Search

KPP (Kolmogorov-Petrovsky-Piskunov) solutions of the reaction-diffusion equation have application in various physical phenomena occurring in biology, ecology, and reacting flows. In particular, these solutions are commonly used in turbulent combustion to scale turbulent burning velocities. Subject to certain conditions on reaction rate profile through the flame brush and turbulent diffusivity, this theory relates the turbulent burning velocity to the derivative of the reaction rate (ω[over ̃]) at the leading edge of the flame brush (dω[over ̃]/dc[over ̃]|_{c[over ̃]=0}). Such waves are often referred to as "pulled fronts." However, turbulent flames never actually satisfy the KPP conditions for a pulled front, as the turbulent flame brush, parametrized here by the thickness δ_{t}, consists of an ensemble of laminar flamelets of thickness δ, where ε=δ/δ_{t}≪1 is very small, but nonzero, and dω[over ̃]/dc[over ̃] tends to zero at the brush leading edge for high activation energy, combustion-type kinetics. This paper analyzes these effects on KPP wave solutions, parametrized by ε=δ/δ_{t} and Zeldovich number Ze focusing on whether turbulent flames retain their pulled front character and what the correction to the KPP wave speed is. Variational solutions of the reaction-diffusion equation show that the solution can be expanded in powers of 1/|lnε|. Both numerical and asymptotic results are presented, showing that the wave still exhibits pulled front solutions but with significant corrections to the KPP result. The leading order correction is of the form |lnε|^{-2} and independent of Ze. Higher order corrections are function of both ε and Ze. However, the dominant factor influencing the wave speed correction is due to the finite ε, with Ze exhibiting a weaker effect.

This paper describes an investigation into the nonlinear response of flames to harmonic forcing. The amplitude dependent response of the heat release to imposed forcing plays a key role in determining conditions where self-excited oscillations occur, and their amplitudes. This paper particularly focuses on the flame response amplitude. Several prior results, both experimental and theoretical have noted the saturating character of the flame response as the disturbance amplitude increases. In addition, some experiments have also shown more complex amplitude dependencies, including inflection points and even non-monotonic response behaviors. In the latter case, the flame response decreases with increasing disturbance amplitude over a range of amplitudes. This paper analyzes the conditions under which such non-monotonic behavior occurs. By expanding the unsteady heat release response in an expansion of the form, Q ( t ) = Q 0 + ϵ Q 1 ( t ) + ϵ 2 Q 2 ( t ) + ϵ 3 Q 3 ( t ) + O ( ϵ 4 ) , we show that non-monotonic behavior occurs when the phase between the O(ϵ3) term and the linear term fall in a certain band. The model is then applied to representative measured data from the literature where non-monotonic behavior is observed, to extract the different flame response contributions. Finally, results are presented from nonlinear premixed flame calculations to explicitly show the parameter regions where such behavior occurs.

This paper analyzes the flame dynamics of near LBO (lean blowoff) swirl stabilized flames, using simultaneous OH and CH2O (formaldehyde) PLIF (planar laser induced fluorescence) measurements. Prior studies have shown that recirculation stabilized flames approach blowoff through two distinct stages – “stage 1” characterized by local extinction, where the overall flame and flow field remain largely unchanged, and “stage 2”, characterized by a fundamental change in the flow field, accompanied by violent flame flapping and wake disruption. This paper quantifies extinction spots along the flame edge, and entrained reactants within the combustion product region to analyze these stages in greater detail. Extinction spots were quantified by the overlapping regions of OH and CH2O – numerous such spots were found near blowoff. The entraining of unburnt reactants into the recirculation zone was quantified by detecting low intensity OH pockets that were not surrounded by CH2O. As expected, the flame near blowoff displayed significantly more entrained reactant pockets relative to a stable flame. Unexpectedly however, the volume of these pockets is tiny compared to the products, even on the edge of blowoff. Once they enter the wake, they are short-lived, suggesting that they are diluted and/or quickly react. This was surprising given the non-trivial baseflow and flame position disruption at these conditions, suggesting a striking similarity between the average composition of the wake, to that of a stable flame.

This paper describes nitrogen oxide (NOx) measurements from a reacting jet in crossflow (RJICF). The NOx production of the RJICF is controlled by jet stoichiometry, crossflow temperature and compos...

This paper describes an analysis of the near-lean blow off (LBO) dynamics of spray flames, including the influence of fuel composition upon these dynamics. It is motivated by the fact that, while reasonable correlations exist for predicting blowoff conditions, the fundamental reasons for why flames supported by flow recirculation actually blow off are not well understood. Prior work on gaseous systems has shown that the blowoff event is a culmination of several intermediate processes, initiating with local extinction of reactions (“stage 1”), followed by large scale changes in flame and flow dynamics (“stage 2”), finally leading to blowoff. In this study, near-LBO dynamics were characterized for ten liquid fuels with widely varying kinetic and physical properties. Results were compared at two air inlet temperatures, 450 and 300 K, as this influences the relative importance of physical and kinetic properties in controlling LBO. Extinction, re-ignition, and recovery of the flame are evident from these data, and grow in frequency as blowoff is approached. Results show that after a near-blowoff event, the flame can move upstream at velocities much faster than the flow velocity, corresponding to re-ignition. Nonetheless, the majority of the flame recovery events appear to be associated with convection of hot products back upstream, not re-ignition. In contrast, downstream motion of the flame faster than the flow, which would correspond to bulk flame extinction, was never observed. This indicates that “extinction events” actually correspond to convection of the flame downstream by the flow when it loses its stabilization point. The dependence of the equivalence ratio when these events appear, their frequency, and event duration were quantified as a function of fuel composition and air inlet temperature. For example, the data shows a higher percentage of recovery from near-blowoff events through re-ignition for high DCN fuels at the 450 K air temperature condition. These extinction/re-ignition results suggest that high DCN fuels are harder to blow off than low DCN fuels through two mechanisms: (1) by delaying the onset of LBO precursor events, and (2) because they are able to recover from these precursor events through re-ignition more often.

