Search

Significant efforts made by the gas turbine industry have helped reduce nitrogen oxides (NOx) emissions considerably. To meet and surpass the increasingly stringent regulations, accurate and robust thermochemical mechanisms are needed to help design future sub-10 ppm combustion systems. Uncertainty in kinetic modeling, however, can result in large prediction uncertainty and significant discrepancy between models that hinder the identification of promising combustors with confidence. Direct reaction rate measurements are seldom available for some reactions, especially when involving short-lived radicals such as methylidyne, CH. As the main precursor to the prompt-NO formation pathway, its large parametric uncertainty directly propagates through the nitrogen chemistry preventing accurate and precise emissions predictions. Recent independent CH concentration measurements obtained at various operating conditions are used as indirect rate measurements to perform statistical, or Bayesian, calibration. A subset of important reactions in the CH chemistry affecting peak-CH concentration is identified through uncertainty-weighted sensitivity analysis to first constrain the parametric space of this prompt-NO precursor. Spectral expansion provides the surrogate model used in the Markov-Chain Monte Carlo method to evaluate the posterior kinetic distribution. The resulting constrained CH-chemistry better captures experimental measurements while providing smaller prediction uncertainty of a similar order as the uncertainty of the measurements, which can increase the confidence in simulation results to identify promising future low-emissions configurations. For the quasi-steady-state species CH, fuel decomposition reactions leading to CH production are constrained while little impact is observed for intermediate reactions within the CH-chemistry. The reduction in prediction uncertainty results mainly from the constrained correlations between parameters which greatly limit the set of feasible reaction rate combinations. Additional independent direct and indirect measurements would be necessary to further constrain rate parameters in the CH chemistry, but this calibration demonstrates that predictions of radical species can be improved by assimilating enough data.

Quantification of aleatoric uncertainties due to the inherent variabilities in operating conditions and fuel composition is essential for designing and improving premixers in dry low-emissions (DLE) combustion systems. Advanced stochastic simulation tools require a large number of evaluations in order to perform this type of uncertainty quantification (UQ) analysis. This task is computationally prohibitive using high-fidelity computational fluid dynamic (CFD) approaches such as large eddy simulation (LES). In this paper, we describe a novel and computationally efficient toolchain for stochastic modeling using minimal input from LES, to perform uncertainty and risk quantification of a DLE system. More specially, high-fidelity LES, chemical reactor network (CRN) model, beta mixture model, Bayesian inference and sequential Monte Carlo (SMC) are integrated into the toolchain. The methodology is applied to a practical premixer of low-emission combustion system with dimethyl ether (DME)/methane–air mixtures to simulate auto-ignition events at different engine conditions. First, the benchmark premixer is simulated using a set of LESs for a methane/air mixture at elevated pressure and temperature conditions. A partitioning approach is employed to generate a set of deterministic chemical reactor network (CRN) models from LES results. These CRN models are then solved at the volume-average conditions and validated by LES results. A mixture modeling approach using the expectation-method of moment (E-MM) is carried out to generate a set of beta mixture models and characterize uncertainties for LES-predicted temperature distributions. These beta mixture models and a normal distribution for DME volume fraction are used to simulate a set of stochastic CRN models. The Bayesian inference approach through SMC method is then implemented on the results of temperature distributions from stochastic CRN models to simulate the probability of auto-ignition in the benchmark premixer. The results present a very satisfactory performance for the stochastic toolchain to compute the auto-ignition propensity for a few events with a particular combination of inlet temperature and DME volume fraction. Characterization of these rare events is computationally prohibitive in the conventional deterministic methods such as high-fidelity LES.

Development of probabilistic modelling tools to perform Bayesian inference and uncertainty quantification (UQ) is a challenging task for practical hydrogen-enriched and low-emission combustion systems due to the need to take into account simultaneously simulated fluid dynamics and detailed combustion chemistry. A large number of evaluations is required to calibrate models and estimate parameters using experimental data within the framework of Bayesian inference. This task is computationally prohibitive in high-fidelity and deterministic approaches such as large eddy simulation (LES) to design and optimize combustion systems. Therefore, there is a need to develop methods that: (a) are suitable for Bayesian inference studies and (b) characterize a range of solutions based on the uncertainty of modelling parameters and input conditions. This paper aims to develop a computationally-efficient toolchain to address these issues for probabilistic modelling of NOx emission in hydrogen-enriched and lean-premixed combustion systems. A novel method is implemented into the toolchain using a chemical reactor network (CRN) model, non-intrusive polynomial chaos expansion based on the point collocation method (NIPCE-PCM), and the Markov Chain Monte Carlo (MCMC) method. First, a CRN model is generated for a combustion system burning hydrogen-enriched methane/air mixtures at high-pressure lean-premixed conditions to compute NOx emission. A set of metamodels is then developed using NIPCE-PCM as a computationally efficient alternative to the physics-based CRN model. These surrogate models and experimental data are then implemented in the MCMC method to perform a two-step Bayesian calibration to maximize the agreement between model predictions and measurements. The average standard deviations for the prediction of exit temperature and NOx emission are reduced by almost 90% using this method. The calibrated model then used with confidence for global sensitivity and reliability analysis studies, which show that the volume of the main-flame zone is the most important parameter for NOx emission. The results show satisfactory performance for the developed toolchain to perform Bayesian inference and UQ studies, enabling a robust and consistent process for designing and optimising low-emission combustion systems.

Multiplication of hydroxyl and formaldehyde planar laser-induced fluorescence signals for turbulent hydrogen-enriched methane–air flames with compositionally inhomogeneous mixtures is investigated experimentally. Hydrogen-enriched methane–air flames with a global fuel–air equivalence ratio of 0.8 and hydrogen-enrichment percentage of 60% are examined. Two nozzles, each containing 4 fuel/air injection lobes are used in the experiments. The lobes of the first nozzle are straight, while those of the second nozzle are not, generating a swirling motion. The fuel is injected through several small diameter holes into the lobes. The amount of injected fuel flow rate varies between the lobes, generating stratified conditions. For each nozzle, two mean bulk flow velocities of 5 and 25 m/s are tested. Simultaneous hydroxyl and formaldehyde planar laser-induced fluorescence as well as separate stereoscopic particle image velocimetry are performed for the tested reacting conditions. For non-reacting flow tests, separate particle image velocimetry and acetone planar laser-induced fluorescence experiments are conducted to study the background turbulent flow characteristics and fuel/air mixing, respectively. The results show that stratification can lead to fragmentation of the flames and generation of islands with noticeable multiplication of hydroxyl and formaldehyde planar laser-induced fluorescence signals. Due to their significantly large number of occurrences, such flame structure can generate relatively large integral of the PLIF signals multiplication.

A comprehensive understanding of uncertainty sources in experimental measurements is required to develop robust thermochemical models for use in industrial applications. Due to the complexity of the combustion process in gas turbine engines, simpler flames are generally used to study fundamental combustion properties and measure concentrations of important species to validate and improve modelling. Stable, laminar flames have increasingly been used to study nitrogen oxide (NOx) formation in lean-to-rich compositions in low-to-high pressures to assess model predictions and improve accuracy to help develop future low-emissions systems. They allow for non-intrusive diagnostics to measure sub-ppm concentrations of pollutant molecules, as well as important precursors, and provide well-defined boundary conditions to directly compare experiments with simulations. The uncertainties of experimentally-measured boundary conditions and the inherent kinetic uncertainties in the nitrogen chemistry are propagated through one-dimensional stagnation flame simulations to quantify the relative importance of the two sources and estimate their impact on predictions. Measurements in lean, stoichiometric, and rich methane-air flames are used to investigate the production pathways active in those conditions. Various spectral expansions are used to develop surrogate models with different levels of accuracy to perform the uncertainty analysis for 15 important reactions in the nitrogen chemistry and the 6 boundary conditions (ϕ, Tin, uin, du/dzin, Tsurf, P) simultaneously. After estimating the individual parametric contributions, the uncertainty of the boundary conditions are shown to have a relatively small impact on the prediction of NOx compared to kinetic uncertainties in these laboratory experiments. These results show that properly calibrated laminar flame experiments can, not only provide validation targets for modelling, but also accurate indirect measurements that can later be used to infer individual kinetic rates to improve thermochemical models.

