Nano organic carbon and soot in turbulent non-premixed ethylene flames

Spectral optical techniques are combined to characterise the distribution of large-molecule soot precursors, nanoparticles of organic carbon, and soot in two turbulent non-premixed ethylene flames with differing residence times. Laser-induced fluorescence, laser-induced incandescence and light scattering are used to define distributions across the particle size distribution. From the scattering and laser-induced emission measurements it appears that two classes of particles are formed. The first ones are preferentially formed in the fuel-rich region of the flame closer to the nozzle, have sizes of the order of few nanometers but are not fully solid particles, because the constituent molecules still maintain their individual identity exhibiting strong broadband fluorescence in the UV. The second class of particles constituted by solid particles, with sizes of the order of tens of nanometers are able to absorb a sufficient number of photons to be heated to incandescent temperatures. These larger particles are formed at larger residence times in the flame since they are the result of slow growth processes such as coagulation or carbonization. The flames are also modeled in order to produce mixture fraction maps. A new discovery is that nanoparticles of organic carbon concentration, unlike soot, does correlate well with mixture fraction, independent of position in the flame. This is likely to be a significant benefit to future modelling of soot inception processes in turbulent non-premixed flames.     

Effect of swirl on combustion dynamics in a lean-premixed swirl-stabilized combustor

The effect of inlet swirl on the flow development and combustion dynamics in a lean-premixed swirl-stabilized combustor has been numerically investigated using a large-eddy-simulation (LES) technique along with a level-set flamelet library approach. Results indicate that when the inlet swirl number exceeds a critical value, a vortex-breakdown-induced central toroidal recirculation zone is established in the downstream region. As the swirl number increases further, the recirculation zone moves upstream and merges with the wake recirculation zone behind the centerbody. Excessive swirl may cause the central recirculating flow to penetrate into the inlet annulus and lead to the occurrence of flame flashback. A higher swirl number tends to increase the turbulence intensity, and consequently the flame speed. As a result, the flame surface area is reduced. The net heat release, however, remains almost unchanged because of the enhanced flame speed. Transverse acoustic oscillations often prevail under the effects of strong swirling flows, whereas longitudinal modes dominate the wave motions in cases with weak swirl. The ensuing effect on the flow/flame interactions in the chamber is substantial.     

Prediction of soot and thermal radiation in a model gas turbine combustor burning kerosene fuel spray at different swirl levels

Combustion of kerosene fuel spray has been numerically simulated in a laboratory scale combustor geometry to predict soot and the effects of thermal radiation at different swirl levels of primary air flow. The two-phase motion in the combustor is simulated using an Eulerian–Lagragian formulation considering the stochastic separated flow model. The Favre-averaged governing equations are solved for the gas phase with the turbulent quantities simulated by realisable k–? model. The injection of the fuel is considered through a pressure swirl atomiser and the combustion is simulated by a laminar flamelet model with detailed kinetics of kerosene combustion. Soot formation in the flame is predicted using an empirical model with the model parameters adjusted for kerosene fuel. Contributions of gas phase and soot towards thermal radiation have been considered to predict the incident heat flux on the combustor wall and fuel injector. Swirl in the primary flow significantly influences the flow and flame structures in the combustor. The stronger recirculation at high swirl draws more air into the flame region, reduces the flame length and peak flame temperature and also brings the soot laden zone closer to the inlet plane. As a result, the radiative heat flux on the peripheral wall decreases at high swirl and also shifts closer to the inlet plane. However, increased swirl increases the combustor wall temperature due to radial spreading of the flame. The high incident radiative heat flux and the high surface temperature make the fuel injector a critical item in the combustor. The injector peak temperature increases with the increase in swirl flow mainly because the flame is located closer to the inlet plane. On the other hand, a more uniform temperature distribution in the exhaust gas can be attained at the combustor exit at high swirl condition.     