Flow data are often decomposed using proper orthogonal decomposition (POD) of the space–time separated form, $\boldsymbol {q}'\left (\boldsymbol {x},t\right )=\sum _j a_j\left (t\right )\boldsymbol {\phi }_j\left (\boldsymbol {x}\right )$ , which targets spatially correlated flow structures in an optimal manner. This paper analyses permuted POD (PPOD), which decomposes data as $\boldsymbol {q}'\left (\boldsymbol {x},t\right )=\sum _j a_j\left (\boldsymbol {n}\right )\boldsymbol {\phi }_j\left (s,t\right )$ , where $\boldsymbol {x}=(s,\boldsymbol {n})$ is a general spatial coordinate system, $s$ is the coordinate along the bulk advection direction and $\boldsymbol {n}=(n_1,n_2)$ are along mutually orthogonal directions normal to the advection characteristic. This separation of variables is associated with a fundamentally different inner product space for which PPOD is optimal and targets correlations in $s,t$ space. This paper presents mathematical features of PPOD, followed by analysis of three experimental datasets from high-Reynolds-number, turbulent shear flows: a wake, a swirling annular jet and a jet in cross-flow. In the wake and swirling jet cases, the leading PPOD and space-only POD modes focus on similar features but differ in convergence rates and fidelity in capturing spatial and temporal information. In contrast, the leading PPOD and space-only POD modes for the jet in cross-flow capture completely different features – advecting shear layer structures and flapping of the jet column, respectively. This example demonstrates how the different inner product spaces, which order the PPOD and space-only POD modes according to different measures of variance, provide unique ‘lenses’ into features of advection-dominated flows, allowing complementary insights.

This paper characterises the steady and time-periodic behaviour of swirling jets using numerical bifurcation analysis. Its objective is to elucidate the dynamics of fully developed, unconfined, laminar swirling jets under variations in the Reynolds number $\textit {Re}$ and swirl ratio $S$ . Within the $(0,0)\leq (\textit {Re},S)\leq (300,3)$ range, the steady, axisymmetric flow exhibits several distinct patterns ranging from a quasi-columnar jet along the central axis at low $S$ to a radial jet attached to the containing wall at high $S$ with various forms of vortex breakdown in between. A cusp bifurcation appears in the steady solution manifold which triggers bistable behaviour due to a competition between inner and outer low pressure regions associated with vortex breakdown and entrainment of the ambient fluid, respectively. Instability of the steady flow is linked to eigenmodes which are singly ( $|m|=1$ ) or doubly ( $|m|=2$ ) azimuthally periodic, although additional instabilities with other azimuthal wavenumbers occur at $(\textit {Re},S)$ -values beyond the leading neutral curves. The various branches of limit cycle solutions stemming from these neutral curves are associated with both super- and sub-critical Hopf bifurcations. The resulting unsteady flow fields exhibit a wide array of rotating, three-dimensional flow structures, and comparisons between the time-averaged and steady flow patterns highlight the role of these unsteady nonlinear interactions on the overall behaviour of swirling jets. Similarities and differences between this laterally unconfined jet and broader classes of swirling flows, including confined swirling jets and unconfined vortex models, are also discussed.

Premixed flames are sensitive to flow disturbances, which can arise from acoustic or vortical fluctuations. For transverse instabilities, it is known that a dominant mechanism for flame response is “injector coupling”, whereby pressure oscillations associated with transverse waves excite axial flow disturbances. These axial flow disturbances then excite heat release oscillations. The objective of this paper is to consider another mechanism — the direct sensitivity of the unsteady heat release to transverse acoustic waves, and to compare its significance relative to the induced axial disturbances, in a linear framework. The rate at which the flame adds energy to the disturbance field is quantified using the Rayleigh criterion and evaluated over a range of control parameters, such as flame length and swirl number. The results show that radial modes induce heat release fluctuations that always add energy to the acoustic field, whereas heat release fluctuations induced by mixed radial-azimuthal modes can add or remove energy. These amplification rates are then compared to the flame response from induced axial fluctuations. For combustor centered flames, these results show that the direct excitation mechanism has negligible amplification rates relative to the induced axial mechanism for radial modes. For transverse modes, the fact that the nozzle is located at a pressure node indicates that negligible induced axial velocity disturbances are excited; as such, the direct mechanism dominates. For flames that are not centered on pressure nodes, the direct mechanism for mixed-modes, dominates for certain nozzle locations and flame angles.

This paper presents measurements of 10 kHz acetone planar laser induced fluorescence (PLIF) to study the behavior of effusion cooling fluid injected into a non-reacting gas turbine combustor flow at elevated pressure. This study was performed as part of a larger effort to understand potential interactions of the swirling flame with the cooling air. The combustor — which was representative of a rich-burn/quick-quench/lean-burn (RQL) configuration — consisted of a swirl nozzle, quench jets, and a modular liner that could be fitted with various effusion cooling panels and optical access windows. Primary air was seeded with acetone, and passed through the swirl nozzle. Unseeded secondary air was passed on the outside of the liner, entering the combustion chamber through the quench jets and effusion panels. The PLIF laser sheet was arranged parallel to the effusion panel at various offset distances to visualize the mixing between the core flow and effusion jets. The PLIF images were analyzed with a POD-based methodology to de-noise the images and identify patterns in the effusion jet characteristics. The results show that high blowing ratios produce individual effusion jets rather than a single, coalesced film. The effusion jets are highly unsteady, interacting strongly with the turbulent flow from the swirl nozzle and dilution jets. Furthermore, the average trajectories of effusion jets are non-uniform across the panel and are shaped by upstream features in the combustor, namely swirl and dilution jets.

A confluence of technology development, policy support, and industry investment trends is accelerating the pace of Hydrogen (H2) technology demonstrations, increasing the likelihood of power sector impacts. In preparation for a large scale power sector shift toward decarbonization for a low carbon future, several major power equipment manufacturers are developing gas turbines that can operate on a high H2 volume fuel. Many have H2 capable systems now that range from 5% to 100% H2. Units with 100% H2 capabilities are either using a diffusion burner or some version of a wet low emissions (WLE) burner. Most dry low emission/dry low NOx (DLE/DLN) technologies are currently limited to approximately 60% H2 or less. Therefore, research is currently underway to develop low NOx gas turbine combustion systems with improved Hydrogen capability. This paper provides an overview of the technical challenges of Hydrogen combustion and the probable technologies with which the manufacturers will respond.