New ignition delay time (IDT) measurements for two natural gas (NG) blends composed of C1–C7n-alkanes, NG6 (C1:60.625%, C2:20%, C3:10%, C4:5%, nC5:2.5%, nC6:1.25%, nC7:0.625%) and NG7 (C1:72.635%, C2:10%, C3:6.667%, C4:4.444%, nC5:2.965%, nC6:1.976%, nC7:1.317%) by volume with methane as the major component are presented. The measurements were recorded using a high-pressure shock tube (HPST) for stoichiometric fuel in air mixtures at reflected shock pressures (p5) of 20–30 bar and at temperatures (T5) of 987–1420 K. The current results together with rapid compression machine (RCM) measurements in the literature show that higher concentrations of the higher n-alkanes (C4–C7) ∼1.327% in the NG7 blend compared to the NG6 blend result in the ignition times for NG7 being almost a factor of two faster than those for NG6 at compressed temperatures of (TC) ≤ 1000 K. This is due to the low temperature chain branching reactions that occur for higher alkane oxidation kinetics in this temperature range. On the contrary, at TC > 1000 K, NG6 exhibits ∼20% faster ignition than NG7, primarily because about 12% of the methane in the NG7 blend is primarily replaced by ethane (∼10%) in NG6, which is significantly more reactive than methane at these higher temperatures. The performance of NUIGMech1.2 in simulating these data is assessed, and it can reproduce the experiments within 20% for all the conditions considered in the study. We also investigate the effect of hydrogen addition to the auto-ignition of these NG blends using NUIGMech1.2, which has been validated against the existing literature for natural gas/hydrogen blends. The results demonstrate that hydrogen addition has both an inhibiting and a promoting effect in the low- and high-temperature regimes, respectively. Sensitivity analyses of the hydrogen/NG mixtures are performed to understand the underlying kinetics controlling these opposite ignition effects. At low temperatures, H-atom abstraction byO˙H radicals from C3 and larger fuels are the key chain-branching reactions consuming the fuel and providing the necessary fuel radicals, which undergo low temperature chemistry (LTC) leading to ignition. However, with the addition of hydrogen to the fuel mixture, the competition by H2 for O˙H radicals via the reaction H2 + O˙H ↔ H˙ + H2O reduces the progress of the LTC of the higher hydrocarbon fuels thereby inhibiting ignition. At higher temperatures, since H˙ + O2 ↔ Ö + O˙H is the most sensitive reaction promoting reactivity, the higher concentrations of H2 in the fuel mixture lead to higher H˙ atom concentrations leading to faster ignition due to an enhanced rate of the H˙ + O2 ↔ Ö + O˙H reaction.

Characterisation of autoignition risk is crucial for designing and optimising low-emission combustion systems as there is an increased demand for highly reactive and novel fuel mixtures. Achieving a residence time to prevent autoignition and obtaining an adequate mixing quality is a challenging trade-off for these fuels in lean-premixed combustion systems. The level of complexity increases further due to low-temperature chemical pathways and pressure-dependent reactions that strongly influence ignition delay at engine operating conditions. Detailed chemical kinetic mechanisms with hundreds of species and thousands of reactions are developed and employed to address this complexity and predict ignition delay accurately, especially for heavier hydrocarbons. However, direct implementation of these kinetic mechanisms is computationally prohibitive in high-fidelity CFD approaches such as large eddy simulation (LES) and stochastic simulation tools that require a large number of evaluations. Advanced stochastic methods are essential tools to quantify uncertainties due to the inherent variabilities in ambient, operating conditions and fuel composition on ignition delay time calculation for practical applications. This study introduces and implements a computationally efficient method based on metamodellig to predict ignition delay time over a wide range of operating conditions and fuel compositions for gas turbine combustion systems. A metamodel or surrogate model is an accurate and quick approximation of the original computational model based on a detailed chemical kinetic mechanism. Polynomial chaos expansion (PCE) as an advanced method is employed to build metamodels using a limited set of runs of the original ignition delay time model based on NUIGMech1.0 chemical kinetic mechanism as the most detailed and state-of-the-art chemical kinetic mechanism for natural gas. Developed metamodels for ignition delay time are valid over operating conditions of P = 20–40 bar and T = 700–900 K for natural gas containing C1 to C7 hydrocarbons at stoichiometric condition. These metamodels provide a fast, robust, and considerably accurate framework instead of a detailed chemical kinetic model that facilitates (a) characterising ignition delay time at different operating conditions and fuel compositions, (b) designing and optimising premixers and burners and (c) conducting uncertainty quantification and stochastic modelling studies.

Quantification of aleatoric uncertainties due to the inherent variabilities in operating conditions and fuel composition is essential for designing and improving premixers in dry low-emissions (DLE) combustion systems. Advanced stochastic simulation tools require a large number of evaluations in order to perform this type of uncertainty quantification (UQ) analysis. This task is computationally prohibitive using high-fidelity computational fluid dynamic (CFD) approaches such as large eddy simulation (LES). In this paper, we describe a novel and computationally-efficient toolchain for stochastic modelling using minimal input from LES, to perform uncertainty and risk quantification of a DLE system. More specially, high-fidelity LES, chemical reactor network (CRN) model, beta mixture model, Bayesian inference and sequential Monte Carlo (SMC) are integrated into the toolchain. The methodology is applied to a practical premixer of low-emission combustion system with dimethyl ether (DME)/methane-air mixtures to simulate autoignition events at different engine conditions. First, the benchmark premixer is simulated using a set of LESs for a methane/air mixture at elevated pressure and temperature conditions. A partitioning approach is employed to generate a set of deterministic chemical reactor network (CRN) models from LES results. These CRN models are then solved at the volume-average conditions and validated by LES results. A mixture modelling approach using the expectation-method of moment (EMM) is carried out to generate a set of beta mixture models and characterise uncertainties for LES-predicted temperature distributions. These beta mixture models and a normal distribution for DME volume fraction are used to simulate a set of stochastic CRN models. The Bayesian inference approach through Sequential Monte Carlo (SMC) method is then implemented on the results of temperature distributions from stochastic CRN models to simulate the probability of autoignition in the benchmark premixer. The results present a very satisfactory performance for the stochastic toolchain to compute the autoignition propensity for a few events with a particular combination of inlet temperature and DME volume fraction. Characterisation of these rare events is computationally prohibitive in the conventional deterministic methods such as high-fidelity LES.

