A C o m p a ra tiv e C o m p u ta tio n a l F lu id D y n a m ic s S tudy on F la m e le t-G e n e ra te d M a n ifo ld and S te a d y L a m in a r F la m e le t M o d e lin g fo r T u rb u le n t F la m e s The laminar flamelet model (LFM) (Peters, 1986, "Laminar Diffusion Flamelet Models in Non-Premixed Combustion," Prog. Energy Combust. Sci., 10, pp. 319-339; Peters, "Laminar Flamelet Concepts in Turbulent Combustion," Proc. Combust. Inst., 21, pp. 1231-1250 represents the turbulent flame brush using statistical averaging of laminar flamelets whose structure is not affected by turbulence. The chemical nonequilibrium effects considered in this model are due to local turbulent straining only. In contrast, the flamelet-generated manifold (FGM) (van Oijen and de Goey, 2000, "Modeling of Pre mixed Laminar Flames Using Flamelet-Generated Manifolds," Combust. Sci. Technol., 161, pp. 113-137) model considers that the scalar evolution; the realized trajectories on the thermochemical manifold in a turbulent flame are approximated by the scalar evolu tion similar to that in a laminar flame. This model does not involve any assumption on flame structure. Therefore, it can be successfully used to model ignition, slow chemistry, and quenching effects far away from the equilibrium. In FGM, ID premixed flamelets are solved in reaction-progress space rather than physical space. This helps better solution convergence for the flamelets over the entire mixture fraction range, especially with large kinetic mechanisms at the flammability limits (ansys flu en t 14.5 Theory Guide Help Document, http:llwww.ansys.com). In the present work, a systematic comparative study o f the FGM model with the LFM for four different turbulent diffusion!premixed flames is presented. The first flame considered in this work is methane-air flame with dilution air at the downstream. The second and third flames considered are jet flames in a coaxial flow of hot combustion products from a lean premixed flame called Cabra lifted H2 and CH4 flames (Cabra, et al., 2002, "Simultaneous Laser Raman-Rayleigh-LIF Measurements and Numerical Modeling Results of a Lifted Turbulent H2IN2 Jet Flame in a Vitiated Coflow," Proc. Combust. Inst., 29(2), pp. 1881-1888; Lifted CH4/AirJet Flame in a Viti ated Coflow, http://www.me.berkeley.edu/cal/vcb/data/VCMAData.htmI) where the react ing flow associated with the central jet exhibits similar chemical kinetics, heat transfer, and molecular transport as recirculation burners without the complex recirculating fluid mechanics. The fourth flame considered is a Sandia flame D (Barlow et al., 2005, "Piloted Methane/Air Jet Flames: Scalar Structure and Transport Effects," Combust. Flame, 143, pp. 433-449), a piloted methane-air jet flame. It is observed that the simula tion results predicted by the FGM model are more physical and accurate compared to the LFM in all the flames presented in this work. The autoignition-controlled flame lift off is also captured well in the cases of lifted flames using the FGM model. • Expt --...

The steady laminar flamelet model (SLFM) [1, 2] has been shown to be reasonably good for the predictions of mean temperature and the major species in turbulent flames [3, 4]. However, the SLFM approach has limitations in the prediction of the slow chemistry phenomena like NO formation [5, 6]. In case of SLFM, the turbulence and chemistry are coupled through a single variable, called scalar dissipation, which is representative of the strain inside the flow. The SLFM model is not able to respond to the steep changes in the scalar dissipation values and generally tend to approach to the equilibrium solution as the strain relaxes [7]. The pollutant like NO is formed in the post flame zones and with a high residence time, where the scalar dissipation diminishes and hence the NO is over predicted using SLFM approach. In order to improve the prediction of slow forming species, a transient history of the scalar dissipation evolution is required. In this work, a multiple unsteady laminar flamelet approach is implemented and used to model the NO formation in two turbulent diffusion flames using detailed chemistry. In this approach, multiple unsteady flamelet equations are solved, where each flamelet is associated with its own scalar dissipation history. The time averaged mean variables are calculated from weighted average contributions from different flamelets. The unsteady laminar flamelet solution starts with a converged solution obtained from steady laminar flamelet modeling approach. The unsteady flamelet equations are therefore solved as a post processing step with the frozen flow field. The domain averaged scalar dissipation for a flamelet at each time step is obtained by solving a scalar transport equation, which represents the probability of occurrence of the considered flamelet. The present work involves the study of the effect of number of flamelets and also the different methods of probability initialization on the accuracy of NO prediction. The current model predictions are compared with the experimental data. It is seen that the NO predictions improves significantly even with a single unsteady flamelet and further improves marginally with an increase in number of unsteady flamelets.