EPRI has been developing a digital twin of simple and combined cycle gas turbines over the last 5+ years to provide owners and operators with improved capabilities that typically reside in the expert domain of OEMs and 3rd party service providers. The digital twin is a digital model, a physics-based representation of the actual asset. The model is thermodynamic and is created with the intent to support 5 M&D areas: • Integrate with existing M&D tools such as advanced pattern recognition (APR) • Power plant performance prediction and trending such as day, week, and month ahead performance prediction for capacity and generation planning • Health Monitoring and Fault Diagnostics to support asset management with additional health scores and virtual instrumentation enabled by the digital twin model • Monitoring and prediction of both base and part-load performance. Many gas turbine tools have been simplified to work only at full load conditions. To be useful and to improve utilization of collected data, part-load conditions should also be considered. • Outage and repair impacts, including “what-if” capability to understand and quantify potential root causes of less than expected performance improvement or recovery after outage and repairs. This paper presents current progress in creating an EPRI Digital Twin applicable to gas turbines. The formulation, methodology, and real-world use cases are presented. To date, digital twins have been created and tested for both E and F class frames. This paper describes the process of generating closed-form equations capable of transforming existing, measured historian data into the health parameters and virtual sensors needed to better track unit health and monitor faulted performance. These equations encapsulate the digital twin physical model and provide end-users with a methodology to calibrate to their specific unit and efficiently use their choice of monitoring software. Tests have been performed using operator data and have shown good accuracy at detecting anomalous operation and predicting week ahead performance with excellent accuracy. Post-outage impact analysis is also assessed. Real-world application cases for the digital twin are also presented. Examples include using the digital twin to identify causes of post-outage emissions and performance issues, expected impact of degradation and fault conditions, and simulating improvements to operation through part repair and upgrades.

This paper presents measurements of 10 kHz OH planar laser induced fluorescence (PLIF) with an objective to study the interaction of effusion cooling with the flame and hot combustion products in the liquid fueled combustor. The combustor rig is a single sector representation a rich-burn/quick-quench/lean-burn (RQL) configuration. It consists of a swirl nozzle, dilution, and effusion jets. The rig is operated under realistic aircraft conditions, including elevated combustor inlet temperature, and elevated pressure. The PLIF laser sheet was arranged perpendicular and parallel to the liner at distinct liner locations. Parametric variations of important parameters, namely equivalence ratio, and effusion cooling air blowing ratio are conducted to investigate their effect on flame-effusion jet interactions. The PLIF images were analyzed using several data reduction techniques to de-noise the images and identify patterns in the effusion jet-flame interactions. Results show that the effusion jets are highly unsteady, interacting strongly with the turbulent flame from the swirl nozzle and the dilution jets. This work is an extension of recent effusion film mixing studies that were performed with acetone PLIF under non-reacting conditions.

This paper develops a formalism for optimizing nozzle location/configuration with respect to combustion stability of high-frequency transverse modes in a can combustor. The stability of these acoustically noncompact flames was assessed using the Rayleigh integral (RI). Several key control parameters influence RI—flame angle, swirling strength, nozzle location, as well as nozzle location with respect to the acoustic mode shape. In this study, we consider a N-around-1 configuration such as typically used in a multi-nozzle can system and study the overall stability of this system for different natural transverse modes. Typically, such nozzles are distributed in a uniformly circular manner for which we study the overall RI, and for cases where RI > 0, we optimize the nozzle distribution that can reduce and minimize RI. For a fixed geometry such as a circular configuration, the analysis shows how the flame's parameters must vary across the different nozzles, to result in a relatively stable system. Additionally, for a fixed set of flame parameters, the analysis also indicates the noncircular distribution of the N nozzles that minimizes RI. Overall, the analysis aims to provide insights on designing nozzle locations around the center nozzle for minimal amplification of a given transverse mode.

This paper analyzes the dynamics of unstable azimuthal thermoacoustic modes in a lean premixed combustor. Azimuthal modes can be decomposed into two counter rotating waves where they can either compete and potentially suppress one of them (spinning) or coexist (standing), depending on the operating conditions. This paper describes experimental results of the dynamical behaviors of these two waves. The experimental data were taken at different mass flow rates as well as different azimuthal fuel staging in a multi-nozzle can combustor. It is shown that at a low flow rate with uniform fuel distribution, the two waves have similar amplitudes, giving rise to a standing wave. However, the two amplitudes are slowly oscillating out of phase to each other, and the phase difference between the two waves also shows oscillatory behavior. For an intermediate flow rate, the dynamics show intermittency between standing and spinning waves, indicating that the system is bistable. In addition, the phase difference dramatically shifts when the mode switches between standing and spinning waves. For a high flow rate, the system stabilizes at a spinning wave most of the time. These experimental observations demonstrate that not only the amplitudes of two waves but also the phase difference plays an important role in the dynamics of azimuthal mode. For non-uniform azimuthal fuel staging, the modal dynamics exhibit only an oscillatory standing wave behavior regardless of the mass flow rate. Compared to the uniform fuel staging, however, the pressure magnitude is considerably reduced, which provides a potential strategy to mitigate and/or suppress the instabilities.

Understanding thermoacoustic instabilities is essential for the reliable operation of gas turbine engines. To complicate this understanding, the extreme sensitivity of gas turbine combustors can lead to instability characteristics that differ across a fleet. The capability to monitor flame transfer functions in fielded engines would provide valuable data to improve this understanding and aid in gas turbine operability from R&D to field tuning. This paper presents a new experimental facility used to analyze performance of full-scale gas turbine fuel injector hardware at elevated pressure and temperature. It features a liquid cooled, fiber-coupled probe that provides direct optical access to the heat release zone for high-speed chemiluminescence measurements. The probe was designed with fielded applications in mind. In addition, the combustion chamber includes an acoustic sensor array and a large objective window for verification of the probe using high-speed chemiluminescence imaging. This work experimentally demonstrates the new setup under scaled engine conditions, with a focus on operational zones that yield interesting acoustic tones. Results include a demonstration of the probe, preliminary analysis of acoustic and high speed chemiluminescence data, and high speed chemiluminescence imaging. The novelty of this paper is the deployment of a new test platform that incorporates full-scale engine hardware and provides the ability to directly compare acoustic and heat release response in a high-temperature, high-pressure environment to determine the flame transfer functions. This work is a stepping-stone towards the development of an on-line flame transfer function measurement technique for production engines in the field.