Significant efforts made by the gas turbine industry have helped reduce nitrogen oxides (NOx) emissions considerably. To meet and surpass the increasingly stringent regulations, accurate and robust thermochemical mechanisms are needed to help design future sub-10 ppm combustion systems. Uncertainty in kinetic modelling, however, can result in large prediction uncertainty and significant discrepancy between models that hinder the identification of promising combustors with confidence. Direct reaction rate measurements are seldom available for some reactions, especially when involving short-lived radicals like methylidyne, CH. As the main precursor to the prompt-NO formation pathway, its large parametric uncertainty directly propagates through the nitrogen chemistry preventing accurate and precise emissions predictions. Recent independent CH concentration measurements obtained at various operating conditions are used as indirect rate measurements to perform statistical, or Bayesian, calibration. A subset of important reactions in the CH chemistry affecting peak-CH concentration is identified through uncertainty-weighted sensitivity analysis to first constrain the parametric space of this prompt-NO precursor. Spectral expansion provides the surrogate model used in the Markov-Chain Monte Carlo method to evaluate the posterior kinetic distribution. The resulting constrained CH-chemistry better captures experimental measurements while providing smaller prediction uncertainty of a similar order as the uncertainty of the measurements, which can increase the confidence in simulation results to identify promising future low-emissions configurations. For the quasi-steady state species CH, fuel decomposition reactions leading to CH production are constrained while little impact is observed for intermediate reactions within the CH-chemistry. The reduction in prediction uncertainty results mainly from the constrained correlations between parameters which greatly limit the set of feasible reaction rate combinations. Additional independent direct and indirect measurements would be necessary to further constrain rate parameters in the CH chemistry, but this calibration demonstrates that predictions of radical species can be improved by assimilating enough data.

Ignition delay time measurements for multi-component natural gas mixtures were carried out using a rapid compression machine at conditions relevant to gas turbine operation, at equivalence ratios of 0.5–2.0 in ‘air’ in the temperature range 650–1050 K, at pressures of 10–30 bar. Natural gas mixtures comprising C1–C7 n-alkanes with methane as the major component (volume fraction: 0.35–0.98) were considered. A design of experiments was employed to minimize the number of experiments needed to cover the wide range of pressures, temperatures and equivalence ratios. The new experimental data, together with available literature data, were used to develop and assess a comprehensive chemical kinetic model. Replacing 1.875% methane with 1.25% n-hexane and 0.625% n-heptane in a mixture containing C1–C5 components leads to a significant increase in a mixture's reactivity. The mixtures containing heavier hydrocarbons also tend to show a strong negative temperature coefficient and two-stage ignition behavior. Sensitivity analyses of the C1–C7 blends have been performed to highlight the key reactions controlling their ignition behavior. The authors would like to acknowledge Science Foundation Ireland for funding via project number16/SP/3829. We also acknowledge funding from Siemens Canada Ltd. peer-reviewed

There has been increased interest in plasma-assisted combustion over the last two decades due to its ability to improve flame stability and combustion performance. However, the effect of nanosecond repetitively pulsed (NRP) discharges is still not well understood since the plasma can act on the thermal, kinetic and hydrodynamic properties of combustion. This study investigates the hydrodynamic effect of NRP discharges produced upstream of a lean premixed methane-air flame at atmospheric pressure. The effect of the plasma pulse repetition frequency (PRF) on the flame was studied with time-resolved imaging and shows that, as the PRF increases, the flame stabilizes further upstream closer to the plasma source. Time-resolved imaging also shows that the flame remains steady and does not relax between consecutive NRP discharges for PRFs ranging between 1 to 5 kHz and that the full relaxation of the flame takes ∼ 100 ms. Particle image velocimetry is used to assess the effect of the plasma on both a cold flow and a stagnation flame. The results show that the NRP discharges reduce the flow velocity upstream of the flame front by up to 20%, which is the main cause of the displacement of the flame. The stretch rate of the flame was also studied, and the stretch induced by the plasma was found to be the main cause of the 12% increase in flame speed. The mechanism through which the plasma causes the observed hydrodynamic effect remains unknown.

Ideally, nitric oxide (NO) production pathways would be measured individually to understand the formation mechanisms at a fundamental level. Unfortunately, the four production routes in hydrocarbon combustion cannot be fully decoupled. Hydrogen combustion at low flame temperatures eliminates prompt-NO and mitigates thermal production, such that only the N2O and NNH pathways remain as significant production routes. The H2/O2 system, whose base chemistry has been studied in great detail, offers an excellent platform to validate nitrogen chemistry by limiting the possibility of error propagation during model calibration. The current work presents measurements of velocity, temperature, and NO concentration in premixed, jet-wall stagnation, hydrogen–air flames at atmospheric pressure, diluted with argon to maintain adiabatic flame temperatures below 1800 K. Measurements of reference flame speeds, Su,ref, obtained with particle tracking velocimetry, highlight the modeling differences in H2/O2 chemistry from a selection of thermochemical mechanisms, especially in lean flames affected by preferential diffusion. Laser induced fluorescence measurements in lean-to-rich flames ( ϕ = 0.7 –1.5) yield concentrations of NO from 2 to 0.5 ppm, respectively. Simulated NO profiles cover one order of magnitude in predicted signal intensity. Fortunately, recent mechanisms with accurate descriptions of the N2O and NNH pathways predict NO concentrations within experimental uncertainties for multiple operating conditions.

Effective flame anchoring maximizes the low-emissions operating window of gas-turbine systems featuring closely spaced, multiple lean flames on single injector units. The complex flame-flow interactions inherent to such designs are not well understood, in particular, at part-load gas turbine conditions. In the present work, large eddy simulation (LES) was used to investigate flames in an optically accessible, confined, linear array of five industrial swirl injector elements. Well-anchored, partially attached and fully lifted flames were observed and three cases corresponding to these were selected for modeling. The dynamically thickened flame (DTF) model was used with a newly developed reduced chemical kinetics scheme that contains low temperature reaction pathways. Chemical explosive mode analysis (CEMA) shows that the flames exhibit thin ignition zones near their base, while a more distributed region of strongly positive modes occur downstream when a transition to partial blow-off takes place. Extinctions are frequent in this region and create broken reaction zones which seem to exhibit both autoignition as well as flamelet characteristics. Further decrease in reactivity leads to a fully lifted flame. It is clear that part-load conditions are best investigated by models that are combustion regime independent.

A novel compute intensification methodology to the construction of low-dimensional, high-fidelity "compact" kinetic models for NOX formation is designed and demonstrated. The method adapts the data intensive Machine Learned Optimization of Chemical Kinetics (MLOCK) algorithm for compact model generation by the use of a Latin Square method for virtual reaction network generation. A set of logical rules are defined which construct a minimally sized virtual reaction network comprising three additional nodes (N, NO, NO2). This NOX virtual reaction network is appended to a pre-existing compact model for methane combustion comprising fifteen nodes. The resulting eighteen node virtual reaction network is processed by the MLOCK coded algorithm to produce a plethora of compact model candidates for NOX formation during methane combustion. MLOCK automatically; populates the terms of the virtual reaction network with candidate inputs; measures the success of the resulting compact model candidates (in reproducing a broad set of gas turbine industry-defined performance targets); selects regions of input parameters space showing models of best performance; refines the input parameters to give better performance; and makes an ultimate selection of the best performing model or models. By this method, it is shown that a number of compact model candidates exist that show fidelities in excess of 75% in reproducing industry defined performance targets, with one model valid to >75% across fuel/air equivalence ratios of 0.5-1.0. However, to meet the full fuel/air equivalence ratio performance envelope defined by industry, we show that with this minimal virtual reaction network, two further compact models are required., arXiv admin note: text overlap with arXiv:2202.08021

