics of particular significance for closure modelling. [Preview Abstract] |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700