In this paper, we discuss the development of a non-intrusive surface temperature sensor based on long-wavelength infrared (LWIR) hyperspectral technology. The LWIR detection enables to minimize optical interferences from hot combustion gases (emission mostly within UV-MWIR region). Utilization of hyperspectral detection allows to further improve temperature measurement accuracy and precision. The developed sensor with fiber coupling provides the required flexibility to be maneuvered around/through combustor hardware. The LWIR fiber probe is fully protected by the custom-designed water-cooled probe housing. This device is designed to sustain temperature of 2400 K at pressure of 50 bar, which enables long-term optical diagnostics inside the practical high-pressure combustion facilities where extreme thermal acoustic perturbation and intense heat fluxes are present. The housing featured a diamond window to selectively measure spectra in the LWIR region to get accurate surface temperature exclusively of the combustor wall. The probe was installed into a RQL style combustor to get surface temperature of both hot and cold side of the combustor wall. Further, pointwise heat flux estimates across the combustion liner wall was derived using the temperature measurements.

This paper analyzes turbulence-chemistry interactions of an n-dodecane-air flame, focusing on the degree to which flame structure and fuel oxidation pathways change in turbulent flames relative to their corresponding laminar flames. This work is based on a lean (ϕ = 0.7) n-dodecane-air flame DNS database from Aspden et al. (2017). The relative roles of dominant reactions that release heat and produce/consume radicals are examined at various turbulence intensities and compared with stretched flame calculations from counterflow flames and perfectly stirred reactors. These results show that spatially integrated chemical pathways are relatively insensitive to turbulence intensity and mimic the behavior of stretched flames. In other words, the contribution of a given reaction to heat release or radical production, integrated over the entire flame, is insensitive to turbulence. Localized analysis conditioned on topological feature of the flame and on temperature is also performed. The former analysis reveals that larger alteration of pathways occurs in the positively-curved regions of the flame. Most significantly, it shows that the thermal structure of the flame is altered, as peak reaction rates and heat release in the low temperature (i.e., below 1200 K) region shift towards higher temperatures with increases in Karlovitz number. This result is particularly interesting given that prior work with lighter fuels (e.g., hydrogen) showed the opposite behavior. Various stretched, laminar flame calculations with altered transport effects were performed for reference. These calculations, using mixture-averaged and Le = 1 transport model, can capture similar shifts towards higher temperatures in laminar flames with increasing stretch. However, the stretch rate and transport values must be tuned to match these shifts in thermal structure differently, depending upon reactions, species, and Ka value. This effect is particularly prominent for low temperature species. Thus, these results show that flames do not simply shift to a Le = 1 thermal structure with high turbulence intensity, as previously suggested, but there is interplay between altered scalar diffusivity and stretch effects.

Recent experimental works have shown that the global equivalence ratio defining lean blow-out (LBO) in model gas turbine combustors correlates with the derived cetane number (DCN) of the tested fuel, which represents the chemical reactivity potential of the fuel, but additional physical and kinetic parameters of the fuel also have influence. The current work explores the significance of preferential vaporization impacts on LBO behaviors; i.e., rather than parameterizing the fuel by overall averaged fuel properties, it looks at DCN correlations based upon distillation properties prior to full vaporization. Preferential vaporization potentials of six fuels are evaluated by measuring the DCN values of five distillation cuts (each of 20% liquid distillation volume recovered). In spite of relatively large disparities in total fuel DCN values (∼9.1), two petroleum-derived jet fuels are found to have nearly the same LBO equivalence ratios, which is attributed to the relatively indiscernible difference of DCN values (∼2) for the initial 20% distillation cut of each fuel. Trade-off impacts between fuel chemical and physical properties are demonstrated by comparing n-dodecane and Gevo-ATJ, which do not have preferential vaporization potential. LBO results suggest that fuel physical properties (particularly fuel boiling characteristics) predominantly control LBO behaviors at low air inlet temperature conditions, whereas fuel chemical properties appear to gain significance with increasing air inlet temperature. Further evidence of preferential vaporization effects on LBO is discussed with two surrogate mixtures formulated to emulate the fully pre-vaporized combustion behaviors of Jet-A, but having drastically different preferential vaporization potentials. Finally, the relationship between DCNs and LBO equivalence ratios is re-examined using the DCN values of initial 20% distillation cuts of all six fuels. The results display a significantly improved correlation, suggesting that the relevance of preferential vaporization on LBO can be significant for fuels that exhibit significant departure of the DCN for high volatile fractions (i.e., the initially vaporized constituents) in comparison to the overall fuel DCN.

Data-driven approaches are increasingly valuable as our ability to store massive amounts of it, the computational power to crunch through it, and the advanced analytics to make sense of it have come to maturity. These opportunities have led to the development of major facilities for aggregating, analyzing, and monetizing data from industrial sources. But the promise of Big Data, machine learning, and data analytics is predicated on access to data. This article delves into four distinct but somewhat overlapping challenges at play in terms of access to data: ownership of data, data nationalism, cybersecurity, and data privacy.