Simultaneous OH-PLIF and PIV measurements in a gas turbine model combustor

In highly turbulent environments, combustion is strongly influenced by the effects of turbulence chemistry interactions. Simultaneous measurement of the flow field and flame is, therefore, obligatory for a clear understanding of the underlying mechanisms. In the current studies simultaneous PIV and OH-PLIF measurements were conducted in an enclosed gas turbine model combustor for investigating the influence of turbulence on local flame characteristics. The swirling CH₄/air flame that was investigated had a thermal power of 10.3 kW with an overall equivalence ratio of ϕ=0.75 and exhibited strong thermoacoustic oscillations at a frequency of approximately 295 Hz. The measurements reveal the formation of reaction zones at regions where hot burned gas from the recirculation zones mixes with the fresh fuel/air mixture at the nozzle exit. However, this does not seem to be a steady phenomenon as there always exist regions where the mixture has failed to ignite, possibly due to the high local strain rates present, resulting in small residence time available for a successful kinetic runaway to take place. The time averaged PIV images showed flow fields typical of enclosed swirl burners, namely a big inner recirculation zone and a small outer recirculation zone. However, the instantaneous images show the existence of small vortical structures close to the shear layers. These small vortical structures are seen playing a vital role in the formation and destruction of reaction zone structures. One does not see a smooth laminar flame front in the instantaneous OH-PLIF images, instead isolated regions of ignition and extinction highlighting the strong interplay between turbulence and chemical reactions. PACS 33.20.-t; 33.50.-j; 47.27.-i; 47.32.Ef; 47.70.Pq; 82.33.Vx; 82.40.-g     

Effects of gas and soot radiation on soot formation in counterflow ethylene diffusion flames

Numerical study of soot formation in counterflow ethylene diffusion flames at atmospheric pressure was conducted using detailed chemistry and complex thermal and transport properties. Soot kinetics was modelled using a semi-empirical two-equation model. Radiation heat transfer was calculated using the discrete-ordinates method coupled with an accurate band model. The calculated soot volume fractions are in reasonably good agreement with the experimental results in the literature. The individual effects of gas and soot radiation on soot formation were also investigated.     

An elementary model for the validation of flamelet approximations in non-premixed turbulent combustion

The fundamental soundness of three flamelet models for non-premixed turbulent combustion is examined on the basis of their performance in an idealized model problem that merges ideas from the laminar asymptotic theory for non-premixed flames and rigorous homogenization theory for the diffusion of a passive scalar. The overall flame configuration is stabilized by a mean gradient in the passive scalar: large Damköhler number asymptotics results are available for the laminar case to quantify the finite-rate effects that cause the flame to depart from its equilibrium state; the same results can also be used to incorporate higher-order corrections in the approximation of the reactive variables in terms of the passive scalar. The use of such flamelet approximations has been extended well beyond the laminar regime as they lie at the core of practical strategies to simulate non-premixed flames in the turbulent regime: the flamelet representation avoids the problem of turbulence closure for the reactive variables by replacing it by the presumably much simpler closure problem for a passive scalar. It is precisely the validity of this substitution outside the laminar regime that is addressed here in the idealized context of a class of small-scale periodic flows for which extensive rigorous results are available for the passive scalar statistics. Results for this simplified problem are reported here for significant wide ranges of Peclet and Damköhler numbers. Asymptotic convergence is observed in terms of the Damköhler number, with a convergence rate that is found to match the laminar predictions and appears relatively insensitive to the Peclet number. The passive scalar dissipation plays a key role in achieving higher-order corrections for the finite-rate case: replacing its pointwise value by an averaged value is convenient practically and can be rigorously motivated for the class of flows studied here, but while it does achieve an overall improvement over the lower-order equilibrium model, the simplification compromises the higher asymptotic convergence observed with the original finite-rate flamelet model with exact local dissipation.(Some figures in this article are in colour only in the electronic version; see www.iop.org)     

Attenuation of combustion instability in a fuel-staged dual-nozzle gas turbine combustor with asymmetric hydrogen composition

The instability attenuation mechanism of fuel staging was investigated in a CH₄/H₂ fueled dual-nozzle gas turbine combustor. Fuel staging was implemented using an asymmetry in fuel composition between the two nozzles. The fuel composition of the upper nozzle was varied while keeping that of the lower nozzle constant. Under these conditions, the self-excited and forced responses of fuel-staged flames were analyzed using OH* chemiluminescence imaging, OH planar laser-induced fluorescence, and particle image velocimetry. In the self-excited measurements, although strong combustion instability was exhibited in the symmetric condition, it weakened gradually with increasing asymmetry in fuel composition. The symmetric flame exhibited significant fluctuations in the heat release rate around the flame tip, which acted as the primary cause of driving combustion instability. However, in asymmetric flames, the H₂ addition induced phase leads in heat release rate fluctuations at the upper region, which damped combustion instability. Thus, our observations revealed a high correlation between the phase leads and the attenuation of combustion instability. Analyses of the forced responses showed that the heat release rate fluctuations were induced by interactions between the flame and the shedding vortex released from the nozzle tip into the downstream. Although these characteristics of shedding vortices did not depend on the H₂ addition, the change in the axial position of the flame caused by the H₂ addition induced the relocation of the site, at which the flame interacted with the vortex. Subsequently, it induced phase leads in the heat release rate fluctuations. The phase difference of heat release rate fluctuations between the two flames due to this phase leads enlarged progressively with increasing asymmetry in fuel composition, leading to the attenuation of combustion instability in asymmetric conditions.     