The minimization of auto-ignition risk is critical to the design of premixers of high power aeroderivative gas turbines as an increased use of highly reactive future fuels (for example, hydrogen or higher hydrocarbons) is anticipated. Safety factors based on ignition delays of homogeneous mixtures are generally used to guide the choice of a residence time for a given premixer. However, auto-ignition chemistry under aeroderivative conditions is fast (0.5–2 ms) and can be initiated within typical premixer residence times. The analysis of what takes place in this short period necessarily involves the study of low-temperature auto-ignition precursor chemistry, but precursors can change with fuel and local reactivity. Chemical explosive modes (CEMs) are a natural alternative to study this as they can provide a measure for auto-ignition risk by considering the whole thermochemical state in the framework of an eigenvalue problem. When transport effects are included by coupling the evolution of the chemical explosive modes to turbulence, it is possible to obtain a measure of spatial auto-ignition risk where both chemical (e.g., ignition delay) and aerodynamic (e.g., local residence time) influences are unified. In this article, we describe a method that couples large eddy simulation (LES) to newly developed, reduced auto-ignition chemical kinetics to study auto-ignition precursors in an example premixer representative of real life geometric complexity. A blend of pure methane and di-methyl ether (DME), a common fuel used for experimental auto-ignition studies, was transported using the reduced mechanism (38 species/238 reactions) under engine conditions at increasing levels of DME concentrations until exothermic auto-ignition kernels were formed. The resolution of species profiles was ensured by using a thickened flame model where dynamic thickening was carried out with a flame sensor modified to work with multistage heat release. This paper is outlined as follows: First, a reduced mechanism is constructed and validated for modeling methane as well as DME auto-ignition. Second, sensitivity analysis is used to show the need for chemical explosive modes. Third, the thickened flame model modifications are described and then applied to an example premixer at 25 bar/890 K preheat. The chemical explosive mode analysis closely follows the large thermochemical changes in the premixer as a function of DME concentrations and identifies where the premixer is sensitive and flame anchoring is likely to occur.

Près de soixante années après l’avènement de la Révolution tranquille, la notion de Grande Noirceur continue de hanter la mémoire du duplessisme. Ce texte examine d’abord les différentes approches que des essayistes, des historiens et des sociologues ont implicitement ou explicitement mises en oeuvre dans leurs travaux ou leurs réflexions sur la notion de Grande Noirceur. On peut, à ce propos, distinguer trois postures fort différentes. Les deux premières, la réitération et la rectification, prennent en quelque sorte la notion au pied de la lettre et cherchent soit à en réaffirmer la validité, soit, tout au contraire, à la contester en montrant sa fausseté. La troisième approche, celle de l’objectivation, consiste à construire la notion de Grande Noirceur en s’inspirant des règles de la méthode dans le domaine des sciences sociales. Dans une telle perspective, l’auteur proposera de déplacer le regard vers les années 1960 et 1970 et de considérer l’idée de Grande Noirceur comme l’une des notions centrales du discours providentialiste québécois. Ce discours s’organise, en effet, à partir de deux notions antithétiques, celles de Grande Noirceur et de Révolution tranquille, dont il est impossible de faire la synthèse. À partir de là, le discours providentialiste se déploie en deux chaînes de significations parfaitement étrangères l’une à l’autre. L’article se termine par un bref retour sur le duplessisme.

Accurate thermochemical mechanisms that can predict the formation of nitrogen oxides (NO x ) are important design tools for low-emissions engines. The lack of accurate direct measurements of reac...

The minimization of autoignition risk is critical to the design of premixers of high power aeroderivative gas turbines as an increased use of highly reactive future fuels (for example, hydrogen or higher hydrocarbons) is anticipated. Safety factors based on ignition delays of homogeneous mixtures, are generally used to guide the choice of a residence time for a given premixer. However, autoignition chemistry at aeroderivative conditions is fast (0.5–2 milliseconds) and can be initiated within typical premixer residence times. The analysis of what takes place in this short period necessarily involves the study of low-temperature autoignition precursor chemistry, but precursors can change with fuel and local reactivity. Chemical Explosive Modes are a natural alternative to study this as they can provide a measure of autoignition risk by considering the whole thermochemical state in the framework of an eigenvalue problem. When transport effects are included by coupling the evolution of the Chemical Explosive Modes to turbulence, it is possible to obtain a measure of spatial autoignition risk where both chemical (e.g. ignition delay) and aerodynamic (e.g. local residence time) influences are unified. In this article, we describe a method that couples Large Eddy Simulation to newly developed, reduced autoignition chemical kinetics to study autoignition precursors in an example pre-mixer representative of real life geometric complexity. A blend of pure methane and dimethyl ether (DME), a common fuel used for experimental autoignition studies, was transported using the reduced mechanism (38 species / 238 reactions) at engine conditions at increasing levels of DME concentration until exothermic autoignition kernels were formed. The resolution of species profiles was ensured by using a thickened flame model where dynamic thickening was carried out with a flame sensor modified to work with multi-stage heat release. The paper is outlined as follows: First, a reduced mechanism is constructed and validated for modeling methane as well as di-methyl ether (DME) autoignition. Second, sensitivity analysis is used to show the need for Chemical Explosive Modes. Third, the thickened flame model modifications are described and then applied to an example premixer at 25 bar / 890K preheat. The Chemical Explosive Mode analysis closely follows the large thermochemical changes in the premixer as a function of DME concentration and identifies where the premixer is sensitive and flame anchoring is likely to occur.

New ignition delay time (IDT) data for stoichiometric natural gas (NG) blends composed of C1 – C5 n-alkanes with methane as the major component were recorded using a high pressure shock tube (ST) at reflected shock pressures (p5) and temperatures (T5) in the range 20–30 bar and 1000–1500 K, respectively. The good agreement of the new IDT experimental data with literature data shows the reliability of the new data at the conditions investigated. Comparisons of simulations using the NUI Galway mechanism (NUIGMech1.0) show very good agreement with the new experimental results and with the existing data available in the literature. Empirical IDT correlation equations have been developed through multiple linear regression analyses for these C1 – C5 n-alkane/air mixtures using constant volume IDT simulations in the pressure range pC = 10–50 bar, at temperatures TC = 950–2000 K and in the equivalence ratio (φ) range 0.3–3.0. Moreover, a global correlation equation is developed using NUIGMech1.0, to predict the IDTs for these NG mixtures and other relevant data available in the literature. The correlation expression utilized in this study employs a traditional Arrhenius rate form including dependencies on the individual fuel fraction, TC, φ and pC.

Accurate low dimension chemical kinetic models for methane are an essential component in the design of efficient gas turbine combustors. Kinetic models coupled to computational fluid dynamics (CFD) provide quick and efficient ways to test the effect of operating conditions, fuel composition and combustor design compared to physical experiments. However, detailed chemical kinetic models are too computationally expensive for use in CFD. We propose a novel data orientated three-step methodology to produce compact models that replicate a target set of detailed model properties to a high fidelity. In the first step, a reduced kinetic model is obtained by removing all non-essential species from the detailed model containing 118 species using path flux analysis (PFA). It is then numerically optimised to replicate the detailed model's prediction in two rounds; First, to selected species (OH,H,CO and CH4) profiles in perfectly stirred reactor (PSR) simulations and then re-optimised to the detailed model's prediction of the laminar flame speed. This is implemented by a purposely developed Machine Learned Optimisation of Chemical Kinetics (MLOCK) algorithm. The MLOCK algorithm systematically perturbs all three Arrhenius parameters for selected reactions and assesses the suitability of the new parameters through an objective error function which quantifies the error in the compact model's calculation of the optimisation target. This strategy is demonstrated through the production of a 19 species and a 15 species compact model for methane/air combustion. Both compact models are validated across a range of 0D and 1D calculations across both lean and rich conditions and shows good agreement to the parent detailed mechanism. The 15 species model is shown to outperform the current state-of-art models in both accuracy and range of conditions the model is valid over., Conference Paper: ASME Turbo Expo 2021