Premixed flames in turbulent, shear flows are disturbed by a superposition of broadband, turbulent fluctuations, and coherent disturbances originating from acoustic disturbances and/or hydrodynamic instabilities. Prior studies have shown nonlinear interactions between such coherent and stochastic disturbances on the flame. This paper investigates the effect of these interactions on the correlation between tangential flame stretch and curvature, an important quantity in turbulent premixed flame physics. High speed particle image velocimetry is used to measure the flow field in an experimental facility where harmonic wrinkles are introduced into a premixed flame that is subjected to varying turbulence intensities, while Mie scattering images are used to extract flame edges. Results show a clear correlation between phase-averaged tangential flame stretch and flame curvature. The magnitude of the fluctuations in both of these quantities decreases with increase in turbulence intensity, even while their overall correlation remains strong. Results from a linearized analysis of the perturbation of the approach flow by a curved, laminar flame are shown to capture the key features of the experimental data. These results suggest that the phase-averaged tangential stretch-curvature correlation is due to gas expansion effects, causing divergence of the approach flow in regions of positive flame curvature, the same mechanism that is responsible for the Darrieus–Landau instability. This result shows that turbulent flames in shear flows with large scale coherent structures, are subjected to not only mean strain, but also coherent modulations in strain rate.

This paper presents analysis of results from optical diagnostics in a high-pressure combustor burning gaseous oxygen (GOX) and liquid kerosene RP-2 fuel through a jet-swirl coflow injector. The obj...

This paper reviews current understanding of the manner in which non-premixed jet flames respond to flow disturbances, and compares these characteristics to premixed flames. Disturbances in flow velocity excite wrinkles on non-premixed flame sheets and perturbations in heat release. These disturbances convect axially along the flame at the local flow velocity and decay in amplitude. Sufficiently large disturbance amplitudes cause opposing sides of the flame to merge and pinch off, leading to multiply connected flame sheets. The space-time characteristics of flame wrinkles on premixed and non-premixed flame sheets are quite similar, as they are dominated by flow disturbances moving the flame around, convection of the wrinkles downstream, and dissipation of wrinkles. In contrast, the heat release dynamics are quite different, as heat release oscillations in premixed flames are nearly uniformly distributed along the flame, whereas in non-premixed flames they are concentrated near the burner outlet where mixture fraction gradients are highest. In addition, flame area fluctuations are the dominant mechanism leading to heat release oscillations in constant burning velocity premixed flames, while mass burning rate oscillations dominate non-premixed flames. The flame transfer function, i.e., the normalized ratio of heat release and flow oscillations, is of O(1) at low frequencies and transitions to a 1/St behavior at high frequencies, for both premixed and non-premixed flames, where St denotes the flame Strouhal number, physically representing the ratio of the convective and the forcing time-scale. While the mechanisms dominating heat release oscillations are quite different for premixed and non-premixed flames, these comparable asymptotic tendencies are actually the result of the same effects; namely, that the heat release is proportional to the instantaneous mass flux of reactants in the quasi-steady, low Strouhal number case, and phase cancellation associated with convecting regions of oscillatory heat release leads to the 1/St behavior in the St ≫ 1 regime. Non-premixed flames have an additional 1 / S t transfer function character at an intermediate range of Strouhal numbers, due to the sharply varying axial dependence of mean heat release near the burner exit.

This study explores the effect of heat release on the growth of the shear layer vortical structures in a reacting jet in crossflow. Jets composed of mixtures of hydrogen, helium and nitrogen were used to independently vary the momentum flux ratio (J), jet to crossflow density ratio (S) and heat release. Velocity fields were obtained from 10 kHz high-speed stereoscopic particle image velocimetry (SPIV) and regions of elevated temperature/combustion products from simultaneous OH planar laser induced fluorescence (OH-PLIF). The shear layer vortices (SLV) originating from instabilities in the windward and leeward shear layers were identified using vortex identification indicator functions in order to track their spatial location and strength. The results show that the asymmetries in shear layer strength between the windward and leeward shear layers are dependent primarily on J, for both reacting and non-reacting flow-fields. The SLV growth rate dependencies on J and S is found to match trends noted by previous studies for non-reacting jets, where SLV growth rates increase with degree of global instability of the JICF. Heat release is also shown to suppress the SLV growth rates relative to non-reacting cases with the same jet parameters. Related to this point, the degree of lifting of the flame also has a significant impact on SLV growth. As flame lifting is directly related to autoignition times, this point shows strong coupling between kinetic rates and jet hydrodynamic stability.

This paper describes NOx measurements from reacting jets in crossflow (RJICF). This work is motivated by interest in axial staging of combustion as a means of reducing NOx emissions at high flame temperatures (>1900 K), where thermal NOx production rates are high. In this approach, the majority of the fuel is burned in a conventional lean-premixed flame, but additional fuel is injected from the combustor walls into the vitiated flow further downstream. The NOx emissions of RJICF are influenced by the secondary fuel jet stoichiometry, jet/crossflow mixing before combustion, as well as secondary combustion product mixing with the bulk product stream. In turn, jet/crossflow mixing is controlled by the hydrodynamic stability of the jet, as well as degree of flame lifting. A key challenge in understanding fundamental factors influencing NOx is decoupling the effect of bulk temperature rise due to the RJICF (∆T), JICF momentum flux ratio (J), and JICF stoichiometry (ϕJet), as they cannot be varied independently. As such, significant effort was made in developing a test matrix to differentiate their effects. Measurements reported here were obtained from rich premixed methane/air jets injected into a varying temperature (1650 K–1800 K) vitiated crossflow, for bulk temperature rises from 20 K–290 K, J values from 1.3–4.4, and ϕJet values from 1–9. These measurements show that NOx emissions monotonically increase with ∆T, as noted in prior studies, but the data reported here are able to differentiate the effects of ∆T and other parameters. In fact, for a given ΔT value, NOx values can vary by 2X depending upon other parameters. For example, the lifting of the flame (LO), which varies with ϕJet and J has significant effects on NOx emissions. These data suggest that the key fundamental JICF parameters influencing NOx emissions are ∆T, ϕJet, J, and LO.

This paper analyzes the time averaged flow structure of a reacting jet in cross flow (RJICF), emphasizing the structure of the counter-rotating vortex pair (CVP) by using simultaneous tomographic particle image velocimetry (TPIV) and hydroxyl radical planar laser induced fluorescence (OH-PLIF). It was performed to determine the extent to which heat release, and the associated effects of gas expansion and baroclinic vorticity production, impact the structure of the CVP. These results show the clear presence of a CVP in the time averaged flow field, whose trajectory lies below the jet centerline on either side of the centerplane. Consistent with other measurements of high momentum flux ratio JICF in nonreacting flows, there is significant asymmetry in strength of the two vortex cores. The strength and structure of the CVP was quantified with vorticity and swirling strength (λci), showing that some regions of the flow with high shear are not necessarily accompanied by large scale bulk flow rotation and vice-versa. The OH PLIF measurement allows for correlation of the flame position with the dominant vortical structures, showing that the leeward flame branch lies slightly above, as well as, in the region between the CVP cores.