Mixing in a model gas turbine combustor studied by panoramic optical techniques

Using planar optical methods based on laser-induced fluorescence and particle image velocimetry instantaneous velocity fields and passive tracer concentration are measured simultaneously in a model of GT-combustor at realistic flow rates. Spatial distributions of velocity pulsations and passive tracer concentration pulsations are measured at air flow rate about 0.4 kg/s. Correlations of velocity and concentration pulsations are measured. The most intense turbulent mass flux in the region of swirling flow mixing layer was observed. The contribution of advective and turbulent components in the transfer of a passive tracer in the axial direction was estimated.     

On the experimental validation of combustion simulations in turbulent non-premixed jets

A Reynolds averaged Navier–Stokes (RANS) based combustion model, which incorporated the conditional source-term estimation (CSE) method for the closure of the chemical source term and the trajectory generated low-dimensional manifold (TGLDM) method for the reduction of detailed chemistry, was applied to predict the OH radical distribution in a combusting non-premixed methane jet. The results of the numerical prediction were compared with the results of a complementary experimental study in which the OH radical fields of combusting non-premixed methane jets were visualized using planar laser induced fluorescence (PLIF). It is well known within the modelling community that RANS based models are unable to capture the stochastic nature of turbulent combustion and autoignition, and are therefore unable to predict individual realizations of the flame. In this study, the agreement between the predicted OH field and a well-converged ensemble average of the experimental results was also shown to be poor. The lack of agreement between the numerical results and the ensemble averaged experimental results expose the potential significance of the known weakness in the RANS method. A statistical analysis of the experimental results was also performed. The results of the analysis showed that a minimum of 100 individual realizations was required to provide a well-converged average OH field for the combusting non-premixed jet under investigation. The significance of this result with respect to the validation of large-eddy simulations (LES) of combusting jets is discussed.     

Experimental investigation of ethylene/air combustion instability in a model scramjet combustor using image-based methods

The combustion instabilities of supersonic combustion were investigated experimentally in a laboratory-scale scramjet combustor with a cavity flame holder. Ethylene was injected transversely from an orifice to the supersonic flow of Mach 2 with a stagnation temperature of 1900 K and a total pressure of 0.37 MPa. The dynamic pressure, CH* chemiluminescence and shadowgraph images were measured with a pressure sensor and a high-speed video camera. Dynamic pressure was analyzed by fast Fourier transform, and time-resolved CH* chemiluminescence images were modally decomposed by the sparsity-promoting dynamic mode decomposition (SP-DMD). The results indicated that two combustion instabilities were observed for cavity shear-layer stabilized combustion and the oscillation between jet-wake stabilized and cavity shear-layer ram combustions for the power spectral density (PSD) of pressure. In the case of the combustion instability of cavity shear-layer stabilized combustion, a dominant peak of approximately 128 Hz was observed for the PSD of pressure. This instability corresponded to an entire flame oscillation of the cavity shear-layer stabilized combustion, which was validated by the SP-DMD and a low rank reproduction with 10 modes. This was driven by a fuel injection oscillation in the injection orifice. In the case of oscillation between the jet-wake stabilized and the cavity shear-layer ram combustions, peaks around 1600 Hz were observed for the PSD of pressure. This mechanism was also explained by the SP-DMD modes and a low rank reproduction using within 10 modes. The DMD and shadowgraph images indicated that the vortex formed by a separation of the boundary layer induced a strong jet-wake flame, resulting in the temporal thermal choke followed by cavity shear-layer stabilized ram combustion. The data-driven approach with SP-DMD clarified the combustion instability mechanisms of the supersonic combustion in detail.     