New ignition delay time measurements of natural gas mixtures enriched with small amounts of n-hexane and n-heptane were performed in a rapid compression machine to interpret the sensitization effect of heavier hydrocarbons on auto-ignition at gas-turbine relevant conditions. The experimental data of natural gas mixtures containing alkanes from methane to n-heptane were carried out over a wide range of temperatures (840–1050 K), pressures (20–30 bar), and equivalence ratios (φ = 0.5 and 1.5). The experiments were complimented with numerical simulations using a detailed kinetic model developed to investigate the effect of n-hexane and n-heptane additions. Model predictions show that the addition of even small amounts (1–2%) of n-hexane and n-heptane can lead to an increase in reactivity by ∼40–60 ms at compressed temperature (TC) = 700 K. The ignition delay time (IDT) of these mixtures decreases rapidly with an increase in concentration of up to 7.5% but becomes almost independent of the C6/C7 concentration beyond 10%. This sensitization effect of C6 and C7 is also found to be more pronounced in the temperature range 700–900 K compared to that at higher temperatures (>900 K). The reason is attributed to the dependence of IDT primarily on H2O2(+M) ↔ 2ȮH(+M) at higher temperatures while the fuel-dependent reactions such as H-atom abstraction, RȮ2 dissociation, or Q˙OOH + O2 reactions are less important compared to 700–900 K, where they are very important.

Increasingly stringent regulations are imposed on nitrogen oxides emissions due to their numerous negative impacts on human health and the environment. Accurate, experimentally validated thermochemical models are required for the development of the next generation of combustors. This paper presents a series of experiments performed in lean, premixed, laminar, jet-wall stagnation flames at pressures of 2, 4, 8, and 16 atm. To target postflame temperatures relevant to gas turbine engines, the stoichiometry of the nonpreheated methane–air mixture is adjusted to an equivalence ratio of 0.7. One-dimensional (1D) profiles of temperature and NO mole fraction are measured via laser-induced fluorescence (LIF) thermometry and NO-LIF, respectively, to complement previously published flame speed data (Versailles et al., 2018, “Measurements of the Reactivity of Premixed, Stagnation, Methane-Air Flames at Gas Turbine Relevant Pressures,” ASME. J. Eng. Gas Turbines Power, 141(1), p. 011027). The results reveal that, as the pressure increases, the maximum postflame temperature stays relatively stable, and the concentration of NO produced through the flame front remains constant within uncertainty. Seven thermochemical models, selected for their widespread usage or recent date of publication, are validated against the experimental data. While all mechanisms accurately predict the postflame temperature, thanks to consistent thermodynamic parameters, important disagreements are observed in the NO concentration profiles, which highlights the need to carefully select the models used as design tools. The lack of pressure dependence of NO formation that many models fail to capture is numerically investigated via sensitivity and reaction path analyses applied to the solution of flame simulations. The termolecular reaction H+O2(+M)↔HO2(+M) is shown to hinder the production of atomic oxygen and to consume hydrogen radicals at higher pressures, which inhibits the formation of nitric oxide through the N2O pathway.

This paper reports on an optimization study of the CO turndown behavior of an axially staged combustor, in the context of industrial gas turbines (GTs). The aim of this work is to assess the optimally achievable CO turndown behavior limit given system and operating characteristics, without considering flow-induced behaviors such as mixing quality and flame spatial characteristics. To that end, chemical reactor network (CRN) modeling is used to investigate the impact of various system and operating conditions on the exhaust CO emissions of each combustion stage, as well as at the combustor exit. Different combustor residence time combinations are explored to determine their contribution to the exhaust CO emissions. The two-stage combustor modeled in this study consists of a primary (Py) and a secondary (Sy) combustion stage, followed by a discharge nozzle (DN), which distributes the exhaust to the turbines. The Py is modeled using a freely propagating flame (FPF), with the exhaust gas extracted downstream of the flame front at a specific location corresponding to a specified residence time (tr). These exhaust gases are then mixed and combusted with fresh gases in the Sy, modeled by a perfectly stirred reactor (PSR) operating within a set tr. These combined gases then flow into the DN, which is modeled by a plug flow reactor (PFR) that cools the gas to varying combustor exit temperatures within a constrained tr. Together, these form a simplified CRN model of a two-stage, dry-low emissions (DLEs) combustion system. Using this CRN model, the impact of the tr distribution between the Py, Sy, and DN is explored. A parametric study is conducted to determine how inlet pressure (Pin), inlet temperature (Tin), equivalence ratio (ϕ), and Py–Sy fuel split (FS), individually impact indicative CO turndown behavior. Their coupling throughout engine load is then investigated using a model combustor, and its effect on CO turndown is explored. Thus, this aims to deduce the fundamental, chemically driven parameters considered to be most important for identifying the optimal CO turndown of GT combustors. In this work, a parametric study and a model combustor study are presented. The parametric study consists of changing a single parameter at a time, to observe the independent effect of this change and determine its contribution to CO turndown behavior. The model combustor study uses the same CRN, and varies the parameters simultaneously to mimic their change as an engine moves through its steady-state power curve. The latter study thus elucidates the difference in CO turndown behavior when all operating conditions are coupled, as they are in practical engines. The results of this study aim to demonstrate the parameters that are key for optimizing and improving CO turndown.

The main motivation of this study is to investigate detailed NOx and CO formation in high-pressure dump combustor fired with lean premixed methane-air mixture using CFD-CRN hybrid approach. Further, this study is extended to investigate the effect of H2 enrichment on emission formation in the same combustor. Three-dimensional steady RANS CFD simulations have been performed using a Flamelet Generated Manifold (FGM) model in Simcenter STAR-CCM+ 2019.2 with the DRM22 reduced mechanism. The CFD simulations have been modelled along with all three heat transfers modes: conduction, convection and radiation. The conjugate heat transfer (CHT) approach and participating media radiation modelling have been used here. Initially, CFD simulations are performed for five lean equivalence ratios (ϕ = 0.43–0.55, Tinlet = 673 K, Vinlet = 40 m/s) of pure methane-air mixture operating at 5 bar. The exit temperature and flame-length are compared with available experimental data. The automatic chemical reactor network has been constructed from CFD data and solved using the recently developed reactor network module of Simcenter STAR-CCM+ 2019.2 in a single framework for each cases. It is found out that the CRNs up to 10,000 PSRs can provide adequate accuracy in exit NOx predictions compared to experiments for pure methane cases. The contribution of NOx formation pathway, changes from N2O intermediate to thermal NO as equivalence ratio increases. Further studies are performed for two equivalence ratios (ϕ = 0.43 and 0.50 to simulate the impact of H2 addition (up to 40% by volume) on NOx formation pathways and CO emission. It is found out here that the contribution from NNH pathway increases for leaner equivalence ratio cases (ϕ = 0.43), while thermal pathway slightly increases for ϕ = 0.50 with increase in H2 content from 0% to 40%.