This paper describes an analysis of the mechanisms of autoignition-controlled flame initiation and flame stabilization in a nonpremixed jet in crossflows, using simultaneous high-speed (10 kHz) tomographic particle image velocimetry, OH-PLIF and line-of-sight flame emissions. Measurements are conducted on a turbulent, transverse, reacting propane jet issued into a crossflow generated by combustion of natural gas at an equivalence ratio of 0.4 with the crossflow velocity of 10 m/s, the crossflow temperature of 1350 K and the jet momentum flux ratio of 41. While several prior studies have analyzed the lifted character of the flame in similar configurations, we show that several dynamic processes precede the leading edge of the lifted diffusion flame, including formation and evolution of “autoignition kernels”, “flame kernels” and “flame fragments”. “Autoignition kernels”, i.e., discrete compact reaction zones with the peak hydroxyl (OH) fluorescence intensity below that of the diffusion flame, initiate preferably at bulges along the jet periphery where the strain rates and the scalar dissipation rates are lower. The autoignition kernel grows in both size and the OH-fluorescence intensity as it convects downstream. An autoignition kernel transitions into a propagating flame kernel, which quickly gets distorted and elongated in the direction of the principal expansion strain rate to form a flame fragment. Neighboring flame fragments merge with each other and with the downstream diffusion flame via edge-flame propagation. Merging of upstream flame fragments with the downstream diffusion flame results in an upstream advancement of the diffusion-flame front. The diffusion flame front is intrinsically unsteady because of the rather random formation and evolution of autoignition kernels, flame kernels and flame fragments, presumably due to the stochastic velocity, the strain rate and mixture-fraction oscillations.

Simultaneous dual-plane stereoscopic particle-imaging velocimetry (DS-PIV) and hydroxyl planar laser-induced fluorescence (OH-PLIF) was performed at 10 kHz repetition rate in a highly turbulent swirling jet flame. The burst-mode laser system provides sufficient 532-nm dual-pulse energy enabling high-speed large-area dual-plane stereo-PIV measurement. The scattered light from the two planes was separated onto respective cameras using orthogonal polarizations. A diode-pumped solid-state Nd: YAG/dye laser system produced the narrowband excitation 283-nm beam for the 10-kHz OH-PLIF imaging. The combined diagnostics produced simultaneous, temporally resolved two-dimensional fields of OH and two-dimensional, three-component velocity fields, facilitating the observation of the interaction of fluid dynamics with flame fronts.

High frequency thermoacoustic instabilities are problematic for lean-premixed gas turbines. Identifying which acoustic mode is being excited is important, in that it provides insight into potential mitigation measures and mechanical stress/life. However, the frequency spacing between modes becomes quite close for high frequency instabilities in a can combustor. This makes it difficult to distinguish between the modes (e.g., the first transverse mode vs. a higher order axial/mixed mode) based upon frequency calculations alone, which inevitably have uncertainties in boundary conditions, temperature profiles, and combustion response. This paper presents a methodology to simultaneously identify the acoustic mode shapes in the axial and azimuthal directions from acoustic pressure measurements. Multiple high temperature pressure transducers, located at distinct axial and azimuthal positions, are flush mounted in the combustor wall. The measured pressure oscillations from each sensor are then used to reconstruct the pressure distributions by using a least squares method in conjunction with a solution of a three dimensional wave equation. In order to validate the methodology, finite element method (FEM) calculations with estimated post-flame temperature is used to provide the candidate frequencies and corresponding mode shapes. The results demonstrate the reconstructed mode shapes and standing/spinning character of transverse waves, as well as the associated frequencies, both of which are consistent with the FEM predictions. Nodal line location was also extracted from the experimental data during the instabilities in the pressure data.

This paper introduces a data-driven framework for combustor-focused, performance-based condition monitoring of gas turbines. Commercial condition monitoring systems typically generate huge amounts of data that make efficient onboard monitoring challenging. This paper focuses on quantifying combustor component degradation, using premixer centerbody degradation in a swirl stabilized combustor as a case study. The input for these analyses is acoustic pressure measurements acquired at various locations on the combustor. The diagnosis methodology is based on a classification framework and consists of 3 steps: 1) Data curation, 2) Feature Engineering, and 3) Diagnosis. Data curation ensures good quality of the data that is passed through the algorithm. Feature engineering deals with the extraction of the most informative features, from the most informative sensors, that can accurately capture the introduced fault. To perform diagnosis, the classification model is trained using experimentally acquired data and is then tested on a separate data set. The framework was able to achieve high classification accuracy (>99%) for training size as low as 30% of the total recorded observations. The low number of features required to achieve this accuracy suggests high potential for integration into existing onboard condition monitoring systems.

This paper considers the relationship between nonlinearly interacting helical flow disturbances and flame area response in a swirling premixed flame. The present study was performed to determine whether there are nonlinear mechanisms through which helical modes ( helical modes.

This paper addresses the mechanisms through which transverse acoustic oscillations excite unsteady heat release. Forced and self-excited transverse acoustic instability studies to date have strong coupling between the transverse and axial acoustic fields near the flame. This is significant, as studies suggest that it is not the transverse disturbances themselves, but rather the induced axial acoustic disturbances, that control the bulk of the heat release response. This paper presents results from an experiment that controls the relative amplitudes of transverse and axial disturbances and measures the flow field and heat release response for an acoustically compact, swirling flame. 5 kHz, simultaneous sPIV and OH-PLIF measured the flow field and flame edge, and OH* chemiluminescence measured the relative heat release. Experiments performed with essentially the same transverse acoustic wave field, but with and without axial acoustics, show that significant heat release oscillations are only excited in the former case. The results show that the axial disturbances are the dominant cause of the heat release oscillations. These observations support the theory that the key role of the transverse motions is to act as the “clock” for the instability, setting the frequency of the oscillations while having a negligible direct effect on the actual heat release fluctuations. They also show that transverse instabilities can be damped by either actively canceling the induced axial acoustics in the nozzle (rather than the much larger energy transverse combustor disturbances), or by passively tuning the nozzle impedance to drive an axial acoustic velocity node at the nozzle outlet.