Effects of OH concentration and temperature on NO emission characteristics of turbulent non-premixed CH4/NH3/air flames in a two-stage gas turbine like combustor at high pressure

This study examined the effects of OH concentration and temperature on the NO emission characteristics of turbulent, non-premixed methane (CH₄)/ammonia (NH₃)/air swirl flames in two-stage combustors at high pressure. Emission data were obtained using large-eddy simulations with a finite-rate chemistry method from model flames based on the energy fraction of NH₃ (E_NH3) in CH₄/NH₃ mixtures. Although NO emissions at the combustor exit were found to be significantly higher than those generated by CH₄/air and NH₃/air flames under both lean and stoichiometric primary zone conditions, these emissions could be lowered to approximately 300 ppm by employing far-rich equivalence ratios (?) of 1.3 to 1.4 in the primary zone. This effect was possibly due to the lower OH concentrations under far-rich conditions. An analysis of local flame characteristics using a newly developed mixture fraction equation for CH₄/NH₃/air flames indicated that the local temperature and NO and OH concentration distributions with local ? were qualitatively similar to those in NH₃/air flames. That is, the maximum local NO and OH concentrations appeared at local ? of 0.9, although the maximum temperature was observed at local ? of 1.0. Both the temperature and OH concentration were found to gradually decrease with the partial replacement of CH₄ with NH₃. Consequently, NO emissions from CH₄/NH₃ flames were maximized at E_NH3 in the range of 20% to 30%, after which the emissions decreased. Above 2100 K, the NO emissions from CH₄/NH₃ flames increased exponentially with temperature, which was not observed in NH₃/air flames because of the lower flame temperatures in the latter. But, the maximum NO concentration in CH₄/NH₃ flames was occurred at a temperature slightly below the maximum temperature, just as in NH₃/air flames. The apparent exponential increase in NO emissions from CH₄/NH₃ flames is attributed to a similar trend in the OH concentration at high temperatures.     

Flame-acoustic coupling of combustion instability in a non-premixed backward-facing step combustor: the role of acoustic-Reynolds stress

Combustion instability in a laboratory scale backward-facing step combustor is numerically investigated by carrying out an acoustically coupled incompressible large eddy simulation of turbulent reacting flow for various Reynolds numbers with fuel injection at the step. The problem is mathematically formulated as a decomposition of the full compressible Navier–Stokes equations using multi-scale analysis by recognising the small length scale and large time scale of the flow field relative to a longitudinal mode acoustic field for low mean Mach numbers. The equations are decomposed into those for an incompressible flow with temperature-dependent density to zeroth order and linearised Euler equations for acoustics as a first order compressibility correction. Explicit coupling terms between the two equation sets are identified to be the flow dilatation as a source of acoustic energy and the acoustic Reynolds stress (ARS) as a source of flow momentum. The numerical simulations are able to capture the experimentally observed flow–acoustic lock-on that signifies the onset of combustion instability, marked by a shift in the dominant frequency from an acoustic to a hydrodynamic mode and accompanied by a nonlinear variation of pressure amplitude. Attention is devoted to flow conditions at two Reynolds numbers before and after lock-on to show that, after lock-on, the ARS causes large-scale vortical rollup resulting in the evolution of a compact flame. As compared to acoustically uncoupled simulations at these Reynolds numbers that show an elongated flame with no significant roll up and disturbance in the upstream flow field, the ARS is seen to alter the shear layer dynamics by affecting the flow field upstream of the step as well, when acoustically coupled.     

Analysis of transient blow-out dynamics in a swirl-stabilized combustor using large-eddy simulations

In the present work, the flame structure and dynamics of a swirl-stabilized methane/air flame near and at blow-out conditions are investigated using large-eddy simulations (LES). To this end, simulations at stable conditions and during a transient blow-out sequence are performed employing a flamelet/progress variable (FPV) model and a thickened flame (TFLES) approach with finite-rate chemistry. Good agreement is obtained for velocities, OH mass fraction profiles, and statistics between LES and measurements at conditions near blow-out. Lift-off height comparisons reveal that the TFLES model predicts a significantly less stable flame compared with the FPV model, which is consistent with the stability limits observed for the two models. Subsequently, transient blow-out simulations are performed and examined. This analysis is guided by employing early warning indicators from dynamical system analysis to identify critical transition points and precursors that trigger the onset of blow-out. It is found that the variance of the integrated heat release is a sensible quantity as an early warning signal in detecting blow-out. Detailed comparisons are then performed between the two models to examine the difference in terms of the blow-out mechanism. The comparison reveals that the different phases of the transient blow-out sequence are qualitatively similar for the two models. However, local extinction phenomena are more significant for the TFLES model, reducing the stabilizing feedback by hot combustion products and leading to a total blow-off time, which is three times shorter compared to FPV.