This paper reports on an optimisation study of the CO turndown behaviour of an axially staged combustor, in the context of industrial gas turbines (GT). The aim of this work is to assess the optimally achievable CO turndown behaviour limit given system and operating characteristics, without considering flow-induced behaviours such as mixing quality and flame spatial characteristics. To that end, chemical reactor network modelling is used to investigate the impact of various system and operating conditions on the exhaust CO emissions of each combustion stage, as well as at the combustor exit. Different combustor residence time combinations are explored to determine their contribution to the exhaust CO emissions. The two-stage combustor modelled in this study consists of a primary (Py) and a secondary (Sy) combustion stage, followed by a discharge nozzle (DN), which distributes the exhaust to the turbines. The Py is modelled using a freely propagating flame (FPF), with the exhaust gas extracted downstream of the flame front at a specific location corresponding to a specified residence time (tr). These exhaust gases are then mixed and combusted with fresh gases in the Sy, modelled by a perfectly stirred reactor (PSR) operating within a set tr. These combined gases then flow into the DN, which is modelled by a plug flow reactor (PFR) that cools the gas to varying combustor exit temperatures within a constrained tr. Together, these form a simplified CRN model of a two-stage, dry-low emissions (DLE) combustion system. Using this CRN model, the impact of the tr distribution between the Py, Sy and DN is explored. A parametric study is conducted to determine how inlet pressure (Pin), inlet temperature (Tin), equivalence ratio (ϕ) and Py-Sy fuel split (FS), individually impact indicative CO turndown behaviour. Their coupling throughout engine load is then investigated using a model combustor, and its effect on CO turndown is explored. Thus, this aims to deduce the fundamental, chemically-driven parameters considered to be most important for identifying the optimal CO turndown of GT combustors. In this work, a parametric study and a model combustor study are presented. The parametric study consists of changing a single parameter at a time, to observe the independent effect of this change and determine its contribution to CO turndown behaviour. The model combustor study uses the same CRN, and varies the parameters simultaneously to mimic their change as an engine moves through its steady-state power curve. The latter study thus elucidates the difference in CO turndown behaviour when all operating conditions are coupled, as they are in practical engines. The results of this study aim to demonstrate the parameters that are key for optimising and improving CO turndown.

New ignition delay time measurements of natural gas mixtures enriched with small amounts of n-hexane and n-heptane were performed in a rapid compression machine to interpret the sensitization effect of heavier hydrocarbons on auto-ignition at gas-turbine relevant conditions. The experimental data of natural gas mixtures containing alkanes from methane to n-heptane were carried out over a wide range of temperatures (840–1050 K), pressures (20–30 bar), and equivalence ratios (φ = 0.5 and 1.5). The experiments were complimented with numerical simulations using a detailed kinetic model developed to investigate the effect of n-hexane and n-heptane additions. Model predictions show that the addition of even small amounts (1–2%) of n-hexane and n-heptane can lead to increase in reactivity by ∼40–60 ms at compressed temperature (TC) = 700 K. The ignition delay time (IDT) of these mixtures decrease rapidly with an increase in concentration of up to 7.5% but becomes almost independent of the C6/C7 concentration beyond 10%. This sensitization effect of C6 and C7 is also found to be more pronounced in the temperature range 700–900 K compared to that at higher temperatures (> 900 K). The reason is attributed to the dependence of IDT primarily on H2O2(+M) ↔ 2ȮH(+M) at higher temperatures while the fuel dependent reactions such as H-atom abstraction, RȮ2 dissociation or Q.OOH + O2 reactions are less important compared to 700–900 K, where they are very important.

Accurate and robust thermochemical models are required to identify future low-NOx technologies that can meet the increasingly stringent emissions regulations in the gas turbine industry. These mechanisms are generally optimized and validated for specific ranges of operating conditions, which result in an abundance of models offering accurate nominal solutions over different parameter ranges. At atmospheric conditions, and for methane combustion, a relatively good agreement between models and experiments is currently observed. At engine-relevant pressures, however, a large variability in predictions is obtained as the models are often used outside their validation region. The high levels of uncertainty found in chemical kinetic rates enable such discrepancies between models, even as the reactions are within recommended rate values. The current work investigates the effect of such kinetic uncertainties in NO predictions by propagating the uncertainties of 30 reactions, that are both uncertain and important to NO formation, through the combustion model at engine-relevant pressures. Understanding the uncertainty sources in model predictions and their effect on emissions at these pressures is key in developing accurate thermochemical models to design future combustion chambers with any confidence. Lean adiabatic, freely-propagating, laminar flames are therefore chosen to study the effect of parametric kinetic uncertainties. A non-intrusive, level 2, nested sparse-grid approach is used to obtain accurate surrogate models to quantify NO prediction intervals at various pressures. The forward analysis is carried up to 32 atm to quantify the uncertainty in emissions predictions to pressures relevant to the gas turbine community, which reveals that the NO prediction uncertainty decreases with pressure. After performing a Reaction Pathway Analysis, this reduction is attributed to the decreasing contribution of the prompt-NO pathway to total emissions, as the peak CH concentration and the CH layer thickness decrease with pressure. In the studied lean condition, the contribution of the pressure-dependent N2O production route increases rapidly up to 10 atm before stabilizing towards engine-relevant pressures. The uncertain prediction ranges provide insight into the accuracy and precision of simulations at high pressures and warrant further research to constrain the uncertainty limits of kinetic rates to capture NO concentrations with confidence in early design phases.

Increasingly stringent regulations are imposed on nitric oxide (NO) due to its numerous, direct and indirect, deleterious effects on human health and the environment. A better control of the post-flame temperature field to contain the thermal (Zel’dovich) route resulted in significant reductions of engine emissions. An improved knowledge of the chemistry and rate of the secondary prompt, NNH, and N2O pathways is now required to decrease nitric oxide emissions further. For this effort, NO laser-induced fluorescence (LIF) measurements are presented for lean ( ϕ = 0.7 ), jet-wall, stagnation, premixed flames at pressures of 2, 4, and 8 atm. For all cases, the NO-LIF signal increases rapidly through the flame front, and relatively slowly in the post-flame region where the temperature is too low to sustain the thermal pathway. Nitric oxide mole fractions are inferred from the measurements and show that the pressure has a very weak, monotonic, adverse effect on NO formation. Reaction pathway analyses are applied to flame simulations performed with a thermochemical model capturing the general trends of the data to assess the contribution of each NO formation route. For all cases, the N2O pathway, which proceeds mostly in the flame front region, dominates. This route produces slightly larger amounts of NO at higher pressures, but the variation appears very limited when considering the termolecular nature of its initiation reaction. The thermal route is predicted to progress slowly in the post-flame region, which causes a shallow increase in the NO mole fraction. The NNH and prompt pathways generate small amounts of NO, and their contributions reduce with the pressure. Overall, the thermochemical model predicts that the formation of NO is relatively unaffected by the pressure, which is consistent with the experiments. The experimental dataset reported in this work is made available for the development of thermochemical models.