The natural eigen-modes of a combustion system are a strong function of its geometric design, temperature profile, and boundary conditions. Many practical systems have multiple, closely-spaced natural mode frequencies. While several prior studies have noted the nonlinear interactions that occur when multiple, linearly unstable modes are present, the nature of these interactions for closely-spaced modes has not been analytically treated. This paper treats this problem theoretically by utilizing a Galerkin expansion of the Euler equations, and the method of averaging to derive a set of amplitude equations for the modal amplitudes. In cases where the frequency spacing is large, many of the oscillatory terms have short time-scales and average out. However, in the case of closely-spaced modes, terms oscillating at the frequency difference between the closely-spaced frequencies correspond to long time-scales and, thus, remain after the averaging process. Results show that frequency spacing between the modes has significant impacts on limit-cycle amplitudes and their stability, even in the case of a static non-linearity. In addition, the conditions under which both modes can co-exist are a strong function of the frequency spacing. There are also certain conditions where one mode is completely suppressed by the other, even if both modes have positive linear growth rates; non-intuitively, the mode that is suppressed could be the one with the larger growth rate. The conditions under which one mode is suppressed are also a strong function of the frequency spacing parameter. Finally, for conditions in which one mode is suppressed, the limit cycle amplitude of the other mode is independent of the frequency spacing. Taken together these results show the difficulty in experimentally inferring the linear stability/instability of (non-dominant) combustor modes based upon “steady state” data taken during limit cycle conditions.

High swirl flows with vortex breakdown are frequently used to create recirculating flow regions for flame stabilization. This paper experimentally characterizes flame stabilization in such flows, motivated by the fact that conditions under which they enable (or do not enable) aerodynamically stabilized flames are not well understood. Such flames are substantially more difficult to experimentally characterize than shear layer stabilized flames, as the flame leading edge moves around significantly and may or may not coincide with a diagnostic laser sheet, in contrast to the near-fixed leading edge of shear layer stabilized flames. This paper characterizes the leading edge conditions, based upon simultaneous stereo Particle Image Velocimetry (sPIV), OH Planar Laser-Induced Fluorescence (PLIF) and OH* chemiluminescence data. Frames are conditionally analyzed when the flame's most upstream point was captured in the laser sheet. These conditioned frames were used to determine the coordinates, local flow velocity and full, three-dimensional, hydrodynamic strain component of flame stretch at the dynamically evolving flame leading edge. Results are taken at two conditions, 70 m/s and 45 m/s axial velocity, referred to as case 1 and 2, respectively. These data show that the leading edge of the flame precesses off-axis, with additional uncorrelated radial and axial motions. The conditioned mean flame stretch rate is positive in both cases, due to the bulk flow deceleration from the high velocity nozzle flow into the larger diameter combustor. Mean strain or velocity values conditioned on the leading flame edge are substantially different from mean values calculated at the same location. For example, the case 1 mean conditioned axial velocity at the flame leading edge is 9 m/s, while the mean axial velocity at the same location is 4 times higher, 36 m/s. In addition, the case 1 mean hydrodynamic strain rate is 3.6 times higher than the corresponding laminar flame extinction stretch rate, κext, calculated from detailed kinetics for the high velocity case, while less than half of κext in case 2. The mean strain rate predicted from bulk velocity scaling changes by a factor of 70/45 between the two cases, a value confirmed by measurements. However, the conditioned mean stretch values differ by a factor of 14. This result illustrates that additional work is needed to understand stretch scalings in such flames. Furthermore, they show that current correlations for flame stability do not capture much of the intermittency that is present in swirl flows with precessing flow features, and illustrates the need for continued work in very basic physics of how and when flames are stabilized aerodynamically.

This paper presents 5 kHz stereo particle image velocimetry and OH planar laser induced fluorescence measurements of transversely forced swirl flames. The presence of transverse forcing on this naturally unstable flow both influences the natural instabilities, as well as amplifies disturbances that may not necessarily manifest themselves during natural oscillations. By manipulating the structure of the acoustic forcing field, both axisymmetric and helical modes are preferentially excited away from the frequency of natural instability. The paper presents a method for spatially interpolating the phase locked $r{-}z$ and $r{-}\unicode[STIX]{x1D703}$ planar velocity and flame position data, extracting the full three-dimensional structure of the helical disturbances. These helical disturbances are also decomposed into symmetric and anti-symmetric disturbances about the jet core, showing the subsequent axial evolution (in magnitude and phase) of each of these underlying disturbances. It is shown that out-of-phase acoustic forcing excites $m=\pm 1$ modes, but the flow field preferentially amplifies the counter-winding, co-rotating helical disturbance over the co-winding, counter-rotating helical disturbance. This causes the flow and flame to transition from a transverse flapping near the jet exit to a precessing motion further downstream. In contrast, in-phase forcing promotes axisymmetric $m=0$ disturbances which dominate the flow field over the entire axial domain. In both cases, the amplitudes of the anti-symmetric disturbances about the jet core grow with downstream distance before saturating and decaying, while the symmetric disturbances appear nearly negligible. It is suggested that this saturation and decay is due to linear effects (e.g. a negative spatial growth rate), rather than nonlinear interactions.

This paper compares flow structures in a single- and triple-nozzle swirl combustor. Realistic combustor hardware usually contains several nozzles; however, test facilities for studying flame dynami...