In the present work, two equation soot models proposed by Moss-Brookes (MB) and Moss-Brookes-Hall (MBH), available in ANSYS FLUENT14.5, are used to study the soot formation in a turbulent kerosene-air flame. The model constants in the original works of MB and MBH model were primarily tuned for the methane-air or other lower hydrocarbon flames. In this work, the emphasis has been given on the applicability of these models in modeling the soot formation in heavy hydrocarbon fuels. The current work is primarily focused on the parametric study of the various modeling constants for calculating the soot inception and oxidation rates. A parametric study is performed to calculate the soot inception rates by considering different soot precursors like C2H2, C2H4, C6H6 and C6H5. Steady laminar flamelet approach with a detailed chemical reaction mechanism (Jet_SurF_2.0), is used for modeling gas phase combustion. The current numerical predictions are compared with experimental results of Young et al. [1] and earlier published numerical results of Wen et al. [2]. The study is further extended to understand the role of chemical reaction mechanism on soot predictions considering detailed versus reduced (JP10revC) chemical mechanisms.

The steady laminar flamelet model (SLFM) (Peters, 1984, “Laminar Diffusion Flamelet Models in Non-Premixed Turbulent Combustion,” Prog. Energy Combust. Sci., 10(3), pp. 319–339; Peters, 1986, “Laminar Flamelet Concepts in Turbulent Combustion,” Symp. (Int.) Combust., 21(1), pp. 1231–1250) has been shown to be reasonably good for the predictions of mean temperature and the major species in turbulent flames (Borghi, 1988, “Turbulent Combustion Modeling,” Prog. Energy Combust. Sci., 14(4), pp. 245–292; Veynante and Vervisch, 2002, “Turbulent Combustion Modeling,” Prog. Energy Combust. Sci., 28(3), pp. 193–266). However, the SLFM approach has limitations in the prediction of slow chemistry phenomena like NO formation (Benim and Syed, 1998, “Laminar Flamelet Modeling of Turbulent Premixed Combustion,” Appl. Math. Model., 22(1–2), pp. 113–136; Heyl and Bockhorn, 2001, “Flamelet Modeling of NO Formation in Laminar and Turbulent Diffusion Flames,” Chemosphere, 42(5–7), pp. 449–462). In the case of SLFM, the turbulence and chemistry are coupled through a single variable called scalar dissipation, which is representative of the strain inside the flow. The SLFM is not able to respond to the steep changes in the scalar dissipation values and generally tends to approach to the equilibrium solution as the strain relaxes (Haworth et al., 1989, “The Importance of Time-Dependent Flame Structures in Stretched Laminar Flamelet Models for Turbulent Jet Diffusion Flames,” Symp. (Int.) Combust., 22(1), pp. 589–597). A pollutant like NO is formed in the post flame zones and with a high residence time, where the scalar dissipation diminishes and hence the NO is overpredicted using the SLFM approach. In order to improve the prediction of slow forming species, a transient history of the scalar dissipation evolution is required. In this work, a multiple unsteady laminar flamelet approach is implemented and used to model the NO formation in two turbulent diffusion flames using detailed chemistry. In this approach, multiple unsteady flamelet equations are solved, where each flamelet is associated with its own scalar dissipation history. The time averaged mean variables are calculated from weighted average contributions from different flamelets. The unsteady laminar flamelet solution starts with a converged solution obtained from the steady laminar flamelet modeling approach. The unsteady flamelet equations are, therefore, solved as a post processing step with the frozen flow field. The domain averaged scalar dissipation for a flamelet at each time step is obtained by solving a scalar transport equation, which represents the probability of occurrence of the considered flamelet. The present work involves the study of the effect of the number of flamelets and also the different methods of probability initialization on the accuracy of NO prediction. The current model predictions are compared with the experimental data. It is seen that the NO predictions improves significantly even with a single unsteady flamelet and further improves marginally with an increase in number of unsteady flamelets.

Laminar Flamelet Model (LFM) [1–2] represents the turbulent flame brush using statistical averaging of laminar flamelets whose structure is not affected by turbulence. The chemical non-equilibrium effects considered in this model are due to local turbulent straining only. In contrast, Flamelet Generated Manifold (FGM) [3] model considers that the scalar evolution, the realized trajectories on the thermo-chemical manifold in a turbulent flame is approximated by the scalar evolution similar to that in a laminar flame. This model does not involve any assumption on flame structure. Therefore, it can be successfully used to model ignition, slow chemistry and quenching effects far away from the equilibrium. In FGM, 1D premixed flamelets are solved in reaction-progress space rather than physical space. This helps better solution convergence for the flamelets over the entire mixture fraction range, especially with large kinetic mechanisms at the flammability limits [4]. In the present work, a systematic comparative study of FGM model with LFM for four different turbulent diffusion/premixed flames is presented. First flame considered in this work is methane-air flame with dilution air at the downstream. Second and third flame considered are jet flames in a coaxial flow of hot combustion products from a lean premixed flame called Cabra lifted H2 and CH4 flames [5–6] where the reacting flow associated with the central jet exhibits similar chemical kinetics, heat transfer and molecular transport as recirculation burners without the complex recirculating fluid mechanic. The fourth flame considered is Sandia flame D [7], a piloted methane-air jet flame. It is observed that the simulation results predicted by FGM model are more physical and accurate compared to LFM in all the flames presented in this work.