An experimental and kinetic modeling study of the influence of NOx (i.e. NO2, NO and N2O) addition on the ignition behavior of methane/‘air’ mixtures is performed. Ignition delay time measurements are taken in a rapid compression machine (RCM) and in a shock tube (ST) at temperatures and pressures ranging from 900–1500 K and 1.5–3.0 MPa, respectively for equivalence ratios of 0.5–2.0 in ‘air’. The conditions chosen are relevant to spark ignition and homogeneous charge compression ignition engine operating conditions where exhaust gas recirculation can potentially add NOx to the premixed charge. The RCM measurements show that the addition of 200 ppm NO2 to the stoichiometric CH4/oxidizer mixture results in a factor of three increase in reactivity compared to the baseline case without NOx for temperatures in the range 600–1000 K. However, adding up to 1000 ppm N2O does not show any appreciable effect on the measurements. The promoting effect of NO2 was found to increase with temperature in the range 950–1150 K, while the sensitization effect decreases at higher pressures. The experimental results measured are simulated using NUIGMech1.2 comprising an updated NOx sub-chemistry in this work. A kinetic analysis indicates that the competition between the reactions ĊH3 + NO2 ↔ CH3Ȯ + NO and ĊH3 + NO2 (+M) ↔ CH3NO2 (+M), the former being a propagation reaction and the latter being a termination reaction governs NOx sensitization on CH4 ignition. Recent calculations by Matsugi and Shiina (A. Matsugi, H. Shiina, J. Phys. Chem. A. 121 (2017) 4218–4224) for the nitromethane formation reaction CH3 + NO2 (+M) ↔ CH3NO2 (+M), together with the recently calculated rate constants for HONO/HNO2 reactions significantly improve ignition delay time predictions in the temperature range 600–1000 K. Furthermore, the experiments with NO addition reveal a non-monotonous sensitization impact on CH4 ignition at lower temperatures with NO initially acting as an inhibitor at low NO concentrations and then as a promoter as NO concentrations increase in the mixture. This non-monotonous trend is attributed to the role of the chain-termination reaction ĊH3 + NO2 (+M) ↔ CH3NO2 (+M) and the impact of NO on the transition to the chain-branching steps CH2O + HȮ2 ↔ HĊO + H2O2, H2O2 (+M) ↔ ȮH + ȮH (+M), HĊO ↔ CO + Ḣ followed by CO + O2 ↔ CO2 + O and Ḣ + O2 ↔ O + ȮH. NUIGMech1.2 is systematically validated against the new ignition delay measurements taken here together with species measurements and high temperature ignition delay time data available in the literature for CH4/oxidizer mixtures diluted with NO2/N2O/NO and is observed to accurately capture the sensitization trends.

As the foundation for detailed hierarchical combustion chemistry models, the accuracy of the hydrogen oxidation mechanism is paramount to achieve truly predictive models. The recent introduction of new chemically termolecular reactions to this system suggests that our understanding of this fundamental chemistry could have initially been plagued with structural inaccuracies that have been carried along in the development of larger thermochemical mechanisms. Consequently, there is a need for additional independent experimental data in various experimental setups to assess the accuracy of model predictions. An atmospheric jet-wall stagnation-flame configuration is used in this work to measure the reactivity of premixed, lean-to-rich hydrogen-air flames diluted with argon. Mixtures with estimated adiabatic flame temperatures of ∼ 1800 K are obtained by argon dilution to maintain the oxygen-nitrogen ratio while allowing for stable, laminar flames over the desired conditions. The accuracy of various thermochemical mechanisms available in the literature is assessed with direct comparisons of the velocity profiles in the stagnation configuration obtained through Particle Tracking Velocimetry to the solution of the quasi–1D flame simulation. Unstretched laminar flame speeds ( S L o ) are subsequently inferred using a direct comparison technique between the experiments and the simulations at multiple stretch values, similarly to a non-linear extrapolation to unstretched conditions. Significant discrepancies are observed between model predictions and experimental measurements, which supports the current efforts in model development. The inaccuracies in kinetic rates of the hydrogen oxidation model must first be resolved to avoid optimization and validation errors in the hierarchical development of increasingly complex hydrocarbon mechanisms.

This study provides fundamental experimental evidence for the influence of exhaust gas recirculation (EGR) on the formation of NO x and reactivity of premixed flames at atmospheric pressure. The experiments are conducted in a twin counterflow apparatus using methane–air premixed flame compositions with and without EGR dilution. The nitric oxide concentration, temperature and flame burning velocity are measured with non-intrusive, laser diagnostic methods. The combined effects of N 2 , CO 2 , and H 2 O dilution are assessed at flame temperatures of 1850 K and 2000 K. These flames are also directly modeled using the experimental boundary conditions and the GRI-Mech 3.0, San Diego 2005, and Combustion Science and Engineering thermochemical mechanisms to assess the model performance in successfully capturing the EGR conditions. These experimental data confirm the view that EGR does reduce NO formation compared to cases without EGR; however, these reductions are realized only for sufficiently long post-flame residence times. This is due to differing NO formation rates, that are fast for the diluted case, but are slow for the undiluted flame compositions. The prediction of these residence times are obscured by high variability (factor of ≈ 4) and uncertainty (up to ≈ 99%) in the present thermochemical mechanisms. A reduced set of reactions which can assist in the optimization of these models is also presented.

Uncertainty quantification (UQ) is becoming an essential attribute for development of computational tools in gas turbine combustion systems. Prediction of emissions with a variety of gaseous fuels and uncertain conditions requires probabilistic modeling tools, especially at part load conditions. The aim of this paper was to develop a computationally efficient tool to integrate uncertainty, sensitivity, and reliability analyses of CO and NOx emissions for a practical swirl-stabilized premixed burner. Sampling-based method (SBM), nonintrusive polynomial chaos expansion (NIPCE) based on point collocation method (PCM), Sobol sensitivity indices, and first-order reliability method (FORM) approaches are integrated with a chemical reactor network (CRN) model to develop a UQ-enabled emissions prediction tool. The CRN model consisting of a series of perfectly stirred reactors (PSRs) to model CO and NOx is constructed in Cantera. Surrogate models are developed using NIPCE-PCM approach and compared with the results of CRN model. The surrogate models are then used to perform global sensitivity and reliability analyses. The results show that the surrogate models substantially reduce the required computational costs by 2 to 3 orders of magnitude in comparison with the SBM to calculate sensitivity indices, importance factors and perform reliability analysis. Moreover, the results obtained by the NIPCE-PCM approach are more accurate in comparison with the SBM. Therefore, the developed UQ-enabled emissions prediction tool based on CRN and NIPCE-PCM approaches can be used for practical combustion systems as a reliable and computationally efficient framework to conduct probabilistic modeling of emissions.

Critical slowing down (CSD) is a phenomenon that is common to many complicated dynamical systems as they approach critical transitions/bifurcations. We demonstrate that pressure signals measured during the onset of thermoacoustic instabilities in a gas turbine engine test exhibit evidence of CSD well before the oscillation amplitude increases. CSD was detected through both the variance and the lag-1 auto-regressive coefficient in a rolling window of the pressure signal. Increasing trends in both metrics were quantified using Kendall's τ, and the robustness and statistical significance of the observed increases were confirmed. Changes in the CSD metrics could be detected several seconds prior to changes in the oscillation amplitude. Hence, real-time calculation of these metrics holds promise as early warning signals of impending thermoacoustic instabilities.

The adiabatic, unstrained, laminar flame speed, SL, is a fundamental combustion property, and a premier target for the development and validation of thermochemical mechanisms. It is one of the leading parameters determining the turbulent flame speed, the flame position in burners and combustors, and the occurrence of transient processes, such as flashback and blowout. At pressures relevant to gas turbine engines, SL is generally extracted from the continuous expansion of a spherical reaction front in a combustion bomb. However, independent measurements obtained in different types of apparatuses are required to fully constrain thermochemical mechanisms. Here, a jet-wall, stagnation burner designed for operation at gas turbine relevant conditions is presented, and used to assess the reactivity of premixed, lean-to-rich, methane-air flames at pressures up to 16 atm. One-dimensional (1D) profiles of axial velocity are obtained on the centreline axis of the jet-wall burner using Particle Tracking Velocimetry, and compared to quasi-1D flame simulations performed with a selection of thermochemical mechanisms available in the literature. Significant discrepancies are observed between the numerical and experimental data, and among the predictions of the mechanisms. This motivates further chemical modeling efforts, and implies that designers in industry must carefully select the mechanisms employed for the development of gas turbine combustors.