This paper describes implementation of simultaneous, high speed (5 kHz) stereo PIV, OH and fuel-PLIF in a pressurized, liquid fueled, swirl stabilized flame. The experiments were performed to characterize the flow field, qualitative heat release and fuel spray distributions, and flame dynamics. Acquiring high speed OH-PLIF in pressurized, liquid fuel systems is difficult due to the strong overlap of the fuel's absorption and emission spectra with the OH fluorescence spectrum. To overcome difficulties associated with the overlap, the OH and fuel fluorescence signals were partially separated by using two cameras with differing spectral filters and data acquisition timing. Upon data reduction, regions containing fuel, OH and a mixture of fuel and OH are identified. Instantaneous and time-averaged results are discussed showing the flow field, flame position and dynamics, and spray distribution from the fuel signal for two multi-component liquid fuels, at two inlet temperatures and three pressures. These results are used to infer several important observations on coupled flow and flame physics. Specifically, the flame is “M-shaped” at higher preheat temperature and higher fuel/air ratio, as opposed to no visible reaction on the inside of the annular fuel/air jet at low temperature and fuel/air ratio conditions. While such fundamentally different flame topologies in gaseous, premixed flames are well known, these results show that there are also different families of flame shapes and heat release distributions in spray flames. In addition, the flame position with respect to the flow is different for the liquid-fueled flame than for gaseous, premixed flames—in premixed flames with this geometry, the flame lies in the low velocity shear layer separating the reactants and the recirculating products. In contrast, the flame location is controlled by the spray location in this spray flame, as opposed to the shear layer. For example, reactions are observed near the nozzle outlet in the core of the high velocity annular jet, something which would not be observed in the premixed flame configuration. Also of interest is the near invariance of the key flow features—such as jet core trajectory or shear layer locations—to the operating condition changes for this study, even as the spray penetration and distribution, and flame position change appreciably.

This paper considers the manner in which turbulent premixed flames respond to a superposition of turbulent and narrowband disturbances. This is an important fundamental problem that arises in most combustion applications, as turbulent flames exist in hydrodynamically unstable flow fields and/or in confined systems with narrowband acoustic waves. This paper presents the first measurements of the sensitivity of the turbulent displacement speed to harmonically oscillating flame wrinkles. The flame is attached to a transversely oscillating, heated wire, resulting in the introduction of coherent, convecting wrinkles on the flame. The approach flow turbulence is varied systematically using a variable turbulence generator, enabling quantification of the effect of turbulent flow disturbances on the harmonic wrinkles. Mie scattering measurements are used to quantify the flame edge dynamics, while high speed particle image velocimetry is used to measure the flow field characteristics. By ensemble averaging the results, the ensemble-averaged flame edge and flow characteristics are recovered. For low turbulence intensities, sharp cusps are present in the negative curvature regions of the ensemble-averaged flame position, similar to laminar flames. These cusps are smoothed out at high turbulence intensities. The coherent, ensemble-averaged flame wrinkle amplitude decays with increasing turbulence intensity and with downstream distance. In addition, the ensemble-averaged turbulent flame speed is modulated in space and time. The most significant result of these measurements is the clear demonstration of the correlation between the ensemble-averaged turbulent flame speed and ensemble-averaged flame curvature, with the phase-dependent flame speed increasing in regions of negative curvature. These results have important implications on turbulent combustion physics and modelling, since quasi-coherent velocity disturbances are nearly ubiquitous in shear driven, high turbulent flows and/or confined systems with acoustic feedback. Specifically, these data clearly show that nonlinear interactions occur between the multi-scale turbulent disturbances and the more narrowband disturbances associated with coherent structures. In other words, conceptual models of the controlling physics in combustors with shear driven turbulence must account for the fundamentally different effects of spectrally distributed turbulent disturbances and more narrowband, quasi-coherent disturbances.

This study of an externally forced, amplifier-type turbulent reacting swirling annular jet presents a low-order model for the flow response to transverse acoustic excitation and compares the model's predictions with experimental measurements. The model is formulated based on linear stability calculations about the turbulent mean flow and eddy viscosity fields obtained from separate measurements of the unforced flow. The stability calculations yield weakly global spatial modes associated with the forcing frequency, which serve as a basis upon which to project the forcing input. Thus, the model constitutes a hydrodynamic transfer function connecting the input forcing to the output coherent flow response through the linearized low Mach number compressible Navier–Stokes equations. Following a detailed presentation of the stability analysis underlying the model, the response predictions are evaluated against previously reported experiments where the jet was transversely excited at both an acoustic pressure node and an antinode. The results reveal excellent agreement between the predicted response and the measured fluctuating fields, suggesting that the low-order linear model based on the turbulent mean flow field captures the essential physics of the mode selection process in this forced configuration. This work provides further evidence that linear hydrodynamics govern the growth and decay of spatiotemporally coherent vortical structures in the swirling, turbulent jet flame, and, in particular, explains the dominance of co-rotating spiral structures.

Axially staged combustors offer the potential for achieving low NO x emissions from gas turbines at the elevated temperatures required to achieve at least 65% combined cycle efficiencies. While there is evidence that premixing between the main burner products and secondary stream reduces NO x , it is important to characterize the specific requirements needed to achieve low emissions at gas turbine conditions. This study examines the sensitivity of NO x to finite-rate, large-scale entrainment of the main and secondary streams under the simplifying assumption of infinitely-fast small-scale mixing. Essentially, this allows us to isolate the effects of limited large-scale entrainment rates by using a homogeneous/uniform condition for the entrained and reacting gases. We use a reduced-order reactor network model to examine a generic staged-combustor whose inputs are physical time scales that embody the finite-rate entrainment characteristics. With simulations conducted over a large parameter space at typical operating conditions (25 atm, 650 K air), the results show that, even when small-scale mixing is infinitely fast and the reaction zone is uniform, entrainment significantly affects NO x emissions due to its influence on the equivalence ratio — and thus the temperature and time — at which the entrained mixture ignites. Fuel-air staging is shown to be vital to NO x reduction in most practical cases where entrainment times exceed roughly 1 ms and the secondary stream finishes entraining before the main burner products. A constrained NO x minimization indicates that using the leanest possible main burner and re-routing as much air as possible to the secondary stage is key to low NO x with finite-rate entrainment; for example, less than 10 ppm NO x can be achieved if the secondary stage contains 20% of the total mass flow.