Many sources of uncertainty exist when emissions are modelled for a gas turbine combustion system. They originate from uncertain inputs, boundary conditions, calibration, or lack of sufficient fidelity in the model. In this paper, a non-intrusive polynomial chaos expansion (NIPCE) method is coupled with a chemical reactor network (CRN) model using Python to rigorously quantify uncertainties of NOx emission in a premixed burner. The first objective of the uncertainty quantification (UQ) in this study is development of a global sensitivity analysis method based on NIPCE to capture aleatory uncertainty due to the variation of operating conditions and input parameters. The second objective is uncertainty analysis of Arrhenius parameters in the chemical kinetic mechanism to study the epistemic uncertainty in the modelling of NOx emission. A two-reactor CRN consisting of a perfectly stirred reactor (PSR) and a plug flow reactor (PFR) is constructed in this study using Cantera to model NOx for natural gas at the relevant operating conditions for a benchmark premixed burner. UQ is performed through the use of a number of packages in Python. The results of uncertainty and sensitivity analysis using NIPCE based on point collocation method (PCM) are then compared with the results of advanced Monte Carlo simulation (MCS). Surrogate models are also developed based on the NIPCE approach and compared with the forward model in Cantera to predict NOx emissions. The results show the capability of NIPCE approach for UQ using a limited number of evaluations to develop a UQ-enabled emission prediction tool for gas turbine combustion systems.

Finite-rate chemical effects at gas turbine conditions lead to incomplete combustion and well-known emissions issues. Although a thin flame front is preserved on an average, the instantaneous flame location can vary in thickness and location due to heat losses or imperfect mixing. Postflame phenomena (slow CO oxidation or thermal NO production) can be expected to be significantly influenced by turbulent eddy structures. Since typical gas turbine combustor calculations require insight into flame stabilization as well as pollutant formation, combustion models are required to be sensitive to the instantaneous and local flow conditions. Unfortunately, few models that adequately describe turbulence–chemistry interactions are tractable in the industrial context. A widely used model capable of employing finite-rate chemistry is the eddy dissipation concept (EDC) model of Magnussen. Its application in large eddy simulations (LES) is problematic mainly due to a strong sensitivity to the model constants, which were based on an isotropic cascade analysis in the Reynolds-averaged Navier–Stokes (RANS) context. The objectives of this paper are: (i) to formulate the EDC cascade idea in the context of LES; and (ii) to validate the model using experimental data consisting of velocity (particle image velocimetry (PIV) measurements) and major species (1D Raman measurements), at four axial locations in the near-burner region of a Siemens SGT-100 industrial gas turbine combustor.

Multiple, interacting flames in DLE systems can increase flame surface area and promote mixing of hot-products into the reactants — leading to an efficient usage of combustion volume and improved injector performance. An optically-accessible, confined, linear array of five swirl nozzles was recently built [1] to investigate flame dynamics and validate computational strategies. The present work focuses on modeling a dataset representative of lean gas turbine conditions, using a flamelet approach. A preheated (500K), premixed fuel-air mixture (ϕ = 0.55, Tflame = 1732K) at atmospheric pressure was injected through the swirlers at 40 m/s into a rectangular chamber. High-speed laser measurements of the flow (3 component velocity field from 10 kHz stereoscopic particle image velocimetry (S-PIV)) and flame (planar laser induced fluorescence of the hydroxyl radical (OH-PLIF)) were used for model validation. The objectives of this work: (1) Evaluate a flamelet-progress variable method based on flamelet-generated manifolds (FGM) and examine its sensitivity to models for micro (scalar dissipation) and large scale mixing (anisotropic RANS vs LES) and (2) Obtain insight into the velocity field and flame stabilization in an interacting system. Computations indicate that high-swirl nozzles produce bluff-body flames anchored to shear-layer vortices due to an arrested flow expansion. The anisotropic RANS turbulence model under-predicts the recirculation zone strength but predicts flow development and Reynolds stress profiles fairly well. While LES is more accurate overall, both models over-predict flow fluctuations in the transitional flow at the end of the recirculation bubble where flow becomes axially positive. The flamelet approach predicts the flame-shape and length correctly but over-predicts the reaction rate in-between swirlers. The effect of including a reactive SDR model is to significantly increase flame-flow interaction (higher scalar variance) but does not appear to influence the overall shape or location of the flame.

Increasingly stringent regulations on emissions in the gas turbine industry require novel designs to minimize the environmental impact of oxides of nitrogen (NOx). The development of advanced low-NOx technologies depends on accurate and reliable thermochemical mechanisms to achieve emissions targets. However, current combustion models have high levels of uncertainty in kinetic rates that, when propagated through calculations, yield significant variations in predictions. A recent study identified and optimized nine elementary reactions involved in CH formation to accurately capture its concentration and improve prompt-NO predictions. The current work quantifies the uncertainty on peak CH concentration and NOx emissions generated by these nine reaction rates only, when propagated through the San Diego mechanism. Various non-intrusive spectral methods are used to study atmospheric alkane-air flames. 1st- and 2nd-order total-order expansions and tensor-product expansions are compared against a reference Monte Carlo analysis to assess the ability of the different techniques to accurately quantify the effect of uncertainties on the quantities of interest. Sparse grids, subsets of the full tensor-product expansion, are shown to retain the advantages of tensor formulation compared to total-order expansions while requiring significantly fewer collocation points to develop a surrogate model. The high resolution per dimension can capture complex probability distributions witnessed in radical species concentrations. The uncertainty analysis of lean to rich flames demonstrated a high variability in NOx predictions reaching up to 400 % of nominal predictions. Wider concentration intervals were observed in rich conditions where prompt-NOx is the dominant contributor to emissions. The high variability and scale of uncertainty in NOx emissions originating from these nine elementary reactions demonstrate the need for future experiments and data assimilation to constrain current models to accurately capture CH for robust NOx emissions predictions.

Journal article The development of accurate chemical kinetic models capable of predicting the combustion of methane and dimethyl ether in common combustion environments such as compression ignition engines and gas turbines is important as it provides valuable data and understanding of these fuels under conditions that are difficult and expensive to study in the real combustors. In this work, both experimental and chemical kinetic model-predicted ignition delay time data are provided covering a range of conditions relevant to gas turbine environments (T = 600-1600 K, p = 7-41 atm, phi = 0.3, 0.5, 1.0, and 2.0 in 'air' mixtures). The detailed chemical kinetic model (Mech_56.54) is capable of accurately predicting this wide range of data, and it is the first mechanism to incorporate high-level rate constant measurements and calculations where available for the reactions of DME. This mechanism is also the first to apply a pressure-dependent treatment to the low-temperature reactions of DME. It has been validated using available literature data including flow reactor, jet-stirred reactor, shock-tube ignition delay times, shock-tube speciation, flame speed, and flame speciation data. New ignition. delay time. measurements are presented for methane, dimethyl ether, and their mixtures; these data were obtained using three different shock tubes and a rapid compression machine. In addition to the DME/CH4 blends, high-pressure data for pure DME and pure methane were also obtained. Where possible, the new data were compared with existing data from the literature, with good agreement. Peter O’Toole acknowledges the financial support of the Irish government under PRTLI Cycle 4. Ultan Burke acknowledges the financial support of the Irish Research Council. Kieran P. Somers acknowledges the support of Science Foundation Ireland under Grant No. [08/IN1./I2055] as part of their Principal Investigator Awards. peer-reviewed